Construction Method and Device of Digital Power System Based on Multi-Functional Agent

Abstract

The present application discloses a construction method and device for a digital power system based on a multi-functional intelligent agent. Feature production factor data and control production factor data are respectively extracted from the power data on the power plant side and the power grid side; a target data vector is constructed based on the power function requirements, and key control production factor data associated with the power function requirements is determined; an artificial intelligence algorithm is used to learn the mapping function between the target data vector and the key control production factor data, a functional operator with the mapping function as the kernel is constructed, and an intelligent agent is constructed based on the functional operator; intelligent agents corresponding to various different power function requirements are added to the digital power system as power function implementation units. This system can realize diversified and digital power functions driven by data, more accurately capture the connection between data and data, and form high-value data assets in the power system.

Description

Technical Field

[0001] The present application relates to the technical field of power systems, and in particular, to a method and device for constructing a digital power system based on multi-functional agents. Background Art

[0002] With the advancement of the industrialization process and the progress and development of science and technology, the international cooperation in electric power energy has been gradually strengthened. The global new power market has emerged: new energy has leaped to the main force, and fossil energy has turned into a supporting role. In this transformation, new entities such as new energy storage, virtual power plants, and intelligent microgrids have emerged one after another, injecting new impetus into the power system. At the same time, the entire power system industry chain also faces many challenges.

[0003] The promotion and development of the global power trading market are inseparable from an accurate understanding and sensitive control of the power system. Therefore, it is urgent to construct an intelligent and digital power system. At present, the construction of the power system faces challenges of complexity and variability. The complexity is reflected in the relatively complex types of power plants connected to the power grid, which poses great challenges to power grid dispatching. The variability is reflected in the uncertainty of some new energy power generation methods (such as photovoltaic, hydropower, etc.). Similarly, this also greatly affects the reliability and stability of power supply in the power system. Currently, it is urgent to combine emerging fields such as new power and intelligent hydropower services to construct an intelligent and digital power system that helps the reform of the domestic power market and global power trading. Summary of the Invention

[0004] The present application provides a method and device for constructing a digital power system based on multi-functional agents, aiming to extract production factor data on the power plant side and the power grid side, construct agents with rich functions, and through agents with diverse functions, assist the power system to realize diverse and digital power functions based on data-driven.

[0005] The first aspect of the present application provides a method for constructing a digital power system based on multi-functional agents, and the method includes:

[0006] Extract characteristic production factor data corresponding to a characteristic production factor group from the power data on the power plant side; and extract control production factor data corresponding to a control production factor group from the power data on the power grid side; the characteristic production factor group includes characteristic production factors at multiple levels; the control production factor group includes control production factors in multiple aspects;

[0007] Based on the power function requirements of the digital power system, use the extracted characteristic production factor data to construct a target data vector, and determine key control production factor data associated with the power function requirements from the extracted control production factor data;

[0008] Based on the target data vector and the key control production factor data, an artificial intelligence algorithm is used to learn the mapping function between the target data vector and the key control production factor data, a functional operator with the mapping function as the kernel is constructed, and an agent is constructed based on the functional operator;

[0009] Agents corresponding to various different power function requirements are added to the digital power system, and each agent is used as a power function implementation unit in the digital power system.

[0010] A second aspect of the present application provides a construction device for a digital power system based on a multi-functional agent, and the device includes:

[0011] A data extraction module, configured to extract feature production factor data corresponding to a feature production factor group from the power data on the power plant side; and extract control production factor data corresponding to a control production factor group from the power data on the grid side; the feature production factor group includes feature production factors at multiple levels; the control production factor group includes control production factors in multiple aspects;

[0012] A vector construction module, configured to construct a target data vector by using the extracted feature production factor data based on the power function requirements of the digital power system;

[0013] A data determination module, configured to determine key control production factor data associated with the power function requirements from the extracted control production factor data;

[0014] An operator construction module, configured to use an artificial intelligence algorithm to learn the mapping function between the target data vector and the key control production factor data based on the target data vector and the key control production factor data, construct a functional operator with the mapping function as the kernel, and construct an agent based on the functional operator;

[0015] An agent addition module, configured to add agents corresponding to various different power function requirements to the digital power system, and use each agent as a power function implementation unit in the digital power system.

[0016] Optionally, in an implementation of the first aspect and the second aspect, the feature production factor group includes first-level feature production factors, second-level feature production factors, and third-level feature production factors corresponding to various different power generation types; among them, the first-level feature production factors are natural feature production factors, the second-level feature production factors are single-system feature production factors, and the third-level feature production factors are plant-system feature production factors; the natural feature production factors are production factors that directly reflect natural features; the single-system feature production factors are production factors related to individual systems in the power system; the plant-system feature production factors are production factors related to the overall system in the power system.

[0017] Optionally, in an implementation of the first aspect and the second aspect, the control production factors of the multiple aspects include: electricity price, inertia and frequency, active and reactive power, carbon emissions from electricity, security and stability, and grid connection and disconnection.

[0018] Extracting control production factor data corresponding to the control production factor group from the power data on the grid side includes:

[0019] Extracting electricity price data, inertia and frequency data, power data, carbon emissions from electricity data, security and stability index data, and grid connection and disconnection impact data from the power data on the grid side.

[0020] Optionally, the power function requirements are: power function requirements for power transactions between the grid side and the power plant side, power function requirements for the grid side to implement dispatching control over the power plant side, or power function requirements for the grid's own regulation.

[0021] Regarding the first aspect and the second aspect, in the first possible implementation, the power function requirement for power transactions between the grid side and the power plant side is specifically calculating the electricity price, the functional operator is the electricity price operator, and the electricity price operator is used to calculate the electricity price; the process of the digital power system applying the agent constructed based on the electricity price operator includes:

[0022] Invoking the agent to obtain the electricity price calculation result of the power plant side through the electricity price operator;

[0023] Verifying the electricity price calculation result through the financial verification model on the grid side; the financial verification model includes: a first verification condition and a second verification condition; the first verification condition is: the actual output of various different types of power plants ≥ grid electricity; the second verification condition is: the sum of the products of the actual output of various different types of power plants, the grid electricity price at the corresponding moment, and the power generation duration ≤ the product of the grid average price and the grid electricity.

[0024] If the electricity price calculation result meets the first verification condition and the second verification condition, it is determined that the electricity price calculation result passes the verification; if the electricity price calculation result does not meet either the first verification condition or the second verification condition, it is determined that the electricity price calculation result fails the verification.

[0025] If the verification of the electricity price calculation result passes, further select the electricity price calculation function that is adapted and feasible on the grid side from the dispatching strategy library, and load the electricity price calculation function into the electricity price operator to update the function of the electricity price operator.

[0026] Obtain the electricity price calculation result on the grid side through the updated electricity price operator.

[0027] Regarding the first aspect and the second aspect, in the second possible implementation manner, the power function requirement for power trading between the grid side and the power plant side is specifically power clearing, the functional operator is a clearing operator, and the clearing operator is used for power clearing; the process of the digital power system using the agent constructed based on the clearing operator includes:

[0028] Perform low-order data abstraction processing according to the power generation element characteristics of each associated power plant of the power grid to obtain the first data vector of each power generation element characteristic.

[0029] Predict the long-term trading electricity price according to the preset long-term electricity price function library and each first data vector to establish a long-term trading stack.

[0030] Establish a spot trading stack based on the spot trading electricity prices of each associated power plant, and determine the first clearing function through the long-term trading stack and the spot trading stack.

[0031] According to the first clearing function, determine the power supply and demand balance state when the power grid clears under the first clearing function.

[0032] Adjust the function of the first clearing function according to the power supply and demand balance state to obtain the second clearing function, and perform power clearing based on the second clearing function.

[0033] Regarding the first aspect and the second aspect, in the third possible implementation manner, the power function requirement for the grid side to implement dispatching control over the power plant side is specifically peak shaving dispatching in the scenario of high-frequency load variation of thermal power, the functional operator is a thermal power peak shaving dispatching operator, and the thermal power peak shaving dispatching operator is used to implement peak shaving dispatching in the scenario of high-frequency load variation of thermal power; the process of the digital power system using the agent constructed based on the thermal power peak shaving dispatching operator includes:

[0034] Calculate the heat storage and release margin of the energy storage device in the thermal power plant; the heat storage and release margin includes a heat storage margin characterization value and a heat release capacity characterization value;

[0035] Send the correlation data among the loads, performances and costs of each thermal power unit in the thermal power plant and the heat storage and release margin to the grid side, so that the grid side determines the target units for expected auxiliary peak shaving and issues peak shaving scheduling instructions based on the change of thermal power load demand, the correlation data provided by each thermal power plant and the heat storage and release margin;

[0036] Receive the peak shaving scheduling instructions issued by the grid side; the peak shaving scheduling instructions carry the unit identification of the target unit and the peak shaving requirement information for the target unit; the peak shaving requirement information includes a peak shaving load curve;

[0037] Execute auxiliary peak shaving services based on the peak shaving requirement information.

[0038] Regarding the first aspect and the second aspect, in the fourth possible implementation manner, the power function requirement for implementing the dispatching control of the grid side over the power plant side is specifically the water resource dispatching of cascade hydropower stations, the functional operator is a water resource dispatching operator, and the water resource dispatching operator is used to implement the water resource dispatching of cascade hydropower stations; the process of the digital power system using the intelligent agent constructed based on the water resource dispatching operator includes:

[0039] Obtain the water resource data of each hydropower station in the cascade hydropower station at the current moment;

[0040] For each hydropower station, judge whether the hydropower station is in a safe operation state according to the water resource data of the hydropower station;

[0041] When it is determined that each hydropower station is in the safe operation state, judge whether to perform water resource dispatching according to the water resource data of each hydropower station, the preset electricity price plan and the target historical data; the target historical data includes the average power generation water consumption and the average power plant revenue corresponding to the month where the current moment is located;

[0042] If it is determined to perform water resource dispatching, then determine a water resource dispatching plan based on the water resource data of each hydropower station and a water resource dispatching model; the water resource dispatching model is a model trained in advance for outputting a water resource dispatching plan; the water resource dispatching plan includes daily dispatching and seasonal dispatching.

[0043] For the first and second aspects, in the fifth possible implementation, the power function requirement for implementing the dispatching control of the grid side over the power plant side is specifically the dispatching of the power grid over the power plant in a large machine and small grid power system. The functional operator is a large machine and small grid dispatching operator, and the large machine and small grid dispatching operator is used to implement the dispatching of the power grid over the power plant in a large machine and small grid power system. The process of the digital power system using the intelligent agent constructed based on the large machine and small grid dispatching operator includes:

[0044] Based on the digital feature extraction of the power grid and the power source points of various power generation types in the power system, determine the accident reserve capacity of various power generation types to be reserved in the power system and configure the accident reserve capacity;

[0045] Based on the output of the power source points of the current various power generation types, determine whether the current power system conforms to the characteristics of a large machine and small grid;

[0046] If it is determined that the current power system conforms to the characteristics of a large machine and small grid, the power grid uses the configured accident reserve capacity of various power generation types to dispatch the power source points in the power system for voltage adjustment based on a multi-round voltage adjustment scheme, and / or dispatch the power source points in the power system for frequency adjustment based on a multi-round frequency adjustment scheme. The multi-round voltage adjustment scheme includes: the multi-round voltage adjustment methods, multi-round voltage adjustment ranges, and multi-round voltage adjustment priority information for various power generation types. The multi-round frequency adjustment scheme includes: the multi-round frequency adjustment methods, multi-round frequency adjustment ranges, and multi-round frequency adjustment priority information for various power generation types;

[0047] Based on the stability control requirements for the frequency dynamic characteristic index of the entire network of the power system, the predicted total network load, and the unit characteristics of each power source point, the power grid generates the output curves of the units of each power source point in the future time period. The frequency dynamic characteristic index is used to numerically represent the power change amount of the entire network required to cause a unit frequency change in the power system;

[0048] The power grid sends a power generation dispatching instruction to each power source point for power adjustment. The power generation dispatching instruction includes the output curve of the unit for the corresponding power source point.

[0049] For the first and second aspects, in the sixth possible implementation, the power function requirement for the self-regulation of the power grid is specifically to adjust the grid structure of the power grid. The functional operator is a grid structure adjustment operator, and the grid structure adjustment operator is used to implement the adjustment of the grid structure of the power grid. The process of the digital power system using the intelligent agent constructed based on the grid structure adjustment operator includes:

[0050] Obtain the current system inertia corresponding to the current grid structure;

[0051] Calculate a frequency change rate based on the current system inertia and the difference between the power supply demand and the actual power supply.

[0052] If the frequency change rate is greater than a preset first threshold, determine a target grid structure and a target setting value corresponding to the target grid structure based on a pre-established deep learning model; the frequency change rate of the target grid structure is less than or equal to the first threshold.

[0053] Modify the current setting value applied in the current grid structure based on the target setting value, and adjust the current grid structure of the power grid to the target grid structure.

[0054] As can be seen from the above technical solutions, the embodiments of the present application have the following advantages:

[0055] In the technical solution of the present application, characteristic production factor data corresponding to a characteristic production factor group and control production factor data corresponding to a control production factor group are respectively extracted from the power data on the power plant side and the power grid side. Thus, the coordination of the characteristic production factors on the power plant side and the control production factors on the power grid side is realized. On the basis of the extracted relevant production factor data, in order to construct an intelligent body with specific power functions, artificial intelligence technology is further used, making full preparations for the application of this technology: based on the power function requirements of the digital power system, a target data vector is constructed using the extracted characteristic production factor data, and key control production factor data associated with the power function requirements is determined from the extracted control production factor data. On the basis of the target data vector and the key control production factor data, an artificial intelligence algorithm is used to learn the mapping function between the target data vector and the key control production factor data, a functional operator with the mapping function as the kernel is constructed, and an intelligent body is constructed based on the functional operator. The intelligent bodies corresponding to various different power function requirements are added to the digital power system, and each intelligent body is used as a power function implementation unit in the digital power system. Thus, the digital power system is equipped with intelligent bodies with diverse functions, and the system can realize diverse and digital power functions based on data driving.

[0056] In the technical solution of this application, since the construction of the intelligent agent depends on the relevant production factor data on the power plant side and the power grid side, and the characteristic production factor group includes multiple levels of characteristic production factors, and the control production factor group includes multiple aspects of control production factors, therefore, it is equivalent to achieving the coordination of multiple levels of characteristic production factors and multiple aspects of control production factors. Based on this, the constructed intelligent agent effectively utilizes the diversified information of power data and realizes the power functions of coordinating multiple production factors, such as the power function for power trading between the power grid side and the power plant side, the power function for realizing the dispatching control of the power grid side over the power plant side, the power function for the power grid's own regulation, etc. The intelligent agent can capture the connection between data and data more accurately, and then obtain more accurate and transferable data analysis results or data calculation results. It can also be understood that through the digital power system constructed by this solution, high-value data assets are formed in the power system, enhancing the analysis value of power data. The constructed digital power system can effectively contribute to the domestic power market reform and global power trading. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of this application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0058] Figure 1A It is a schematic flow chart of constructing an intelligent agent based on data in this application;

[0059] Figure 1B It is a schematic diagram of constructing an operator through deep learning and reinforcement learning provided by an embodiment of this application;

[0060] Figure 1C It is a structural diagram of data in the construction of the digital power system provided by an embodiment of this application;

[0061] Figure 1D It is an implementation architecture diagram of digital power association provided by an embodiment of this application;

[0062] Figure 1E It is a flow chart of a method for constructing a digital power system based on a multi-functional intelligent agent provided by an embodiment of this application;

[0063] Figure 1F It is a schematic diagram of a power supply scenario provided by an embodiment of this application;

[0064] Figure 1G It is a schematic diagram of a production factor division architecture provided by an embodiment of this application;

[0065] Figure 1H Schematic diagram of the relationship among an agent, an operator, a function, and production factor data provided by an embodiment of the present application;

[0066] Figure 1I Flowchart of implementing the construction of a power analysis agent provided by an embodiment of the present application;

[0067] Figure 1J Another schematic diagram of the relationship among an agent, an operator, a function, and production factor data provided by an embodiment of the present application;

[0068] Figure 2A Flowchart on the grid side based on a power price operator;

[0069] Figure 2B Flowchart on the power plant side based on a power price operator;

[0070] Figure 2C Signaling diagram among the power plant side, the grid side, and the agent;

[0071] Figure 3A Flowchart of a grid clearing method provided by an embodiment of the present application;

[0072] Figure 3B Signaling interaction diagram of a grid clearing method provided by an embodiment of the present application;

[0073] Figure 3C Another flowchart of a grid clearing method provided by an embodiment of the present application;

[0074] Figure 3D Schematic diagram of a long-term trading stack provided by an embodiment of the present application;

[0075] Figure 3E Another flowchart of a grid clearing method provided by an embodiment of the present application;

[0076] Figure 3F Schematic diagram of parameters representing the thermal performance of a unit under full load conditions provided by an embodiment of the present application;

[0077] Figure 3G Schematic diagram of comprehensive consideration parameters for coal price and long-term trading power price provided by an embodiment of the present application;

[0078] Figure 3H Schematic diagram of the stack structure of a spot trading stack and a long-term trading stack provided by an embodiment of the present application;

[0079] Figure 3I Flowchart of a method for establishing a spot trading stack provided by an embodiment of the present application;

[0080] Figure 4A A schematic diagram of power generation prediction based on a wind power plant provided by an embodiment of the present application;

[0081] Figure 4B A schematic diagram of a spot trading stack provided by an embodiment of the present application;

[0082] Figure 4C A schematic diagram of a curve of a grid clearing function provided by an embodiment of the present application;

[0083] Figure 4D A schematic diagram of the process of another grid clearing method provided by an embodiment of the present application;

[0084] Figure 4E A schematic diagram of the process of adjusting the first clearing function when the target grid is in an over-generation state provided by an embodiment of the present application;

[0085] Figure 4F A schematic diagram of calculating a second data vector through a deep learning model provided by an embodiment of the present application;

[0086] Figure 4G A schematic diagram of the process of adjusting the first clearing function when the target grid is in an under-generation state provided by an embodiment of the present application;

[0087] Figure 4H A schematic diagram of fitting a power generation curve of a target grid provided by an embodiment of the present application;

[0088] Figure 5A A schematic diagram of a power supply scenario provided by an embodiment of the present application;

[0089] Figure 5B A flowchart of a peak shaving scheduling method in a thermal power high-frequency variable load scenario provided by an embodiment of the present application;

[0090] Figure 5C A signaling interaction diagram of a peak shaving scheduling method in a thermal power high-frequency variable load scenario provided by an embodiment of the present application;

[0091] Figure 5D A flowchart of implementing a method for determining a target unit for expected auxiliary peak shaving provided by an embodiment of the present application;

[0092] Figure 5E A flowchart of fault judgment and working condition judgment provided by an embodiment of the present application;

[0093] Figure 5F A flowchart of another peak shaving scheduling method in a thermal power high-frequency variable load scenario provided by an embodiment of the present application;

[0094] Figure 6AFlowchart of a water resource scheduling method provided by an embodiment of the present application;

[0095] Figure 6B Schematic diagram of historical power generation water consumption corresponding to each month of an upstream hydropower station provided by an embodiment of the present application;

[0096] Figure 6C Schematic diagram of historical power generation water consumption corresponding to each month of a downstream hydropower station provided by an embodiment of the present application;

[0097] Figure 6D Schematic diagram of historical power plant revenue corresponding to each month of an upstream hydropower station provided by an embodiment of the present application;

[0098] Figure 6E Schematic diagram of historical power plant revenue corresponding to each month of a downstream hydropower station provided by an embodiment of the present application;

[0099] Figure 6F Schematic diagram of water assets corresponding to each month of an upstream hydropower station provided by an embodiment of the present application;

[0100] Figure 6G Schematic diagram of water assets corresponding to each month of a downstream hydropower station provided by an embodiment of the present application;

[0101] Figure 6H Schematic diagram of a water resource regulation scheme for cascade hydropower stations provided by an embodiment of the present application;

[0102] Figure 6I Schematic diagram of the interaction between a power plant and a power grid provided by an embodiment of the present application;

[0103] Figure 6J Flowchart of another water resource scheduling method provided by an embodiment of the present application;

[0104] Figure 7A Schematic diagram of an application scenario of a power grid scheduling method in a large machine and small network power system provided by an embodiment of the present application;

[0105] Figure 7B Example flowchart of a power dispatching method in a large machine and small network power system provided by an embodiment of the present application;

[0106] Figure 7C Schematic diagram of a multi-round voltage adjustment provided by an embodiment of the present application;

[0107] Figure 7D Schematic diagram of a multi-round frequency adjustment provided by an embodiment of the present application;

[0108] Figure 7E Another example flowchart of a power dispatching method in a large machine and small network power system provided by an embodiment of the present application;

[0109] Figure 7F A power dispatching signaling diagram in a large machine and small network power system provided by an embodiment of the present application;

[0110] Figure 8A A flowchart of a method for adjusting a power grid grid structure provided by an embodiment of the present application;

[0111] Figure 8B Another flowchart of a method for adjusting a power grid grid structure provided by an embodiment of the present application;

[0112] Figure 8C A flowchart of a method for determining a target grid structure in the case of a sudden change in power generation provided by an embodiment of the present application;

[0113] Figure 8D A schematic diagram of a binary tree structure provided by an embodiment of the present application;

[0114] Figure 9 An interaction diagram of the power generation dispatching process of a power system provided by an embodiment of the present application;

[0115] Figure 10 A schematic diagram of the relationship between functions and operators provided by an embodiment of the present application. Detailed implementation manners

[0116] Currently, international cooperation in electric power energy is becoming increasingly extensive. Building digital artificial intelligence technology that drives power intelligent agents with numbers, promoting the unified construction of the global power trading market, and formulating top-level design rules to guide the marketization process are the key demands faced by the current electric power industry. Based on the practical experience in the domestic and international electric power industries, the inventor proposes a construction plan for a digital power system based on multi-agent. Based on the demands of the cost-benefit of the entire industrial chain under the global new power market and the digital characteristics of the production factor data included in the power system, these complex factors (coal consumption, light, wind speed, electricity price mechanism, investment cost, etc.) are transformed into data or high-level data vectors to promote cross-industry information circulation and efficient utilization, and new life is given to the data with the help of AI technology.

[0117] In this application, the inventors integrate forward-looking thinking and face the technical and industrial challenges across the entire industrial chain. They emphasize the integration of technology and economy, the digitization of the value chain, and the simultaneous progress of marketization and intelligentization, aiming to provide intelligent and automated solutions for power problems in the energy industry. Combining emerging fields such as new power and intelligent hydropower services, the technical solution of this application provides a method and device for constructing a digital power system based on multi-functional agents, which can serve the global energy Internet and form digital assets. Transforming data into digital assets requires both the data itself and the deepest and most essential understanding of the operation mode of the power industry in terms of the essence of artificial intelligence. With the deepening of the power market reform, the innovation of various technologies and mechanisms has become the key to promoting the reform. Constructing an underlying design framework for the deep integration of digitization and artificial intelligence, digital power based on data-driven, algorithms based on digital power, and agents based on power systems, especially in key links such as dispatching, settlement, and pricing, abstracting the low-order, medium-order, and high-order features of data from different dimensions to achieve a double leap in efficiency and intelligentization. The technical concept of the technical solution of this application can promote the transformation and application of related technologies, thus accelerating the process of power market reform.

[0118] The following introduces some inventive concept ideas involved in the technical solution of this application:

[0119] Symbiosis: At the data level, after digitizing a class of analogous things, they have certain associated and connected properties under a certain abstraction. Symbiosis symbiotically integrates the production factors in nature or the information of production and control of production factors into data, making it possible for internal association and transmission in data. Symbiosis also enables the abstraction of data features to exist in different operators or agents, and different operators and agents can coexist simultaneously, and the characteristics also exist in multiple operators and agents. Essentially, different types of things may have the characteristics of being seemingly scattered but actually similar in essence. Finding the essence and connotation of symbiosis can find the same nature. In the technical solution of this application, the inventors, based on their accumulated knowledge of power systems, use symbiosis to symbiotically integrate the concepts in neural networks into the power system, realizing the ingenious application of artificial intelligence technology in the power field.

[0120] Migration: Transfer the characteristics of a class of things to the characteristics of another related class of things, so that an association is generated and the two classes of things can be interchanged at the data asset level. By using the nesting and invocation of agents, the migration function can be realized, enabling the multi-functional agent to require not too many iterations, and using operators to achieve specific functions. Using migration enables most functions in the power system to be carried from one agent to another.

[0121] Emergence: When the data scale is large enough and there are enough logics within the intelligent body, the phenomenon that the data and logics can regenerate and iterate with each other is the demarcation point between intelligence and wisdom. When artificial intelligence develops to the emergence stage, wisdom will be generated. Recall that humans use tools to change productivity. Tools are means of production. First, in ancient times, and then continuously through the expansion of the input data scale, in order to continuously promote the development of productivity, production tools are invented, and then the laws are summarized and abstracted through disciplines such as mathematics, physics, and chemistry, and finally logics and formula theorems are obtained; in modern society, the development of informatics has changed the carrier of the third industrial revolution of mankind, and electricity, computers, etc. have emerged, which are also tools for iterating information in informatics. However, in the fourth industrial revolution, artificial intelligence has fundamentally upgraded computers, etc. from simple to complex. Using neural networks and statistical principles, it approaches the production function again, enabling it to directly drive from data in a simple and straightforward manner. In the next 10 years, the development of intelligent bodies and their internal logics is equivalent to the development of production tools in previous societies. Human understanding is always deepening step by step. Many phenomena that cannot be explained have internal connections and deeper theorems and concepts. However, the high-order abstraction of data exists and is easier to obtain. The key is to analyze the relevant logics within the intelligent body when the time is ripe, more like waiting for a certain opportunity.

[0122] Intelligent body: An intelligent body is a unit with one or more functions. It makes the power system and artificial intelligence more closely combined. The intelligent body has a protective effect on the power system and the outside world, similar to a cell, but there are smaller kinetic units such as operators inside. The intelligent body has special data characteristics in different scenarios, which has a more profound significance for generating emergence functions and studying the digital essence of the power system. The intelligent body is a core component of artificial intelligence in the power industry, similar to the thermal control logic diagram or relay protection involving logic in the power industry during the third industrial revolution. For example, the logic inside the "five-prevention" of the switchyard can be symbolically represented. The logic inside the intelligent body can also be symbolically represented, which is used to more accurately and deterministically represent the key logics and controls within the power system based on big data and large models. The general logic of the intelligent body conforms to the key theorems and common sense within the power system, conforms to the energy conversion company and relevant rules, and its research objects are production characteristic elements and control production elements, such as illuminance, wind speed, water level, heat consumption, load rate, system inertia, frequency, etc. The interface of the intelligent body is directly connected to the input layer or output layer (output function) of deep learning and reinforcement learning, and is also connected to the intermediate layer when needed. In the technical solution of this application, the first half of the neural network abstracts the natural production elements into calibrated transmitted data, enhancing the parallel computing ability, that is, the fast ability. The operator function and intelligent body in the second half of the output reflect the ability of mathematical abstraction and the physical laws within the industry, and can be combined and reconstructed, enabling the intelligent body to possess the abilities and functions of various industries.

[0123] In the technical solution of this application, an algorithm is constructed throughout the text by focusing on data and information flow. After extracting the features of the data, the data needs to be calibrated, that is, a reasonable data structure is given to the data. At the same time, the internal logic of the operator controls the information flow, and different data structures and information flows corresponding to the digitization of power plants and power grids are also demonstrated through different embodiments. It is better to use different neural networks for the algorithm. For the operators involved, there are differences in data structure and information flow mode. For example, the power grid is suitable to use binary trees for power grid splitting and grid connection, and stacks for clearing electricity prices, while power plants are suitable to gradually establish information flow modes, internal safety logics, etc. from the first-level indicators of nature. Combining the applicable objects, the conditions and directions of information flow are determined, which also determines the selection of convolutional neural networks, recurrent neural networks, graph neural networks, attention networks, etc. The following issues will be discussed: I. Digital feature extraction, II. Mathematical tools, III. Binary trees and stacks in data structures, IV. Logical core, V. Reasoning ability and information classification under the underlying logic.

[0124] I. Digital feature extraction

[0125] Digital features are divided into power plant and power grid digital features. Power grid feature variables are established, such as the real part of the voltage of each node, the imaginary part of the voltage, the active power flowing into each node, the reactive power flowing into each node, system inertia, system frequency, the ratio of system power change rate to system frequency change rate, etc. Power plant feature variables are established, such as generator power angle, voltage of each generator, electromagnetic power of each generator, reactive power of each generator, etc. According to the above production factor digital features from the first-level feature production factors of nature, such as illuminance, coal type and quality, and water level, first, the digital features of each level are extracted. For example, the power plant side extracts the digital features of low-order feature production factors, such as the seasonal volatility of water level and the efficiency stability of illuminance, and the power grid side extracts the digital features of high-order control production factors, such as extracting the feature of system inertia to frequency change rate and the feature of frequency change rate to power change rate. Different means are also adopted. For the digital features of low-order production factors such as water level and photovoltaic illuminance, digital features can be extracted by involving clustering or Euclidean distance method, while the digital features of high-order control production factors require modeling or more advanced mathematical tools to extract digital features.

[0126] II. Mathematical tools

[0127] Mathematical tools can be selected from convolutional neural networks, recurrent neural networks, graph neural networks, or attention networks according to the data and information flow characteristics of the object. The characteristics of the above neural networks are as follows: 1. Convolutional neural network: For regular network data, the information flow is to local areas; 2. Recurrent neural network: Data is input in order, and the information flows in a sequence; 3. Graph neural network: Data is in a fixed graph structure, and the information flows along fixed edges; 4. Attention network: Data is in an unordered set, and the information flow is dynamically controlled by the neural network. These neural networks target the logical core: operators, agents. In the process of constructing a digital power system corresponding to the (operators) in a multi-functional agent, typical cases include: grid connection and disconnection operators of the power grid (Example 6 below), hydropower seasonal scheduling operators (Example 4 below), power source and power grid control operators for large generators and small grids (Example 5 below), clearing function operators (Example 2 below), etc. The data research and information flow characteristics in these operators all deeply reflect the characteristics of the above neural networks. By analyzing the data characteristics and information characteristics respectively from the data feature extraction to the rules of information sequence flow (physical and equipment mechanisms), and selecting the appropriate neural network for each, the algorithms of artificial intelligence can be highly compatible with the power system.

[0128] When using mathematical tools, the specific tools can be determined in combination with the data scale and sample size. The following Figure 1A introduces the example process of constructing an agent based on data in this application. As Figure 1A shown, this process includes:

[0129] (1) Data collection and input.

[0130] (2) Determine the data scale, small sample or large sample. If it is a large sample, step (3) can be omitted; if it is a small sample, directly execute.

[0131] (3) Data fitting, using mathematical tools or formulas to regress the data to a certain convergence value.

[0132] (4) Data cleaning, deleting or processing abnormal data according to the fitted data.

[0133] (5) Data feature extraction, confirming whether it is a production factor feature or a control production factor feature.

[0134] (6) Agent construction, writing key processes or constraints according to the internal logic of the power industry, laying out deep learning and reinforcement learning interfaces for the constructed agent, receiving the data extracted from the data features, and further performing deep learning or reinforcement learning according to its features and classifications.

[0135] (7) Create an output layer output function for the results of deep learning or reinforcement learning, and perform the backpropagation algorithm on the function to adjust and correct the intermediate results of each layer.

[0136] (8) Generalize the output function into control logic, introduce it into the agent, and nest the agent and its key internal data elements into the control logic of the actuator to achieve data-driven.

[0137] Deep learning and reinforcement learning enable information to have a more representative and error-correcting mechanism in the compressed feature transfer. Deep learning has specific algorithms or logics in extracting data features, making the data features more regular in subsequent extractions, facilitating the establishment of ultra-high-level mechanisms or data structures based on data cleaning using data characteristics; maintaining less information loss of data during data transfer while ensuring the data calculation and transmission speed. In reinforcement learning, through mechanisms such as rewards, continuously verify the correctness of the learned logic and determine the priority, enabling the logic to be generated through continuous comparison and calibration.

[0138] Combined with the process introduced above, in the technical solution of this application, digitization is not a single digitization method or the use of a single mathematical tool. There are differences in large and small samples. For digitizing large sample data, deep learning methods (such as using convolutional neural networks, etc.) need to be used to compare the measured data with the samples to form an operation flow, and repeatedly deduce and act in reverse according to the logic flow in the agent to simulate human reasoning. The convolution algorithm can obtain features through deep learning and give them to the operator function. The convolutional neural network adopts an "end-to-end" method, which can automatically identify data features and has functions such as right sharing, hierarchical abstraction of low-level, middle-level, and high-level data features. For small samples, data features are directly extracted from the data.

[0139] For the research object of small samples, the concept of the agent is suitable for application. In the power system, when to use large models (corresponding to large samples) and when to use small samples depend on the characteristics of the data. Currently, weather prediction involves a large amount of data and is suitable for using large models. Regarding dispatching, grid connection and disconnection are also suitable for using large models because the data volume is large. Therefore, the research on agents, especially the innovation of operator logic design of deep learning and reinforcement learning in the power system. Because computing power involves advanced processes, GPUs, HBMs, and data transmission rates, network throughput, etc. are limited, so the logic in the uncertain agent must adapt to GPU high-speed operations very quickly. Therefore, the selected algorithm must adapt to the characteristics of data and logical information flows, capture logical and digital certainty in the changing natural production factors of uncertainty. In the stage from artificial intelligence to AGI\ASI, the application of agents makes small sample data more advantageous than large sample data.

[0140] For small sample data, currently, binary trees and stacks can be used as data structures to achieve the encapsulation of data features.

[0141] III. Binary Trees and Stacks in Data Structures

[0142] The encapsulation of data features using binary trees or stacks enables the data features to be more appropriately represented or easily shared.

[0143] In the embodiments described below, Example 2 shows an example of using the stack method as a data structure. After abstracting the production factors, they are placed into the above data structure to obtain the implementation of its parallel computing and agent logic. Example 2 shows the stack method with the priority gradually decreasing from the bottom to the top. The structure and order characteristics of the stack, where the elements are pushed in first and popped out last, conform to the clearing logic of the power grid and can also achieve the dynamic orderliness of data. For operations such as electricity price clearing for high-order production factors, if high-speed calibration of data parallel computing is required, multiple stacks or hierarchical processing are needed. This increases the pressure of parallel computing but reduces the time for logical judgment, which is a typical case driven by data.

[0144] Example 6 shows an example of using a binary tree as a data structure. By leveraging the inherent characteristics of the data structure, such as a binary tree, grid disconnection and connection can be achieved. Practical data tools are not limited to functions and also include graph theory, etc.

[0145] IV. Logical Core

[0146] Build the associations among 3 major groups, 2 major production factors, and 4 data structures. Through the logical core related to the above content, the logical core connects production factors, operators, agents, etc. After the data is cleaned and calibrated, the data features representing the production factors enter the operators to achieve specific functions, such as controlling electricity price, clearing, frequency, etc. This enables the agents containing these operators to also obtain such multiple functions, realizing symbiosis and migration. And it realizes the underlying logic of data-driven and operators and the subsequent reasoning process of agents, enabling emergence.

[0147] I. Structured and Systematic Logic System

[0148] 1. Build a characteristic production factor group centered on the power plant side, the content of which includes 1 - 3 level indicators, namely natural indicators, single-system indicators, and plant-system indicators. The internal calculation is the parallel calculation of data vectors. Because the parallel calculation speed of data vectors is high enough and only the features of the 1 - 3 level indicators need to be extracted, the data can reflect the characteristics of each part, transmit the features, and perform simple analysis and statistics. Refer to Figure 1D the right part of the implementation architecture diagram of a digital power association shown

[0149] 2. Build a control production factor group centered on the power grid, the content of which includes important indicators of the power grid, namely electricity price, inertia and frequency, active and reactive power, electricity-carbon, security and stability, network combination and disconnection, etc. Refer toFigure 1D The left part of the implementation architecture diagram of a digital power correlation as shown. It is a function for the underlying data logic. According to the system rules within the intelligent body through the data vector, a function mapped by the data vector is summarized through reinforcement learning. The important rules inside the power grid are expressed through function calculation. For example, through the calculation of system inertia, the rate of change of frequency is obtained, so as to call the intelligent body interface for the control production factors of the power grid, and the intelligent body is processed under reinforcement learning to make the power grid balanced and stable. Others such as electricity prices are also based on the principle of digital power correlation, mapping various different types of data vectors to functions on the power grid side, and continuously iterating the function through reinforcement learning to obtain the optimal solution of the control production factors on the power grid side.

[0150] 3. Construct multifunctional intelligent bodies as shown in Embodiment 1 to Embodiment 6. The intelligent body is the key logic body that combines artificial intelligence and electricity to achieve specific functions.

[0151] II. Dataization and functionalization

[0152] 4. The abstract features of the data reflect the primary indicators and the indicators in nature. These indicators need to be abstracted using mathematical tools and then digitized, normalized, and abstracted under the condition of allowing signal loss. Then, taking advantage of the characteristics of parallel computing, data iteration in the transmission layer is carried out. Finally, at the output layer, it is functionalized again, and the intelligent body embodies the relevant rules of nature or the industry such as rules and constraints. The operator is used as a function to implement specific functions, and this operator is nested in the intelligent body and has certain functions, and these functions are symbiotic with the intelligent body. The combination and splitting of the intelligent body are equivalent to the combination and splitting of cells. In this process, the operator is equivalent to protein, and its different combinations and reconstructions make the protein have certain functions, and the data that transmits protein information or nutrients is the data in the transmission layer. It is the data that is generalized into an operator through a function, and these data come from the feature abstraction of production factors. For example Figure 1B , this figure is a schematic diagram of constructing an operator through deep learning and reinforcement learning provided by an embodiment of the present application.

[0153] Nature has established the above-mentioned connection with the Second Machine Revolution based on neural networks and data annotation. Under the self-attention mechanism, this structure and interface enable the probability method to improve the hit rate of artificial intelligence. Coupled with the emergence of GPU parallel computing power, it is necessary to do a good job in the mathematical abstraction, that is, the design of the operator, and the development of the intelligent body in the power industry or all walks of life. The characteristic production factors are the carriers of the characteristics that abstract various indicators. The relevant power plant network indicators are digitized into data vectors to represent their characteristics, and after feature extraction, the data is sent to the operator.

[0154] 5. In the data structure, the digital vector is the key to reflecting the power index quantities on the power plant side. For example, the measurement of the impact of haze on production starts from the data vectors of each power plant (number of haze days, power affected by haze, etc.), and then through the intelligent agent interface, deep learning is used for automatic feature extraction and its association. Similarly, queue sorting is preferably used for the scheduling of thermal power plants participating in variable load peak shaving, and data structure applications such as stack-pointer-electricity price are reflected in the priority and real-time electricity price scheduling methods.

[0155] 6. The function concept of the operator: In the data structure, the function reflects electricity price, active and reactive power, inertia and frequency, electricity-carbon, safety and stability, network combination and disconnection, which is embodied through the operator. The parallel computing function unit that realizes a specific function in a certain power system within the intelligent agent, such as electricity price, system inertia, and frequency. As long as it is a function, it can exist in the form of an operator. For example, the fitting curve function, which corresponds to mapping and summarizing the function according to the data vector, and then strengthening the function association of the logic through the intelligent agent interface.

[0156] After the extraction of data features, the data is encapsulated with the above tools to form a new logical core as the operator. Multiple operators form an intelligent agent, and multiple intelligent agents constitute a digital power system. In this kind of operator, the logic relationship based on data features is applied to the physical and equipment mechanisms (plant network control and equipment) within the power system, and the rules for the formation of information flow are detailed in the flowcharts, signaling diagrams, etc. involved in the following embodiments.

[0157] Based on the three operators of power generation control, clearing, and internal control logic, an association is established. Inspired by the feature extraction in the text, high-dimensional data vectors are used for comparison to realize the data comparison between the power grid and the power plant side, and the optimal features with large correlation weights are found. It fully reflects the association in the operator and the "source-network" matching degree as the underlying logic and algorithm for safety consideration.

[0158] The three operators, namely the power generation control operator, the clearing operator, and the internal control logic operator, respectively describe the functions from aspects such as the power generation mechanism on the power supply side, the power grid clearing principle and logic, and safety consideration. It describes their respective technical solutions, which conform to the creative characteristics of the invention. At the same time, it creatively proposes the association of these three operators, leading to many deeper topics regarding the construction of a digital power system by multiple intelligent agents, such as studying their symbiosis, migration, until more intelligent agents appear and generate emergence.

[0159] The method for constructing a digital power system by multiple intelligent agents. Essentially, this method gives many operations on data and information flow, including data feature extraction, comparison, and matching degree research. The information in the information flow controls the flow direction according to the mechanism and constructs intelligent agents to realize the various functions of the digital power system. The key lies in constructing appropriate operators within the intelligent agent to enable the construction of indicators such as "matching degree" to measure various physical mechanisms of the power grid and the power plant.

[0160] The production factors of the power grid are reflected in the operators and belong to the power grid. For example, in the function operators, there is the electricity price operator, such as the electricity price function, the power grid electricity price equilibrium point, and the internal control operator, and the operator corresponding to the electricity charge in the final function. According to the digital power plant composition of different types of units (including the electricity price equilibrium point of the power plant), and combined with the actual clearing function of the power grid, the power grid electricity price operator is constructed, and at the same time, the power grid electricity price equilibrium point is established. The equilibrium point is the operating point that balances the power plant cost and the power grid clearing transaction electricity price according to the power grid electricity price. Its significance lies in being a key reference for the power grid to formulate electricity prices.

[0161] III. Digitalization and Functionality

[0162] The manifestation of digitalization and data-driven. Multiple embodiments of this application are deeply involved in digitalization. For example, in Embodiment 4, its digitalization is not a completely single digitalization method. There are differences in large and small sample environments. For the digitalization of large samples, deep learning methods must be used to compare the measured data with the samples to form an operation flow, and through repeated deduction and reverse action according to the logic flow in the intelligent agent, it simulates human reasoning.

[0163] The data collected in Embodiment 4 is a small sample. By studying the distribution of water levels, the seasonal cycles and laws of the upstream water levels are found. Only after digitizing it and finding the laws can quarterly regulation scheduling be implemented in the dispatching strategy to maximize higher-order data power generation benefits, unit power generation water consumption, and water assets. Only by finding the correlations between the data, and at the same time finding the main factors in the correlations and analyzing the internal logic in the market and environment can the most core key data be found in the data-driven approach to adapt to the relevance and key nature of each item inside the selected data vector in the data-driven approach. In addition, other similar low- and medium-order data of water levels can be found, and the abstraction of this data feature can be used to create a data-driven method.

[0164] Another example is the second embodiment. In this scheme, for units of various power types, their impact on the power and frequency of the power grid is given by the ratio K of the high-order power difference to the frequency difference. The curve is derived from the quotation and power provided by the initial power point and the physical characteristics of its power type. At the same time, the power curve needs to be further fitted by a power source such as photovoltaics. The data fitting involves a large amount of real-time data. According to the degree of fitting, the next step of the process is selected under data drive. If the fitting is not good, it is necessary to adjust the output power of other units, and even adjust the K value of the entire power grid for the safety of the power grid, so that the K value or the power output value of each power source can be dynamically adjusted and find a balance point under the stable frequency regulation. These are all interacting with each other in the intelligent body. The bottom layer is the association of data and the data drive after abstraction. It is the method of parallel operation that enables this coordination to respond quickly and achieve flexibility that human judgment cannot achieve. The abstract data law replaces the back-and-forth process of human beings abstracting with mathematical formulas and then approaching the function result, which simplifies the process of discovering the law of things. Digitally transmitting laws and information directly to subsequent calculations ensures the accuracy and speed of information.

[0165] Figure 1C The structural diagram of data in the construction of the digital power system provided in the embodiment of the present application focuses on the data aspect. Figure 1C As shown in the figure, from data elements, data vectors, data functionalization to electricity prices, and finally to scheduling algorithms, it directly affects the digitization of the entire industry chain. The connotation of the digital power system revolves around data, whether it starts from the data vector of the digitization of production factors, to the normalization of production factors, to the electricity prices and scheduling strategies after the operation, all of which reflect the underlying logic of the data.

[0166] Quantization of safety data is discussed in Example 4, hydropower cascade scheduling; digitization of protection settings is discussed in Example 6, grid structure change; digitization of production factors is discussed below, reflecting the digitization and coordination of multiple production factors; quantization of other data is discussed in Example 3; scheduling algorithms are discussed in Example 2, clearing function, Example 4, hydropower scheduling, and Example 5; artificial intelligence processing of complex scheduling is discussed in Example 6, Example 4, and Example 5; normalization of production factors is discussed in Example 2; digital functionization of production and settlement is discussed in Example 3, Example 2, and Example 4; electricity price and scheduling functionization is discussed in Example 2 and Example 3; digitization of the entire industrial chain is discussed in Example 1.

[0167] V. Reasoning Ability and Classification of Information Underlying Logic

[0168] The production factors on the grid side highlight the synergistic internal connections. Specifically, the production factors on the power plant side are controlled, and the digital logic of high-order control production factors such as hydropower (inertia control) and photovoltaic (active and reactive power control) is incorporated.

[0169] In the first embodiment, the production factors on the power plant side and the grid side are described, highlighting the synergistic internal connections. The essence lies in establishing the control of computing power within different functions, such as the computing power group of active and reactive power functions, the operator group of grid inertia functions, the computing power group of grid load and electricity price functions, the computing power group of grid security and stability, and the computing power group of grid carbon footprint. There are direct high-speed data connections between each computing power cluster, and they can also be nested and called with each other, enabling decentralized control of the blockchain. These computing power groups are all data-driven based on the previous production or control of production factors. Thus, the dataization of the operator functions within the intelligent body is formed, which is the core of digital power.

[0170] Taking the computing power cluster as the starting point for digital power means deeply considering and grasping digital power. Professionally, by calculating and measuring the computing power of electricity and the digital world, a thinking ability similar to 0PENAIO1 is established. It is to establish a synergistic relationship of function mapping between the intelligent body on the grid side and the power plant side, and establish an associated data vector relationship in terms of data, so as to widely associate vectors with functions, groups (production factors and characteristics) with intelligent bodies (electricity price, active and reactive power, inertia, security, electric carbon footprint).

[0171] Figure 1D This is an implementation architecture diagram for the association of digital power provided by the embodiments of the present application. The digital power system of the multi-functional intelligent body in this solution is also built based on the concept of this implementation architecture diagram.

[0172] Next, through the accompanying drawings and embodiments, the construction method of the digital power system based on the multi-functional intelligent body and the implementation of related devices in the technical solution of the present application will be specifically described. Figure 1E This is a flowchart of a construction method for a digital power system based on a multi-functional intelligent body provided by the embodiments of the present application. As Figure 1E shown, the method includes:

[0173] S101. Extract the characteristic production factor data corresponding to the characteristic production factor group from the power data on the power plant side; and extract the control production factor data corresponding to the control production factor group from the power data on the grid side.

[0174] At present, the analysis of power data in the power system generally focuses on isolated analysis. For example, when analyzing a certain indicator, only a very small number of data directly related to this indicator are considered, thus ignoring the influence of some other factors, resulting in insufficient accuracy of the power analysis results, or the content being too single and the value of the analysis results being low. In this case, it is easy to cause ineffective waste of power resources. In order to build a digital power system based on multi-agent, this application proposes to coordinate the data characteristics on the power plant side and the data characteristics on the grid side. Specifically, it coordinates the characteristic production factor data on the power plant side and the control production factor data on the grid side, constructs a functional operator containing the mapping function of the relationship between the two based on this, and builds an agent accordingly. See S102 - S103 for details. This concept solves the existing problems of incomplete analysis and overly single data considered, can capture the connection between data and data in the power system more accurately and sensitively, and form data assets to meet diverse power function requirements. Thus, it can assist relevant personnel to perform scheduling and control in the power system more efficiently and maximize the utility of the digital power system.

[0175] Figure 1F This is a schematic diagram of a power supply scenario provided by an embodiment of this application. As Figure 1F shown, in practical applications, various different types of power plants may communicate with the grid side and have power transmission. These different types of power plants can be collectively referred to as the power plant side. In the embodiments of this application, the power plant side can report its own data to the grid side and give timely feedback; the grid side can dispatch the power plant side, such as peak shaving and frequency modulation dispatching, etc. Affected by various factors such as power generation cost, the performance of generator sets, the uncertainty of power generation of different types of power plants, and the change of load demand, the power plant side often needs to execute corresponding actions according to the instructions of the grid side after reaching a "consensus" with the grid side. Taking a thermal power plant as an example, the thermal power plant can further increase or release the heat stored in the thermal power plant according to the instructions of the grid side; or the thermal power plant can adjust the follow-up mode of the electric-motor-boiler according to the load requirements. Taking a wind power plant as an example, the fan can increase or decrease the load according to the instructions of the grid side and adjust the chamfer angle of the fan. Taking a photovoltaic power plant as an example, it can control the angle of the thyristor according to the load dispatching instructions of the grid side, and then change the power generation amount. Taking a hydropower plant as an example, it can calculate the water volume and water use cost according to the dispatching curve based on the grid side, and then control the water turbine power generation. Combining Figure 1FIt can be seen that in the actual power production process, close scheduling is required between the grid side and the power plant side to better and more stably achieve power supply and maintain the cost demands of both sides. Combining the above actual needs, a digital power system that collaboratively constructs a multi-functional intelligent body with multiple production factors is proposed in this application. This system has a relatively broad application prospect and has the potential to be applied in the power system for a long time. By collaborating and digitally analyzing multiple production factors, it can improve the accuracy and reliability of power analysis in the current power system. By collaborating multiple production factors, the data barriers between the power plant side and the grid side are broken through, and the artificial intelligence technology is used to capture the data connections between multiple production factors, so as to achieve digital power analysis in the field of power system in a more intelligent and automated manner.

[0176] For ease of understanding, the production factor division framework in the technical solution of this application is introduced below in combination with Figure 1G the following. Figure 1G FIG. 5 is a schematic diagram of a production factor division framework provided by an embodiment of this application. In the embodiment of this application, the production factors are divided according to the power plant side and the grid side respectively. The basis for dividing the production factors on the power plant side is called the characteristic production factor group; the basis for dividing the production factors on the grid side is called the control production factor group.

[0177] In the embodiment of this application, the characteristic production factor group includes multiple levels of characteristic production factors. Specifically, these characteristic production factors can be further divided into: primary characteristic production factors, secondary characteristic production factors, and tertiary characteristic production factors. Among them, the primary characteristic production factors are natural characteristic production factors (or called: natural indicators), which are production factors that can directly reflect natural characteristics. The secondary characteristic production factors are single-system characteristic production factors (or called: single-system indicators), which are production factors related to individual systems in the power system. The tertiary characteristic production factors are plant-system characteristic production factors (or called: plant-system indicators), which are production factors related to the overall system in the power system.

[0178] As shown in combination with Figure 1F FIG. 6, there are various types of power plants, such as thermal power plants, photovoltaic power plants, wind power plants, hydropower plants, etc. Their power generation methods are different, and the natural resources and internal system structures used are also different. In order to more accurately achieve digital power analysis, in the technical solution of this application, for various different power generation types, primary characteristic production factors, secondary characteristic production factors, and tertiary characteristic production factors can be further divided. Thus, for each type of power plant, data of the corresponding level of characteristic production factors can be extracted subsequently.

[0179] The following exemplarily introduces several primary feature production factors in combination with different power generation types. Taking photovoltaic power generation as an example, the primary feature production factors may include: photovoltaic irradiance. Taking hydropower generation as an example, the primary feature production factors may include: precipitation (inflow), etc. Taking thermal power generation as an example, the primary feature production factors may include: the lower calorific value of coal type, coal consumption, heat consumption, etc. Taking wind power generation as an example, the primary feature production factors may include: wind speed, etc.

[0180] For thermal power plants, a single system can be a boiler system, an electrical system, etc. For hydropower plants, a single system can be a turbine system, a speed regulation system, a butterfly valve layer system, etc. For photovoltaic power plants, a single system can be a photovoltaic panel area, a substation area, etc. Additionally, these different power plants also share a power transmission and transformation system. Correspondingly, the secondary feature production factors can be production factors reflecting a single system, such as boiler efficiency, steam turbine efficiency, etc. The tertiary feature production factors can be production factors reflecting the overall situation of multiple systems, such as photovoltaic efficiency, hydropower efficiency, thermal power unit efficiency, plant heat efficiency (specific heat index), etc. As can be seen from Figure 1G this, from the primary feature production factors, secondary feature production factors, to tertiary feature production factors, the level of mining for data features gradually deepens. It can be seen that in the embodiments of the present application, relying on the division framework of the feature production factors in the feature production factor group, data of natural features, single-system features, and plant-system features can be extracted collaboratively, and digital power analysis can be realized starting from features at multiple levels.

[0181] In the embodiments of the present application, the control production factor group includes control production factors in multiple aspects. Control production factors are production factors reflecting the performance of the power grid, which are composed of some key physical quantities in the power system. Control production factors help provide an effective theoretical basis for power system control. In the embodiments of the present application, for example, it may include but is not limited to the following six aspects of control production factors: (1) electricity price; (2) inertia and frequency; (3) active and reactive power; (4) carbon emissions from electricity; (5) safety and stability; (6) grid connection and disconnection. It should be noted that the above six aspects of control production factors exemplified are only defined from a macroscopic perspective. In specific implementation, each aspect of control production factors can be further subdivided or derived into more specific indicators, and no limitations are imposed on the quantity and content of control production factors here. In the embodiments of the present application, relying on the division framework of the control production factors in the control production factor group, data in multiple aspects such as electricity price, inertia and frequency, active and reactive power, carbon emissions from electricity, safety and stability, and grid connection and disconnection can be extracted collaboratively, and digital power analysis can be realized starting from multiple aspects concerned by the power grid.

[0182] As introduced above, the technical solution of this application pre-plans a characteristic production factor group and a control production factor group. Among them, the characteristic production factor group includes characteristic production factors at multiple levels, such as first-level characteristic production factors, second-level characteristic production factors, and third-level characteristic production factors for various different power generation types. The control production factor group includes control production factors in multiple aspects, such as electricity price, inertia and frequency, active and reactive power, carbon emissions from electricity, safety and stability, and grid connection and disconnection. Based on the above characteristic production factor group and control production factor group, in this step, production factor data is extracted by combining the power data on the power plant side and the grid side.

[0183] In an optional implementation manner, from the power data on the power plant side, characteristic production factor data corresponding to the characteristic production factor group is extracted, including: from the power data of a thermal power plant, first-level characteristic production factor data, second-level characteristic production factor data, and third-level characteristic production factor data corresponding to thermal power generation are extracted; from the power data of a photovoltaic power plant, first-level characteristic production factor data, second-level characteristic production factor data, and third-level characteristic production factor data corresponding to photovoltaic power generation are extracted; from the power data of a hydropower plant, first-level characteristic production factor data, second-level characteristic production factor data, and third-level characteristic production factor data corresponding to hydropower generation are extracted; from the power data of a wind power plant, first-level characteristic production factor data, second-level characteristic production factor data, and third-level characteristic production factor data corresponding to wind power generation are extracted.

[0184] In an optional implementation manner, from the power data on the grid side, control production factor data corresponding to the control-related production factor group is extracted to form a data vector to represent a specific function, such as power system stability, etc., including: from the power data on the grid side, electricity price data, inertia and frequency data, power data, carbon emissions data from electricity, safety and stability index data, and grid connection and disconnection impact data are extracted.

[0185] As an example, the electricity price data may include electricity prices at different times and different load demands.

[0186] Inertia and frequency data can include system inertia and frequency. The relationship between system inertia and frequency is an important aspect in power system stability analysis. System inertia generally refers to the moment of inertia of generator sets in a power system, which reflects the system's resistance to frequency changes. The following are some key points that outline the connection between system inertia and frequency. Inertia support function: When large-scale load changes or sudden addition or withdrawal of generator sets occur in a power system, the system inertia can slow down the rate of change of the system frequency. This inertia support function helps maintain the frequency stability of the power system. Rate of change of frequency: The magnitude of system inertia directly affects the rate of change of frequency. The larger the inertia, for a given power imbalance, the smaller the rate of change of the system frequency. A smaller rate of change of frequency helps avoid under-frequency load shedding or other stability control measures triggered by too rapid frequency changes.

[0187] Voltage and active / reactive power data can include power at different voltages, such as in the form of (voltage, active power, reactive power). In a power system, voltage and frequency have no direct relationship, but they are related to power (active and reactive respectively), resulting in an indirect relationship. When reactive power is insufficient, the voltage drops; when reactive power is excessive, the voltage rises; when active power is insufficient, the frequency drops; when active power is excessive, the frequency rises.

[0188] Electric carbon emission data can include carbon footprint, such as the carbon dioxide emissions per kilowatt-hour of electricity generated.

[0189] Safety and stability index data can include the mean time between failures (MTBF) of the power grid. MTBF is an important indicator for measuring the reliability of products (especially electronic products), which represents the average working time of a product between two adjacent failures under specified conditions and within a specified time range.

[0190] Power grid connection and disconnection image data can include the degree of influence of each network on other networks.

[0191] In specific implementation, after extracting the control production factor data, it can be further converted into a digital vector representation to facilitate network recognition, operation, and processing. The obtained digital vector representation can contain data on control production factors in one aspect or multiple aspects. The dimension and data source of the digital vector representation can be set according to requirements and are not limited here.

[0192] S102. Based on the power function requirements of the digital power system, construct a target data vector using the extracted characteristic production factor data, and determine the key control production factor data associated with the power function requirements from the extracted control production factor data.

[0193] In an embodiment of the present application, an agent is constructed based on the power function requirements and in combination with the characteristic production factor data and control production factor data extracted in the previous step. It can be understood that in the previous step, production factor data at multiple levels and in multiple aspects was extracted. However, some of this data may be highly relevant to the power function requirements, and this data helps to construct an agent capable of meeting specific power function requirements. However, there may also be a part of the extracted data that has a very weak correlation with the power function requirements. For example, if the power function requirements are related to electricity prices, the carbon emission data in the data is currently of little significance for the calculation and analysis of electricity prices. Therefore, to avoid the negative and adverse effects of redundant and massive data on agent construction, the present application extracts and refines the obtained massive data, so as to conveniently and accurately capture the connections between the data. As the correlation of data features in reinforcement learning is strengthened in the future, and the carbon emission data gradually shows an increasing trend in data-drivenness, the convolutional neural network algorithm will automatically extract this data feature and include it in the electricity price operator, which also reflects the mechanism of future deduction and emergence capabilities.

[0194] In an embodiment of the present application, based on the extracted characteristic production factor data of each feature, and based on the specific requirements reflected in the power function requirements and the connection between the specific requirements and the characteristic production factors, one or more characteristic production factor data with a relatively high correlation with the specific requirements are further extracted from the extracted characteristic production factor data of each feature (as the target characteristic production factors), and then a data vector is constructed. For the convenience of distinction, this data vector is called the target data vector. When constructing the target data vector, the vector can be constructed based on certain construction rules. For example, the dimension of the vector and the level or name of the characteristic production factor corresponding to each dimension can be set in the construction rules.

[0195] Table 1 shows some of the characteristic production factor data extracted from the power data of a thermal power plant. In Table 1, SHRW represents the weighted specific heat index, with the unit kJ / (kW·h). Ea represents the power supply (or: generated electricity), with the unit kW·h. EP represents the electricity bill, with the unit US dollar. SHRcc is the specific heat index, with the unit kJ / (kW·h). U is the coal price, with the unit USD / t. As an example, U = 69 US dollars per ton. E represents the heat consumption rate of the unit for power generation, that is, the heat consumption per degree of electricity of the unit, with the unit kcal / kWh. N represents the specific heat index, specifically the heat required per kilowatt-hour of electricity, with the unit kcal / kWh, and AUX represents the auxiliary power rate. ECRm is the electricity charge per kilowatt-hour, with the unit USD / (kW·h). ECRm represents the electricity charge ratio, that is, the coal consumption cost per degree of electricity produced, with the unit USD / (kW·h).

[0196] Table 1

[0197]

[0198] If the specific requirement reflected in the power function requirement is to calculate the electricity price, then multiple target data vectors can be extracted according to the data of each characteristic production factor shown in Table 1. For example, each row of data in Table 1 is extracted to construct a target data vector with a structure of (operating condition, SHRW, ECRm, Ea, CERm, EP, actual coal consumption per degree of electricity, actual coal cost per degree of electricity, overall coal cost). For example, using the second row of data in Table 1 to construct target data vector 1 as (100%, 2138.4, 0.0293, 9267053600, 0.0311, 287772056.8, 450.0942, 0.0311, 287802236), and using the third row of data in Table 1 to construct target data vector 2 as (95%, 2152.69, 0.0295, 8803700920, 0.0311, 273383454, 453.5357, 0.0313, 275502693). And so on.

[0199] In addition, in the specific implementation, the corresponding relationship between the control production factors and the power function requirements can be pre-constructed. Furthermore, according to the specific power function requirements, the corresponding control production factors can be determined in a timely and efficient manner as the key control production factors, and based on this, the key control production factor data associated with the power function requirements can be determined from the extracted massive control production factor data. For the extracted key control production factor data, similarly, it can also be transformed into a digital vector representation.

[0200] S103. Based on the target data vector and the key control production factor data, use an artificial intelligence algorithm to learn the mapping function between the target data vector and the key control production factor data, construct a functional operator with the mapping function as the kernel, and construct an intelligent agent based on the functional operator.

[0201] Figure 1H Schematic diagram of the relationship between an intelligent agent, an operator, a function, and production factor data provided by an embodiment of the present application. Combined with Figure 1H The implementation process of this step will be described. As Figure 1HAs shown, in the embodiments of the present application, based on the characteristic production factor data, the target data vector is obtained, and based on the control production factor data, the key control production factor data is determined. Next, it is necessary to rely on the artificial intelligence algorithm to learn based on the target data vector and the key control production factor data, and infer and analyze the potential data connection between the characteristic production factor data and the key control production factor data behind the target data vector. Through the application of the artificial intelligence algorithm, the connection between the two can be expressed through a mapping function.

[0202] In the embodiments of the present application, it is proposed to construct a functional operator with the mapping function as the kernel. In Figure 1H the example of Agent 1, it contains multiple functional operators, namely: Operator 1, Operator 2, and Operator 3. Each functional operator loads the mapping function. It can be understood that the mapping functions loaded in different functional operators are different because the power function requirements for driving the construction of each functional operator may be in different aspects, and thus the characteristic production factor data and control production factor data used may also be different. By executing this step, the constructed agent corresponds to the aforementioned power function requirements. For example, according to Power Function Requirement 1, Agent 1 is finally constructed; according to Power Function Requirement 2, Agent 2 is finally constructed.

[0203] In the embodiments of the present application, the artificial intelligence algorithms used may be diverse. For example, all the open-source artificial intelligence algorithms currently on the market can be based on the above technical concept to infer and analyze the internal connection between the two based on the target data vector and the key control production factor data, and then obtain an accurate mapping function. The specific type of the artificial intelligence algorithm used here is not limited.

[0204] Combined with Figure 1H it can be seen that in the embodiments of the present application, an agent can contain one or more functional operators. Taking Agent 1 as an example, it contains 3 functional operators. The functional operators within the same agent can be independent of each other. In addition, the functional operators within the same agent can also have a technical connection or a logical connection.

[0205] In Figure 1H it also shows other agents besides Agent 1, such as Figure 1HAgents 2, Agent N, etc. shown in [the figure]. After these agents are successively constructed, an automated device integrating the wisdom of multiple agents is formed. Therefore, this device can also be regarded as a "robot". It has rich knowledge related to the power system and can solve various power system analysis problems by applying the functional operators in each agent. With the increasingly extensive international cooperation in the power industry, the construction method of the agents and the architecture of the "robot" proposed in the technical solution of this application are exactly breakthrough technologies for coping with international power cooperation.

[0206] S104. Add agents corresponding to various different power function requirements to the digital power system, and use each agent as a power function implementation unit in the digital power system.

[0207] In practical applications, the power function requirement can be expressed through a power analysis request. The power analysis request can be initiated by the power plant side, the power grid side, or a third party other than the power plant side and the power grid side. As long as the relevant entity or device initiating the power analysis request has the permission or qualification to start the data analysis of the power analysis agent, the execution device (such as a server or a terminal) of the method for creating a digital power system based on a multi-functional agent proposed in this solution can respond to the power analysis request and analyze the matching degree between the power function requirement carried in the request and one or more agents that have been constructed. If no matching agent is found, it is necessary to call the steps in the above method to establish an agent or establish a corresponding operator for this power function requirement.

[0208] It can be understood that if it is analyzed and determined that the power function requirement carried in the power analysis request matches a certain agent that has been constructed, then call this agent and use this agent as a power function implementation unit to implement the response to the power analysis request through the functional operators inside the agent. For example, if the requirement information carried in the power analysis request indicates calculating the electricity price, then call the electricity price operator (whose function is to calculate the electricity price) in the power analysis agent and perform the electricity price calculation based on the relevant data necessary for the current calculation provided to this operator.

[0209] In the technical solution of the present application, since the construction of the intelligent body depends on the relevant production factor data on the power plant side and the power grid side, and the characteristic production factor group includes characteristic production factors of multiple levels, and the control production factor group includes control production factors of multiple aspects, it is equivalent to realizing the coordination of characteristic production factors of multiple levels and control production factors of multiple aspects. The intelligent body constructed based on this effectively utilizes the diversified information of power data to realize the power function of coordinating multiple production factors, such as the power function for power trading between the power grid side and the power plant side, the power function for realizing the dispatching control of the power grid side on the power plant side, and the power function for the power grid to regulate itself. The intelligent body can more accurately capture the connection between data and data, and then obtain more accurate and transferable data analysis results or data calculation results. It can also be understood that the digital power system constructed by this solution forms high-value data assets in the power system and enhances the analytical value of power data. The constructed digital power system can effectively contribute to the domestic power market reform and global power trading.

[0210] In practical applications, it is possible that a single operator cannot meet a relatively large analysis goal. In this case, a possible solution is to call multiple operators to meet an analysis goal. This type of situation is introduced below.

[0211] To deal with the above problems, the power analysis demand information can be parsed in the initial stage of constructing the intelligent agent to obtain multiple sub-demand information and the logical dependency relationship or digital conversion relationship between the multiple sub-demand information. By parsing and disassembling the power analysis demand information, functional operators can be built according to the sub-demand information in a targeted manner. Assume that by parsing the power analysis demand information, multiple sub-demand information is obtained, including the first sub-demand information and the second sub-demand information.

[0212] Figure 1I This is a flowchart for implementing a power analysis intelligent entity provided in an embodiment of the present application. Figure 1I As shown, in a specific implementation, the method for constructing a digital power system based on a multifunctional intelligent agent provided in an embodiment of the present application further includes:

[0213] A first target data vector is constructed based on the first sub-demand information and a first key control production factor data associated with the first sub-demand information is determined; and a second target data vector is constructed based on the second sub-demand information and a second key control production factor data associated with the second sub-demand information is determined.

[0214] for Figure 1IIn the specific example shown, based on the target data vector and the key control production factor data, an artificial intelligence algorithm is used to learn the mapping function between the target data vector and the key control production factor data, a functional operator with the mapping function as the kernel is constructed, and a power analysis intelligent agent is constructed based on multiple functional operators. Specifically, it may include:

[0215] Based on the first target data vector and the first key control production factor data, an artificial intelligence algorithm is used to learn the first operator; and, based on the second target data vector and the second key control production factor data, an artificial intelligence algorithm is used to learn the second operator; based on the logical dependency relationship or the numerical conversion relationship, the transfer function relationship between the first operator and the second operator is constructed; based on the transfer function relationship, the first operator and the second operator, a power analysis intelligent agent is constructed.

[0216] So far, a multi-operator power analysis intelligent agent including the first operator and the second operator has been constructed. Based on this, when a power analysis request is received and the demand information carried therein also matches the first sub-demand information and the second sub-demand information, the power analysis intelligent agent can be called to trigger the first operator and the second operator to work based on the transfer function relationship to generate the final analysis response result for the power analysis request. In this way, the collaborative work of multiple functional operators in the power analysis intelligent agent is realized. By constructing multiple functional operators with logical dependency relationships or numerical conversion relationships and constructing a power analysis intelligent agent accordingly, the power analysis intelligent agent can handle more complex and diverse power data analysis work.

[0217] As mentioned above, based on the logical dependency relationship or the numerical conversion relationship, the transfer function relationship between the first operator and the second operator is constructed. Therefore, in the intelligent agent, there are also corresponding logical dependency relationships or numerical conversion relationships between different operators. The following provides several dependency relationships existing in the power system: reactive power and voltage, heat consumption and coal consumption, system inertia and frequency, electricity price and coal price, etc.

[0218] For example, the numerical conversion relationship between different operators may mean that the functions of different operators are basically the same, but there are differences at the transaction level or the pricing level. In order to meet the intercommunication requirements of transactions or pricing, it is necessary to call the operator to perform digital conversion at the transaction level or the pricing level. As an example, operator A calculates the electricity price of the power grid in country A, and operator B calculates the electricity price of the power grid in country B. When it is necessary to perform electricity price conversion for the electricity prices of country A and country B, it is necessary to operate according to the digital conversion relationship.

[0219] In the embodiments of the present application, each agent is configured with multiple types of interfaces to complete communication or data exchange with other entities. As an example, the agent is configured with a power plant side interface, a power grid side interface, and an agent interface. Among them, the power plant side interface is used for the agent to dock with the power plant; the power grid side interface is used for the agent to dock with the power grid; the agent interface is used for the agent to dock with other agents.

[0220] In a possible implementation manner, the agent may be specifically configured with multiple power plant side interfaces, and different power plant side interfaces are respectively used to dock with different types of power plants, or different power plants of the same type. The present application does not limit the number of power plant side interfaces configured for an agent.

[0221] In a possible implementation manner, the agent may be specifically configured with multiple agent interfaces, and different agent interfaces are respectively used for the agent to dock with different other agents. The present application does not limit the number of agent interfaces configured for an agent.

[0222] By invoking the agent, an analysis response result for the power analysis request can be generated. Based on the above introduction of the interfaces configured on the agent, in the technical solution of the present application, the agent can send the analysis response result to the power plant through the power plant side interface; the agent can send the analysis response result to the power grid through the power grid side interface; the agent can send the analysis response result to other agents through the agent interface. It can be seen that through the configuration of the above multiple types of interfaces, the agent has a way to communicate and transmit the analysis results with various external objects through the interfaces.

[0223] In addition, in the embodiments of the present application, there may also be a scenario where, after the agent obtains the analysis response result, it needs to send a control instruction to the outside world. With the configuration of multiple types of interfaces, the agent can transmit the control instruction. In an exemplary implementation manner, the agent generates a first control instruction for the power plant, a second control instruction for the power grid, or a third control instruction for other agents based on the analysis response result. It should be noted that the above first control instruction, second control instruction, and third control instruction are generated according to specific control requirements. The agent can send the first control instruction to the power plant through the power plant side interface; or send the second control instruction to the power grid through the power grid side interface; or send the third control instruction to other agents through the agent interface.

[0224] Figure 1J This is another schematic diagram showing the relationship between an agent, an operator, a function, and production factor data provided by the embodiments of the present application. In Figure 1J On the left side, objects at the same level as concepts such as data, function, operator, and agent in the technical solution of the present application are shown from a biological perspective; andFigure 1J On the right side, the concepts of data, functions, operators, intelligent agents, etc. mentioned in the digital power analysis method for coordinating multiple production factors proposed in this application are displayed. Figure 1J The left and right sides are connected by an arrow marked with the word "symbiosis", which reflects the similarity of the roles or functions of the concepts or objects on the left and right sides.

[0225] Combination Figure 1J As shown on the left, the construction of proteins depends on mitochondria, the construction of cells depends on proteins, and the operation of physiological functions such as human joints (or organs) depends on cells. A complete human body cannot be separated from diverse and multifunctional joints and organs. It is understandable that mitochondria provide key biological characteristics, such as the genetic material deoxyribonucleic acid (DNA) in mitochondria. DNA carries the genetic information necessary for the synthesis of RNA and proteins and is an essential biological macromolecule for the development and normal operation of organisms. The performance of human joints (or organs) is also inseparable from the genetic information carried by DNA. Figure 1J The data, functions, operators, agents and robots shown on the right also have a similar progressive relationship. The formation of the mapping function is inseparable from the extraction of characteristic production factor data and the extraction of control production factor data. The mapping function depends on these data, the operator uses the function as the core, and the agent contains one or more operators to realize the analysis function of the power system data. The robot finally formed has the ability to handle complex and diverse power analysis needs because of its diverse agents. And the foundation of all this depends on Figure 1G The production factor division framework is shown.

[0226] The above text describes in detail the construction process of the construction method of the digital power system based on the multifunctional intelligent agent. In the technical solution of the present application, a plurality of intelligent agents corresponding to different power function requirements are added to the constructed digital power system. In order to facilitate the understanding of the responsiveness of the constructed digital power system to the diversified power function requirements, the present application also combines the following six embodiments to explain the use of operators to realize power functions.

[0227] In the technical solution of this application, the power function requirement can be: the power function requirement for power trading between the power grid and the power plant, the power function requirement for realizing the dispatching control of the power grid on the power plant, or the power function requirement for the power grid to regulate itself. In the technical solution described below in this application:

[0228] In the first embodiment, the power function requirement for power trading between the grid side and the power plant side is specifically to calculate the electricity price. The functional operator is the electricity price operator, and the electricity price operator is used to calculate the electricity price. The first embodiment introduces the process of the digital power system using an agent constructed based on the electricity price operator. By using the agent constructed by the electricity price operator, a method for calculating the electricity price is realized. Combining the following first embodiment and Figures 2A to 2C description.

[0229] In the second embodiment, the power function requirement for power trading between the grid side and the power plant side is specifically power clearing. The functional operator is the clearing operator, and the clearing operator is used for power clearing. The second embodiment introduces the process of the digital power system using an agent constructed based on the clearing operator. The difference between the electricity price operator and the clearing operator is that the clearing operator includes a part for calculating the electricity price, but after calculating the electricity price, it will implement the function of the clearing curve according to the priority or the priority of the stack, making the function of the clearing operator function more complex. By using the agent constructed by the clearing operator, a method for grid clearing is realized. Combining the following second embodiment, Figures 3A to 3I and Figures 4A to 4H description.

[0230] In the third embodiment, the power function requirement for realizing the dispatching control of the grid side over the power plant side is specifically peak shaving dispatching in the scenario of high-frequency load variation of thermal power. The functional operator is the thermal power peak shaving dispatching operator, and the thermal power peak shaving dispatching operator is used to realize peak shaving dispatching in the scenario of high-frequency load variation of thermal power. The third embodiment introduces the process of the digital power system using an agent constructed based on the thermal power peak shaving dispatching operator. By using the agent constructed by the thermal power peak shaving dispatching operator, a method for peak shaving dispatching in the scenario of high-frequency load variation of thermal power is realized. Combining the following third embodiment and Figures 5A to 5F description.

[0231] In the fourth embodiment, the power function requirement for realizing the dispatching control of the grid side over the power plant side is specifically the water resource dispatching of cascade hydropower stations. The functional operator is the water resource dispatching operator, and the water resource dispatching operator is used to realize the water resource dispatching of cascade hydropower stations. The fourth embodiment introduces the process of the digital power system using an agent constructed based on the water resource dispatching operator. By using the agent constructed by the water resource dispatching operator, a method for water resource dispatching is realized. Combining the following fourth embodiment and Figures 6A to 6J description.

[0232] In Embodiment 5, the power function requirement for realizing the dispatching control of the grid side over the power plant side is specifically the dispatching of the power grid over the power plant in a large - machine and small - grid power system. The functional operator is the large - machine and small - grid dispatching operator, which is used to realize the dispatching of the power grid over the power plant in the large - machine and small - grid power system. Embodiment 5 introduces the process of the digital power system using an agent constructed based on the large - machine and small - grid dispatching operator. By using the agent constructed by the large - machine and small - grid dispatching operator, a power grid dispatching method in a large - machine and small - grid power system is realized. Combined with Embodiment 5 below and Figures 7A to 7F description.

[0233] In Embodiment 6, the power function requirement for the self - regulation of the power grid is specifically to adjust the grid structure of the power grid. The functional operator is the grid - structure adjustment operator, which is used to realize the adjustment of the grid structure of the power grid. Embodiment 6 introduces the process of the digital power system using an agent constructed based on the grid - structure adjustment operator. By using the agent constructed by the grid - structure adjustment operator, a method for adjusting the grid structure of the power grid is realized. Combined with Embodiment 6 below and Figures 8A to 8D description.

[0234] Example 1

[0235] In Embodiment 1, the power function requirement for power trading between the grid side and the power plant side is specifically to calculate the electricity price. The functional operator is the electricity - price operator, which is used to calculate the electricity price. Embodiment 1 introduces the process of the digital power system using an agent constructed based on the electricity - price operator.

[0236] In practical applications, the power plant side needs to calculate the electricity price and then provide the electricity price to the power plant side. Subsequently, if the power plant side and the grid side reach an agreement on the electricity price, power trading can be carried out.

[0237] Next, combined with Figure 2A and Figure 2B From the perspectives of the grid side and the power plant side respectively, the grid - side process and the power - plant - side process based on the electricity - price operator are introduced.

[0238] Such as Figure 2AAs shown, the figure shows the grid-side process based on the electricity price operator. First, it is necessary to extract the key control production factor data related to electricity price calculation from the grid-side power data. Optionally, these key control production factor data can be represented as digital vectors. Then, based on the target data vectors obtained by feature extraction from the power data on the power plant side, it can be understood that such data vectors are also related to electricity price calculation. In an example, the data vector A related to the performance efficiency of Unit A is represented as (coal consumption 1, heat rate 1, unit efficiency 1), and the data vector B related to the performance efficiency of Unit B is (coal consumption 2, heat rate 2, unit efficiency 2). For these data vectors, alignment is performed, and the comprehensive production factor data vector (coal consumption 1 + coal consumption 2, heat rate 1 + heat rate 2, unit efficiency 1 + unit efficiency 2) is statistically obtained. Next, based on the data vectors extracted from the power plant side and the grid side above, an operator function computing power group is obtained by using artificial intelligence algorithms. The operator function computing power group contains many functional operators with functions as the core. Having these operator function computing power groups can also be understood as constructing a power analysis intelligent agent.

[0239] Taking the electricity price calculation as an example, the operator function computing power group constructed can be called the grid load and electricity price function computing power group. In addition, considering other possible power analysis requirements, a grid inertia frequency computing power group, an active and reactive power function computing power group, a grid carbon footprint computing power group, a grid voltage and current function computing power group, a grid security and stability computing power group, a grid connection and disconnection computing power group, etc. can also be established. Among them, the active and reactive power function computing power group includes functions for reflecting the performance of active and reactive power in the grid; the grid inertia function computing power group includes functions for reflecting the performance of the stable frequency transformation of the grid; the grid load and electricity price function computing power group includes functions for calculating the correlation between the grid electricity price and the power generation load of the power plant; the grid security and stability computing power group includes functions for calculating the grid step and power flow; the grid carbon footprint computing power group includes functions for calculating the carbon consumption in the production of equipment such as hydropower, thermal power, photovoltaic power, and wind power in the grid.

[0240] Taking the grid load and electricity price function computing power group as an example, the mapping function as the core of the electricity price operator can be expressed as: CERm = U * N / E(1 - AUX). CERm is the grid compensation coefficient for paying for 1 kWh of electricity. U represents the coal price, which is also the fuel cost subsidy, in US dollars per ton. E represents the unit power generation heat rate of the unit, that is, the heat consumption per kWh of electricity of the unit, in kcal / kWh. N represents the specific heat index, in kcal / kWh. AUX represents the plant electricity rate. This formula reflects the relationship between the grid compensation coefficient for paying for 1 kWh of electricity and characteristic production factors such as coal price, unit power generation heat consumption, specific heat index, and plant efficiency.

[0241] After obtaining the electricity price operator by applying artificial intelligence algorithms, when the agent containing this operator is called next, the electricity price calculation result on the power plant side is obtained through this electricity price operator. It should be noted that since the construction of this operator depends on the power data on the power plant side, therefore, this electricity price calculation result actually reflects more of the cost considerations and profit requirements of the power plant side itself. Subsequently, the grid side still needs to further consider this electricity price calculation result.

[0242] As Figure 2A shown, after obtaining the electricity price calculation result on the power plant side through the electricity price operator, the electricity price calculation result is verified through the financial verification model on the grid side. The main purpose of verifying the electricity price calculation result through the financial verification model on the grid side is to ensure the balance and stability of the power plant and the grid. In one example, the financial verification model on the grid side includes: the first verification condition and the second verification condition; the first verification condition is: the actual output of various different types of power plants ≥ the grid electricity quantity; the second verification condition is: the sum of the products of the actual output of various different types of power plants, the grid electricity price at the corresponding moment, and the power generation duration ≤ the product of the grid average price and the grid electricity quantity. If the electricity price calculation result meets the first verification condition and the second verification condition, it is determined that the electricity price calculation result passes the verification; if the electricity price calculation result does not meet any one of the first verification condition and the second verification condition, it is determined that the electricity price calculation result fails the verification.

[0243] The expression of the first verification condition is:

[0244] Q 火电 +M 风电 +Q 水电 +M 光伏 ≥N 电量

[0245] The expression of the second verification condition is:

[0246] Σ(H*Q 火电 *N 各时刻电网电价 )+ΣM 风电 *N 各时刻电网电价 +Σ(H*Q 水电 *N 各时刻电网电价 )+ΣM 光伏 *N 各时刻电网电价 ≤P 电网均价 *N 电量

[0247] In the above expressions, Q 火电 、M 风电 、Q 水电 、M 光伏 are the actual outputs of thermal power plants, wind power plants, hydropower plants, and photovoltaic power plants respectively. N 电量 is the grid electricity quantity. H represents the power generation duration. N 各时刻电网电价is the grid electricity price at the corresponding moment. P 电网均价 represents the average grid price.

[0248] For the calculation, both thermal power and hydropower are physical power generations with controllable resources, which are deterministic; while for photovoltaic and wind power, the scenery cannot be controlled, and the natural elements are highly uncertain. Here, different expressions are used for the processing, and Q and M are used for distinction. Thus, it represents the difference in data abstraction of the two types of production factors. The production factors of thermal and hydropower generation have strong certainty and weak data abstraction, which is represented by Q; while the production factors of wind and photovoltaic power generation have strong uncertainty and strong data abstraction, and a higher-order data operator function M is required to represent them.

[0249] If the verification of the electricity price calculation result is passed, then further select the electricity price calculation function that is adapted to and feasible for the grid side from the dispatching strategy library. The dispatching strategy library can be responsible for by a special operator, and this operator provides diversified calculation functions. As mentioned before, the construction of the already constructed electricity price operator depends on the power data on the power plant side. Therefore, this electricity price calculation result actually reflects more of the cost considerations and profit demands of the power plant side itself. Subsequently, the grid side still needs to further consider this electricity price calculation result. The specific consideration method is to use the dispatching strategy library. The grid side also needs to select an electricity price calculation function that is adapted to the grid side and technically feasible from the dispatching strategy library, and then load the electricity price calculation function into the electricity price operator to update the function of the electricity price operator; obtain the electricity price calculation result of the grid side through the updated electricity price operator. In this way, the grid side not only refers to the electricity price calculation result of the power plant side, but also updates the electricity price operator based on its own cost and profit demands, and realizes the collaborative calculation of the electricity price between the power plant side and the grid side in this process.

[0250] If an electricity price calculation function that is adapted to the grid side and technically feasible cannot be selected from the dispatching strategy library, then it is necessary to deeply learn the grid energy consumption average efficiency operator, the average electricity price operator and other important production factor operators, return the learning result as a sample, and use artificial intelligence algorithms to learn the electricity price operator.

[0251] In an alternative implementation, after loading the electricity price calculation function into the electricity price operator to update the function of the electricity price operator, the collaborative multi-production factor digital power analysis method provided by the embodiments of the present application further includes:

[0252] The grid side collects the applied transaction power and electricity price reported by the power plant side to the grid side. It is ready to conduct transactions between the power plant side and the grid side. At this time, the grid side completes the electricity price calculation function for the power plant side. Then, based on the judgment of the grid side on whether the grid is stable, it is determined whether to activate the virtual power plant function. The purpose of activating the virtual power plant function is to call the storage capacity, then give the cleared power generation of each power generation according to the principle of optimal cost, and finally issue an instruction in advance according to the artificial intelligence calculation. The grid calls the storage capacity of the virtual power plant. When there is a power shortage, the electricity price is high, and the virtual power plant (battery) is called to discharge; when there is a power surplus, the electricity price is low, and the virtual power plant (battery) is charged.

[0253] In the technical solution of this application, the grid side can record the respective clearing curves and dispatching curves under different loads or in emergency situations as the input data for deep learning, continuously verify, and learn and imitate the laws and characteristics of this grid as the reference basis for the next instruction issuance.

[0254] If it is determined that there is no need to activate the virtual power plant function, the grid side sends dispatching instructions for the clearing electricity price and electricity volume to various types of power plants based on the electricity price calculation result on the grid side. If it is determined that the virtual power plant function needs to be activated, the virtual power plant function is activated to call the storage capacity. And return by combining the operator of the virtual power plant function, and perform the steps of obtaining the electricity price operator using the artificial intelligence algorithm.

[0255] As Figure 2B shown, this figure shows the power plant side process based on the electricity price operator. First, collect various characteristic production factor data on the power plant side to form a data vector. Then combine the key control production factor data on the grid side to train the electricity price operator. The power plant side receives the clearing electricity price given by the grid side; combine the power plant cost and the clearing electricity price, and use the financial verification model on the power plant side for verification. In an optional implementation manner, the financial verification model on the power plant side includes: a third verification condition; the third verification condition is: the sum of the products of the actual output of the power plant, the grid electricity price at the corresponding moment, and the power generation duration > the product of the power plant cost and the actual output of the power plant. Taking a thermal power plant as an example, the expression of the third verification condition is:

[0256] Σ(H*Q 火电 *N 各时刻电网电价 )>Σ(H*Q 火电 *P 电厂成本 )

[0257] In the above expression of the third verification condition, Q 火电 is the actual output of the thermal power plant. H represents the power generation duration. N 各时刻电网电价 is the grid electricity price at the corresponding moment. Apply artificial intelligence to use the grid electricity price N 各时刻电网电价 at each moment as the data input, and at the same time apply the power plant cost P 电厂成本 measured by this power plant system under different loadsSubstitute it into the formula. The expression of this verification condition reflects that the cumulative operation of the actual output of thermal power generation production factors is the electricity bill settled by the power grid. It fluctuates with the power grid, but for a single power plant, the cost production factors vary with the load factor, and finally map to the cost production factors of the power plant.

[0258] The same type of power plant may have multiple different power sources. For example, Figure 2B As shown in the figure, if the verification result indicates that the verification is passed, it means that it generally meets the verification conditions. Then, it is necessary to further determine whether the different power sources of the same type of power plant can be cleared based on the electricity quantity and electricity price. If each power source can be cleared, the cleared electricity quantity is executed. Subsequently, the actions corresponding to the adjusted load of the thermal power plant unit are taken. If there is a situation where a power source cannot be cleared, it is also necessary to calculate the production cost of the power plant or adjust the cleared electricity price of the power plant, and then judge again whether each power source can be cleared after verification.

[0259] The above first verification condition, second verification condition, and third verification condition are mathematical logics, all of which express some physical constraints in the power system. Through these verification conditions, the power system is made stable.

[0260] Figure 2C It is a signaling diagram between the power plant side, the power grid side, and the agent. In Figure 2C The significance of peak shaving in the power system is as follows: within the electricity price range specified by the power grid, find the power plants suitable for peak shaving electricity prices, so that the revenue and expenditure of the power grid are balanced and within the budget. At the same time, estimate the cleared electricity quantity and price of various types of power plants, and clear them from two buses or algorithms, that is, the technical clearing points of the electricity quantity that each power plant can provide at the basic electricity price and the clearing curve. After the total electricity price of the power grid is allocated to various types of power plants, the minimum value that satisfies the above inequality is the electricity price and electricity quantity clearing point of the power plant.

[0261] When digital power analysis is required, first digitize the measured data. Secondly, (based on the power data on the power plant side to be analyzed and / or the power data on the power grid side to be analyzed), perform data cleaning and form data vectors for subsequent parallel computing. Generalize the output result of the computing into an operator (function) based on the data model. This operator (function) completes specific functions. After the agent uses the above operator to complete the analysis, the entire agent logic process is completed. Use the established mapping function to output and generate power analysis results, or bring the operator output into a new agent nested iteration. The operator is the most important part of the agent, realizing specific functions of the power system, such as frequency operator, active and reactive power operator, electricity price operator. The electricity price operator is: The operator is a function. The agent is a process or logic that contains many operators and realizes many functions.

[0262] After the production factor data is digitized, data parallel computing is performed. Due to the characteristics of the GPU, after the same type of data vectors are aligned, the data can be quickly operated in parallel. At the same time, selecting the neural network algorithm will lose some characteristics of its symbolic artificial intelligence algorithm. However, the data tool of the graph neural network is sufficient to retain and operate the data features in the input and transmission layers under the self-attention mechanism in the probability method, reflecting the characteristics of data-driven. To make up for the loss of symbolic artificial intelligence and to enable more intelligent agent functional computing and control, the concept of an operator is introduced at the output. Multiple operators are put into the intelligent agent logic, enabling the intelligent agent to obtain corresponding functions. The digital power analysis capabilities also revolve around the specific capabilities of these operators, making the intelligent agent the basic unit for the power system to perform power data analysis.

[0263] Example 2

[0264] In Embodiment 2, the power function requirement for power trading between the grid side and the power plant side is specifically power clearing, and the functional operator is the clearing operator, which is used for power clearing. Embodiment 2 introduces the process of the digital power system using an intelligent agent constructed based on the clearing operator.

[0265] In the current power grid clearing process, it is often necessary to involve the electricity price prediction for different power plants and the power distribution process. However, in actual application scenarios, the prediction of power plant electricity prices and the power distribution process are often affected by various factors, resulting in poor accuracy of power grid clearing. To solve the above problems, the embodiments of the present application implement power clearing through a clearing operator, which can also be understood as implementing power grid clearing. First, perform low-order data abstraction processing based on the power generation element characteristics of each associated power plant, and quantify various factors affecting power grid clearing into a first data vector, thereby reducing the data complexity of power generation element characteristics. Subsequently, based on the first data vector and a preset long-term electricity price function library, predict the long-term trading electricity price to predict the corresponding long-term trading electricity price according to the actual power generation element characteristics of each associated power plant, and establish a corresponding long-term trading stack, and improve the accuracy of power grid clearing by ensuring the accuracy of the long-term trading electricity price. Further, establish a spot trading stack based on the spot trading electricity prices of each associated power plant, and determine a first clearing function through the long-term trading stack and the spot trading stack. Correspondingly, according to the determined first clearing function, determine the power supply and demand balance state when the target power grid clears under the first clearing function, and perform real-time function adjustment on the first clearing function according to the actual power supply and demand balance state of the target power grid to obtain a second clearing function. In this way, it is possible to dynamically adjust the clearing function of the target power grid according to the supply and demand state of the target power grid, establish a specific clearing function through the trading stack constructed in real time by the long-term trading electricity price and the spot trading electricity price, so that the power grid clearing can effectively consider the impact of power generation element characteristics on the power generation of associated power plants and the changes in the spot trading electricity prices of each associated power plant, and then determine a more accurate clearing function, thereby achieving the effect of improving the accuracy of power grid clearing.

[0266] Next, a power grid clearing method provided by the embodiments of the present application will be introduced in combination with specific embodiment drawings.

[0267] See Figure 3A and Figure 3B , Figure 3A is a schematic flowchart of a power grid clearing method provided by the embodiments of the present application, Figure 3B is a signaling interaction diagram of a power grid clearing method provided by the embodiments of the present application. In Figure 3A , the power grid clearing method is implemented on the target power grid side, while in Figure 3B , the overall implementation of the solution is introduced through the interaction signals between the associated power plant side and the target power grid side.

[0268] As Figure 3A shown, the power grid clearing method specifically includes the following steps:

[0269] S301: Perform low-order data abstraction processing based on the power generation element characteristics of each of the associated power plants to obtain the first data vectors of the power generation element characteristics of each of the associated power plants.

[0270] The power generation element characteristics of the associated power plants are data of various natural elements and production elements that affect the power generation performance of the power plants, such as weather conditions (illuminance, temperature, wind speed, etc.), hydrological inflow, coal type coal consumption, and so on. In actual power generation scenarios, the power generation performance of the associated power plants is often affected by multiple factors such as weather environment, unit equipment, and coal type coal consumption.

[0271] Therefore, in order to determine the power generation performance of the associated power plants as accurately as possible, in the grid clearing method provided in this application, it is first necessary to perform low-order data abstraction processing on the power generation element characteristics of each associated power plant, so as to abstract the power generation element characteristics that have a greater impact on the power generation performance of the power plants from the complex power generation element characteristics, and integrate these power generation element characteristics into the first data vector. At the same time, the first data vectorization of the power generation element characteristics can provide a data basis for the subsequent adjustment of the preset long-term electricity price function library. Through the preset long-term electricity price function library and the first data vector, the long-term trading electricity price of the associated power plants can be predicted more accurately in the subsequent process, thereby improving the accuracy of grid clearing.

[0272] Specifically, taking a thermal power plant as an example of an associated power plant, after low-order data abstraction processing, the first data vector of the thermal power plant can be the data characteristics of its corresponding thermal performance, such as (unit heat rate, coal consumption, auxiliary power consumption rate). Correspondingly, when the associated power plant is a photovoltaic power plant, its corresponding first data vector can also be the photovoltaic conversion efficiency characteristics, such as (illuminance, power generation, photovoltaic conversion efficiency).

[0273] Next, in combination with the specific embodiment drawings, the process of performing low-order data abstraction processing on the power generation element characteristics in this step will be introduced.

[0274] See Figure 3C , which is a schematic flowchart of another grid clearing method provided by the embodiment of this application, specifically including the following steps:

[0275] S1101: Establish an energy consumption data model for each of the associated power plants according to the power generation element characteristics; the energy consumption data model is used to evaluate the power generation performance of the associated power plants.

[0276] Specifically, in the process of performing low-level data abstraction processing on the power generation element characteristics of each associated power plant, it is necessary to convert various collected power generation element characteristics into first-level indicators (natural indicators), and based on the first-level indicators, establish an energy consumption data model for secondary and tertiary indicators related to the associated power plant. Among them, these indicators include parameter indicators such as energy consumption and power generation efficiency. By constructing the energy consumption data model of each associated power plant, the power generation performance of the associated power plant under different conditions can be reflected, so as to comprehensively evaluate the impact of various power generation element characteristics on the power generation performance of its associated power plant.

[0277] At the same time, by constructing an energy consumption data model based on the power generation element characteristics of each associated power plant, a mapping from the data corresponding to each power generation element characteristic to a specific function in the preset long-term electricity price function library can be established, so as to initially construct the association between the data and the function.

[0278] S1102: Perform low-level data abstraction processing on each of the power generation element characteristics through the energy consumption data model to obtain the first data vector of each of the power generation element characteristics.

[0279] After constructing the energy consumption data model of each associated power plant, perform low-level data abstraction processing on each power generation element characteristic through the energy consumption data model, so as to extract from the complex power generation element characteristics the power generation element characteristics that can accurately characterize the power generation performance of the associated power plant, and integrate these power generation element characteristics into the first data vector uniquely corresponding to the associated power plant.

[0280] The first data vector generated thereby can accurately describe the key characteristics of the associated power plant during the power generation process. At the same time, it will also be used to adjust the preset long-term electricity price function library and be applied in the long-term electricity price prediction, so as to achieve dynamic and rapid long-term electricity price prediction, so as to quickly adjust the clearing function of the target power grid.

[0281] In a possible implementation manner, after constructing the energy consumption data model of each associated power plant, set the weights of each indicator in its energy consumption data model to highlight the importance of different power generation elements in the power output, so as to effectively improve the accuracy when adjusting the function of the preset electricity price function library subsequently.

[0282] The above is the specific process of performing low-level data abstraction processing on the power generation element characteristics in step S301. Next, the grid clearing method shown in Figure 3A will be introduced.

[0283] S302: Perform long-term trading electricity price prediction according to the preset long-term electricity price function library and each of the first data vectors to establish a long-term trading stack.

[0284] In the preset long-term electricity price function library, multiple operator functions for calculating the long-term trading electricity price of associated power plants are stored. By using the electricity price operator functions for different power plants in the preset long-term electricity price function library and combining with the first data vector determined by the characteristics of each power generation element, the index parameters represented by the first data vector can be imported into the corresponding operator functions, thereby calculating the long-term trading electricity prices of each associated power plant, and thus completing the prediction of the long-term trading electricity prices of the associated power plants.

[0285] For example, taking a hydropower plant as an associated power plant as an example, its corresponding long-term electricity price operator function is:

[0286] PTOPE (long-term trading electricity price) = A + B + C + D + E = Am + Bm + Cm + Dm + Em = (TOPEm - PCm) * CCRm + (TOPEm - PCm) * FOMRm + (TOPEm - PCm) * HFCm + NEOm * VOMRm + (TOPEm - PCm) * CCRTm = (TOPEm - PCm) * (CCRm + FOMRm + HFCm + CCRTm) + NEOm * VOMRm

[0287] In the formula, A represents the capital cost recovery element, B represents the fixed operation and maintenance cost element, C represents the hydropower facility element, D represents the variable operation and maintenance cost element, E represents the special facility element, and Am, Bm, Cm, Dm, Em represent the elements (rupees) of each annual billing period;

[0288] TOPEm represents the take-or-pay electricity volume (kWh) in the billing period, which is obtained through TOPE of each year, and TOPE needs to be confirmed by the target power grid and both parties of the associated power plant;

[0289] CCRm represents the capital cost recovery charging rate (Rp / kWh) applicable to the billing period;

[0290] PCm = the payment credit (kWh) of the billing period. There is only payment credit PCm when the end of each billing period and the net output in the billing period is less than TOPEm;

[0291] Here: PCm = TOPEm - NEOm - DCm - WLCm;

[0292] Among them, NEOm represents the net output in the billing period; NEOm in the billing period should not exceed TOPEm;

[0293] DCm represents the transmission credit in the billing period, which should be equal to the electricity volume (kWh) produced by non-power plants according to the dispatching instructions of the target power grid plus the considered transmitted electricity volume or generated electricity volume; if the power supply system of the target power grid cannot accept all the electricity produced by the associated power plant, the target power grid should issue appropriate dispatching instructions to the power plant;

[0294] FOMRm = Recovery charge rate (Rp / kWh) for the fixed operation and maintenance costs applicable to the billing period (expressed in rupees), and HFCm represents the allowances for the following three aspects: (i) part of the operation and maintenance costs of the hydropower facility, (ii) basin river coordination contribution, and (iii) water charges. All the above three aspects are expressed in Rp / kWh.

[0295] VOMRm represents the recovery charge rate (Rp / kWh) for the variable operation and maintenance costs during the billing period, that is:

[0296] VOMRm = VOMRFm + VOMRLm;

[0297] In the formula, VOMRFm represents the recovery charge rate (Rp / kWh) for the variable operation and maintenance costs applicable to the billing period (expressed in non-rupees), and VOMRLm represents the recovery charge rate (Rp / kWh) for the variable operation and maintenance costs applicable to the billing period (expressed in rupees);

[0298] CCRTm represents the recovery charge rate (Rp / kWh) for the capital costs of special facilities applicable to the billing period.

[0299] As can be seen from the above formula, by importing the first data vector containing the power generation element characteristics of the hydropower plant into its corresponding long-term electricity price operator function, the long-term trading electricity price of the hydropower plant within a certain period in the future can be predicted. By presetting the long-term trading electricity price operator functions corresponding to each associated power plant in the long-term electricity price function library and importing the first data vector of each power plant into it, the long-term trading electricity prices of each power plant can be predicted.

[0300] It can be understood that in the actual long-term trading electricity price prediction scenario, there are differences in the power generation element characteristics of associated power plants in different regions and different countries, and their corresponding electricity price calculation formulas will also be different. This embodiment will not elaborate on this.

[0301] Furthermore, by sorting the long-term trading electricity prices of each power plant, the stack priority in the long-term trading stack can be determined, thereby establishing the long-term trading stack. Exemplarily, the long-term trading stack can be referred to Figure 3D as shown in the schematic diagram of a long-term trading stack. As can be seen from the figure, the long-term trading stack is sorted based on the long-term trading electricity prices predicted for each associated power plant.

[0302] It can be understood that in actual application scenarios, the long-term trading electricity volume between the power grid and its associated power plants is determined by the long-term trading contracts pre-signed between the two. In the process of constructing the long-term trading stack, the long-term trading electricity volume indicated in the long-term trading contract can also be used as the weight of the stack. The more the trading electricity volume, the lower the long-term trading electricity price, and the higher the priority of the stack. In the case of the same long-term trading electricity price, in order to ensure the stable clearing of the power grid, the long-term trading electricity volume needs to occupy more weight than the long-term trading electricity price, and the amount of trading electricity volume is used as the key factor in constructing the long-term trading stack. This embodiment will not elaborate on this.

[0303] Thus, through the first data vectors corresponding to each associated power plant and the preset long-term electricity price function library, the specific long-term trading prices of each associated power plant under the influence of the current power generation factor characteristics can be predicted, and then the long-term trading stack can be established through the long-term trading electricity prices of each associated power plant.

[0304] Next, in combination with the specific embodiment drawings, the establishment process of the long-term trading stack in step S302 will be introduced. Refer to Figure 3E , which is a schematic flow chart of another power grid clearing method provided by the embodiment of the present application, specifically including the following steps:

[0305] S2201: Through the first data vector, perform data-driven adjustment on the function indicators of the preset long-term electricity price function library to obtain a dynamic electricity price function library.

[0306] As can be seen from the foregoing, operator functions for calculating the long-term trading electricity prices of each associated power plant are pre-stored in the preset electricity price function library. In actual application scenarios, the long-term trading electricity price of an associated power plant is determined by its own power generation cost, and there are significant differences in the power generation costs of different associated power plants in different regions, different time periods, and even different seasons. Therefore, in order to accurately predict the long-term trading electricity price of an associated power plant, it is necessary to perform data-driven adjustment on the function indicators in the preset electricity price function library through the first data vector generated by the power generation factor characteristics, so as to adjust the function indicators in each operator function, thereby improving the prediction accuracy of the long-term trading electricity price.

[0307] Here, taking a thermal power plant as an example of an associated power plant, taking the first data vector of the thermal power plant as (unit heat rate, coal consumption, auxiliary power consumption rate, electricity price) as an example, its first data vector is: (Pm, SHRW, GCVs, AUX, CERm), and its corresponding long-term trading electricity price operator function is:

[0308] CERm = Pm * SHRW / GCVs(1 - AUX) = 0.031053242 USD / KWh(1)

[0309] Wherein, Pm represents the coal price in USD, SHRW represents the unit heat capacity index of the unit in Kcal / KWh, GCVs represents the high calorific value of coal in Kcal / Kg, AUX represents the auxiliary power consumption rate of the plant, in %, CERm represents the electricity price, in USD / KWh, and SHRW = 2138.4 Kcal / KWh is set in formula (1);

[0310] Adjust the electricity price operator function through the first data vector, and the adjusted electricity price operator function is:

[0311] CERm = Pm * SHRW(1 + AGE) / GCVs(1 - AUX) = 0.031947846 USD / KWh (2)

[0312] Wherein, AGE represents the unit loss, in %, and the others are the same as above. When AGE is set to 5.4%, SHRW = 2210.14 Kcal / KWh;

[0313] Specifically, reference can be made to Figure 3F and Figure 3G where Figure 3F shows a parameter representing the thermal performance of the unit under full load conditions provided by an embodiment of the present application, Figure 3G is Figure 3F Based on the data shown, it is a schematic diagram of the comprehensive consideration parameter for coal price and long-term trading electricity price. It can be seen that after the unit performance decays and the specific heat capacity index changes, the loss needs to be considered in the unit electricity price change formula, so that (2) replaces formula (1), and the electricity price changes from 0.031 USD / KWh to 0.032 USD / KWh.

[0314] It can be seen that through the first data vector determined in real time by each associated power plant, the long-term electricity price operator function for each associated power plant in the preset long-term electricity price function library can be adaptively adjusted, so as to ensure the prediction accuracy of the long-term trading electricity price.

[0315] S2202: Perform long-term trading electricity price prediction according to the dynamic electricity price function library and the characteristics of each power generation element to obtain the long-term trading predicted electricity price of each associated power plant.

[0316] Correspondingly, after dynamically adjusting the preset long-term electricity price function library, taking the characteristics of each power generation element as the input data in the adjusted dynamic electricity price function library, the long-term trading predicted electricity price of each associated power plant can be calculated.

[0317] S2203: Establish the long-term trading stack through the long-term trading predicted electricity price of each associated power plant.

[0318] It can be seen that the first data vector generated from the actual power generation element characteristics of each associated power plant can effectively adjust the function operators in the preset electricity price function library, so as to predict the actual long-term trading electricity price for the associated power plants through the adjusted dynamic electricity price function library, ensure the prediction accuracy of the long-term trading electricity price, and obtain the long-term trading stack.

[0319] S303: Establish a spot trading stack based on the spot trading electricity prices of each of the associated power plants, and determine a first clearing function through the long-term trading stack and the spot trading stack.

[0320] During the establishment of the long-term trading stack, the target power grid will synchronously receive the spot trading electricity prices sent by the associated power plants. The target power grid can establish a corresponding spot trading stack based on the magnitudes of the spot trading electricity prices of each associated power plant. In this way, the first clearing function of the target power grid can be determined through the spot trading stack and the long-term trading stack.

[0321] Regarding the spot trading stack and the long-term trading stack, specific reference can be made to Figure 3H , Figure 3H which is a schematic diagram of the stack structure of a spot trading stack and a long-term trading stack provided by an embodiment of the present application. As can be seen from the figure, its stack is divided into three levels according to the electricity price, and is gradually incremented according to the electricity price and divided into different levels at the top, middle, and bottom. Among them, each layer considers different energy types (such as photovoltaic, wind power) and market types (spot trading, long-term trading). The adjustment mechanism for the stack is shown beside the stack. By reducing or increasing the trading volume, a balance is achieved between spot trading and long-term trading, so as to optimize the overall electricity price and electricity quantity distribution.

[0322] In actual application scenarios, the spot trading electricity prices of each associated power plant often have multiple types, such as hourly quotes, block quotes, etc. The spot trading electricity price is calculated by a spot trading electricity price calculation formula set by the power plant itself. Taking the calculation of the hourly quote of a hydropower plant as an example, the calculation formula for its hourly quote is:

[0323] Hydropower plant:

[0324]

[0325] Q m3 =PR 实 ×P m (1.2)

[0326] In the formula, m is the m-th hydropower station in the cascade hydropower stations, and i is the i-th unit in the m-th hydropower station; both m and i are integers not less than 1; Q m2 in formula (1.1) represents the actual revenue corresponding to the potential electric energy of the hydropower station at the current moment; Q in (1.2) in the target formulam3 The actual power generation revenue corresponding to the actual power generation of the hydropower station at the current moment, G i额 Denote the rated power of the i-th unit, H im Denote the water level duration of the i-th unit, F iJm Denote the derated power of the i-th unit, H iZm Denote the interruption duration within the water level period of the i-th unit, P im The power generation amount of the i-th unit within the water level period, PR 实 Denote the electricity price actually traded by all units, P m The power generation amount of all units within the water level period.

[0327] As can be seen from the foregoing, in the preset electricity price function library, operator functions for calculating the spot trading electricity prices of each associated power plant are pre-stored. In an actual application scenario, the spot trading electricity price of an associated power plant is determined by its own power generation cost, and there are significant differences in the power generation costs of different associated power plants in different regions, different time periods (the performance of the unit decays with the service life), and even different seasons. Therefore, in order to accurately predict the long-term trading electricity price of an associated power plant, it is necessary to perform data-driven adjustment on the function indicators in the preset electricity price function library through the first data vector generated from the power generation element characteristics, so as to adjust the function indicators in each operator function, thereby improving the prediction accuracy of the long-term trading electricity price.

[0328] In addition, in the process of establishing the spot trading stack, it is necessary to construct it through two dimensions of the spot trading electricity price and the power generation amount of the associated power plant. Next, with reference to the specific embodiment drawings, the process of establishing the spot trading stack will be introduced. Refer to Figure 3I , which is a schematic flow chart of a method for establishing a spot trading stack provided by an embodiment of the present application, specifically including the following steps:

[0329] S3031: Perform power generation prediction based on the first data vectors of the respective associated power plants to obtain the predicted power generation amounts of the respective associated power plants.

[0330] As can be seen from the foregoing, in the first data vector corresponding to the associated power plant, multiple power generation element characteristics that affect the power generation efficiency of the associated power plant are aggregated. Based on the first data vector for power generation prediction, the power generation amount of the associated power plant within a certain time period can be predicted.

[0331] Specifically, in the process of performing power generation prediction based on the first data vector, a neural network model can be used to predict the power generation amount of the associated power plant within a certain time period. Here, a wind power plant is taken as an example, and specifically, reference can be made to Figure 4A , Figure 4A which is a schematic diagram of power generation prediction based on a wind power plant provided by an embodiment of the present application.

[0332] As can be seen from the figure, among them, the global forecast data, regional actual data, and site observation data of the wind power plant in the historical time period are used as the training data of the model, so as to predict the power generation of the wind power plant through the trained neural network model. Correspondingly, the model can output the meteorological prediction results and corresponding power prediction results of the wind power plant. Through the meteorological data and power prediction data within a certain period of time, the power generation of the wind power plant can be effectively predicted according to the main characteristics affecting the power generation performance of the wind power plant.

[0333] In addition, for the electricity quantity prediction of associated power plants, a time series prediction model (ARIMA) or a machine learning method (such as random forest, XGBoost) can be used. For the first data vector with a complex data structure in this application, the temporal relationship in the sequential data can be obtained through a deep learning model such as LSTM (Long Short-Term Memory).

[0334] LSTM (Long Short-Term Memory) is a special recurrent neural network (RNN) architecture designed to address the vanishing gradient and exploding gradient problems faced by ordinary RNNs when dealing with long sequence data. By introducing memory cells and gating mechanisms, LSTM can effectively capture long-term dependencies in the sequence.

[0335] S3032: Establish a preliminary spot trading stack based on the predicted electricity quantities of the associated power plants.

[0336] After obtaining the predicted electricity quantities of the associated power plants, the call priority of each associated power plant during power call can be determined by the magnitudes of the predicted electricity quantities of the associated power plants, thereby establishing a preliminary spot trading stack. For example, taking wind power, photovoltaic power, hydropower, thermal power, and nuclear power as associated power plants, assume the predicted electricity quantities of each power plant are (100, 50, 30, 60, 20) respectively. Through the predicted electricity quantities of the associated power plants, a preliminary spot trading stack can be established based on the magnitudes of the predicted electricity quantities, that is, (wind power, thermal power, photovoltaic power, hydropower, nuclear power).

[0337] S3033: Adjust the preliminary spot trading stack based on the spot trading electricity prices of the associated power plants to obtain the spot trading stack.

[0338] On the basis of establishing a preliminary spot trading stack based on the predicted electricity quantities of associated power plants, the spot trading electricity price calculated through the spot electricity price calculation formula within each associated power plant is used to further adjust the preliminary spot trading stack, so that the two-dimensional preliminary spot trading stack becomes a three-dimensional spot trading stack including the spot trading electricity price. In this way, when there is a situation where the power generation of an associated power plant is insufficient or excessive, following the method of taking out from the upper layer of the stack and reconstructing the stack, starting from the power generation of the associated power plant and the actual spot trading electricity price, the spot trading stack is dynamically adjusted, thereby improving the accuracy of grid clearing.

[0339] Exemplarily, the spot trading stack can be referred to Figure 4B as shown in the schematic diagram of the spot trading stack.

[0340] It should be particularly noted that in the scenario of establishing the spot trading stack, the weight occupied by the predicted electricity quantity in establishing the stack is greater than the weight occupied by the spot trading electricity price. That is, when the spot trading electricity prices of two associated power plants are the same, the associated power plant with a larger predicted electricity quantity is placed in a relatively higher priority position in the spot trading stack, so as to ensure the power supply stability of the target power grid.

[0341] S104: Determine the power supply-demand balance state of the target power grid when clearing under the first clearing function.

[0342] Correspondingly, based on the determined first clearing function to control the target power grid for power clearing, the power supply-demand balance state of the target power grid can be determined according to the total power generation and total load of the target power grid when clearing based on the first clearing function. When the total power generation of the target power grid is greater than the total load, it indicates that the target power grid is in an over-generation state, which means that the current power generation of the target power grid exceeds the actual power demand, and there is a problem of power surplus. When the total power generation of the target power grid is less than the total load, it indicates that the target power grid is in an under-generation state, which means that the current power generation of the target power grid cannot meet the actual power demand, and power supply needs to be replenished in a timely manner.

[0343] Specifically, it can be referred to Figure 4C as shown in the curve schematic diagram of a power grid clearing function. In the figure, the horizontal coordinate axis represents the change in power supply, and the vertical coordinate axis represents the change in electricity price. The power dispatch priorities shown in the figure are: wind power / solar power > hydropower with storage capacity > nuclear power > coal power > gas power.

[0344] When the power supply curve is higher than the power demand curve, it indicates that the electricity quantity supplied by the power grid exceeds the demand at this time, which may lead to a reduction in electricity price. Correspondingly, when the power supply curve is lower than the power demand curve, the power demand exceeds the supply, which may lead to an increase in electricity price.

[0345] By determining the power supply-demand balance state of the target power grid under the first clearing function, the under-generation state or over-generation state of the target power grid under the current first clearing function can be characterized, so as to facilitate subsequent control of the long-term trading stack and the spot trading stack according to the actual power supply-demand balance state of the target power grid, thereby realizing the clearing adjustment for the power grid and ensuring the clearing accuracy rate of the power grid.

[0346] S305: Adjust the first clearing function according to the power supply-demand balance state to obtain a second clearing function, and perform power clearing based on the second clearing function.

[0347] Finally, according to the over-generation state or under-generation state of the target power grid under the first clearing function, the first clearing function is adjusted, so that the clearing function of the target power grid can be adjusted according to the actual power generation situation of the target power grid, thereby improving the clearing accuracy rate of the power grid.

[0348] Next, in combination with the specific embodiment drawings, the adjustment methods for the first clearing function in different power supply-demand balance states will be introduced. See Figure 4D and Figure 4E , Figure 4D which is a schematic flowchart of another power grid clearing method provided by an embodiment of the present application, Figure 4E and

[0349] S3051: When the power supply-demand balance state is the over-generation state, perform high-order abstraction processing on the characteristics of each power generation element through a preset deep learning model to obtain second data vectors of each associated power plant.

[0350] In an actual application scenario, the long-term trading electricity price is often determined by the long-term contract between the target power grid and the associated power plants, while the spot trading electricity price is often affected by the immediate supply-demand situation. Therefore, the long-term trading electricity price is often lower than the spot trading electricity price. When the power supply-demand balance state of the target power grid is the over-generation state, at this time, the total power generation of the target power grid is greater than the total load demand of the target power grid.

[0351] In order to minimize the cost of power dispatching and improve the accuracy of power grid clearing as much as possible, when the target power grid is in the over-generation state, since the total power generation of the target power grid is greater than the total charge at this time, by increasing the first long-term trading volume in the first clearing function, the proportion of the long-term trading volume in the clearing function can be increased, thereby effectively reducing the cost of power clearing and improving the accuracy of power clearing.

[0352] Therefore, in order to ensure the accuracy of the adjustment for long-term trading volume, it is necessary to perform high-order data abstraction processing on the power generation element characteristics of associated power plants through a pre-set deep learning model, so as to obtain the second data vector that affects the power generation efficiency of each associated power plant.

[0353] Different from the low-order data abstraction processing of power generation element characteristics through the energy consumption data model, the low-order data abstraction processing only involves feature screening and combination of power generation element characteristics, while the high-order data abstraction processing involves combinatorial calculation between the characteristics of each power generation element, so as to calculate the data vector that plays a core role in the power generation performance of associated power plants.

[0354] Specifically, reference can be made to Figure 4F The schematic diagram of calculating the second data vector through a deep learning model as shown. In the figure, a hydropower plant is taken as an example of an associated power plant. Among them, parameters such as power generation amount and water consumption for power generation are the power generation element characteristics of the hydropower plant, while the Top, Qm', and average power plant revenue of the hydropower plant are the parameters calculated through the power generation amount and water consumption for power generation. The process of calculating these parameters through the power generation element characteristics and then obtaining the relevant parameters is the above-mentioned process of performing high-order abstraction processing on the power generation element characteristics based on the pre-set reinforcement learning model to obtain the second data vector. In the pre-set reinforcement learning model provided in the embodiment of the present application, a variety of different types of parameter calculation formulas are pre-set. When performing high-order data abstraction processing on the power generation element characteristics through the pre-set reinforcement learning model, parameter calculation is performed through the pre-set calculation formula and the power generation element characteristics, and the calculated results are integrated to obtain the second data vector.

[0355] Compared with the linear model, deep learning has stronger non-linear modeling ability, and it can simulate the complex non-linear relationships between the power generation element characteristics in the associated power plants. Therefore, through the high-order data abstraction processing of the power generation element characteristics by the pre-set deep learning model, complex features can be further extracted from the original data to obtain the second data vector that can accurately represent the power generation performance of the associated power plants.

[0356] S3052: Determine the long-term trading volume increment according to each of the second data vectors and the pre-set long-term electricity price function library.

[0357] S3053: Adjust the first long-term trading volume according to the long-term trading volume increment to obtain the second clearing function.

[0358] Further, after determining the second data vectors of each associated power plant, the function indicators in the preset long-term electricity price function library are adjusted based on the second data vectors, which can change the specific indicator values of the long-term electricity price function, so as to determine the new long-term trading volume of the target power grid. By comparing this new long-term trading volume with the long-term trading volume in the first clearing function, the increase in the long-term trading volume can be determined, and the corresponding second clearing function can be obtained.

[0359] As Figure 4D can be seen, when the target power grid is in an over-generation state, through the preset deep learning model to perform high-order data abstraction processing on the power generation element characteristics of the associated power plants, the increase in the long-term trading volume can be determined, and then the subsequent long-term trading stack can be adjusted, so as to realize the adjustment of the first clearing function and obtain the second clearing function.

[0360] Next, in combination with the specific embodiment drawings, the process of adjusting the first clearing function when the target power grid is in an under-generation state will be introduced. Refer to Figure 4G , which is a schematic flowchart of the adjustment method of the first clearing function when the target power grid provided by the embodiment of the present application is in an under-generation state, specifically including the following steps:

[0361] S3054: When the power supply and demand balance state is the under-generation state, for the first power generation curve when the target power grid clears based on the first clearing function, the first power generation curve is fitted through a preset fitting function to obtain a fitted power generation curve.

[0362] When the target power grid is in an under-generation state, it indicates that the total power generation of the target power grid is lower than its own total load. It can be understood that in the dual trading stack composed of the long-term trading stack and the spot trading stack, since the long-term trading electricity price in the long-term trading stack is often slightly lower than the spot trading electricity price in the spot trading stack, the long-term trading electricity in the long-term trading stack can be preferentially dispatched. In this case, if the target power grid is still in an under-generation state, it means that the target power grid still cannot meet the load demand even when fully invoking the long-term trading electricity in the long-term trading stack.

[0363] Therefore, in this case, it is necessary to increase the spot trading volume to meet the power load demand. To ensure the accuracy of the increase in the spot trading volume, before calculating the increase in the spot trading volume, it is first necessary to perform fitting processing on the first power generation curve when the target power grid clears based on the first clearing function through a preset fitting function, so as to convert complex and discrete power generation data into a continuous function model, thereby simplifying the data storage and analysis process and improving the data processing efficiency. At the same time, through the fitted power generation curve obtained by fitting, the trends and patterns in the power generation data can be more clearly identified, so as to facilitate the subsequent optimization of the clearing function.

[0364] Specifically, the formula of the preset fitting function is as follows:

[0365] PR*S = P;

[0366] PR = A + BH T -CH T 2 ;

[0367] In the formula, PR represents the power generation conversion efficiency of the power plant, H T represents the power generation efficiency adjustment factor, S represents the primary index affecting the power generation efficiency of the associated power plant, P represents the fitting coefficient, and A, B, and C represent the characteristic constants calculated by the deep learning model, which are used to characterize the performance characteristics corresponding to different power plants.

[0368] In the above formula, the power of H T characterizes the influence degree of this characteristic on the power generation performance of the associated power plant. Correspondingly, the positive or negative sign in front of H in the formula T indicates the positive or negative influence of the corresponding characteristic on the power generation efficiency of the power plant. Thus, through the above formula, the fitting accuracy of the first power generation curve can be effectively improved, thereby ensuring the accuracy of grid clearing.

[0369] It can be understood that for different associated power plants, the data types corresponding to S will also be different. For example, when PR represents the power generation conversion efficiency of a photovoltaic power plant, S is related to the illuminance of the photovoltaic power plant. Similarly, when PR represents the power generation conversion efficiency of a thermal power plant, S is related to the coal quality of the thermal power plant.

[0370] By calculating the fitting coefficient P corresponding to all associated power plants of the target grid, the first power generation curve of the target grid can be effectively fitted, thereby improving the accuracy of the first power generation curve.

[0371] Specifically, reference can be made to Figure 4H , Figure 4H , which is a schematic diagram of fitting the power generation curve of the target grid provided by the embodiment of the present application. As shown in the figure, the topmost power generation curve in the figure is the preset power generation curve, and the curve that closely fits the preset power generation curve is the fitted power generation curve after fitting when the target grid clears based on the first clearing function. It can be seen that after the first power generation curve is fitted by the preset fitting function, the fitted power generation curve is closer to the preset power generation curve. By comparing the preset power generation curve and the fitted power generation curve, the spot trading volume that needs to be supplemented can be accurately determined, thereby ensuring the stability of grid clearing.

[0372] S3055: Based on the fitted power generation curve and the pre-designed planned power generation curve, determine whether the first spot trading volume meets the pre-designed planned power generation curve.

[0373] Subsequently, by comparing the fitted power generation curve with the pre-set planned power generation curve, it can be determined whether the first spot trading volume in the first clearing function can meet the pre-designed planned power generation curve. Among them, the pre-designed planned power generation curve is used to represent the power generation curve expected by the target power grid for the current power load. By comparing the two, it can be determined whether the first spot trading volume of the target power grid in the first clearing function is sufficient.

[0374] See Figure 4H , as shown in the figure, the fitted power generation curve in the figure is lower than the preset power generation curve, that is, the power generation amount of the fitted power generation curve is lower than that of the preset power generation curve. At this time, there is a certain gap between the fitted power generation curve and the preset power generation curve. It can be determined that the power generation curve of the target power grid under the first clearing function cannot meet the requirements of the preset power generation curve. Therefore, it can be determined that the first spot trading volume in the first clearing function cannot meet the pre-designed planned power generation curve, and additional spot trading volume is needed to supplement.

[0375] S3056: When the first spot trading volume does not meet the pre-designed planned power generation curve, correct the function of the preset long-term electricity price function library through a preset reinforcement learning model to determine the long-term trading volume reduction value.

[0376] Correspondingly, when the first spot trading volume does not meet the pre-designed planned power generation curve, it indicates that the first spot trading volume of the target power grid at this time cannot meet the current power load. Therefore, in order to meet the current power load by increasing the spot trading volume, it is necessary to correct the function of the preset long-term electricity price function based on the preset reinforcement learning model to obtain the long-term trading volume reduction value. In this way, by reducing the proportion of the long-term trading volume, the increase value of the spot trading volume can be determined to ensure the smooth clearing of the power grid.

[0377] Among them, the preset reinforcement learning model can adopt MARL (Multi-agent Reinforcement Learning). MARL is a method for dealing with reinforcement learning problems in a multi-agent environment. Each participant (such as different associated power plants) can be regarded as an agent, and these agents need to adjust their bidding strategies according to the real-time data of the market and the behaviors of other agents. MARL can help these agents learn and adapt under complex and dynamically changing market conditions to improve the clearing accuracy of the power grid.

[0378] Specifically, in the process of correcting the functions in the preset long-term electricity price function library, through the screening mechanism of the reinforcement learning model, the first long-term trading electricity price function that has the greatest impact on the overall long-term trading electricity price can be screened out from the preset long-term electricity price function library.

[0379] Subsequently, by adjusting the function indicators in the first long-term trading electricity price function and increasing the function indicators therein to increase the long-term trading electricity price, the effect of reducing the long-term trading volume can be achieved, and the long-term trading volume reduction value can be obtained.

[0380] Here, taking a thermal power plant as an example, as can be seen from the previous text, the long-term trading electricity price calculation function of the thermal power plant is:

[0381] By adjusting the function indicators in the long-term trading electricity price calculation function and increasing the long-term trading electricity price, the effect of reducing the long-term trading volume can be achieved, and the long-term trading volume reduction value can be obtained.

[0382] S3057: Adjust the proportion of the spot trading volume in the first clearing function according to the long-term trading volume reduction value to obtain the spot trading volume increase value.

[0383] S3058: Adjust the first spot trading volume according to the spot trading volume increase value to obtain the second clearing function.

[0384] On the basis of determining the long-term trading volume reduction value, by increasing the proportion of the spot trading volume in the first clearing function, the spot trading volume value can be obtained. Thus, the first spot trading volume in the first clearing function can be adjusted by the spot trading volume increase value to obtain the second clearing function.

[0385] An embodiment of the present application provides a power grid clearing method. In this method, first, low-order data abstraction processing is performed according to the power generation element characteristics of each associated power plant, and various factors affecting power grid clearing are quantified into a first data vector, thereby reducing the data complexity of the power generation element characteristics. Subsequently, long-term trading electricity price prediction is performed based on the first data vector and a preset long-term electricity price function library to predict the corresponding long-term trading electricity price according to the actual power generation element characteristics of each associated power plant, and a corresponding long-term trading stack is established to improve the accuracy of power grid clearing by ensuring the accuracy of the long-term trading electricity price. Further, a spot trading stack is established based on the spot trading electricity prices of each associated power plant, and a first clearing function is determined through the long-term trading stack and the spot trading stack. Correspondingly, according to the determined first clearing function, the power supply and demand balance state when the target power grid clears under the first clearing function is determined, and the first clearing function is adjusted in real time according to the actual power supply and demand balance state of the target power grid to obtain a second clearing function. In this way, the clearing function of the target power grid can be dynamically adjusted according to the supply and demand state of the target power grid, and a specific clearing function can be established through the trading stack constructed in real time by the long-term trading electricity price and the spot trading electricity price, so that the power grid clearing can effectively consider the impact of power generation element characteristics on the power generation of associated power plants and the changes in the spot trading electricity prices of each associated power plant, and then determine a more accurate clearing function, thereby achieving the effect of improving the accuracy of power grid clearing. In this solution, in order to better illustrate its data-driven nature, measured data is used to fit the power prediction curve of the dispatching. The clearing operator quantifies electricity prices and electricity quantities, as well as long-term and short-term electricity prices, etc., making the algorithms in the data and information flow clear. This intelligent agent needs to be constructed to realize the determination of electricity prices and clearing. Algorithms need to be selected according to the characteristics of the data and information flow, that is, neural networks.

[0386] Example 3

[0387] In Embodiment 3, the power function requirements for realizing the dispatching control of the power grid side over the power plant side are specifically the peak shaving dispatching in the scenario of high-frequency load variation of thermal power. The functional operator is the thermal power peak shaving dispatching operator, and the thermal power peak shaving dispatching operator is used to realize the peak shaving dispatching in the scenario of high-frequency load variation of thermal power. Embodiment 3 introduces the process of the digital power system using the intelligent agent constructed based on the thermal power peak shaving dispatching operator.

[0388] In the scenario where thermal power generation is parallel with new energy generation, due to the instability of new energy generation, the load of thermal power generation changes frequently. Under the condition of high-frequency load changes in the thermal power scenario, there is currently a lack of a supporting peak shaving and dispatching scheme for digital calculation costs, making it difficult to implement peak shaving and dispatching safely and stably, and easily affecting the lifespan of the power system. How to ensure the lifespan of the power system is a key issue. In view of this, the inventor proposes a peak shaving and dispatching method in the scenario of high-frequency variable load of thermal power based on a digital power system with multi-functional agents, which uses a thermal power peak shaving and dispatching operator. Based on various data fed back by thermal power plants, it can comprehensively consider the load capacity, performance, cost of thermal power units and the change of thermal power load demand. By using digital technology means, it analyzes the change relationship between the plant cost and the thermal power load, so as to ensure the safety and stability of peak shaving and dispatching.

[0389] Figure 5A Shows a power supply scenario. As Figure 5A shown, this power supply scenario involves multiple power generation methods, including thermal power generation and photovoltaic power generation. Among them, photovoltaic power generation, as a new energy generation method, is gradually increasing its proportion in the energy consumption structure due to its characteristics such as zero emissions, zero pollution, and sustainability. However, compared with thermal power generation, photovoltaic power generation has more prominent instability. Therefore, in the Figure 5A power supply scenario shown, thermal power plants that generate electricity through thermal power need to change their loads frequently. In specific implementation, thermal power plants need to cooperate with the peak shaving and dispatching on the grid side. Taking Figure 5A as an example, Figure 5A Three thermal power plants are shown on the left. Each thermal power plant can store and release electrical energy. The electrical energy released by the thermal power plants is transmitted through their electrical connections with the grid, and this electricity ultimately serves thousands of households that purchase electricity from the grid.

[0390] Each thermal power plant generally has multiple thermal power units, and there are often differences in the equipment performance, supporting perfection, and maintenance costs of each thermal power plant. Furthermore, there are often differences in the equipment durability, safety attributes, ramp-up ability to cooperate with peak shaving and dispatching, and power generation costs of each thermal power unit. For thermal power plants, when cooperating with peak shaving and dispatching, they need to pay attention to whether the thermal power units will result in operating losses due to cooperating with peak shaving and dispatching; in addition, they also need to determine whether they can cooperate with the peak shaving working conditions. For the grid side, it not only needs to combine the quotations of each unit in the thermal power plant and the capabilities and performance of the units themselves, but also needs to consider various aspects in dispatching to ensure the safe and stable realization of peak shaving and dispatching.

[0391] Figure 5B This is a flowchart of a peak shaving and dispatching method in the scenario of high-frequency variable load of thermal power provided by an embodiment of this application. Figure 5C This is a signaling interaction diagram of a peak shaving and dispatching method in the scenario of high-frequency variable load of thermal power. InFigure 5B In this section, the implementation of the solution is mainly introduced from the perspective of the thermal power plant. In Figure 5C it, the implementation of the solution is introduced interactively from both sides (the power plant side and the grid side).

[0392] Such as Figure 5B shown, a peak shaving and dispatching method in a high-frequency variable load scenario of thermal power provided by an embodiment of the present application includes:

[0393] S501. Calculate the heat storage and release margin of the energy storage device in the thermal power plant.

[0394] Each thermal power plant is configured with an energy storage device. The energy storage device stores the heat generated by the coal combustion of the thermal power unit and releases it when needed. The heat is converted into electrical energy and transmitted to the grid. In a possible implementation manner, the energy storage device in the thermal power plant includes, but is not limited to: boilers, steam turbine systems, heaters, deaerators, etc.

[0395] The heat storage and release margin of each type of energy storage device is divided into two different types. One is the heat storage margin characterization value, and the other is the heat release capacity characterization value. The heat storage margin characterization value is used to represent the ability of the energy storage device to further store heat. The heat release capacity characterization value is used to represent the ability of the energy storage device to further release heat. The heat storage margin characterization value can be represented by a specific heat value or by a percentage compared with the rated stored heat. Similarly, the heat release capacity characterization value can be represented by a specific heat value or by a percentage compared with the pre-rated stored heat. Both the heat storage margin and the heat release capacity are related to the size and performance of the unit.

[0396] The heat storage and release margin of the energy storage device in the thermal power plant reflects the output capacity of the current thermal power plant from the two aspects of "storage" and "retrieval". In specific implementation, according to the actual requirements of the grid side, the heat storage margin can be statistically calculated according to two different dimensions. For example, it is statistically calculated according to the dimensions of "day-ahead" and "intra-day". The thermal power plant provides the day-ahead heat storage and release margin of the energy storage device, which is convenient for the grid side to prepare in advance according to the day-ahead heat storage and release margin and carry out orderly peak shaving and dispatching. The thermal power plant provides the intra-day heat storage and release margin of the energy storage device, which is convenient for the grid side to achieve accurate, flexible and reliable peak shaving and dispatching based on high-real-time data according to the intra-day heat storage and release margin.

[0397] In a possible implementation, the heat storage and extraction margin sent by a thermal power plant to the grid side can be specifically refined to the heat storage and extraction margins corresponding to each thermal power unit of the thermal power plant. It should be noted that the heat storage and extraction margin can be calculated by the thermal power plant based on the coal type parameters of the selected target coal type and the thermal system design. The selection of the coal type affects the heat consumption and coal consumption. Table 2 is a data table of a digital coal type selection model. In this Table 2, various coal types such as GEB 5200, GEB 4800, GEB 4700, etc. are shown for the detection items. Here, the detection items can also be regarded as the coal type parameters mentioned above. These coal type parameters include but are not limited to total moisture, moisture in air-dried basis, ash in as-received basis, volatile matter in dry ash-free basis, carbon in as-received basis, hydrogen in as-received basis, total sulfur, lower heating value, etc. In Table 2, LHV represents the lower heating value of the coal type, and HHV represents the higher heating value of the coal type.

[0398] Table 2

[0399]

[0400] The digital coal type selection model can realize the selection of the coal type by extracting features from the data shown in Table 2. Because the part of H (hydrogen) in carbon combines with oxygen in the physical combustion of the boiler to produce water, and this part of water absorbs a large amount of heat called "latent heat of vaporization", so it is a relatively important factor when selecting the coal type. This hydrogen content is measured by the percentage difference between the associated HHV and LHV, as shown in the last row of data in Table 2. The smaller this value is, the less heat is lost due to the latent heat of vaporization, the more heat is generated per unit of coal for power generation, and the better the economic benefits. In practical applications, by using the digital coal type selection model, according to the higher heating value and lower heating value corresponding to each of the multiple candidate coal types, the percentage difference between the higher and lower heating values of each candidate coal type (this value can be understood as a low-level abstract feature) is obtained; the coal type with the smallest percentage difference between the higher and lower heating values among the multiple candidate coal types is determined as the target coal type. Taking Table 2 as an example, the percentage difference between the higher and lower heating values of the coal type GEB 5200 is the smallest, with a specific value of 5.90%, so the coal type GEB 5200 can be selected as the target coal type. By selecting the coal type with the smallest percentage difference between the higher and lower heating values as the target coal type, the heat generated for power generation per unit of coal can be maximized.

[0401] The calculation formulas for the higher heating value HHV and lower heating value LHV of the coal type are as follows:

[0402] HHV = LHV + 25 * M + 25 * 9 * H

[0403] Let M be the total moisture content of coal and H be the hydrogen content on a dry basis. In the above formula, the units of HHV and LHV are both kcal / kg. In one example, M = 23 and H = 4.21. It should be noted that in the above formula, when converting between the heat units kcal and kJ, the values of M and H need to be converted according to the conversion coefficients of different heat units.

[0404] S502. Send the correlation data among the loads, performances, and costs of each thermal power unit in the thermal power plant and the heat storage and extraction margin to the grid side, so that the grid side can determine the target units for expected auxiliary peak shaving and issue peak shaving scheduling instructions based on the changes in thermal power load demand, the correlation data provided by each thermal power plant, and the heat storage and extraction margin.

[0405] To facilitate the grid side's access to, analysis of, and peak shaving scheduling for the data of each thermal power plant and each thermal power unit, in this application, each thermal power plant needs to provide its own data to the grid side. Specifically, when implemented, the thermal power plant sends the correlation data among the loads, performances, and costs of each thermal power unit to the grid side, and also sends the heat storage and extraction margin calculated in step S501. It should be noted that in the embodiments of this application, the correlation data sent by the thermal power plant to the grid side is also calculated based on the coal type parameters of the selected target coal type.

[0406] Table 3 shows the correlation parameter table of heat consumption, coal consumption, and electricity price under different loads. As shown in the first column of Table 3, various different plant loads are presented (such as 100% load, 95% load, 65% load, 45% load, etc.). As shown in the second to twenty-fourth columns of Table 3, various heat consumption, coal consumption, and system parameters under different loads are presented. As shown in combination with Table 3, under the scheduling of high-frequency variable loads, the actual plant power consumption rate (the seventh column of Table 3) of the thermal power plant changes continuously. For example, at 100% plant load, the actual value of the plant power consumption rate is 5.71%, and at 45% plant load, the actual value of the plant power consumption rate is 7.37%.

[0407] By reporting the data shown in Table 3 to the grid side, the grid has the load conditions and thermal energy conditions of each thermal power plant, and can better classify and aggregate the data, so that different weights and priority orders can be given to the units according to the unit conditions and the heat storage and extraction margin. Since the calculations in this process are affected by multiple variables, the features extracted from the data shown in Table 3 can be called medium-order data abstraction features (abbreviated as medium-order features). Transmitting the medium-order data abstraction features in the chain rule reflects the energy consumption level and thermal energy reserve level of the power plant, and assists in realizing peak shaving scheduling in the scenario of high-frequency variable loads of thermal power.

[0408] Table 3

[0409]

[0410] Table 4 shows the electricity price and cost data of a selected target coal type under different operating conditions (power plant load). The grid compensation coefficient for 1 kWh payment is the electricity price given by the grid (excluding limestone cost), which is the fuel part of the capacity price and generally accounts for more than 50% of the entire electricity price.

[0411] An example calculation formula for the grid compensation coefficient for 1 kWh payment is as follows:

[0412] CERm = U * N / E(1 - AUX)

[0413] Among them, U represents the coal price, which is also the fuel cost subsidy, in US dollars per ton (expressed as USD / t). As an example, U = 69 USD / t. E represents the heat consumption rate of the unit for power generation, that is, the heat consumption per kWh of the unit, in kcal / kWh. N represents the specific heat index, specifically the heat required per kWh of electricity, in kcal / kWh, and AUX represents the auxiliary power consumption rate, in %.

[0414] In Table 4, SHRW represents the weighted specific heat index, in kJ / (kW·h). Ea represents the power supply (or: generated electricity), in kW·h. Ep represents the electricity charge, in USD.

[0415] SHRcc is the specific heat index, in kJ / (kW·h). U is the coal price, in USD / t. ECRm is the electricity charge per kWh, in USD / (kW·h). ECRm represents the electricity charge ratio, that is, the coal cost per kWh of electricity produced, in USD / (kW·h). It should be noted that if the load is evenly distributed every day in a month, the weighted specific heat index is equal to the weighted index, that is, SHRW = SHRcc.

[0416] The calculation formula for ECRm is:

[0417] ECRm = SHRcc × (1 / HHV) × U / 1000

[0418] The calculation formula for Ep is:

[0419] Ep = NEOm * CERx

[0420] In the above formula, NEOm represents the net power generation, in kW·h. CERx represents the electricity price, in USD / (kW·h).

[0421] As shown in Table 4, through the correlation between electricity prices and loads with different values, the costs reflected by the fuels of each thermal power plant in the power grid can be obtained, enabling the power grid to extract the unit performance, environmental protection situation, and fuel cost situation of the power plants under different load conditions based on this data. The feature affected by multiple variables (including the coal cost price) extracted in this way is called a high-order abstract feature. The linear high-order feature generated by integrating the coal price cost and other factors on the basis of the middle-order feature includes the correlation between fuel, electricity, and electricity price.

[0422] That is to say, in the technical solution of this application, the grid side determines the target unit for expected auxiliary peak shaving and issues a peak shaving dispatching instruction based on the change of thermal power load demand, the correlation data provided by each thermal power plant, and the heat storage and extraction margin, including: the grid side extracts middle-order features based on the correlation data provided by each thermal power plant and the heat storage and extraction margin, and the middle-order features include features reflecting the energy consumption level and heat energy reserve level of the power plant; the grid side integrates the coal price cost in the correlation data and the middle-order features to obtain high-order features, and the high-order features include the relationship between fuel, electricity, and electricity price; the grid side determines the target unit for expected auxiliary peak shaving and issues a peak shaving dispatching instruction based on the change of thermal power load demand, the middle-order features, and the high-order features.

[0423] It can be seen from Table 4 that the higher the load, the smaller the value of the grid compensation coefficient for 1 kWh of electricity payment, indicating the lower the fuel cost. Under the same working conditions, the lower the fuel cost, the better the unit performance and the smaller the coal consumption. Therefore, the grid side can know the performance advantages and disadvantages of different thermal power units by comparing the fuel costs of different thermal power units under the same working conditions (such as the grid compensation coefficient for 1 kWh of electricity payment shown in Table 4). For example, for thermal power unit A at 100% load, the grid compensation coefficient for 1 kWh of electricity payment is 0.0311 USD / (Kw·h); for thermal power unit B at 100% load, the grid compensation coefficient for 1 kWh of electricity payment is 0.0319 USD / (Kw·h), and the coal types selected by the two units are the same. By comparing this coefficient, it can be known that the performance of thermal power unit A is better than that of thermal power unit B.

[0424] Table 4

[0425]

[0426] Based on the data reported by each thermal power plant (including the correlation data among the loads, performances, and costs of each thermal power generation unit in the thermal power plant, as well as the heat storage and extraction margin, etc., where the correlation data can refer to the examples in Table 3 and Table 4), the grid side can have a full understanding of the thermal power generation units of each thermal power plant. Based on this, the grid side can exercise the peak shaving scheduling function. Specifically, when implementing, the grid side can determine the target units for expected auxiliary peak shaving based on the changes in thermal power load demand, the correlation data provided by each thermal power plant, and the heat storage and extraction margin, and then issue a peak shaving scheduling instruction to the target units.

[0427] Before the grid side determines the target units for expected auxiliary peak shaving scheduling, the grid side can issue a peak shaving load curve for each thermal power generation unit to each thermal power plant. In one example implementation, the peak shaving load curve is in a two-dimensional form, with the horizontal axis representing time and the vertical axis representing power. The peak shaving load curves for different thermal power generation units may be different. The peak shaving load curve essentially expresses the load requirements for the thermal power generation units.

[0428] The thermal power plant can determine the thermal power generation units that meet the requirements of the corresponding peak shaving load curve according to the peak shaving load curve for each thermal power generation unit, and calculate the quotation information for these thermal power generation units to provide auxiliary peak shaving services. Since peak shaving scheduling is an additional auxiliary task provided by the thermal power generation units, a certain fee needs to be charged to the grid side. The amount of the fee is shown in the form of a quotation. The thermal power plant combines the performance of the thermal power generation units, the requirements of the peak shaving load curve, and combines fuel information such as the type of coal burned, coal consumption, and heat consumption, fuel costs, equipment costs, etc., to calculate the quotation information for providing auxiliary peak shaving services. The quotation information is sent to the grid side, so that the grid side can compare and determine the thermal power generation units for expected auxiliary peak shaving based on the quotation information. These units are called target units. The number of target units in a thermal power plant is uncertain. For example, Thermal Power Plant A may have 3 thermal power generation units determined as target units; Thermal Power Plant B may have 1 thermal power generation unit determined as target units; Thermal Power Plant C may have no determined target units. Combining Table 3 and Table 4, the grid side can, through the functions of data-driven and deep learning, based on the changes in thermal power load demand, the correlation data provided by each thermal power plant, the heat storage and extraction margin, and the quotation information reported by each thermal power plant, use data-driven to balance costs and indicators to determine the target units for expected auxiliary peak shaving services.

[0429] For example: Each thermal power plant provides the data in Table 3 of its own to the power grid, and the power grid makes a comparison and selection based on the Table 3 uploaded by each thermal power plant. For example, the power grid learns that there is Plant A with a 95% load and Plant B with a 60% load, and the heat release capacity characterization values of Plant A and Plant B are 70% and 80% respectively. Then, according to the deep learning of the power grid under this operating condition, it is necessary to issue a dispatching instruction focusing on whether the heat release capacity characterization value is prioritized or the load margin is prioritized, and select a unit to generate electricity and connect to the grid. This selection strategy is based on data-driven and historical data-driven deep learning to avoid potential system safety hazards.

[0430] In a possible implementation manner, the power grid side can analyze and study the load increase and decrease change data of thermal power units under various different operating conditions, as well as the correlation data among the load, performance and cost of each thermal power unit, and establish a set of typical load curves after analysis. The peak shaving load curve of the thermal power unit by the power grid side is selected from the pre-established set of typical load curves. The thermal power plant calculates the quotation information for the thermal power unit that provides the auxiliary peak shaving service to meet the requirements of the corresponding peak shaving load curve according to the peak shaving load curve of each thermal power unit. More specifically, it can be to generate the quotation information for the thermal power unit that provides the auxiliary peak shaving service to meet the requirements of the corresponding peak shaving load curve according to the peak shaving load curve of each thermal power unit and the power supply quality. Considering the power supply quality while referring to the peak shaving load curve can make the quotation match the power supply quality and avoid the problem of high quotation but poor power supply quality.

[0431] Figure 5D This is a flowchart for implementing a method for determining a target unit for expected auxiliary peak shaving provided by an embodiment of the present application. Combining Figure 5D In order to determine the target unit for expected auxiliary peak shaving, the technical solution of the present application can combine the ramp-up coefficients and quotation information of each thermal power unit, and through data-driven to balance costs and indicators, select a thermal power unit with an appropriate quotation as the target unit from multiple thermal power units for expected auxiliary peak shaving. As Figure 5D shown, this process includes:

[0432] S5021. The power grid side generates a peak shaving load curve for each thermal power unit based on the change of thermal power load demand, the associated data provided by each thermal power plant, and the heat storage and access margin, and calculates the ramp-up coefficient of each thermal power unit.

[0433] It can be understood that during thermal power generation, since thermal power generation is parallel to new energy generation, the demand for thermal power load may often fluctuate. For example, the demand for thermal power load is relatively low at noon on the same day and relatively high in the evening; the demand for thermal power load is relatively high during rainy days and relatively low on sunny days. By combining the correlation data of load, performance, and cost provided by each thermal power plant, it is possible to analyze and generate the peak shaving load curve for each thermal power unit through feature extraction and other methods using artificial intelligence algorithms. This peak shaving load curve relies on the extraction and analysis of multi-level and multi-step data, realizing a data-driven peak shaving scheduling scheme, and its specific manifestation includes the generated peak shaving load curve.

[0434] Furthermore, in the technical solution of the present application, in order to select a target unit on the grid side, the climbing coefficient of each thermal power unit can be calculated first.

[0435] In the technical solution of the present application, the climbing coefficient is a parameter used to measure the performance of a thermal power unit. The higher the climbing coefficient, the better the performance (climbing ability) of the thermal power unit, and it can better complete the auxiliary peak shaving work. An exemplary calculation method for the climbing coefficient of a thermal power unit is provided below.

[0436] For a certain thermal power unit, in order to calculate its climbing coefficient, the grid side can extract the climbing speed, steam pressure, and rated power generation capacity from the correlation data among the load, performance, and cost of the thermal power unit. Then, according to the climbing speed, steam pressure, rated power generation capacity, heat storage margin of the thermal power unit, and the required climbing time for the thermal power unit by the power plant side, the climbing coefficient of the thermal power unit is calculated. An exemplary formula for calculating the climbing coefficient of a thermal power unit is shown below.

[0437] P = S * T * Q * C / R

[0438] In the formula, P is the ramping coefficient, S is the ramping speed, T is the ramping time required by the power grid, Q is the steam pressure, C is the heat storage margin, and R is the rated power generation capacity. Table 5 exemplarily shows the operating conditions, ramping speeds, ramping times required by the power grid, steam pressures, heat storage margins, and calculated ramping coefficients of three different thermal power units (W1, W2, and W3). Among them, the steam pressure Q is expressed as a percentage of the rated pressure. Combining the above formula, the calculation method of the ramping coefficient of the thermal power unit shown in this formula is as follows: calculate the product of the ramping speed, ramping time, steam pressure, and heat storage margin of the thermal power unit; calculate the ratio of the product to the rated power generation capacity of the thermal power unit, and use the ratio as the ramping coefficient of the thermal power unit. The calculation of the ramping coefficient enables the quantification and measurement of the capacity levels of each thermal power unit and their horizontal comparison with each other. The data-driven method of quantitatively comparing thermal power units enriches the theoretical basis for peak shaving scheduling from the data level and assists in achieving more reliable and orderly high-frequency peak shaving scheduling.

[0439] Taking Table 5 as an example, the heat storage margin in Table 5 is represented by C. As one of the parameters for calculating the ramping coefficient of a thermal power unit, the C value can specifically be used to represent the heat release capacity characterization value in the heat storage margin. Taking the example shown in Table 5, the C values of Unit W1, Unit W2, and Unit W3 are 90%, 50%, and 20% respectively, which means that: the heat release capacity characterization value of Unit W1 is 90%, and the heat storage margin characterization value is 10%; the heat release capacity characterization value of Unit W2 is 50%, and the heat storage margin characterization value is 50%; the heat release capacity characterization value of Unit W3 is 20%, and the heat storage margin characterization value is 80%. That is to say, for the same unit, the sum of its heat release capacity characterization value and heat storage margin characterization value is 100%.

[0440] Table 5

[0441]

[0442] S5022. Initially determine multiple thermal power units expected to assist in peak shaving according to the ramping coefficients of each thermal power unit.

[0443] In the embodiments of the present application, the grid side can initially determine multiple units according to the relative magnitudes of the ramping coefficients of the thermal power units, and expects these units to assist in peak shaving. As shown in the last column of Table 5, different thermal power units with different ramping coefficients can be rated (peak shaving scheduling levels) according to the magnitudes of the ramping coefficients. For example, thermal power units with a ramping coefficient greater than or equal to 0.8 are classified as Class A; thermal power units with a ramping coefficient less than 0.8 but greater than or equal to 0.5 are classified as Class B; thermal power units with a ramping coefficient less than 0.5 are classified as Class C. In this way, the corresponding relationship between thermal power units with different ramping coefficients and peak shaving scheduling levels is established.

[0444] In specific implementation, according to the peak shaving scheduling levels corresponding to each thermal power unit, multiple thermal power units expected to assist in peak shaving are initially determined. For example, on the grid side, thermal power units of Class A can be preferably selected. If the Class A thermal power units are still insufficient for scheduling, thermal power units of Class B can be further selected as the thermal power units expected to assist in peak shaving.

[0445] S5023. Determine whether there is a thermal power unit among the multiple thermal power units expected to assist in peak shaving whose quoted price information is less than or equal to K times the threshold quoted price. If so, proceed to S5024; if not, proceed to S5025.

[0446] S5024. This thermal power unit is determined as the target unit expected to assist in peak shaving.

[0447] S5025. This thermal power unit is not temporarily selected as the target unit expected to assist in peak shaving.

[0448] It can be understood that if each thermal power unit needs to cooperate with the peak shaving service, it may generate heat consumption and energy consumption outside the expected power generation tasks, thus bringing a certain cost expenditure. For this reason, the thermal power unit can calculate the quoted price and report it to the grid side. The grid side compares the quoted price information reported by each thermal power unit with K times the threshold quoted price. If the quoted price is less than or equal to K times the threshold quoted price, the grid side believes that the quoted price is relatively reasonable and acceptable. As shown in step S5024, such units can be designated as target units to assist in peak shaving. On the contrary, if the quoted price is higher than K times the threshold quoted price, the grid side believes that the quoted price is too high and such units are no longer preferably selected as target units.

[0449] Among them, the threshold quoted price is the quoted price reference value obtained by the grid side through deep learning of historical quoted price information. In practical applications, the grid side can comprehensively learn the historical quoted price information in the high-frequency variable load scenarios of thermal power, and combine the historical quoted price information, the performance and cost of the thermal power unit itself, and use data-driven methods to balance costs and indicators to analyze a reasonable quoted price reference value. K is a coefficient greater than 1. In one example, K takes values in the range of 1.1 to 1.3. For example, when K is taken as 1.3, that is, if the quoted price information of the thermal power unit exceeds the threshold quoted price by more than 30%, it is considered not eligible to be selected as a target unit. On the contrary, if the quoted price information of the thermal power unit exceeds the threshold quoted price by no more than 30%, it is considered eligible to be selected as a target unit.

[0450] For the target unit, the grid side can send a peak shaving scheduling instruction to the thermal power unit to which the unit belongs.

[0451] S503. Receive the peak shaving scheduling instruction issued by the grid side.

[0452] The peak shaving scheduling instruction carries the unit identifier of the target unit and the peak shaving requirement information for the target unit. For example, if the peak shaving scheduling instruction carries the unit identifier W1, after receiving this peak shaving scheduling instruction, the thermal power plant can know that this instruction is issued for the thermal power unit with the unit identifier W1. If a thermal power plant contains multiple target units, the grid side can send the peak shaving scheduling instructions for each thermal power unit in parallel. This group-based scheduling improves the peak shaving efficiency. Since the peak shaving scheduling instruction carries the unit identifier, the thermal power plant can accurately forward the peak shaving requirement information.

[0453] S504. Execute the auxiliary peak shaving service based on the peak shaving requirement information.

[0454] As mentioned above, the peak shaving scheduling instruction carries the peak shaving requirement information for the target unit. Specifically, the peak shaving requirement information includes the peak shaving load curve. In a possible implementation, the peak shaving requirement information further includes the ramp rate requirement information. The peak shaving load curve sets requirements for the target unit from the perspective of load, while the ramp rate requirement information sets requirements for the target unit from the level of ramp rate. When the target unit receives the peak shaving requirement information, it needs to make a judgment in combination with its actual situation to confirm whether it can meet the above requirements.

[0455] Figure 5E This is a flowchart of a fault judgment and operating condition judgment provided for the embodiments of the present application. The following combines Figure 5E to exemplarily introduce a possible implementation of this step.

[0456] As Figure 5E shown, after identifying the peak shaving requirement information in the peak shaving scheduling instruction, first, the target unit of the thermal power plant judges whether there are equipment failures and communication failures, and judges whether the target unit meets the ramp rate requirement information. Specifically, it can be as Figure 5E shown, first judge whether there are equipment failures and communication failures. After the judgment is completed, on the premise that there are no equipment failures and no communication failures, further judge whether the ramp rate requirement information is met. If there are no equipment failures and no communication failures in the target unit (that is, neither equipment failures nor communication failures), and the target unit meets the ramp rate requirement information, then execute the auxiliary peak shaving service according to the peak shaving load curve and the ramp rate requirement information in the peak shaving requirement information. If there are equipment failures or communication failures in the target unit, it means that from the hardware conditions or communication conditions, this target unit cannot cooperate to complete the auxiliary peak shaving service. Then the thermal power plant needs to report the fault message of this thermal power unit to the grid side, so that the grid side can timely learn about the fault status of this target unit. If the target unit does not meet the ramp rate requirement information, a prompt message indicating that the operating condition is not met is reported to the grid side.

[0457] It should be noted that in the embodiments of the present application, the grid side can respond to the prompt information of the non - meeting operating conditions fed back by the thermal power plant and adjust the peak - shaving scheduling instruction. For example, according to the prompt information of the non - meeting operating conditions, adjust the peak - shaving load curve in the peak - shaving requirement information and / or adjust the value in the ramp - rate requirement information. By adjusting the peak - shaving load curve (for example, reducing the ordinate value of the load curve) or reducing the value in the ramp - rate requirement information, the threshold requirement for the target unit can be reduced, enabling more units to become the target units serving peak - shaving scheduling.

[0458] In addition, in the embodiments of the present application, the grid side can also re - learn according to the prompt information of the non - meeting operating conditions fed back by each thermal power plant to increase the threshold bid, so that more units can be used as the target units for expected auxiliary peak - shaving. Incorporating more units into the queue of target units for expected auxiliary peak - shaving expands the range of selected units and avoids the problem of insufficient capacity to provide auxiliary peak - shaving services among units with appropriate bids. Incorporating more units into the queue of target units allows more units to have the opportunity to participate in auxiliary peak - shaving work and also avoids the problem that due to the restriction of the transaction price, the thermal power units with better performance cannot exert their full capacity in auxiliary peak - shaving work. Continuously deep - learning the threshold bid enables the threshold bid to be updated in a timely manner according to the conditions and bids of thermal power units, making peak - shaving scheduling dynamic and flexible.

[0459] As mentioned above, in the present application, thermal power units, especially the selected target units, should cooperate with the grid - side scheduling to assist in peak - shaving. In the peak - shaving scheduling method in the high - frequency variable - load scenario of thermal power provided in the present application, after the grid side preliminarily determines multiple thermal power units for expected auxiliary peak - shaving, the method further includes:

[0460] The grid side sends a heat pre - scheduling instruction to the multiple thermal power units for expected auxiliary peak - shaving. The purpose of sending the heat pre - scheduling instruction is to enable each thermal power unit for expected auxiliary peak - shaving to make preparations for the upcoming load - adjustment work in terms of heat. Specifically, the heat pre - scheduling may require the energy storage device to store more heat or release some heat. The heat pre - scheduling instruction can be sent by the grid side to the thermal power unit after knowing the heat storage and release margin of the thermal power unit. On the premise of knowing the heat storage and release margin, when the grid side conducts peak - shaving scheduling and issues the heat pre - scheduling instruction, the content of the issued heat pre - scheduling instruction and the heat adjustment amplitude can be made more compatible with the performance and capacity of the thermal power unit.

[0461] As mentioned earlier, the grid side will determine the target unit for the expected auxiliary peak load regulation based on the quotation information. For the target unit, it is considered that the thermal power plant and the grid side have reached an agreement on the quotation. The target unit selected by the grid side will cooperate with the heat storage device in the thermal power plant to which it belongs to to perform heat pre-dispatch after receiving the heat pre-dispatch instruction. For example, the heat pre-dispatch instruction requires the target unit to further store 12% of heat, and the target unit further stores 12% of heat according to the heat pre-dispatch instruction. Therefore, when auxiliary peak load regulation service is required, the corresponding load adjustment can be completed through the pre-stored heat.

[0462] Figure 5C The peak load dispatching method process under the high-frequency load variation scenario of thermal power is presented, showing the interaction content and interaction sequence from both the thermal power plant and the power grid side. Figure 5C As shown, the thermal power plant selects the type of coal and calculates the heat storage and access margin. The calculated heat storage and access margin and the associated data (specifically, the associated data between the load, performance and cost of each thermal power unit in the thermal power plant) are sent to the power grid side. The specific types of associated data can be referred to Table 3, Table 4, etc. The power grid side digitally processes the data received from each thermal power plant, performs multi-level abstract extraction and analysis on the data, and uses artificial intelligence technology to capture the connection between the data. Then, based on the above extraction and analysis, the power grid side generates a peak load curve for each thermal power unit and sends the curve to the corresponding thermal power plant. In addition, heat pre-dispatch instructions can also be issued. Based on the actual situation of each thermal power unit, the thermal power plant calculates the quotation information that meets the requirements of the peak load curve to assist in peak regulation, and reports the quotation information to the power grid side. At this point, the power grid side can obtain the corresponding quotation information of each thermal power unit in each thermal power plant, and compares it horizontally in combination with the quotation information and the calculated climbing coefficient of the thermal power unit, and selects the target unit by weighing the cost and indicators through data drive. The power grid issues peak-shaving dispatch instructions to the thermal power plants, and the target units cooperate to execute heat pre-dispatching and perform auxiliary peak-shaving services based on the peak-shaving requirement information in the peak-shaving dispatch instructions.

[0463] The technical solution of this application has the following outstanding advantages:

[0464] As the proportion of renewable energy generation in the power system increases, thermal power plants, from design to operation, must pay attention to the storage of heat to meet the needs of auxiliary peak-shaving services for the power grid. The quotation of auxiliary peak-shaving services needs to be market-oriented and coordinated. To this end, the technical solution of this application can meet the needs of market-oriented coordination through the above-mentioned method.

[0465] In addition, in the technical solution of this application, data analysis is introduced into the market-oriented scheduling of high-frequency load changes in thermal power. Through the quantitative Tables 2 to 4, the grid side (combining Tables 3 to 4) can perform digital scheduling, and the power plant side (combining Table 2) can perform digital coal selection. The key lies in feature extraction based on data. The data-driven scheduling process makes the execution of peak shaving scheduling more standardized and digital, with better collaborative effects, making the costs clearer. The grid can rank and set priorities based on the performance conditions reflected by the costs and indicators, so as to better select the units suitable for peak shaving based on data as a theoretical basis, making the peak shaving effect more stable, safe, and ideal.

[0466] Moreover, the technical solution of this application also considers grid security such as the ramp coefficient and the peak shaving load curve. Such key data can all guide the pricing of power quality, with better economy. The pricing of high-frequency peak shaving scheduling and auxiliary peak shaving services takes into account the costs of the grid and power plants under marketization.

[0467] Through the collection and digital tool processing of the natural production factors of thermal power in this embodiment, their data characteristics are abstracted. Then, according to the thermal power price formula, the heat consumption, coal consumption, and electricity charge measurement models of thermal power under full load conditions are obtained. Through the digital research of the models, data-driven digital logic can be involved, which is applied to the power industry to digitalize thermal power scheduling and operation. At the same time, by studying the control production factors on the grid side, such as electricity prices, the production factors of thermal power are associated with electricity prices and complex production factors. By setting low-order, medium-order, and high-order data-abstracted production factors, the intermediate results or production factors may not have physical meanings in reality, such as sub-frequencies and low-voltage overexcitation. Then these data-abstracted production factors are the keys to unlocking artificial intelligence (digital) power.

[0468] Figure 5F It is a flowchart of another peak shaving scheduling method in the scenario of high-frequency load changes in thermal power provided by the embodiment of this application. The following Figure 5F will be briefly described.

[0469] As Figure 5F shown in the process, the thermal power plant first selects coal types in combination with the hydrogen content of each coal type, then calculates the heat storage margin of the thermal power plant, and digitalizes the associated data among the loads, performances, and costs of each thermal power unit in the thermal power plant. The thermal power plant sends this data to the grid side, and the grid uses data-driven methods to weigh costs and indicators to determine the target units for expected auxiliary peak shaving.

[0470] In specific implementation, the grid side generates the peak shaving load curves for each thermal power unit based on the changes in thermal power load demand and the corresponding data provided by each thermal power plant, and sends them to the thermal power units. And calculate the ramp-up coefficients of each thermal power unit, preliminarily determine multiple thermal power units expected to provide auxiliary peak shaving, and send them heat pre-scheduling instructions. The thermal power plant calculates the price of the auxiliary peak shaving service based on the peak shaving load curve and sends it to the grid side. The grid side continues to consider cost factors and the performance data of the thermal power plant, and measures around the price information and the K-fold threshold price to select the target units. If all the prices exceed the K-fold threshold price, it means that the threshold price is too low, which means that it is necessary to re-learn the threshold price to increase the passing rate of the thermal power units with the load price threshold.

[0471] The target units cooperate with the heat pre-scheduling, and after the pre-scheduling is completed, report the heat storage to the grid again. The grid side judges whether there are equipment failures and communication failures in the target units, and judges whether the ramp-up speed requirement information is met. If there are equipment failures or communication failures, it needs to be reported to the grid; if there is a situation where the target units do not meet the ramp-up speed requirement information, it also needs to be reported to the grid. As Figure 5F shown by the dotted line in the figure, for the situation where the working conditions are not met, measures such as re-learning the threshold price can be taken, or the peak shaving requirement information in the peak shaving scheduling instructions sent to the target units can be adjusted.

[0472] Example 4

[0473] In Embodiment 4, the power function requirements for implementing the dispatching control of the grid side over the power plant side are specifically the water resource dispatching of cascade hydropower stations. The functional operator is the water resource dispatching operator, and the water resource dispatching operator is used to implement the water resource dispatching of cascade hydropower stations. Embodiment 4 introduces the process of applying the intelligent agent constructed based on the water resource dispatching operator in the digital power system.

[0474] There are a series of issues that need to be considered in the water resource dispatching of cascade hydropower stations. For example, cascade hydropower stations are generally composed of multiple single-stage power stations connected in series. The upper and lower hydropower stations are connected through reservoir backwater. The control range can extend from the confluence to the entire basin. The risks of some out-of-control situations in the upstream power stations are transmitted through water flow fluctuations, superimposed and accumulated in the downstream hydropower stations, which may cause serious risk events; another example is that during the operation of each hydropower station in the cascade hydropower station, the reservoir safety (the water level should be appropriate) and the unit safety of each hydropower station need to be considered; another example is that during the operation of each hydropower station in the cascade hydropower station, the unit operation safety caused by unexpected grid events (such as failure to receive power from the interconnection point and sudden load shedding) needs to be considered; another example is how to improve the power supply quality, promote the safe operation of the grid, improve the power supply revenue, and realize the full utilization of water resources on the premise of reservoir safety and the safety of hydropower generation equipment.

[0475] Based on the above problems, the present application proposes a water resource scheduling method for cascade hydropower stations by using a water resource scheduling operator in a digital power system based on multi-functional agents. The implementation of this method includes: obtaining the water resource data of each hydropower station in the cascade hydropower station at the current moment; for each hydropower station, judging whether the hydropower station is in a safe operation state according to the water resource data of the hydropower station; when it is determined that each hydropower station is in a safe operation state, judging whether to perform water resource scheduling according to the water resource data of each hydropower station, the preset electricity price plan and the target historical data; if it is determined to perform water resource scheduling, determining a water resource scheduling plan based on the water resource data of each hydropower station and the water resource scheduling model. Through the solution in the present application, when it is determined that each hydropower station is in a safe operation state, a water resource scheduling plan that can obtain as high a power generation benefit as possible can be directly generated based on the water resource scheduling model trained from historical data and the water resource data of each hydropower station at the current moment, realizing the reasonable scheduling of water resources.

[0476] Figure 6A It is a flowchart of a water resource scheduling method provided by an embodiment of the present application. Combining Figure 6A As shown, the disclosed water resource scheduling method of the present application includes:

[0477] S601, obtaining the water resource data of each hydropower station in the cascade hydropower station at the current moment.

[0478] The water resource data of the hydropower station in the present application includes but is not limited to: the measured water level of the reservoir where the hydropower station is located (abbreviation: "measured water level"), the measured speed rise rate of the water flow velocity in the reservoir where the hydropower station is located (abbreviation: "measured speed rise rate"), the measured turbine static pressure (also known as "measured turbine static water pressure"), the rated speed of the hydrogenerator, the working power of the hydrogenerator, the installed capacity of the power station, the maximum / rated / minimum water head, the maximum runaway speed, the rated output of the water turbine, and the penstock diameter, etc.

[0479] Table 6 is a table of the inflow of the upstream hydropower station provided by the present application. Table 6 gives the inflow of the upstream hydropower station from January to December in different years, and the unit of the water volume is 10,000 m 3 .

[0480] For example, the following data can be read from Table 6: the inflow of the upstream hydropower station in January 2016 was 7.75 million m 3 ; the inflow of the upstream hydropower station in February 2023 was 12.57 million m 3 .

[0481] It should be noted that the upstream hydropower station refers to the hydropower station in the upstream position in the cascade hydropower station.

[0482] Table 6

[0483] Month 2016 2017 2018 2019 2020 2023 2024 January 775 3227 3263 3136 1750 2894 1274 February 210 1114 2476 2305 746 1257 3128 March 106 1650 3324 736 1939 2478 798 April 2047 13740 6802 3011 1410 3692 1784 May 8280 23682 35177 3642 5481 5304 2478 June 11174 22343 52766 14295 8700 16633 10618 July 35240 19435 90578 25659 16089 43929 83502 August 30913 22702 115595 73712 33239 49479 25684 September 23290 18653 38499 84177 46228 56418 34185 October 40846 26580 35633 21007 54775 37106 26915 November 16734 11114 10911 7187 12512 18177 None December 6348 5774 4833 2386 2870 4250 None

[0484] Table 7

[0485]

[0486] Table 7 is a table showing the water volume changes of an upstream hydropower station provided in this application. The table shows the daily water volume changes of the upstream hydropower station from January 23 to January 30. Specifically, it includes: current water storage volume (m 3 ), current reservoir energy storage (100 million kWh), current water level (m), average inflow discharge yesterday (m 3 / s), average outflow discharge yesterday (m 3 / s), and water consumption rate for power generation yesterday (m 3 / kWh).

[0487] S602. For each of the hydropower stations, based on the water resource data of the hydropower station, determine whether the hydropower station is in a safe operating state.

[0488] In this application, the safe operating state mainly refers to the water conservation safety state and the vibration safety state.

[0489] When determining whether a hydropower station is in a water conservation safety state in this application, two indicators, namely the measured turbine static water pressure and the measured speed rise rate, are mainly considered.

[0490] Among them, the measured turbine static water pressure, also known as the volute pressure, refers to the actual net water pressure at the inlet of the turbine volute; it is usually affected by the water flow velocity, water flow direction, turbine structure, and reservoir water level. The measured turbine static water pressure is of great significance for evaluating the operating efficiency of the turbine, predicting the wear of the turbine, and formulating the turbine maintenance plan.

[0491] Among them, the measured speed rise rate refers to the actual rising speed of the reservoir water level within a certain period of time, expressed as the water level change amount per unit time; it is usually affected by the inflow water volume, outflow water volume, rainfall situation, and reservoir regulation capacity. The measured speed rise rate is of great significance for evaluating the water storage capacity of the reservoir, predicting future water level changes, and formulating the reservoir operation plan.

[0492] Generally, both the measured speed rise rate and the measured turbine static pressure can be obtained through corresponding sensors; the theoretical speed rise rate and the theoretical turbine static water pressure can be obtained from the technical documentation of the hydropower station.

[0493] It should be emphasized that the theoretical speed rise rate in this application is the maximum actual rising speed of the reservoir water level allowed during the safe operation of the hydropower station within a certain period of time; the theoretical turbine static water pressure in this application is the maximum actual net water pressure allowed at the inlet of the turbine volute during the safe operation of the hydropower station.

[0494] Table 8

[0495]

[0496]

[0497] Table 8 is a table of the spiral case pressures measured on different dates provided by this application. In Table 8, the unit output (MW), upstream water level (m), downstream water level (m), spiral case pressure (Mpa) of each unit, and the values of guide vane opening detected at 09:04 am, 15:00 pm, and 18:25 pm on August 18, 2018, as well as at 09:25 am and 13:50 pm on August 19, 2018 are given.

[0498] In an alternative implementation, it is possible to determine whether each hydropower station in a cascade hydropower station is in a water conservation and safety state through the following steps, specifically:

[0499] For each hydropower station, obtain the measured speed increase rate and the measured turbine static water pressure from the water resource data of this hydropower station; then, compare the measured increase rate of this hydropower station with the theoretical speed increase rate, and compare the measured turbine static water pressure of this hydropower station with the theoretical turbine static water pressure.

[0500] If the measured speed increase rate is less than the theoretical speed increase rate, and the measured turbine static water pressure is less than the theoretical turbine static water pressure, then determine that this hydropower station is in a water conservation and safety state; if the measured speed increase rate is greater than or equal to the theoretical speed increase rate, or the measured turbine static water pressure is greater than or equal to the theoretical turbine static water pressure, then determine that this hydropower station is not in a water conservation and safety state.

[0501] It can be understood that equipment in a hydropower station, such as a hydrogenerator, will generate vibrations during operation. The amplitude and frequency of these vibrations are important indicators for evaluating the operating condition of the equipment. Excessive vibrations may mean that there are problems such as wear, looseness, imbalance, or other potential issues with the equipment. In this application, when determining whether a hydropower station is in a vibration safety state, the actual working power of the hydroengine and the vibration conditions of each component of the hydroengine are mainly considered.

[0502] In an alternative implementation, it is possible to determine whether each hydropower station in a cascade hydropower station is in a vibration safety state through the following method, specifically:

[0503] The first step: For each hydropower station, obtain the power (actual working power) of the hydroengine from the water resource data of this hydropower station, denoted as the first power.

[0504] It should be noted that during the process of the power of the water turbine engine rising from 0 to the maximum power (full load power), in some power intervals (referred to as preset power intervals in this application), due to various reasons such as frequency resonance, there will be a situation where the vibration amounts of various components of the water turbine engine are very large. To avoid damage to the water turbine engine due to excessive vibration or vibration-triggered protection tripping, the actual working power of the water turbine engine at the current moment (i.e., the first power in this application) will be concerned in this application.

[0505] Second step, determine whether the first power is within the preset power interval.

[0506] It should be noted that the preset power interval can be obtained through a variable power vibration measurement test.

[0507] Third step, if the first power is within the preset power interval, quickly adjust the first power to the second power; where the second power is outside the preset power interval.

[0508] If it is determined that the first power is within the preset power interval, it means that the vibration amounts of various parts of the current water turbine generator are relatively large, and it is easy to cause damage to the water turbine generator. At this time, quickly adjust the working power of the water turbine generator. For example, the first power can be quickly adjusted to the second power outside the preset power interval to avoid excessive vibration of the water turbine generator due to reasons such as frequency resonance.

[0509] Fourth step, if the first power is outside the preset power interval, determine whether the hydropower station is in a vibration-safe state.

[0510] If it is determined that the first power is outside the preset power interval, it can be determined that the water turbine generator will not vibrate excessively due to reasons such as frequency resonance at this first power. At this time, it is necessary to further determine whether the hydropower station is in a vibration-safe state.

[0511] In an alternative implementation, the process of determining whether the hydropower station is in a vibration-safe state is as follows:

[0512] First, for each hydropower station, obtain the rotational speed of the water turbine engine of the hydropower station and the actual vibration values of different parts of the water turbine engine from the water resource data of the hydropower station.

[0513] Then, after determining the actual rotational speed of the water turbine engine, compare the actual vibration values of different parts of the water turbine engine measured at this rotational speed with the corresponding vibration allowable values of these parts.

[0514] For each hydropower station, if the actual vibration values of different parts of the water turbine engine at the current rotational speed of the water turbine engine are all less than the corresponding vibration allowable values, it is determined that the hydropower station is in a vibration-safe state; if the actual vibration value of a certain part of the water turbine engine is greater than or equal to the corresponding vibration allowable value of this part, it is determined that the hydropower station is not in a vibration-safe state and there is a vibration risk.

[0515] Table 9

[0516]

[0517] Table 9 is a table of the vibration allowable values for each part of a hydro-generator provided by this application. As shown in Table 9, the water turbine engine has different vibration allowable values at different rotational speeds. For example, when the rotational speed is less than 100 r / min, the vibration allowable value for the vertical vibration of the vertical unit of the hydro-generator with a thrust bearing bracket is 0.1 mm. When the actual vibration value of the vertical vibration of the thrust bearing bracket measured at a rotational speed less than 100 r / min is greater than 0.1 mm, it is considered that there is a risk of equipment vibration in the vertical unit of the hydro-generator.

[0518] When it is determined that each hydropower station in the cascade hydropower stations is in a water conservation safety state and a vibration safety state, that is, when it is determined that each hydropower station is in a safe operation state, the steps of the subsequent embodiments of this application can be continued.

[0519] It can be understood that if it is determined that each hydropower station in the cascade hydropower stations is not in a safe operation state and there are water conservation safety problems or vibration safety risks, it is necessary to stop the machine for maintenance. After the problems are repaired, the power generation facilities in each hydropower station are restarted; after restarting, after verifying that each hydropower station is in a safe operation state, the steps of the subsequent embodiments of this application are executed.

[0520] S603. When it is determined that each of the hydropower stations is in the safe operation state, based on the water resource data of each of the hydropower stations, a preset electricity price plan, and target historical data, determine whether to perform water resource scheduling.

[0521] It can be understood that for hydropower generation manufacturers, when they can safely operate the power generation equipment of the power plant, it is necessary to precisely adjust the water level of the reservoir corresponding to the power station. In this way, higher power generation benefits can be obtained while meeting the power consumption requirements of the power grid.

[0522] Specifically, first, based on the water resource data of each hydropower station and the preset electricity price plan, determine the actual power generation water consumption and the actual power plant revenue of each hydropower station as the current actual data of the hydropower station. Then, based on the difference between the current actual data and the target historical data, determine whether water resource scheduling is required. Among them, the target historical data includes the average power generation water consumption and the average power plant revenue in the month where the current moment is located.

[0523] If the difference between the current actual data and the target historical data is too large and exceeds the preset difference threshold, it is determined that water resource scheduling is required; otherwise, it is determined that water resource scheduling is not required. In this application, the specific value of the preset difference threshold is not limited, and those skilled in the art can set the preset difference threshold according to actual needs.

[0524] In the subsequent embodiments of this application, the acquisition process of the current actual data and the acquisition process of the target historical data will be introduced in detail.

[0525] In an alternative implementation, the following method can be used to determine the actual power generation water consumption and the actual power plant revenue of each hydropower station based on the water resource data of each hydropower station and the preset electricity price plan as the current actual data of the hydropower station. Specifically:

[0526] In the first step, for each hydropower station, obtain the actual power generation and the actual water consumption in the target time period from the water resource data of the hydropower station. Among them, the target time period is a period of time in the past with the current moment as the end point.

[0527] For example, from the water resource data of Hydropower Station A obtained at the current moment, obtain the actual power generation and the actual water consumption in the past 1 hour. Among them, the past 1 hour is the target time period.

[0528] Among them, the actual power generation and the actual water consumption can both be obtained through corresponding detection devices. For example, the actual water consumption in the target time can be measured by an ultrasonic flow tester.

[0529] In the second step, for each hydropower station, based on the actual power generation and the actual water consumption in the target time period, obtain the actual power generation water consumption in the target time period.

[0530] Exemplarily, after obtaining the actual water consumption and the actual power generation of Hydropower Station A in the most recent 1 hour, divide the actual water consumption by the actual power generation and then divide by the target time period to obtain the actual power generation water consumption.

[0531] The actual power generation water consumption in this application reflects the characteristic of the ability of a hydropower station to require water resources per unit time and per unit power generation. The actual power generation water consumption is high-order abstract data, which can be compared with the specific heat index of a thermal power station and the wind resource utilization coefficient of a wind power station at a high order, so as to judge the degree and performance of the utilization of resources by a power source node; it reflects the specific function of the power station's resource energy consumption.

[0532] In the third step, for each hydropower station, based on the corresponding sectional electricity price plan for the target time period and the actual water level of the hydropower station, determine the actual power plant revenue for the target time period.

[0533] The sectional electricity price plan (abbreviated as the electricity price plan) in this application is an electricity fee plan related to the month and the reservoir water volume.

[0534] For the convenience of understanding, taking the sectional electricity price plan corresponding to a cascade hydropower station including an upstream power station (also called an upstream reservoir power station) and a downstream power station (also called a downstream reservoir power station) as an example, introduce the specific form of the sectional electricity price plan:

[0535] (a) Upstream reservoir power station:

[0536] The upstream reservoir power station implements quarterly regulation and prices according to different water levels in different seasons.

[0537] From January to March, the water level is between 250 - 260 m above sea level. If the water level of the upstream reservoir is above 255 m, the on-grid electricity price is J - 0.001 / m; if the water level of the upstream reservoir is below 255 m, the on-grid electricity price is J + 0.001 / m; the electricity price charging method for other water levels within the interval is: if the water level is higher than the upper limit, it is priced according to the upper limit, and if the water level is lower than the lower limit, it is priced according to the lower limit; where J in this application is the preset basic value of the on-grid electricity price.

[0538] From April to June, the water level is between 210 - 250 m above sea level. If the water level of the upstream reservoir is above 230 m, the on-grid electricity price is J - 0.001 / m; if the water level of the upstream reservoir is below 230 m, the on-grid electricity price is J + 0.001 / m; the electricity price charging method for other water levels within the interval is: if the water level is higher than the upper limit, it is priced according to the upper limit, and if the water level is lower than the lower limit.

[0539] From July to September, the water level is between 250 - 260 m above sea level. If the water level of the upstream reservoir is above 255 m, the on-grid electricity price is J - 0.001 / m; if the water level of the upstream reservoir is below 255 m, the on-grid electricity price is J + 0.001 / m; the electricity price charging method for other water levels within the interval is: if the water level is higher than the upper limit, it is priced according to the upper limit, and if the water level is lower than the lower limit.

[0540] From October to December, the water level is between 258 - 262 m above sea level. If the water level of the upstream reservoir is above 260 m, the on-grid electricity price is J - 0.001 / m; if the water level of the upstream reservoir is below 260 m, the on-grid electricity price is J + 0.001 / m; for other water levels within the range, the electricity price charging method is: if the water level is higher than the upper limit, it is priced according to the upper limit, and if the water level is lower than the lower limit, it is priced according to the lower limit.

[0541] When the water levels of the upstream and downstream reservoirs are between 262 - 265 m and 115 - 120 m, the electricity price is the full electricity price of $0.073 per degree. When the water levels of the upstream and downstream reservoirs are below 235 m and 100 m, they are safety water levels and the turbines cannot generate electricity when stopped.

[0542] Table 10

[0543]

[0544] Table 10 is a table showing the relationship between the reservoir water level and electricity price of an upstream hydropower station provided by this application. The upstream reservoir power station can directly give the price increase or decrease according to the water level of each quarter according to the content shown in Table 10, and obtain the actual electricity price PR_actual corresponding to the month at the current moment.

[0545] In Table 10, the water level is represented by the letter L; the units of both the water level and the median water level are meters; the unit of the actual electricity price PR_actual is ($ / degree). When the month is from January to March, the water level is in the range of 250 - 260 meters, and the median water level is 255 meters; within this month range, if the water level L is less than the median water level of 255 meters, the value of PR_actual is generated based on the full electricity price of $0.073 per degree and the floating electricity price of $0.001. In each month of Table 10, when the water level L is greater than the median water level, PR_actual is ($0.073 + $0.001) per degree; when the water level L is less than the median water level, PR_actual is ($0.073 - $0.001) per degree.

[0546] In this application, by analyzing the water resource data, it is found that the data of each quarter has the characteristic of clustering, that is, the quarterly water levels in Table 10 show a segmented characteristic, which reflects the hydrological and water level characteristics of the hydropower station. Only by making better use of the natural production factors (referred to as the primary indicators in this application) can the efficiency of water resource utilization be maximized and the market-oriented analysis of the water use cost of hydropower be carried out. Therefore, in this application, the median water level is targeted, and by comparing with the median value of the historical water level (the historical water level is one of the data samples for deep learning), the measured incoming water volume and the amount of water stored in the reservoir are judged to determine a reasonable power generation electricity price plan.

[0547] It is understandable that in the analysis of historical water resource data or other data, if other patterns are found, it may further guide the water level scheduling strategy adopted in this application, further improve the underlying logic emphasized in this application, which is data-based and drives the scheduling agent, and avoid errors caused by artificially abstracting data and specifying water resource scheduling plans.

[0548] (b) Downstream reservoir power station:

[0549] The downstream reservoir implements day-ahead scheduling, regardless of season.

[0550] The water level is between 105 - 110 m above sea level. If the water level of the downstream reservoir is above 107 m, the on-grid electricity price is J - 0.001 / m. If the water level of the downstream reservoir is below 107 m; the electricity price billing method for other water levels within the range is: the water level above the upper limit is priced according to the upper limit, and the water level below the lower limit is priced according to the lower limit.

[0551] Table 11 is a table showing the relationship between the reservoir water level and electricity price of a downstream hydropower station provided in this application. The downstream reservoir power station can directly give the increased or decreased price according to the water level in each quarter as shown in Table 11 to obtain the actual electricity price PR_actual corresponding to the current month. In Table 11, the water level is represented by the letter L, and the units of both the water level and the water level median are meters, and the unit of the actual electricity price PR_actual is (US dollars / degree). In Table 11, when the water level L is greater than the water level median, PR_actual is (0.073 + 0.001) (US dollars / degree); when the water level L is less than the water level median, PR_actual is (0.073 - 0.001) (US dollars / degree).

[0552] Table 11

[0553]

[0554] For each hydropower station in the cascade hydropower station, after obtaining the corresponding segmented electricity price plan for the target time period and the actual water level of the hydropower station, the actual power plant revenue for the target time period can be determined based on the following formula. The target formula is:

[0555]

[0556]

[0557] Q3 = PR 实 ×P 实 (2.3)

[0558] The meanings of each letter in Formula (2.1), Formula (2.2) and Formula (2.3) are as follows:

[0559] Q1 represents the fee that the power grid should pay to the hydropower station when there is a power grid failure and the hydropower station cannot generate electricity normally (take-or-pay); Q2 represents the actual revenue corresponding to the potential electric energy at the current moment; Q3 represents the actual power generation revenue corresponding to the actual generated electric energy at the current moment. The sum value of formula (2.1) and formula (2.3) represents the actual power plant revenue of the hydropower station; the sum value of formula (2.1), formula (2.2) and formula (2.3) represents the water assets of the hydropower station.

[0560] D represents the number of hydro-generators included in the hydropower station, and D is a positive integer greater than 1; PR_actual is the actual electricity price corresponding to the current month; PR_average represents the average electricity price within the target time period (i.e., within the target pricing time period); i is a hydro-generator in the hydropower station, i = 1, 2... D; P_actual is the power generation amount (actual electric energy) within the target time period; G_i_nominal is the rated power of the i-th hydro-generator in the hydropower station; H_i is the continuous power generation time of the i-th hydro-generator in the hydropower station at the rated power within the target pricing time period; F_i_J is the derated power of the i-th hydro-generator in the hydropower station; H_i_z is the interruption duration during the high water level period of the i-th hydro-generator in the hydropower station.

[0561] Among them, the high water level is the water level at which the hydropower station can generate electricity. If it is stipulated in the contract that the hydropower station can generate electricity when the water level of the hydropower station is within the range of 240 - 265m, then 240 - 265m is the high water level in this application.

[0562] Through deep learning of water level data, the changing rules of seasonal water levels can be discovered and extended research can be carried out; general rules can be obtained through the data characteristics of water levels (or other quantities); through the target formula or generalized electricity price scheme in the foregoing embodiments, based on basic data such as electricity charges and power generation water consumption, associated data such as power plant power generation revenue can be finally obtained; then, the characteristics and correlation relationships of data such as water levels, electricity charges, power plant power generation revenue, power plant power generation costs, and power generation water consumption are extracted to establish a model of the median of water levels and electricity prices; through deep learning and correlation research, a reasonable scheduling strategy for hydropower station revenue and water resource assets is obtained. Among them, the scheduling strategy is not limited to upstream quarterly power generation.

[0563] Table 12 is the data calculation table of the cascade hydropower stations provided by this application. The cascade hydropower stations in Table 12 include an upstream hydropower station and a downstream hydropower station, a total of two hydropower stations. The data in Table 12 are the data calculated based on the water resource data of each hydropower station in the cascade hydropower station at the current moment, formula (2.1), formula (2.2) and formula (2.3).

[0564] Those skilled in the art can calculate the power generation benefits for other months in 2023 by referring to the electricity bills for different quarters and water levels provided in the foregoing embodiments, Formula (2.1), Formula (2.2), and Formula (2.3).

[0565] It should be noted that in Table 12, the upper dam refers to the upstream reservoir; the lower dam refers to the downstream reservoir; the upper power plant refers to the power plant building corresponding to the upstream reservoir; the lower power plant refers to the power plant building corresponding to the downstream reservoir.

[0566] The unit of water volume in Table 12 is 10,000 m 3 ; the unit of water level is m; the unit of flow rate is m 3 / s; the unit of rainfall is mm; the unit of electricity is 10,000 kw.h. For the meanings of the letters in Table 6, refer to the description in the foregoing embodiments, which will not be elaborated here.

[0567] Table 12

[0568]

[0569]

[0570] In an alternative implementation, the method for obtaining the target historical data of each hydropower station is as follows:

[0571] First, count the historical power generation water consumption and historical power plant benefits for each month in the past N years. Here, N is an integer greater than or equal to 1.

[0572] Exemplarily, the water inflow of Hydropower Station A in a specific month, such as January 2024, can be obtained; the reservoir capacity of Hydropower Station A in January 2024 can be obtained; by subtracting the reservoir capacity from the water inflow, the water consumption of Hydropower Station A in January 2024 can be obtained. Count the power generation and power generation duration of Hydropower Station A in January 2024; then divide the water consumption of Hydropower Station A in January 2024 by the power generation of Hydropower Station A in January 2024 and then divide by the power generation duration of Hydropower Station A in January 2024 to obtain the historical power generation water consumption of Hydropower Station A in January 2024.

[0573] Figure 6B This is a schematic diagram of the historical power generation water consumption corresponding to each month of an upstream hydropower station provided by the embodiments of the present application. Figure 6B The unit of the abscissa in it is cubic meters per kilowatt-hour; the unit of the ordinate is month. Figure 6B It shows the historical power generation water consumption for each month from January 2023 to December 2023 and from January 2024 to December 2024 of the upstream hydropower station. Among them, the data for November 2024 and December 2024 are the historical power generation water consumption calculated based on historical data.

[0574] Figure 6C This is a schematic diagram of the historical power generation water consumption corresponding to each month of a downstream hydropower station provided by an embodiment of the present application. Figure 6C The unit of the abscissa in it is cubic meters per kilowatt-hour; the unit of the ordinate is month. Figure 6C It shows the historical power generation water consumption of each month from January 2023 to December 2023 and from January 2024 to December 2024 of the downstream hydropower station. Among them, the data for November and December 2024 are the historical power generation water consumption calculated based on historical data.

[0575] Based on Figure 6B and Figure 6C the comparison of the power generation water consumption, it can be determined that it is more suitable to implement quarterly regulation upstream and day-ahead regulation (also known as day-ahead dispatching) downstream.

[0576] As introduced in the foregoing embodiments, the sum of formula (2.1) and formula (2.3) represents the actual power plant revenue of the hydropower station. Therefore, substituting the historical relevant data into formula (2.1) and formula (2.3) can obtain the historical power plant revenue.

[0577] Figure 6D This is a schematic diagram of the historical power plant revenue corresponding to each month of an upstream hydropower station provided by an embodiment of the present application. Figure 6D The unit of the abscissa in it is US dollars; the ordinate is month. Figure 6D It shows the historical power plant revenue of each month from January 2023 to December 2023 and from January 2024 to December 2024 of the upstream hydropower station. Among them, the data for November and December 2024 are the historical power plant revenue calculated based on historical data.

[0578] Figure 6E This is a schematic diagram of the historical power plant revenue corresponding to each month of a downstream hydropower station provided by an embodiment of the present application. Figure 6E The unit of the abscissa in it is US dollars; the ordinate is month. Figure 6E It shows the historical power plant revenue of each month from January 2023 to December 2023 and from January 2024 to December 2024 of the downstream hydropower station. Among them, the data for November and December 2024 are the historical power plant revenue calculated based on historical data.

[0579] Furthermore, the water asset data of each power station for each month can also be calculated. The water asset of each power plant in the present application refers to the sum of formula (2.1), formula (2.2) and formula (2.3).

[0580] Figure 6F It is a schematic diagram of the water assets corresponding to each month of an upstream hydropower station provided by an embodiment of the present application. Figure 6F The abscissa in it represents the quantity of water assets, with the unit of US dollars; the unit of the ordinate is month. Figure 6F It gives the historical water assets of each month from January 2023 to December 2023 and from January 2024 to December 2024 of the upstream hydropower station. Among them, the data for November 2024 and December 2024 are historical water assets calculated based on historical data.

[0581] Figure 6G It is a schematic diagram of the water assets corresponding to each month of a downstream hydropower station provided by an embodiment of the present application. Figure 6G The abscissa in it represents the quantity of water assets, with the unit of US dollars; the unit of the ordinate is month. Figure 6G It gives the historical water assets of each month from January 2023 to December 2023 and from January 2024 to December 2024 of the downstream hydropower station. Among them, the data for November 2024 and December 2024 are historical water assets calculated based on historical data.

[0582] For easy understanding, the present application Figure 6B , Figure 6C , Figure 6D , Figure 6E , Figure 6F and Figure 6G give the historical power generation water consumption, historical power plant revenue and historical water assets (revenue) of each hydropower station in each month of a cascade hydropower station including a total of two hydropower stations, namely an upstream hydropower station and a downstream hydropower station. Those skilled in the art can obtain the historical power generation water consumption, historical power plant revenue and historical water assets (revenue) of each hydropower station in each month of other cascade hydropower stations by referring to the methods introduced in the foregoing embodiments.

[0583] Then, for each month, based on the N historical power generation water consumptions corresponding to that month, a graph is plotted to obtain the historical power generation water consumption graph corresponding to that month.

[0584] For example, for January, the historical power generation water consumptions in January 2010, January 2011, January 2012, and even January 2024 can be obtained; a graph with the year on the abscissa and the historical power generation water consumption on the ordinate is generated as the historical power generation water consumption graph corresponding to January.

[0585] Similarly, for February, the historical power generation water consumption for February 2010, February 2011, February 2012, and even February 2024 can be obtained; a graph with the year on the x-axis and the historical power generation water consumption on the y-axis is generated as the corresponding historical power generation water consumption graph for February.

[0586] For each month, based on the N historical power plant revenues corresponding to that month, the corresponding historical power plant revenue graph for that month is obtained.

[0587] For example, for January, the historical power plant revenues for January 2010, January 2011, January 2012, and even January 2024 can be obtained; a graph with the year on the x-axis and the historical power plant revenue on the y-axis is generated as the historical power plant revenue graph for January.

[0588] Among them, the historical power generation water consumption graph represents the relationship between historical power generation water consumption and time; the historical power plant revenue graph represents the relationship between historical power plant revenue and time.

[0589] Finally, the average power generation water consumption corresponding to the historical power generation water consumption graph for the month where the current moment is located, and the average power plant revenue corresponding to the historical power plant revenue graph for the month where the current moment is located are used as the target historical data.

[0590] Among them, the average power generation water consumption is the mean value of the historical power generation water consumption included in the historical power generation water consumption graph. The average power plant revenue is the mean value of the historical power plant revenues included in the historical power plant revenue graph.

[0591] After obtaining the current actual data and the target historical data, based on the difference between the current actual data and the target historical data, it is determined whether to perform water resource scheduling. Specifically:

[0592] It is determined whether the difference between the actual power generation water consumption and the average power generation water consumption is within the preset allowable value of power generation water consumption difference (the first judgment condition), and it is determined whether the difference between the actual power plant revenue and the average power plant revenue is within the preset allowable value of power plant revenue difference (the second judgment condition).

[0593] If both the first judgment condition and the second judgment condition are satisfied, then water resource scheduling is not required, and directly return to S601; if the above two judgment conditions are not simultaneously satisfied, it is considered that water resource scheduling is required, and enter S604.

[0594] Further, it is also possible to calculate the actual water assets of the hydropower station corresponding to the current month of the hydropower station (the sum value of Formula (2.1), Formula (2.2) and Formula (2.3)); obtain the average water assets of the hydropower station corresponding to the current month; determine whether the difference between the actual water assets and the average water assets corresponding to the current month is within the preset allowable value of water asset difference (the third judgment condition), and determine whether water resource scheduling is required.

[0595] That is, when the first judgment condition, the second judgment condition and the third judgment condition are satisfied, it is determined that water resource scheduling is not required, and S601 is directly returned; otherwise, it is considered that water resource scheduling is required, and S604 is entered.

[0596] It should be noted that the values of the preset allowable difference in power generation water use, the preset allowable difference in power plant revenue, and the preset allowable difference in water assets in this application are not limited, and those skilled in the art can limit the above values according to needs.

[0597] S604, if it is determined to perform water resource scheduling, then based on the water resource data and the water resource scheduling model of each hydropower station, a water resource scheduling plan is determined.

[0598] Among them, the water resource scheduling model is a model that is trained in advance and used to output a water resource scheduling plan.

[0599] In an alternative implementation, the training steps of the water resource scheduling model include:

[0600] The first step is to obtain training data.

[0601] The training data in this application includes the sample water resource data obtained by each hydropower station in the cascade hydropower station at the sample time point, the electricity price plan at the sample time point, and the sample water resource scheduling plan at the sample time point.

[0602] It should be noted that the sample time point in this application is any one of multiple time points when the hydropower station operates under full conditions. In other words, the sample time point can be the time point when the hydropower station is operating normally, and the data corresponding to the sample time point is the data corresponding to the normal operation time of the hydropower station; the sample time point can also be the time point when the hydropower station conducts load rejection experiments, water and soil conservation experiments or vibration experiments, and the data corresponding to the sample time point is the data corresponding to the hydropower station under the above experimental states.

[0603] The sample water resource data includes historical power generation water consumption, historical power plant revenue, and historical average water assets; the sample water resource scheduling plan includes daily scheduling and seasonal scheduling.

[0604] It is understandable that the water resource scheduling of each hydropower station in a cascade hydropower station is a very complex process. At different time points, different water resource scheduling schemes are adopted under different conditions such as the water level of the reservoir where each hydropower station is located, the actual operation parameters of the power generation facilities, and the precipitation.

[0605] It is understandable that the storage capacity of the upstream hydropower station in a cascade hydropower station is greater than that of the downstream hydropower station; the water volume regulation space of the upstream hydropower station is greater than that of the downstream hydropower station. For the above reasons, among the water resource scheduling schemes included in the training data given in this application, the upstream hydropower station adopts the quarterly regulation method (also known as quarterly scheduling and quarterly regulation), and the downstream hydropower station adopts the day-ahead regulation method (also known as day-ahead scheduling and day-ahead regulation).

[0606] Among them, quarterly regulation means storing a part of the runoff during the wet season of the year by adjusting the reservoir capacity of the hydropower station for use during the dry season. This regulation method enables the hydropower station to achieve effective reallocation of water resources within a year, especially between seasons with uneven water resource distribution. Quarterly regulation can store excess water during seasons with sufficient water volume and release the stored water during seasons with scarce water volume to ensure the stable operation and continuous power generation of the hydropower station.

[0607] Among them, day-ahead scheduling refers to the reallocation of runoff within one day and night. This regulation method is mainly used to cope with water resource changes and power demand fluctuations in the short term (such as within one day). Through day-ahead scheduling, the hydropower station can flexibly adjust its power generation plan and water resource allocation according to the real-time water resource status and power demand to ensure the stability and reliability of power supply.

[0608] Quarterly scheduling mainly focuses on the balanced distribution of water resources between seasons to achieve long-term stable operation and power generation benefits; while day-ahead regulation pays more attention to the dynamic balance of water resources and power demand in the short term to ensure the stability and reliability of power supply. The hydropower station comprehensively adopts these two regulation methods according to factors such as its own reservoir capacity conditions, water resource status, and power market demand to achieve the efficient utilization of water resources and the optimization of power supply.

[0609] Precisely because of the day-ahead scheduling downstream and seasonal scheduling upstream, the calculation of water assets is meaningful. The pursuit of both is the balance of cost and economy. The downstream focuses on power generation, and the upstream focuses on controlling the water volume and water level of the entire basin. For example Figure 6B and Figure 6C in the comparison of day-ahead scheduling and quarterly scheduling between upstream and downstream, the water consumption per kilowatt-hour is 0.018m 3 / kWh and 0.026m 3 / kWh. Since more attention has been paid to monthly profit recently, it shows that the power generation efficiency in the downstream is better. The upstream quarterly scheduling pays more attention to the overall situation of the whole basin, taking power generation into account. By comparing the monthly accumulations of water assets in the upstream and downstream, it can be clearly seen that the details of water assets in the upstream are more than those in the downstream, which further illustrates the importance of adopting quarterly conditions in the upstream.

[0610] Through the high-order feature abstraction and analysis of data, especially the research on power generation water consumption, water assets, etc., under the premise of ensuring the safety of intelligent hydropower stations, that is, all working conditions meet the requirements of load rejection, variable power vibration measurement, and water conservation safety, intelligent agents using the two scheduling functions are driven by data to implement different scheduling strategies for the upstream and downstream of the hydropower station respectively.

[0611] Figure 6H It is a schematic diagram of a water resource regulation scheme for a cascade hydropower station provided by an embodiment of the present application. The cascade hydropower station includes an upstream hydropower station and a downstream hydropower station. Figure 6H The meanings of the letters in it are as follows: J1 represents the cost electricity price; S1 represents the measured water level of the upstream hydropower station; S2 represents the measured water level of the downstream hydropower station; R represents the lowest water level allowed for power generation in the downstream hydropower station; Q represents the lowest water level allowed for power generation in the upstream hydropower station; J represents the online trading electricity price at the current moment, which is also the basic value of the preset online electricity price mentioned in the foregoing embodiment.

[0612] Combined with Figure 6H As shown, the water resource regulation scheme of the cascade hydropower station is specifically as follows:

[0613] S401, determine the cost electricity price J1, the measured water level S1 of the upstream hydropower station, and the measured water level S2 of the downstream hydropower station.

[0614] S402, judge whether S2 < R.

[0615] Judge whether the measured water level of the downstream hydropower station is less than the lowest water level allowed for power generation in the downstream hydropower station. If S2 is less than R, enter S403; otherwise, enter S406.

[0616] S403, judge whether S1 > Q and the cost electricity price < the online electricity price.

[0617] That is, judge whether the measured water level of the upstream hydropower station is greater than the lowest water level allowed for power generation in the upstream hydropower station, and judge whether the cost electricity price is less than the online electricity price (the online trading electricity price at the current moment). If so, enter S404; otherwise, enter S407.

[0618] S404, perform quarterly regulation on the upstream hydropower station.

[0619] S405, perform daily regulation on the downstream hydropower station.

[0620] S406. Wait for the grid shutdown instruction in case of a fault or a power plant water shortage.

[0621] S407. Determine whether the on-grid electricity price changes, J = J ± 0.001, and the cost electricity price J1 < J.

[0622] If so, go to step S404; otherwise, go to S408.

[0623] S408. Regulate the water level at high frequency and change the water level.

[0624] It should be emphasized that Figure 6H the water resource regulation scheme shown in is only for easy understanding, and is a water resource scheduling scheme that may be adopted in the actual working process. For different cascade hydropower stations, at different times, and when the water resource data of each hydropower station is different, targeted different water resource regulation schemes will be adopted.

[0625] In the second step, input the training data into the deep learning model to obtain the predicted water resource debugging scheme corresponding to the training data.

[0626] Exemplarily, in this application, a convolutional neural network, a long short-term memory network, a gated recurrent unit, etc. can be used as the deep learning model.

[0627] Input the sample water resource data obtained by each hydropower station in the cascade hydropower station in the training data at the sample time point and the electricity price scheme at the sample time point into the deep learning model, and the deep learning model inputs the predicted water resource debugging scheme.

[0628] In the third step, taking the sample water resource scheduling scheme as the true value, determine the backpropagation error between the sample water resource deep debugging scheme and the predicted water resource debugging scheme through the loss function, and update the network parameters of the deep learning model based on the backpropagation error until the backpropagation error reaches the set value, and obtain the trained water resource scheduling model.

[0629] In this application, the sample water resource scheduling scheme at the sample time point in the training data is used as the true value, calculate the difference between the sample water resource deep debugging scheme and the predicted water resource debugging scheme through the loss function, that is, the backpropagation error, take the backpropagation error as the loss value, and update the network parameters in the deep learning model, such as weights and biases, through the loss value to reduce the future backpropagation error. When the latest backpropagation error reaches the set value, stop training to obtain the trained water resource debugging model.

[0630] The loss function is the basis for guiding the model to learn effectively. Based on different tasks, different loss functions can be selected or designed so that the model can extract valuable information from the data. Here we introduce several common loss functions. Common loss functions include: squared loss function, cross-entropy loss, mean squared error, and mean absolute error, etc. The specific form of the loss function is not limited in this application. The loss function and the optimization algorithm are two important components of machine learning. Based on the loss function, the selected optimization algorithm can also be chosen according to actual needs. For example, the gradient descent algorithm is adopted. This algorithm utilizes gradient information and continuously iteratively adjusts the parameters to find a suitable solution....

Claims

1. A construction method of a digital power system based on multi-functional agents, characterized in that Including: extracting characteristic production factor data corresponding to a characteristic production factor group from power data on the power plant side; and extracting control production factor data corresponding to a control production factor group from power data on the power grid side; the characteristic production factor group includes characteristic production factors at multiple levels, namely natural characteristic production factors, single-system characteristic production factors, and plant-system characteristic production factors; the control production factor group includes control production factors in multiple aspects; from the natural characteristic production factors to the single-system characteristic production factors and then to the plant-system characteristic production factors, the mining level of data characteristics deepens step by step; based on the power function requirements of the digital power system, constructing a target data vector using the extracted characteristic production factor data, and determining key control production factor data associated with the power function requirements from the extracted control production factor data; on the basis of the target data vector and the key control production factor data, using an artificial intelligence algorithm to learn the mapping function between the target data vector and the key control production factor data, constructing a functional operator with the mapping function as the kernel, and constructing an intelligent agent based on the functional operator; adding intelligent agents corresponding to various different power function requirements to the digital power system, and using each intelligent agent as a power function implementation unit in the digital power system; the various different power function requirements include at least two of the following: calculating electricity prices, power clearing, peak shaving scheduling in the scenario of high-frequency variable load of thermal power plants, water resource scheduling of cascade hydropower stations, grid dispatching of power plants in a large machine and small grid power system, or adjusting the grid grid structure.

2. The method according to claim 1, wherein the characteristic production factor group includes first-level, second-level, and third-level characteristic production factors corresponding to various different power generation types; among them, the first-level characteristic production factors are natural characteristic production factors, the second-level characteristic production factors are single-system characteristic production factors, and the third-level characteristic production factors are plant-system characteristic production factors; the natural characteristic production factors are production factors directly reflecting natural characteristics; the single-system characteristic production factors are production factors involving monomer systems in the power system; the plant-system characteristic production factors are production factors involving the overall system in the power system.

3. The method according to claim 1, wherein the control production factors in multiple aspects include: electricity price, inertia and frequency, active and reactive power, carbon emissions from electricity, safety and stability, and grid connection and disconnection; the extracting control production factor data corresponding to the control production factor group from the power data on the power grid side includes: extracting electricity price data, inertia and frequency data, power data, carbon emissions from electricity data, safety and stability index data, and grid connection and disconnection impact data from the power data on the power grid side.

4. The method according to claim 1, wherein The method further includes: configuring a power plant side interface, a power grid side interface, and an intelligent agent interface for the intelligent agent; the intelligent agent communicates with the power plant through the power plant side interface; the intelligent agent communicates with the power grid through the power grid side interface; the intelligent agent communicates with other intelligent agents through the intelligent agent interface.

5. The method according to any one of claims 1 to 4, characterized in that, The power function requirements are as follows: Power function requirements for power trading between the grid side and the power plant side, power function requirements for realizing the dispatching control of the grid side over the power plant side, or power function requirements for the self-regulation of the grid.

6. The method according to claim 5, characterized in that The power function requirements for power trading between the grid side and the power plant side are specifically to calculate the electricity price. The functional operator is the electricity price operator, and the electricity price operator is used to calculate the electricity price. The process of the digital power system using the agent constructed based on the electricity price operator includes: Invoking the agent to obtain the electricity price calculation result on the power plant side through the electricity price operator; Verifying the electricity price calculation result through the financial verification model on the grid side. The financial verification model includes: the first verification condition and the second verification condition. The first verification condition is: the actual output of various different types of power plants ≥ the grid electricity quantity; the second verification condition is: the sum of the products of the actual output of various different types of power plants, the grid electricity price at the corresponding moment, and the power generation duration ≤ the product of the grid average price and the grid electricity quantity; If the electricity price calculation result meets the first verification condition and the second verification condition, it is determined that the electricity price calculation result passes the verification; if the electricity price calculation result does not meet any of the first verification condition and the second verification condition, it is determined that the electricity price calculation result fails the verification; If the verification of the electricity price calculation result passes, then further select the electricity price calculation function that is suitable and feasible for the grid side from the dispatching strategy library, and load the electricity price calculation function into the electricity price operator to update the function of the electricity price operator; Obtain the electricity price calculation result on the grid side through the updated electricity price operator.

7. The method according to claim 6, characterized in that After loading the electricity price calculation function into the electricity price operator to update the function of the electricity price operator, the process of using the agent constructed based on the electricity price operator further includes: The grid side collects the applied trading electricity quantity and electricity price reported by the power plant side to the grid side; Based on the grid side, judge whether the grid is stable to determine whether to start the virtual power plant function; If it is determined that there is no need to start the virtual power plant function, the grid side sends out the dispatching instructions of the clearing electricity price and electricity quantity to each type of power plant based on the electricity price calculation result on the grid side; If it is determined that the virtual power plant function needs to be started, start the virtual power plant function to call its storage capacity; The power plant side receives the clearing electricity price given by the grid side; Combined with the power plant cost and the clearing electricity price, use the financial verification model on the power plant side for verification. The financial verification model on the power plant side includes: the third verification condition. The third verification condition is: the sum of the products of the actual output of the power plant, the grid electricity price at the corresponding moment, and the power generation duration > the product of the power plant cost and the actual output of the power plant; If the verification result indicates that the verification passes, then further judge whether each power source point of the same type of power plant can be cleared based on the electricity quantity and electricity price; If each power source point can be cleared, execute the cleared electricity quantity.

8. The method according to claim 5, characterized in that, The specific power functional requirement for power trading between the grid side and the power plant side is power clearing. The functional operator is the clearing operator, and the clearing operator is used for power clearing. The process of the digital power system using the agent constructed based on the clearing operator includes: Performing low-order data abstraction processing on the power generation element characteristics of each associated power plant of the power grid to obtain the first data vector of each power generation element characteristic; Predicting the long-term trading electricity price based on the preset long-term electricity price function library and each first data vector to establish a long-term trading stack; Establishing a spot trading stack based on the spot trading electricity prices of each associated power plant, and determining the first clearing function through the long-term trading stack and the spot trading stack; According to the first clearing function, determining the power supply and demand balance state when the power grid clears under the first clearing function; Adjusting the function of the first clearing function according to the power supply and demand balance state to obtain a second clearing function, and performing power clearing based on the second clearing function.

9. The method according to claim 8, characterized in that, The power supply and demand balance state includes: over-generation state; the first clearing function includes: the first long-term trading volume; The adjusting the function of the first clearing function according to the power supply and demand balance state to obtain a second clearing function includes: When the power supply and demand balance state is the over-generation state, performing high-order data abstraction processing on each power generation element characteristic through a preset deep learning model to obtain the second data vector of each associated power plant; Determining the increase value of the long-term trading volume according to each second data vector and the preset long-term electricity price function library; Adjusting the first long-term trading volume according to the increase value of the long-term trading volume to obtain the second clearing function.

10. The method according to claim 9, characterized in that, The power supply and demand balance state includes: under-generation state; the first clearing function includes: the first spot trading volume; The adjusting the function of the first clearing function according to the power supply and demand balance state to obtain a second clearing function includes: When the power supply and demand balance state is the under-generation state, performing fitting processing on the first power generation curve when the power grid clears based on the first clearing function through a preset fitting function to obtain a fitted power generation curve; Based on the fitted power generation curve and the pre-designed planned power generation curve, judging whether the first spot trading volume meets the pre-designed planned power generation curve; When the first spot trading volume does not meet the pre-designed planned power generation curve, correcting the function of the preset long-term electricity price function library through a preset reinforcement learning model to determine the decrease value of the long-term trading volume; Adjusting the proportion of the spot trading volume in the first clearing function according to the decrease value of the long-term trading volume to obtain the increase value of the spot trading volume; Adjusting the first spot trading volume according to the increase value of the spot trading volume to obtain the second clearing function.

11. The method according to claim 10, wherein The correcting the function of the preset long-term electricity price function library through a preset reinforcement learning model to determine the decrease value of the long-term trading volume when the first spot trading volume does not meet the pre-designed planned power generation curve includes: Function screening is performed on the preset long-term electricity price function library through a preset reinforcement learning model to obtain a first long-term trading electricity price function; Index adjustment is performed on the function indexes of the first long-term trading electricity price function to determine the long-term trading volume reduction value.

12. The method according to claim 8, wherein The long-term trading electricity price prediction based on the preset long-term electricity price function library and each of the first data vectors to establish a long-term trading stack includes: Through the first data vector, data-driven adjustment is performed on the function indexes of the preset electricity price function library to obtain a dynamic electricity price function library; Based on the dynamic electricity price function library and each of the power generation element characteristics, long-term trading electricity price prediction is performed to obtain the long-term trading predicted electricity prices of each associated power plant; Through the long-term trading predicted electricity prices of each associated power plant, the long-term trading stack is established.

13. The method according to claim 8, wherein The establishment of the spot trading stack based on the spot trading electricity prices of each associated power plant includes: Power consumption prediction is performed according to the first data vector of each associated power plant to obtain the predicted power consumption of each associated power plant; Based on the predicted power consumption of each associated power plant, a preliminary spot trading stack is established; Based on the spot trading electricity prices of the associated power plants, the preliminary spot trading stack is adjusted to obtain the spot trading stack.

14. The method according to claim 8, wherein The low-order data abstraction processing of each of the power generation element characteristics to obtain the first data vector of each of the power generation element characteristics includes: According to each of the power generation element characteristics, an energy consumption data model of each associated power plant is established; the energy consumption data model is used to evaluate the power generation performance of the associated power plant; Through the energy consumption data model, low-order data abstraction processing is performed on each of the power generation element characteristics to obtain the first data vector of each of the power generation element characteristics.

15. The method according to claim 5, wherein The specific power function requirement for realizing the dispatching control of the grid side over the power plant side is the peak regulation dispatching in the scenario of high-frequency load change of thermal power, the functional operator is the thermal power peak regulation dispatching operator, and the thermal power peak regulation dispatching operator is used to realize the peak regulation dispatching in the scenario of high-frequency load change of thermal power; the process of the digital power system using the intelligent agent constructed based on the thermal power peak regulation dispatching operator includes: Calculate the heat storage and access margin of the energy storage device in the thermal power plant; the heat storage and access margin includes a heat storage margin characterization value and a heat release capacity characterization value; Send the correlation data between the load, performance and cost of each thermal power unit in the thermal power plant and the heat storage and access margin to the grid side, so that the grid side determines the target unit for expected auxiliary peak regulation and issues a peak regulation dispatching instruction based on the change of thermal power load demand, the correlation data provided by each thermal power plant and the heat storage and access margin; Receive the peak regulation dispatching instruction issued by the grid side; the peak regulation dispatching instruction carries the unit identification of the target unit and the peak regulation requirement information for the target unit; the peak regulation requirement information includes a peak regulation load curve; Execute the auxiliary peak regulation service based on the peak regulation requirement information.

16. The method according to claim 15, wherein The process of using the intelligent agent constructed based on the thermal power peak regulation dispatching operator also includes: Before determining the target unit for the expected auxiliary peak load regulation, according to the peak load curve of each thermal power unit, calculating the quotation information of the auxiliary peak load regulation service provided by the thermal power unit that meets the requirements of the corresponding peak load curve; Sending the quotation information to the power grid side; The grid side determines the target unit for expected auxiliary peak load regulation based on the change of thermal power load demand, the associated data provided by each thermal power plant, and the heat storage and access margin, specifically: The grid side determines the target unit for expected auxiliary peak regulation by weighing costs and indicators through data-driven means based on changes in thermal power load demand, the associated data provided by each thermal power plant, the heat storage and access margin, and the quotation information.

17. The method according to claim 16, wherein The grid side determines the target unit for expected auxiliary peak load regulation by weighing costs and indicators through data-driven based on the change of thermal power load demand, the associated data provided by each thermal power plant, the heat storage and access margin and the quotation information, including: The grid side generates a peak load curve for each thermal power unit based on the change in thermal power load demand, the associated data provided by each thermal power plant, and the heat storage and access margin, and calculates the climbing coefficient of each thermal power unit; Preliminarily determine multiple thermal power units expected to assist in peak load regulation based on the ramp coefficients of each thermal power unit; If there is a thermal power unit whose bid information is less than or equal to K times the threshold bid among the multiple thermal power units expected to assist in peak load regulation, the thermal power unit whose bid information is less than or equal to K times the threshold bid is determined as the target unit for expected auxiliary peak load regulation; the K is within the range of 1.1 to 1.3; Among them, the threshold quotation is a quotation reference value obtained by deep learning historical quotation information on the power grid side.

18. The method according to claim 17, wherein The process of using the intelligent agent constructed based on the thermal power peak-shaving dispatching operator further includes: after the multiple thermal power units expected to assist in peak-shaving are preliminarily determined, the power grid side sends a heat pre-dispatching instruction to the multiple thermal power units expected to assist in peak-shaving; The target unit among the multiple thermal power units expected to assist in peak load regulation performs heat pre-scheduling in cooperation with the heat storage device in the thermal power plant to which it belongs based on the heat pre-scheduling instruction.

19. The method according to claim 17, wherein The peak shaving requirement information also includes climbing speed requirement information; and performing auxiliary peak shaving service based on the peak shaving requirement information includes: Determine whether there is equipment failure and communication failure in the target unit, and determine whether the target unit meets the climbing speed requirement information; If there is no equipment failure and communication failure in the target unit, and the target unit meets the climbing speed requirement information, then according to the peak load curve and the climbing speed requirement information in the peak load requirement information, perform auxiliary peak load shaving service; If there is a device failure or a communication failure in the target unit, a fault message is reported to the power grid side; If the target unit does not meet the ramp speed requirement information, reporting prompt information that the operating condition is not met to the power grid side; The power grid side adjusts the peak load requirement information in the peak load scheduling instruction according to the prompt information that the working condition is not satisfied; Based on the prompt information fed back by each thermal power plant that the working conditions are not met, the grid side re-learns to increase the threshold offer price, so that more units can be used as the target units for expected auxiliary peak shaving.

20. The method according to claim 17, wherein The calculation method of the ramp rate coefficient of a thermal power unit is as follows: Extract the ramp rate, steam pressure, and rated power generation capacity from the correlation data among the load, performance, and cost of the thermal power unit; Calculate the ramp rate coefficient of the thermal power unit according to the ramp rate, steam pressure, rated power generation capacity, heat storage margin of the thermal power unit, and the ramp time required by the power plant side for the thermal power unit.

21. The method according to claim 20, wherein The heat storage margin is expressed as a percentage; the calculation of the ramp rate coefficient of the thermal power unit according to the ramp rate coefficient, steam pressure, rated power generation capacity, heat storage margin of the thermal power unit, and the ramp time required by the power plant side for the thermal power unit includes: Calculate the product of the ramp rate, ramp time, steam pressure, and heat storage margin of the thermal power unit; Calculate the ratio of the product to the rated power generation capacity of the thermal power unit, and use the ratio as the ramp rate coefficient of the thermal power unit.

22. The method according to any one of claims 15-21, characterized in that, The heat storage margin and the correlation data are calculated based on the coal type parameters of the selected target coal type; the selection method of the target coal type includes: According to the higher heating value and lower heating value corresponding to each of the multiple candidate coal types, obtain the percentage difference between the higher and lower heating values of each candidate coal type; Determine the coal type with the smallest percentage difference between the higher and lower heating values among the multiple candidate coal types as the target coal type.

23. The method according to claim 22, wherein The percentage difference between the higher and lower heating values is a low-order feature obtained by the thermal power plant based on the coal type parameters of the corresponding coal type for feature extraction; Based on the change of thermal power load demand, the correlation data provided by each thermal power plant, and the heat storage margin, the grid side determines the target units for expected auxiliary peak shaving and issues peak shaving scheduling instructions, including: Based on the correlation data provided by each thermal power plant and the heat storage margin, the grid side extracts intermediate features, and the intermediate features include features reflecting the energy consumption level and thermal energy reserve level of the power plant; the grid side fuses the coal price cost in the correlation data and the intermediate features to obtain high-order features, and the high-order features include the relationship between fuel, electricity, and electricity price; Based on the change of thermal power load demand, the intermediate features, and the high-order features, the grid side determines the target units for expected auxiliary peak shaving and issues peak shaving scheduling instructions.

24. The method according to claim 5, characterized in that, The specific power function requirement for the grid side to realize the dispatching control of the power plant side is the water resource dispatching of cascade hydropower stations. The functional operator is the water resource dispatching operator, and the water resource dispatching operator is used to realize the water resource dispatching of cascade hydropower stations; the process of the digital power system using the agent constructed based on the water resource dispatching operator includes: Obtain the water resource data of each hydropower station in the cascade hydropower station at the current moment; For each hydropower station, judge whether the hydropower station is in a safe operation state according to the water resource data of the hydropower station; When it is determined that each of the hydropower stations is in the safe operation state, based on the water resource data of each hydropower station, a preset electricity price plan, and target historical data, it is judged whether to conduct water resource scheduling; the target historical data includes the average power generation water consumption and the average power plant revenue corresponding to the month in which the current moment is located. If it is determined to conduct water resource scheduling, then based on the water resource data of each hydropower station and a water resource scheduling model, a water resource scheduling plan is determined; the water resource scheduling model is a pre-trained model used to output a water resource scheduling plan; the water resource scheduling plan includes daily scheduling and seasonal scheduling.

25. The method according to claim 24, wherein The safe operation state includes a water and soil conservation safety state. For each of the hydropower stations, judging whether the hydropower station is in the safe operation state according to the water resource data of the hydropower station includes: For each of the hydropower stations, obtain the measured speed rise rate and the measured turbine static water pressure from the water resource data of the hydropower station. For each of the hydropower stations, if the measured speed rise rate is less than the theoretical speed rise rate and the measured turbine static water pressure is less than the theoretical turbine static water pressure, it is determined that the hydropower station is in the water and soil conservation safety state. If the measured speed rise rate is greater than or equal to the theoretical speed rise rate, or the measured turbine static water pressure is greater than or equal to the theoretical turbine static water pressure, it is determined that the hydropower station is not in the water and soil conservation safety state.

26. The method according to claim 24, wherein The safe operation state includes a vibration safety state. For each of the hydropower stations, judging whether the hydropower station is in the safe operation state according to the water resource data of the hydropower station includes: For each of the hydropower stations, obtain the power of the water turbine engine from the water resource data of the hydropower station, denoted as the first power. Judge whether the first power is within a preset power range. If the first power is within the preset power range, adjust the first power to a second power; the second power is outside the preset power range. If the first power is outside the preset power range, judge whether the hydropower station is in the vibration safety state.

27. The method according to claim 26, wherein The judging whether the hydropower station is in the vibration safety state includes: For each of the hydropower stations, obtain the rotational speed of the water turbine engine and the actual vibration values of different parts of the water turbine engine from the water resource data of the hydropower station. For each of the hydropower stations, if the actual vibration value of each part is less than the vibration allowable value corresponding to the rotational speed of that part, it is determined that the hydropower station is in the vibration safety state. Otherwise, it is determined that the hydropower station is not in the vibration safety state.

28. The method according to claim 24, wherein Before obtaining the water resource data of each hydropower station in the cascade hydropower station at the current moment, the process of applying the agent constructed based on the water resource scheduling operator further includes: Conduct a load rejection test on the cascade hydropower station to judge whether the operation parameters of each hydropower station in the cascade hydropower station meet the requirements of the preset load rejection test index.

29. The method according to claim 24, characterized in that, The judging whether to conduct water resource scheduling according to the water resource data of each hydropower station, a preset electricity price plan, and target historical data includes: Based on the water resource data of each hydropower station and the preset electricity price scheme, determine the actual power generation water consumption of each hydropower station and the actual power plant revenue of each hydropower station as the current actual data; Based on the difference between the current actual data and the target historical data, determine whether to conduct water resource scheduling.

30. The method according to claim 29, wherein The determining of the actual power generation water consumption of each hydropower station and the actual power plant revenue of each hydropower station based on the water resource data of each hydropower station and the preset electricity price scheme includes: For each hydropower station, obtain the actual power generation and actual water consumption in the target time period from the water resource data of the hydropower station; the target time period is a period of time in the past ending at the current moment; For each hydropower station, based on the actual power generation and actual water consumption in the target time period, obtain the actual power generation water consumption in the target time period; For each hydropower station, based on the electricity price scheme corresponding to the target time period and the actual water level of the hydropower station, determine the actual power plant revenue in the target time period.

31. The method according to any one of claims 24-30, characterized in that, The method for obtaining the target historical data includes: Statistically calculate the historical power generation water consumption and historical power plant revenue for each month in the historical N years; N is an integer greater than or equal to 1; For each month, based on the N historical power generation water consumptions corresponding to the month, obtain the historical power generation water consumption map corresponding to the month; and based on the N historical power plant revenues corresponding to the month, obtain the historical power plant revenue map corresponding to the month; the historical power generation water consumption map represents the relationship between historical power generation water consumption and time; the historical power plant revenue map represents the relationship between historical power plant revenue and time; Take the average power generation water consumption corresponding to the historical power generation water consumption map of the month where the current moment is located and the average power plant revenue corresponding to the historical power plant revenue map of the month where the current moment is located as the target historical data.

32. The method according to any one of claims 24-30, characterized in that, The training steps of the water resource scheduling model include: Obtain training data; the training data includes the sample water resource data of each hydropower station in the cascade hydropower station at the sample time point, the electricity price scheme at the sample time point, and the sample water resource scheduling scheme at the sample time point; the sample time point is any one of multiple time points when the hydropower station operates under full conditions; the sample water resource data includes historical power generation water consumption, historical power plant revenue, and historical average water assets; the sample water resource scheduling scheme includes day-ahead scheduling and seasonal scheduling; Input the training data into the deep learning model to obtain the predicted water resource debugging scheme corresponding to the training data; Taking the sample water resource scheduling scheme as the true value, determine the backpropagation error between the sample water resource deep debugging scheme and the predicted water resource debugging scheme through the loss function, and update the network parameters of the deep learning model based on the backpropagation error until the backpropagation error reaches the set value to obtain the water resource scheduling model.

33. The method according to claim 5, wherein The specific power function requirements for implementing the dispatching control of the grid side over the power plant side are the dispatching of the power plant by the grid in a large generator - small grid power system. The functional operator is the large generator - small grid dispatching operator, which is used to implement the dispatching of the power plant by the grid in the large generator - small grid power system. The process of the digital power system using the agent constructed based on the large generator - small grid dispatching operator includes: Based on the extraction of the digital characteristics of the grid and the power source points of various power generation types in the power system, determine the accident reserve capacity of various power generation types to be reserved in the power system and configure the accident reserve capacity. Based on the output of the power source points of the current various power generation types, determine whether the current power system conforms to the large generator - small grid characteristics. If it is determined that the current power system conforms to the large generator - small grid characteristics, the grid uses the configured accident reserve capacity of various power generation types to dispatch the power source points in the power system for voltage adjustment based on the multi - round voltage adjustment scheme, and / or dispatch the power source points in the power system for frequency adjustment based on the multi - round frequency adjustment scheme. The multi - round voltage adjustment scheme includes: the multi - round voltage adjustment methods, voltage adjustment ranges, and voltage adjustment priority information for various power generation types. The multi - round frequency adjustment scheme includes: the multi - round frequency adjustment methods, frequency adjustment ranges, and frequency adjustment priority information for various power generation types. Based on the stability control requirements of the frequency dynamic characteristic index of the entire network of the power system, the predicted total network load, and the unit characteristics of each power source point, the grid generates the output power curve of each power source point unit in the future time period. The frequency dynamic characteristic index is used to numerically represent the power change amount of the entire network required to cause a unit frequency change in the power system. The grid sends power generation dispatching instructions to each power source point for power adjustment. The power generation dispatching instructions include the output power curve of the unit corresponding to the power source point.

34. The method according to claim 33, wherein The multi - round voltage adjustment range includes: photovoltaic, hydropower, nuclear power, thermal power, and energy storage. The voltage adjustment priority information includes: photovoltaic > hydropower > nuclear power > thermal power > energy storage. The multi - round voltage adjustment for voltage includes: the voltage adjustment of four basic rounds corresponding to photovoltaic, hydropower, nuclear power, and thermal power, and the voltage adjustment of one accident round corresponding to the fault situation. The multi - round voltage adjustment methods include: the action voltage and delay of the first basic round, the action voltage step and delay of the remaining basic rounds, and the action voltage and delay of the accident round.

35. The method according to claim 34, characterized in that, The action voltage of the first basic round is 0.84 pu, and the delay is 15 - 20 s. The action voltage step of the remaining basic rounds is 0.01 pu, and the delay is 0.2 s. The action voltage of the accident round is 0.8 pu, and the delay is 0.1 s. Using the configured accident reserve capacity of various power generation types to perform voltage adjustment based on the multi - round voltage adjustment scheme, including: If the voltage of the power system drops to 0.84 pu, start the voltage adjustment of the first basic round corresponding to photovoltaic and continue for 15 - 20 s. During the voltage adjustment of the first basic round, a part of the accident reserve capacity of the photovoltaic is injected into the power grid; If the voltage adjustment effect of the first basic round does not reach the expected effect, when the voltage of the power system drops to 0.83 pu, start the voltage adjustment of the second basic round corresponding to hydropower and continue for 0.2 s; during the voltage adjustment of the second basic round, inject a part of the accident reserve capacity of hydropower into the power grid; If the voltage adjustment effect of the second basic round does not reach the expected effect, when the voltage of the power system drops to 0.82 pu, start the voltage adjustment of the third basic round corresponding to nuclear power and continue for 0.2 s; during the voltage adjustment of the third basic round, inject a part of the accident reserve capacity of nuclear power into the power grid; If the voltage adjustment effect of the third basic round does not reach the expected effect, when the voltage of the power system drops to 0.81 pu, start the voltage adjustment of the fourth basic round corresponding to thermal power and continue for 0.2 s; during the voltage adjustment of the fourth basic round, inject a part of the accident reserve capacity of thermal power into the power grid; If the voltage adjustment effect of the fourth basic round does not reach the expected effect, and it is monitored that the voltage of the power system drops to 0.8 pu, start the voltage adjustment of the accident round, cut off the load or start the energy storage system; during the voltage adjustment of the accident round, inject all or part of the remaining accident reserve capacity of each power generation type into the power grid.

36. The method according to claim 33, characterized in that, The multi-round frequency adjustment range includes: thermal power, hydropower, nuclear power and energy storage; the multi-round frequency adjustment priority information includes: thermal power > hydropower > nuclear power > energy storage; The multi-round frequency adjustment for frequency includes: the frequency adjustment of three basic rounds corresponding to thermal power, hydropower and nuclear power, and the frequency adjustment of one accident round corresponding to the fault situation; The multi-round frequency adjustment method includes: the action frequency and delay of the first basic round, the action frequency step difference and delay of the remaining basic rounds, and the action frequency and delay of the accident round.

37. The method according to claim 36, wherein The action frequency of the first basic round is 49.8 Hz, and the delay is 0.2 s; The action frequency step difference of the remaining basic rounds is 0.2 Hz, and the delay is 0.2 s; The action frequency of the accident round is 49.2 Hz, and the delay is 0.1 s; Using the accident reserve capacity of multiple configured power generation types, perform frequency adjustment based on the multi-round frequency adjustment scheme, including: If the frequency of the power system drops to 49.8 Hz, start the frequency adjustment of the first basic round corresponding to thermal power and continue for 0.2 s; During the frequency adjustment of the first basic round, inject a part of the accident reserve capacity of thermal power into the power grid; If the frequency adjustment effect of the first basic round does not reach the expected effect, when the frequency of the power system drops to 49.6 Hz, start the frequency adjustment of the second basic round corresponding to hydropower and continue for 0.2 s; during the frequency adjustment of the second basic round, inject a part of the accident reserve capacity of hydropower into the power grid; If the frequency adjustment effect of the second round of basic round fails to meet the expected effect, when the frequency of the power system drops to 49.4 Hz, start the frequency adjustment of the third round of basic round corresponding to nuclear power and continue for 0.2 s; during the frequency adjustment of the third round of basic round, inject a part of the accident reserve capacity of nuclear power into the power grid; If the frequency adjustment effect of the third round of basic round fails to meet the expected effect, when the frequency of the power system drops to 49.2 Hz, start the frequency adjustment of the accident round, cut off the load or start the energy storage system; during the frequency adjustment of the accident round, inject all or part of the accident reserve capacity of the remaining power generation types into the power grid.

38. The method according to claim 33, wherein Based on the stability control requirements of the frequency dynamic characteristic indexes of the entire power system, the expected total network load and the unit characteristics of each power source point, the power grid generates the output curves of the units at each power source point in the future time period, including: Determine the target value of the frequency dynamic characteristic index; Based on the target value, the frequency disturbance amount of the power system monitored, the expected total network load and the unit characteristics of each power source point, generate the output curves of the units at each power source point in the future time period, so as to control the power of each unit through the output curves, and make the deviation between the actual value and the target value of the frequency dynamic characteristic index within the preset index floating range; the unit characteristics include the power change amount that the unit of this power source point needs to provide for a unit frequency change.

39. The method according to claim 38, wherein Based on the target value, the frequency disturbance amount of the power system monitored, the expected total network load and the unit characteristics of each power source point, generating the output curves of the units at each power source point in the future time period, including: Based on the target value, the frequency disturbance amount of the power system monitored, and the unit characteristics of each power source point, determine the power change amount that each power source point of each power generation type should provide so that the deviation between the actual value and the target value of the frequency dynamic characteristic index is within the preset index floating range; Generate the output curves of the units at each power source point in the future time period according to the expected total network load in the future time period and the determined power change amount that each power source point of each power generation type should provide.

40. The method according to any one of claims 33 - 39, characterized in that, Based on the output of the power source points of the current multiple power generation types, judge whether the current power system conforms to the characteristics of large generators and small power grids, including: If the sum of the outputs of any two units among the power source points of any current power generation type exceeds 15% of the total capacity of the current power system, it is judged that the power system conforms to the characteristics of large generators and small power grids.

41. The method according to claim 5, wherein The specific power function requirement for the power grid's own regulation is to adjust the power grid grid structure, the functional operator is the power grid structure adjustment operator, and the power grid structure adjustment operator is used to realize the adjustment of the power grid grid structure; The process of the digital power system using the intelligent agent constructed based on the power grid structure adjustment operator includes: Obtain the current system inertia corresponding to the current grid structure; Calculate the frequency change rate based on the current system inertia and the difference between the power supply demand and the actual power supply; If the rate of change of the frequency is greater than a preset first threshold, determine a target grid structure and a target set value corresponding to the target grid structure based on a pre-established deep learning model; the rate of change of the frequency of the target grid structure is less than or equal to the first threshold; Modify the current set value applied in the current grid structure based on the target set value, and adjust the current grid structure of the power grid to the target grid structure.

42. The method according to claim 41, wherein Determining a target grid structure based on a pre-established deep learning model includes: Based on a pre-established deep learning model, calculate the simulated system inertia of all optional grid structures in parallel, and determine the optional grid structure with the largest simulated system inertia as the target grid structure.

43. The method according to claim 41, characterized in that, Determining a target grid structure based on a pre-established deep learning model includes: Based on the operating parameters of all optional grid structures, calculate the rate of change of frequency and the change amount of the difference between power generation and power consumption demand corresponding to multiple adjustment schemes for switching from the current grid structure to each optional grid structure respectively through a pre-established deep learning model, and obtain the scheme types to which the multiple adjustment schemes belong; the scheme types include stable type and aggressive type; If there is a stable adjustment scheme, determine the target grid structure from the optional grid structures corresponding to the stable adjustment scheme.

44. The method according to claim 41, characterized in that, Before determining the target grid structure and the target set value corresponding to the target grid structure based on a pre-established deep learning model, the method further includes: If there is a faulty node, disconnect the faulty node from the current grid structure; the node includes at least one of a power plant, a substation, and a dispatching control center.

45. The method according to claim 41, wherein Determining a target grid structure based on a pre-established deep learning model includes: Based on a pre-established deep learning model, determine a target grid structure that satisfies the condition of power generation and power consumption demand balance according to the matching situation between power generation and power consumption demand in the current grid structure.

46. The method according to claim 45, characterized in that, The determining, based on a pre-established deep learning model, of a target grid structure that satisfies the condition of power generation and power consumption demand balance according to the matching situation between power generation and power consumption demand in the current grid structure includes: Obtain the rate of change of power generation of each power generation grid node in the current grid structure; If there is a grid node with a decrease in power generation and the rate of change of power generation is greater than a preset rate-of-change threshold, determine, based on a pre-established deep learning model, a target grid structure for adding a power generation grid grid on the basis of the current grid structure according to the matching situation between power generation and power consumption demand in the current grid structure; If there is a grid node with an increase in power generation and the rate of change of power generation is greater than a preset rate-of-change threshold, determine, based on a pre-established deep learning model, a target grid structure for adding a load grid grid on the basis of the current grid structure according to the matching situation between power generation and power consumption demand in the current grid structure.

47. The method according to claim 41, wherein Determining a target grid structure based on a pre-established deep learning model includes: Based on a pre-established deep learning model, determine a target network structure with stability higher than a preset limit according to the safety priority level of the current grid structure.

48. The method according to claim 41, wherein Determining a target grid structure based on a pre-established deep learning model includes: Based on a pre-established deep learning model and a grid transfer function, obtain the influence degrees of multiple adjustment schemes for switching the current grid structure to each optional grid structure on the performance of other grids in the power system respectively; Based on the pre-established deep learning model and the influence degrees of performance, determine the target grid structure.

49. The method according to claim 41, characterized in that, Based on the pre-established deep learning model, determine the target setting value corresponding to the target grid structure, including: Through the pre-established deep learning model, based on the typical value library and the target grid structure, determine the target setting value corresponding to the target grid structure; the typical value library at least includes the correspondence between historical setting value data and grid structures.

50. The method according to claim 41, characterized in that, The adjustment of the current grid structure of the power grid to the target grid structure includes: Based on the current grid structure and the target grid structure, determine the target actions that multiple relays need to execute; the target actions are opening or closing. Based on the action time information of each relay, control multiple relays to execute the target actions to achieve the adjustment from the current grid structure to the target grid structure; the action time information includes the transmission time of the target setting value, the time required to write the target setting value, the duration of the protection action, the time required for the relay to close, the time required for the relay to open, and the time required for the return status information.

51. The method according to claim 41, wherein The adjustment of the current grid structure of the power grid to the target grid structure includes: Based on the parallel calculation of the target setting value stored in a binary tree structure, obtain the action information of multiple relays that need to execute target actions during the process of adjusting the current grid structure to the target grid structure; the target actions are opening or closing; the action information includes the time nodes when the relays are allowed to act and the sequence of actions of multiple relays. Based on the action information of multiple relays respectively, control multiple relays to execute target actions to adjust the current grid structure to the target grid structure.

52. A construction device for a digital power system based on a multi-functional intelligent agent, characterized in that, Include: A data extraction module, configured to extract feature production factor data corresponding to the feature production factor group from the power data on the power plant side; And extract control production factor data corresponding to the control production factor group from the power data on the power grid side; The feature production factor group includes multiple levels of feature production factors, namely natural feature production factors, single-system feature production factors, and plant-system feature production factors; the control production factor group includes control production factors in multiple aspects; from the natural feature production factors to the single-system feature production factors and then to the plant-system feature production factors, the mining level of data features gradually deepens; A vector construction module, configured to construct a target data vector based on the power function requirements of the digital power system by using the extracted feature production factor data; A data determination module, configured to determine key control production factor data associated with the power function requirements from the extracted control production factor data; An operator construction module, which is used to, based on the target data vector and the key control production factor data, use an artificial intelligence algorithm to learn a mapping function between the target data vector and the key control production factor data, construct a functional operator with the mapping function as the kernel, and construct an agent based on the functional operator; An agent addition module, which is used to add agents corresponding to various different power function requirements to the digital power system respectively, and use each agent as a power function implementation unit in the digital power system; the various different power function requirements include at least two of the following: calculating electricity prices, power clearing, peak shaving scheduling in the high-frequency variable load scenario of thermal power plants, water resource scheduling of cascade hydropower stations, grid dispatching of power plants in a large machine and small grid power system, or adjusting the grid grid structure.

53. The device according to claim 52, characterized in that, The power function requirement is: The power function requirement for power trading between the grid side and the power plant side, the power function requirement for realizing the dispatching control of the grid side over the power plant side, or the power function requirement for the grid's own regulation.

54. The device according to claim 53, characterized in that, The power function requirement for power trading between the grid side and the power plant side is specifically calculating electricity prices, the functional operator is an electricity price operator, and the electricity price operator is used to calculate electricity prices; the process of the digital power system using the agent constructed based on the electricity price operator includes: Invoking the agent to obtain the electricity price calculation result on the power plant side through the electricity price operator; Verifying the electricity price calculation result through the financial verification model on the grid side; the financial verification model includes: a first verification condition and a second verification condition; the first verification condition is: the actual output of various different types of power plants ≥ grid electricity; the second verification condition is: the sum of the products of the actual output of various different types of power plants, the grid electricity price at the corresponding moment, and the power generation duration ≤ the product of the grid average price and the grid electricity; If the electricity price calculation result meets the first verification condition and the second verification condition, it is determined that the electricity price calculation result passes the verification; if the electricity price calculation result does not meet any of the first verification condition and the second verification condition, it is determined that the electricity price calculation result fails the verification; If the electricity price calculation result passes the verification, further select a grid-side adapted and feasible electricity price calculation function from the dispatching strategy library, and load the electricity price calculation function into the electricity price operator to update the function of the electricity price operator; Obtain the electricity price calculation result on the grid side through the updated electricity price operator.

55. The device according to claim 53, characterized in that, The power function requirement for power trading between the grid side and the power plant side is specifically power clearing, the functional operator is a clearing operator, and the clearing operator is used for power clearing; the process of the digital power system using the agent constructed based on the clearing operator includes: Performing low-order data abstraction processing on the power generation element characteristics of each associated power plant of the grid to obtain a first data vector of each power generation element characteristic; Predicting the long-term trading electricity price according to the preset long-term electricity price function library and each first data vector to establish a long-term trading stack; Establish a spot trading stack based on the spot trading electricity prices of each of the associated power plants, and determine a first clearing function through the long-term trading stack and the spot trading stack; According to the first clearing function, determine the power supply-demand balance state of the power grid when clearing under the first clearing function; Adjust the first clearing function according to the power supply-demand balance state to obtain a second clearing function, and perform power clearing based on the second clearing function.

56. The device according to claim 53, characterized in that, The power function requirement for realizing the dispatching control of the grid side over the power plant side is specifically the peak shaving dispatching in the scenario of high-frequency variable load of thermal power. The functional operator is the thermal power peak shaving dispatching operator, and the thermal power peak shaving dispatching operator is used to realize the peak shaving dispatching in the scenario of high-frequency variable load of thermal power. The process of the digital power system applying the intelligent agent constructed based on the thermal power peak shaving dispatching operator includes: Calculate the heat storage and access margin of the energy storage device in the thermal power plant; the heat storage and access margin includes a heat storage margin characterization value and a heat release capacity characterization value; Send the correlation data between the loads, performances and costs of each thermal power unit in the thermal power plant and the heat storage and access margin to the grid side, so that the grid side determines the target units for expected auxiliary peak shaving and issues peak shaving dispatching instructions based on the change of thermal power load demand, the correlation data provided by each thermal power plant and the heat storage and access margin; Receive the peak shaving dispatching instructions issued by the grid side; the peak shaving dispatching instructions carry the unit identifier of the target unit and the peak shaving requirement information for the target unit; the peak shaving requirement information includes a peak shaving load curve; Execute the auxiliary peak shaving service based on the peak shaving requirement information.

57. The device according to claim 53, characterized in that, The power function requirement for realizing the dispatching control of the grid side over the power plant side is specifically the water resource dispatching of cascade hydropower stations. The functional operator is the water resource dispatching operator, and the water resource dispatching operator is used to realize the water resource dispatching of cascade hydropower stations. The process of the digital power system applying the intelligent agent constructed based on the water resource dispatching operator includes: Obtain the water resource data of each hydropower station in the cascade hydropower station at the current moment; For each hydropower station, judge whether the hydropower station is in a safe operation state according to the water resource data of the hydropower station; When it is determined that each hydropower station is in the safe operation state, judge whether to perform water resource dispatching according to the water resource data of each hydropower station, the preset electricity price scheme and the target historical data; the target historical data includes the average power generation water consumption and the average power plant revenue corresponding to the month where the current moment is located; If it is determined to perform water resource dispatching, then determine a water resource dispatching scheme based on the water resource data of each hydropower station and the water resource dispatching model; the water resource dispatching model is a pre-trained model for outputting a water resource dispatching scheme; the water resource dispatching scheme includes day-ahead dispatching and seasonal dispatching.

58. The device according to claim 53, characterized in that, The specific power function requirements for realizing the dispatching control of the grid side over the power plant side are specifically the dispatching of the power grid over the power plant in a large machine - small network power system. The functional operator is the large machine - small network dispatching operator, and the large machine - small network dispatching operator is used to realize the dispatching of the power grid over the power plant in the large machine - small network power system. The process of the digital power system applying the intelligent agent constructed based on the large machine - small network dispatching operator includes: Based on the extraction of the digital characteristics of the power grid and power source points of various power generation types in the power system, determine the accident reserve capacities of various power generation types to be reserved in the power system and configure the accident reserve capacities. Based on the output of the power source points of the current various power generation types, determine whether the current power system conforms to the large machine - small network characteristics. If it is determined that the current power system conforms to the large machine - small network characteristics, then the power grid uses the configured accident reserve capacities of various power generation types to dispatch the power source points in the power system for voltage adjustment based on a multi - round voltage adjustment scheme, and / or dispatch the power source points in the power system for frequency adjustment based on a multi - round frequency adjustment scheme. The multi - round voltage adjustment scheme includes: the multi - round voltage adjustment methods, voltage adjustment ranges, and voltage adjustment priority information for various power generation types. The multi - round frequency adjustment scheme includes: the multi - round frequency adjustment methods, frequency adjustment ranges, and frequency adjustment priority information for various power generation types. Based on the stability control requirements for the frequency dynamic characteristic index of the entire network of the power system, the predicted total network load, and the unit characteristics of each power source point, the power grid generates the output curves of the units of each power source point in the future time period. The frequency dynamic characteristic index is used to numerically represent the amount of power change of the entire network required to cause a unit frequency change in the power system. The power grid sends power generation dispatching instructions to each power source point for power adjustment. The power generation dispatching instructions include the output curves of the units of the corresponding power source points.

59. The device according to claim 53, characterized in that, The specific power function requirements for the power grid's own regulation are specifically to adjust the grid network structure. The functional operator is the grid structure adjustment operator, and the grid structure adjustment operator is used to realize the adjustment of the grid network structure. The process of the digital power system applying the intelligent agent constructed based on the grid structure adjustment operator includes: Obtain the current system inertia corresponding to the current grid structure. Based on the current system inertia and the difference between the power supply demand and the actual power supply, calculate the frequency change rate. If the frequency change rate is greater than a preset first threshold, then determine the target grid structure and the target value corresponding to the target grid structure based on a pre - established deep learning model. The frequency change rate of the target grid structure is less than or equal to the first threshold. Modify the current value applied in the current grid structure based on the target value and adjust the current grid structure of the power grid to the target grid structure.

Images