Generation method and device of operation scene information of power grid and processor

Abstract

The invention discloses a method and device for generating operation scene information of a power grid and a processor. The method comprises the steps that power grid operation data of a power grid are acquired, and the power grid operation data are at least used for representing the operation state of the power grid in the operation process; a data driving model is used for analyzing the power grid operation data to obtain first operation scene information, the data driving model is used for determining an operation mode and an operation rule of a power grid, and the first operation scene information is used for representing an operation state of the power grid under a normal operation condition; simulating the operation state of the power grid by using the physical model to generate second operation scene information which is used for representing the operation state of the power grid under the abnormal operation condition; and fusing the first operation scene information and the second operation scene information to obtain target operation scene information. The technical problem of low accuracy of operation scene generation of the power grid is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power grid scenario generation, in particular to a method and device for generating power grid operation scenario information and a processor. BACKGROUND

[0002] At present, a single data-driven method is often used to generate power grid operation scenario information (such as power grid control business scenario information) in the related art. The data-driven method highly depends on the richness and diversity of historical data. Based on statistical analysis and machine learning technology of historical data, the data-driven method can effectively capture the mode and rule of power grid operation. However, the data-driven method has the following disadvantages: due to the scarcity of extreme scenario data of the power grid, the data-driven method is difficult to generate such operation scenario information. In addition, in the face of high-dimensional random variables in power grid operation, such as the volatility and intermittency of new energy power generation, the data-driven method is difficult to effectively simulate the time sequence change of these random characteristics and cannot fully learn the complex interaction between related variables, resulting in deviation.

[0003] Therefore, there is still the technical problem of low accuracy of power grid operation scenario generation. SUMMARY

[0004] The embodiments of the present application provide a method and device for generating power grid operation scenario information and a processor to at least solve the technical problem of low accuracy of power grid operation scenario generation.

[0005] According to an aspect of the embodiments of the present application, a method for generating power grid operation scenario information is provided, comprising: obtaining power grid operation data of a power grid, wherein the power grid operation data is used to at least represent the operation state of the power grid in the operation process; analyzing the power grid operation data by using a data-driven model to obtain first operation scenario information, wherein the data-driven model is used to determine the operation mode and operation rule of the power grid, and the first operation scenario information is used to represent the operation state of the power grid under normal operation conditions; simulating the operation state of the power grid by using a physical model to generate second operation scenario information, wherein the physical model is used to generate operation scenario information that meets the operation rule and physical constraint of the power grid, and the second operation scenario information is used to represent the operation state of the power grid under abnormal operation conditions; and fusing the first operation scenario information and the second operation scenario information to obtain target operation scenario information, wherein the target operation scenario information is used to represent the operation state of the power grid under different operation conditions.

[0006] Optionally, the first operation scene information and the second operation scene information are fused to obtain target operation scene information, including: fusing the first operation scene information and the second operation scene information according to a fusion strategy to obtain a fusion result, wherein the fusion strategy is used to represent a rule of fusing the first operation scene information and the second operation scene information, and the fusion strategy includes series fusion and / or parallel fusion; and determining the target operation scene information based on the fusion result.

[0007] Optionally, the target operation scene information is determined based on the fusion result, including: screening the fusion result to obtain a set of operation scene information, wherein the set of operation scene information includes different operation scene information of the power grid; and sorting the different operation scene information in the set of operation scene information to obtain the target operation scene information, wherein the target operation scene information is operation scene information in the set of operation scene information that meets an operation scene information threshold.

[0008] Optionally, the first operation scene information is obtained by analyzing power grid operation data using a data-driven model, including: classifying the power grid operation data to obtain a classification result; screening the classification result to obtain key data in the power grid operation data, wherein the key data is used to represent an operation state and an operation trend of the power grid under normal operation conditions; identifying the key data using the data-driven model to obtain an identification result; predicting the operation state of the power grid under the normal operation conditions using the identification result; and generating the first operation scene information in response to a prediction result meeting a prediction result threshold; the method further includes: evaluating the first operation scene information to obtain a first evaluation result; and updating the data-driven model based on the first evaluation result.

[0009] Optionally, the second operation scene information is generated by simulating an operation state of the power grid using a physical model, including: simulating the operation state of the power grid based on a topology structure and an operation constraint condition of the power grid in the physical model to obtain a simulation result, wherein the topology structure is used to represent a connection rule and a physical position relationship between a plurality of components inside the power grid, and the operation constraint condition is used to represent a rule that needs to be met by the power grid during operation; extracting key information from the simulation result, wherein the key information is used to represent an operation state of the power grid under abnormal operation conditions; and generating the second operation scene information based on the key information; the method further includes: evaluating the second operation scene information to obtain a second evaluation result; and updating the physical model based on the second evaluation result.

[0010] Optionally, after the power grid operation data of the power grid is acquired, the method further includes: performing data cleaning processing and / or consistency checking processing on the power grid operation data respectively to obtain processed power grid operation data; performing standardization processing on the processed power grid operation data to obtain standardized processed power grid operation data, wherein the standardization processing is used to convert the power grid operation data into a unified scale range; and performing analysis on the standardized processed power grid operation data to determine the central tendency and dispersion degree of the processed power grid operation data in the unified scale range.

[0011] Optionally, the method further includes: collecting feedback information, wherein the feedback information is used to represent the deviation of the target operation scenario information from the predicted operation scenario information, wherein the predicted operation scenario information is used to represent the expected operation scenario information of the operation state of the power grid in a future period according to the power grid operation data before the target operation scenario information is applied; determining a parameter adjustment value corresponding to the feedback information, and adjusting the parameters in the data-driven model and the parameters in the physical model according to the parameter adjustment value; and generating the first operation scenario information and the second operation scenario information based on the adjusted data-driven model and the physical model.

[0012] According to another aspect of the embodiments of the present application, a device for generating operation scenario information of a power grid is also provided, which includes: an acquisition unit configured to acquire power grid operation data of the power grid, wherein the power grid operation data is used to at least represent the operation state of the power grid in an operation process; a first generation unit configured to analyze the power grid operation data by using a data-driven model to obtain first operation scenario information, wherein the data-driven model is used to determine the operation mode and operation law of the power grid, and the first operation scenario information is used to represent the operation state of the power grid under normal operation conditions; a second generation unit configured to simulate the operation state of the power grid by using a physical model to generate second operation scenario information, wherein the physical model is used to generate operation scenario information that conforms to the operation law and physical constraints of the power grid, and the second operation scenario information is used to represent the operation state of the power grid under abnormal operation conditions; and a fusion unit configured to fuse the first operation scenario information and the second operation scenario information to obtain target operation scenario information, wherein the target operation scenario information is used to represent the operation state of the power grid under different operation conditions.

[0013] According to another aspect of the embodiments of the present application, a processor is also provided, which is used to run a program, wherein the program is executed to perform any one of the above methods when the program is run by the processor.

[0014] According to another aspect of the embodiments of the present application, a computer readable storage medium is also provided, which stores a plurality of instructions, wherein the instructions are adapted to be loaded and executed by a processor to perform any one of the above methods.

[0015] According to another aspect of the embodiments of the present application, an electronic device is provided, comprising a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the computer program to perform any of the above methods.

[0016] According to another aspect of the embodiments of the present application, a computer program product is provided, comprising a computer program, wherein the computer program, when executed by a processor, implements any of the above methods.

[0017] In the embodiments of the present application, if it is necessary to generate the operation scenario of the power grid, the power grid operation data of the power grid can be acquired, wherein the power grid operation data is used at least to represent the operation state of the power grid in the operation process; the data-driven model can be used to analyze the power grid operation data to obtain first operation scenario information, wherein the data-driven model is used to determine the operation mode and operation law of the power grid, and the first operation scenario information is used to represent the operation state of the power grid under normal operation conditions; the physical model can be used to simulate the operation state of the power grid to generate second operation scenario information, wherein the physical model is used to generate operation scenario information conforming to the operation law and physical constraints of the power grid, and the second operation scenario information is used to represent the operation state of the power grid under abnormal operation conditions; and the first operation scenario information and the second operation scenario information can be fused to obtain target operation scenario information, wherein the target operation scenario information is used to represent the operation state of the power grid under different operation conditions. In the embodiments, through the mechanism of double driving of the data-driven model and the physical model, the data-driven model can learn the mode and operation law of the power grid operation, and generate representative scenario information (the first operation scenario information) based on the existing power grid operation data. The physical model is simulated according to the physical characteristics of the power grid, and even in the case of lacking historical data, the operation scenario information (the second operation scenario information) conforming to the physical constraints can be generated, which makes up for the deficiency of the data-driven model, thereby solving the technical problem of low accuracy of the operation scenario generation of the power grid, and achieving the technical effect of improving the accuracy of the operation scenario generation of the power grid. BRIEF DESCRIPTION OF DRAWINGS

[0018] The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiments of the application and serve to explain the principles of the application. In the drawings:

[0019] Figure 1 is a flowchart of a method for generating operation scenario information of a power grid according to an embodiment of the present application;

[0020] Figure 2 is a flowchart of a method for generating power grid regulation and control business scenarios based on data-model double driving according to an embodiment of the present application;

[0021] Figure 3 is a structural schematic diagram of a power grid operation scenario information generation device according to an embodiment of the present application;

[0022] Figure 4 is a schematic diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0023] In order to enable persons skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by persons skilled in the art without creative efforts should fall within the protection scope of the present application.

[0024] It should be noted that the terms "first", "second", and the like in the specification and claims of the present application and the above-described drawings are used to distinguish similar objects, and do not necessarily indicate a specific order or a chronological sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product, or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but can include other steps or units that are not clearly listed or inherent to the process, method, product, or device.

[0025] Embodiment 1

[0026] According to an embodiment of the present application, a power grid operation scenario information generation method embodiment is provided. It should be noted that the steps shown in the flowchart of the drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described herein can be executed in an order different from that shown herein.

[0027] Figure 1 is a flowchart of a power grid operation scenario information generation method according to an embodiment of the present application, as shown in Figure 1 the method comprises the following steps:

[0028] Step S102, obtaining power grid operation data of a power grid.

[0029] In the technical solution provided in the step S102 in the embodiment of the present application, the power grid operation data is used to at least represent the operation state of the power grid in the operation process. The power grid operation data can include historical load data, power generation data, equipment state monitoring data, meteorological data, and new energy prediction data, which are only used for illustration and are not specifically limited herein.

[0030] In this embodiment, in the generation process of the operation scene information of the power grid, such as in the power grid regulation service scene, the power grid operation data of the power grid can be acquired to ensure that the subsequently generated operation scene information is accurate and comprehensive.

[0031] Optionally, the historical load data can reflect the power demand mode of the historical time period. The power generation data includes the power generation amount of the historical time period and the current time period, the operation state of the power generation unit, the power generation type (such as thermal power, hydropower, wind power, and solar power), and the change situation thereof. The equipment state monitoring data provides the real-time operation state and the historical operation state of each device (such as a transformer, a circuit breaker, and a cable) in the power grid, such as the temperature, the pressure, the vibration, and the maintenance record of the device, which can be used to evaluate the health condition of the device, predict the service life and the fault risk of the device. The meteorological data covers environmental factors such as temperature, humidity, precipitation, wind speed, and wind direction. For the operation of the power grid, the meteorological data directly affects the load demand (such as the increase of air conditioning power consumption in summer due to high temperature), affects the output of the power generation unit (such as the influence of wind power and sunlight intensity on new energy power generation), and may induce natural disasters (such as thunderstorms and typhoons) to destroy the power grid. The new energy prediction data includes the power generation prediction of new energy (such as wind energy and solar energy) in the future time period.

[0032] Optionally, the historical load data can be acquired through a data acquisition and monitoring control system or an energy management system. The power generation data can be acquired through a data acquisition and monitoring control system or an energy management system, or can be acquired through an internal database of a power plant. The equipment state monitoring data can be acquired through a sensor. The meteorological data can be acquired through an Application Programming Interface (API) interface provided by a meteorological service agency, or can be acquired through satellite and remote sensing technology. The new energy prediction data can be acquired from new energy equipment manufacturers and operators.

[0033] It should be noted that the specific content and acquisition method of the power grid operation data are only used for illustration and are not specifically limited herein.

[0034] In step S104, a data-driven model is used to analyze the power grid operation data to obtain first operation scene information.

[0035] In the technical solution provided in step S104 of this embodiment of the invention, after obtaining the power grid operation data, a data-driven model can be used to analyze the power grid operation data to obtain first operation scenario information. The data-driven model is used to determine the power grid's operation mode and operation rules, and the first operation scenario information is used to represent the power grid's operating state under normal operating conditions.

[0036] In this embodiment, a suitable data-driven model can be selected based on the nature and objective of the power grid operation data to be analyzed. For example, for time-series data, such as historical load data and renewable energy forecast data, time-series analysis models can be used, such as the Autoregressive Integrated Moving Average (ARIMA) model and the Long Short-Term Memory (LSTM) network. For non-time-series data, machine learning models, such as decision trees, random forests, support vector machines, or neural networks, can be used.

[0037] Optionally, after determining the data-driven model, the model can be used to analyze current or predicted power grid operation data to generate first-stage operation scenario information. This includes, for example, predictions of the load level the power grid may face under normal operating conditions, the output of power generation units, the operating status of equipment, and the expected output of renewable energy generation. This first-stage operation scenario information is a quantitative description of the power grid's operating status and can be used for decision support in areas such as power grid dispatching, load forecasting, and renewable energy integration.

[0038] In this embodiment of the invention, the data-driven model can learn complex operating modes and patterns from historical data to predict the future operating state of the power grid, while the first operating scenario information is a description of the expected operating state of the power grid under normal operating conditions. The power grid operating state analysis based on the data-driven model generates first operating scenario information reflecting normal operating conditions, laying the foundation for subsequent operating scenario information fusion.

[0039] Step S106: Using a physical model, the operating state of the power grid is simulated to generate second operating scenario information.

[0040] In the technical solution provided in step S106 of this embodiment of the invention, after analyzing the power grid operation data using a data-driven model to obtain the first operation scenario information, a physical model can be used to simulate the operation state of the power grid and generate the second operation scenario information. The physical model is used to generate operation scenario information that conforms to the operation rules and physical constraints of the power grid, and the second operation scenario information is used to represent the operation state of the power grid under abnormal operating conditions.

[0041] In this embodiment, unlike the data-driven model which learns the pattern from historical data, the physical model deduces the response and state of the power grid under certain conditions based on the physical laws and equipment characteristics of the power grid. Optionally, according to the physical characteristics of the power grid, a corresponding physical model can be defined, such as a power system flow calculation model, a short-circuit fault analysis model, a device thermodynamic model, an electromagnetic transient model, etc. Each type of model focuses on describing a specific aspect of power grid operation, such as power flow, device temperature, fault current, etc. Since the accuracy and reliability of the physical model depend on the accuracy of the physical model parameters. Therefore, before the physical model is applied, the physical model parameters need to be calibrated according to the actual operation data and equipment specifications of the power grid, such as determining the rated capacity, resistance, reactance, conductance, etc. electrical parameters of the equipment, as well as the physical size, material characteristics, cooling system, etc. thermal parameters of the equipment.

[0042] Optionally, by using the physical model to simulate the operating state of the power grid, the voltage, current, power and equipment temperature, etc. state variables of the power grid can be calculated according to the set conditions to generate a series of possible power grid operating scenario information, that is, the second operating scenario information. The second operating scenario information can cover the response of the power grid under various abnormal conditions, such as voltage drop, current surge, etc. under the conditions of equipment short circuit, lightning strike, overload, etc.

[0043] In the embodiments of the present application, the advantage of the physical model is that it can accurately simulate the physical process of the power grid, especially the response under complex conditions and extreme situations, which helps to improve the reliability of power grid regulation. By generating the second operating scenario information reflecting the power grid under abnormal operating conditions, the physical accuracy and operating constraints that the data-driven model may lack are supplemented, providing more comprehensive and accurate operating scenario information prediction for power grid regulation.

[0044] Step S108, fusing the first operating scenario information and the second operating scenario information to obtain target operating scenario information.

[0045] In the technical solution provided by step S108 in the embodiments of the present application, after using the physical model to simulate the operating state of the power grid and generating the second operating scenario information, the first operating scenario information and the second operating scenario information can be fused to obtain the target operating scenario information. The target operating scenario information is used to represent the operating state of the power grid under different operating conditions.

[0046] In this embodiment, the first operation scene information generated by fusing the data-driven model and the second operation scene information generated by the physical model can fuse the advantages of the two models, and ensure that the generated target operation scene information not only conforms to the historical data trend, but also meets the physical constraint condition. The generation of the target operation scene information enables the power grid control personnel to predict the operation state of the power grid in advance under normal conditions and abnormal conditions, so as to take more effective control and response measures and ensure the safe and stable operation of the power grid.

[0047] Optionally, the fusion of the first operation scene information and the second operation scene information can be performed by weighted average fusion, for example, a weight is assigned to each operation scene information based on the credibility and accuracy of the first operation scene information and the second operation scene information, and the information with high credibility obtains a higher weight; the first operation scene information and the second operation scene information are weighted and averaged according to their respective weights. It can also be fused by rules, that is, a set of rules is prepared in advance to specify which operation scene information to use under which conditions, or how to dynamically adjust the fusion method according to the conditions.

[0048] In steps S102 to S108 of the embodiment of the present application, if it is necessary to generate the operation scene of the power grid, the power grid operation data of the power grid can be obtained, wherein the power grid operation data is used to at least represent the operation state of the power grid in the operation process; the data-driven model can be used to analyze the power grid operation data to obtain the first operation scene information, wherein the data-driven model is used to determine the operation mode and operation law of the power grid, and the first operation scene information is used to represent the operation state of the power grid under normal operation conditions; the physical model can be used to simulate the operation state of the power grid to generate the second operation scene information, wherein the physical model is used to generate the operation scene information that conforms to the operation law and physical constraint of the power grid, and the second operation scene information is used to represent the operation state of the power grid under abnormal operation conditions; the first operation scene information and the second operation scene information can be fused to obtain the target operation scene information, wherein the target operation scene information is used to represent the operation state of the power grid under different operation conditions. In this embodiment, through the dual driving mechanism of the data-driven model and the physical model, the data-driven model can learn the mode and operation law of the power grid operation, and generate representative scene information (the first operation scene information) based on the existing power grid operation data. The physical model simulates according to the physical characteristics of the power grid, and even in the case of lack of historical data, it can also generate the operation scene information (the second operation scene information) that conforms to the physical constraint, making up for the deficiency of the data-driven model, thereby solving the technical problem of low accuracy of the generated operation scene of the power grid, and achieving the technical effect of improving the accuracy of the generated operation scene of the power grid.

[0049] The embodiment of the present application will be described in detail below in combination with the above steps.

[0050] As an optional embodiment, in step S108, the first operation scene information and the second operation scene information are fused to obtain target operation scene information, including: fusing the first operation scene information and the second operation scene information according to a fusion strategy to obtain a fusion result, wherein the fusion strategy is used to represent a rule of fusing the first operation scene information and the second operation scene information, and the fusion strategy includes series fusion and / or parallel fusion; and determining the target operation scene information based on the fusion result.

[0051] In this embodiment, in the process of fusing the first operation scene information and the second operation scene information to obtain the target operation scene information, the first operation scene information and the second operation scene information can be fused according to series fusion and / or parallel fusion to obtain a fusion result, and the target operation scene information is determined based on the fusion result.

[0052] Optionally, in series fusion, the power grid operation data is processed using a data-driven model to generate the first operation scene information, which reflects the trend prediction of the power grid under normal conditions based on historical data. Then, the first operation scene information is input as an input together with related operation conditions and parameters into a physical model for a second-stage simulation and constraint check. The physical model generates the second operation scene information based on this, focusing on ensuring the physical rationality of the scene, such as satisfying the power balance, voltage stability and other constraint conditions, and the output of the physical model is regarded as the fusion result, i.e., the target operation scene information.

[0053] Optionally, in parallel fusion, the data-driven model and the physical model independently predict the power grid operation state to generate the first operation scene information and the second operation scene information, respectively. This parallel operation allows both methods to exert their respective advantages, with the data-driven model emphasizing the flexibility and historical trend of prediction, and the physical model focusing on physical laws and constraint conditions. Next, the first operation scene information and the second operation scene information are merged, possibly through weighted averaging, voting mechanism, Bayesian fusion or other statistical methods. The fusion process needs to consider the reliability and correlation of each piece of information to determine appropriate weights and fusion rules. Based on the fusion result, the target operation scene information is determined, which is the best prediction that integrates historical data trends and physical constraint conditions, considering both typical patterns of power grid operation and physical operation laws of the power grid, providing a more comprehensive basis for decision-making for power grid regulation.

[0054] Optionally, the selection of the fusion strategy can be determined according to the specific needs of power grid operation, data quality, and the maturity of the physical model. Series fusion can ensure the physical reasonableness of scenario prediction, while parallel fusion can better balance the flexibility of data-driven models and the accuracy of physical models. The final target operation scenario information will be able to more comprehensively and accurately reflect the state of the power grid under different operating conditions, providing strong data support and decision assistance for power grid regulation business.

[0055] As an optional embodiment, based on the fusion result, the target operation scenario information is determined, including: screening the fusion result to obtain a set of operation scenario information, wherein the set of operation scenario information includes different operation scenario information of the power grid; and sorting the different operation scenario information in the set of operation scenario information to obtain the target operation scenario information, wherein the target operation scenario information is operation scenario information in the set of operation scenario information that meets an operation scenario information threshold.

[0056] In this embodiment, in the process of determining the target operation scenario information based on the fusion result, the fusion result can be screened to obtain a set of operation scenario information of different operation scenario information of the power grid, and the different operation scenario information in the set of operation scenario information can be sorted to obtain the target operation scenario information that meets the operation scenario information threshold.

[0057] Optionally, based on the characteristics of power grid operation, screening criteria can be defined, such as the credibility of the scenario, the potential impact on the power grid system, the complexity of the scenario, etc. These criteria can be quantified through the operation scenario information threshold to ensure that the selected scenario information has high credibility and importance.

[0058] Optionally, after screening, the operation scenario information in the set of operation scenario information can be sorted by priority to obtain the most urgent or important operation scenario information as the target operation scenario information. The operation scenario information threshold can be set according to the strategy and target of power grid regulation, such as a credibility threshold, an impact degree threshold, etc. This is only an example and the operation scenario information threshold is not limited here.

[0059] As an optional embodiment, in step S104, the power grid operation data is analyzed by using the data-driven model to obtain first operation scenario information, including: classifying the power grid operation data to obtain a classification result; screening the classification result to obtain key data in the power grid operation data, wherein the key data is used to represent the operation state and operation trend of the power grid under normal operation conditions; identifying the key data by using the data-driven model to obtain an identification result; predicting the operation state of the power grid under normal operation conditions by using the identification result; and generating the first operation scenario information in response to the prediction result satisfying a prediction result threshold. The method further includes: evaluating the first operation scenario information to obtain a first evaluation result; and updating the data-driven model based on the first evaluation result.

[0060] In this embodiment, in the process of analyzing the power grid operation data by using the data-driven model to obtain the first operation scenario information, the power grid operation data can be classified to determine the structure and mode of the power grid operation data. For example, the most relevant features to the state and trend of the power grid are selected from the power grid operation data, such as load, power generation, equipment operation state, etc., and appropriate machine learning algorithms are used for classification, such as K-means clustering, decision tree, random forest, etc., to reveal the potential structure of the power grid operation data, and the power grid operation data is divided into several categories, each of which represents a specific mode or state of the power grid operation.

[0061] Optionally, after classification, the classification result can be screened to obtain key data that can reflect the normal operation state and trend of the power grid and reduce the interference of noise and irrelevant information. Then, the key data can be analyzed and identified by using the data-driven model to predict the operation state and trend of the power grid under normal operation conditions. By comparing the prediction result with the actual operation data, the prediction accuracy of the data-driven model is evaluated to determine whether the prediction result threshold is satisfied. In the case where the prediction result satisfies the prediction result threshold, the first operation scenario information is generated.

[0062] Optionally, after the first operation scenario information is generated, the first operation scenario information can be evaluated, such as the credibility of the scenario, the prediction accuracy, the degree of agreement with historical data, etc., and the evaluation result is a first evaluation result. Based on the first evaluation result, if it is found that there is a deviation in the model prediction or the quality of the scenario information does not meet the requirements, the data-driven model is updated, which can involve model parameter adjustment, structure optimization or data set retraining, etc.

[0063] As an optional embodiment, in step S106, the operation state of the power grid is simulated by using the physical model to generate second operation scenario information, including: in the physical model, the operation state of the power grid is simulated based on the topology structure of the power grid and the operation constraint condition to obtain a simulation result, wherein the topology structure is used to represent the connection rule and the physical position relationship between the plurality of components in the power grid, and the operation constraint condition is used to represent the rule to be met by the power grid in the operation process; key information is extracted from the simulation result, wherein the key information is used to represent the operation state of the power grid under the abnormal operation condition; and the second operation scenario information is generated based on the key information; the method further includes: evaluating the second operation scenario information to obtain a second evaluation result; and updating the physical model based on the second evaluation result.

[0064] In this embodiment, in the process of simulating the operation state of the power grid by using the physical model to generate the second operation scenario information, the operation state of the power grid can be simulated based on the topology structure of the power grid, such as the connection rule and the physical position relationship between the elements such as the transmission line, the transformer, the generator, and the load, and the physical constraint condition of the power grid operation, such as the voltage stability, the frequency stability, and the power flow limit, to obtain a simulation result.

[0065] Optionally, after obtaining the simulation result, key information in the simulation result can be extracted, and the above key information focuses on the state of the power grid under abnormal operation conditions, such as fault state, device overload, and unstable operation, and all scene information under abnormal operation conditions is summarized to form complete second operation scenario information.

[0066] Optionally, after generating the second operation scenario information, the effectiveness and the reliability of the second operation scenario information can be evaluated, and the physical model is optimized accordingly to ensure that the physical model can continuously and accurately simulate the state of the power grid under abnormal operation conditions.

[0067] As an optional embodiment, after obtaining the power grid operation data of the power grid, the method further includes: respectively performing data cleaning processing and / or consistency checking processing on the power grid operation data to obtain processed power grid operation data; performing standardization processing on the processed power grid operation data to obtain standardized processed power grid operation data, wherein the standardization processing is used to convert the power grid operation data into a unified scale range; and analyzing the standardized processed power grid operation data to determine the central tendency and the dispersion degree of the processed power grid operation data in the unified scale range.

[0068] In this embodiment, after obtaining the grid operation data of the power grid, preliminary arrangement can be performed on the grid operation data, such as checking the integrity of the data to ensure that there is no missing important information, classifying and labeling the data to facilitate subsequent processing, and converting data in different formats to a unified format to facilitate data fusion and analysis. For example, load data from different sources is unified into the same time series format, meteorological data and new energy prediction data are spatio-temporally matched, and the like. Through these arrangement work, the availability and consistency of the data can be ensured, and a good foundation is laid for subsequent data cleaning and standardization processing.

[0069] Optionally, after the preliminary arrangement of the grid operation data, data cleaning and / or consistency checking can be performed on the grid operation data. The purpose of data cleaning is to remove noise and outliers in the data, correct inconsistent data records, and fill in missing data points, thereby improving the quality of the data. In this process, preliminary checks can be performed on the data to identify obvious errors and outliers. For example, some data points may obviously exceed the reasonable range, or there may be obvious entry errors. For these problems, various methods can be used for processing. For outliers, statistical methods such as the three-sigma principle (3σ principle) or box plot analysis can be used to identify and eliminate them; machine learning algorithms such as Isolation Forest or Local Outlier Factor can also be used to detect and handle complex abnormal situations. For missing values, records containing missing values can be deleted, mean or median values can be used for filling, or more advanced interpolation algorithms such as linear interpolation, spline interpolation, etc. can be used to generate more reasonable filling values.

[0070] Optionally, consistency checks can also be performed on the data to correct inconsistent records. For example, check whether the records of different data sources at the same time point are consistent, or check whether the time series of the data is continuous. Through the above operations, the accuracy and reliability of the data can be significantly improved.

[0071] Optionally, after the data cleaning and / or consistency checking of the grid operation data, the processed grid operation data can be standardized to convert data from different sources, different dimensions, and different dimension units to a unified scale range. For example, standardization can be performed through Z-score standardization, min-max standardization, and decimal scaling standardization. Among them, Z-score standardization converts data points to a distribution with a mean of 0 and a standard deviation of 1, which is suitable for cases where data is normally distributed; min-max standardization linearly transforms data to a specified interval, such as [0, 1], which is suitable for cases where data distribution is relatively uniform; decimal scaling standardization scales data to a specified range by moving the decimal point, which is suitable for cases where data is large or the data range is wide.

[0072] Optionally, after the processed power grid operation data is standardized, the standardized processed power grid operation data can be analyzed to determine the concentration trend and dispersion degree of the processed power grid operation data in a unified scale range.

[0073] As an optional embodiment, the method further comprises collecting feedback information, wherein the feedback information is used to represent the deviation of the target operation scenario information and the predicted operation scenario information, wherein the predicted operation scenario information is used to represent the expected operation scenario information of the operation state of the power grid in the future period according to the power grid operation data before the target operation scenario information is applied; determining the parameter adjustment value corresponding to the feedback information, and adjusting the parameters in the data-driven model and the parameters in the physical model according to the parameter adjustment value; generating the first operation scenario information and the second operation scenario information based on the adjusted data-driven model and the physical model.

[0074] In this embodiment, the main purpose of feedback collection and analysis is to collect feedback information in actual application, evaluate the actual effect of the scenario generation method, and provide basis for subsequent optimization.

[0075] Optionally, feedback information from power grid control personnel, operation and maintenance personnel, and relevant management departments is collected. These feedback information includes evaluation of the accuracy, practicality and reliability of the generated scenario information, as well as problems and challenges encountered in actual application. For example, control personnel may feedback that some scenario information deviates from the actual operation situation, or that the guidance of the scenario is not clear enough in specific situations. At the same time, quantitative analysis of application effect is carried out, such as comparing the difference between the dispatching decision based on the generated scenario information and the actual operation result, evaluating the influence of scenario information generation on power grid operation efficiency, new energy consumption rate, fault response speed and other indicators. Through these feedback and analysis, the performance of the generated operation scenario information in actual application can be deeply understood, and a clear direction is provided for subsequent optimization work.

[0076] Optionally, if the prediction accuracy of the model is found to be insufficient in certain specific scenarios, the parameters, structure or training algorithm of the model can be adjusted to improve its performance in these scenarios. For data-driven models, the model can be retrained, more relevant data can be introduced or more advanced machine learning algorithms can be used; for physical models, the model parameters can be recalibrated, and the boundary conditions and constraint settings of the model can be improved. At the same time, the strategy and method of scenario generation can be optimized. For example, the fusion strategy of data-driven and model-driven scenarios can be adjusted, and the index system of scenario screening and evaluation can be optimized to generate more comprehensive scenarios that meet the actual needs. In addition, according to the problems in actual application, the resolution, time scale, spatial coverage, etc. of the scenario can be adjusted and optimized. Through these optimization measures, the quality and effect of the scenario generation method can be continuously improved, and it can better serve the actual needs of the power grid control business, and provide more powerful support for the safe and stable operation and optimal scheduling of the power grid.

[0077] In the embodiment of the application, if it is necessary to generate a running scenario of a power grid, power grid running data of the power grid can be acquired, wherein the power grid running data is used at least to represent a running state of the power grid in a running process; a data-driven model can be used to analyze the power grid running data to obtain first running scenario information, wherein the data-driven model is used to determine a running mode and a running law of the power grid, and the first running scenario information is used to represent a running state of the power grid under a normal running condition; a physical model can be used to simulate the running state of the power grid to generate second running scenario information, wherein the physical model is used to generate running scenario information that meets the running law and physical constraints of the power grid, and the second running scenario information is used to represent a running state of the power grid under an abnormal running condition; and the first running scenario information and the second running scenario information can be fused to obtain target running scenario information, wherein the target running scenario information is used to represent a running state of the power grid under different running conditions. In this embodiment, through the dual driving mechanism of the data-driven model and the physical model, the data-driven model can learn the mode and the running law of the power grid, and generate representative scenario information (the first running scenario information) based on existing power grid running data. The physical model simulates according to the physical characteristics of the power grid, and even in the case of lack of historical data, it can generate running scenario information (the second running scenario information) that meets the physical constraints, making up for the shortcomings of the data-driven model, thereby solving the technical problem of low accuracy of the running scenario generation of the power grid, and achieving the technical effect of improving the accuracy of the running scenario generation of the power grid.

[0078] Embodiment 2

[0079] The following will be described in detail in combination with another optional specific embodiment.

[0080] Currently, in the field of power grid regulation business scenario generation method, the existing methods can be mainly divided into data-driven method and model-driven method. Although the existing research has made certain progress, there are still many deficiencies. The data-driven method is highly dependent on historical data, but the extreme scenario data of the power grid is scarce, which leads to insufficient reliability and accuracy of the generated scenario. At the same time, these methods have limitations in dealing with high-dimensional complex random characteristics, and it is difficult to effectively simulate the time sequence change details under extreme scenarios. In addition, part of the data-driven model (such as variational autoencoder) assumes that the conditional distribution is a normal distribution, which is difficult to accurately restore the real probability distribution, resulting in poor detail feature restoration ability of the generated scenario. The model-driven method has strong interpretability, but in a complex power grid system, the model complexity is high, and it is difficult to effectively handle multi-dimensional uncertainty. At the same time, this kind of method needs a large amount of historical sample data to construct the model, but the extreme scenario sample is scarce, which leads to low reliability of the generated scenario. In addition, the existing model-driven method usually only considers single or a few source-side factors, and lacks consideration of the correlation of multiple source-side factors and the "source-load" correlation, so there is a big gap between the generated scenario and the real scenario.

[0081] However, the present application proposes a power grid regulation business scenario generation method based on data-model double driving, which solves the problems of existing power grid regulation business scenario generation methods in extreme scenario generation, complex condition adaptability and data and model fusion, improves the accuracy, reliability and flexibility of scenario generation, provides more accurate decision support for power grid regulation, plays an important role in ensuring the safe and stable operation of power grid, helping new energy consumption and optimizing the efficiency of power grid operation.

[0082] The method will be further introduced below.

[0083] In this embodiment, Figure 2 is a flow chart of a power grid regulation business scenario generation method based on data-model double driving according to an embodiment of the present application, as Figure 2 shown, the method comprises the following steps:

[0084] Step S201, determine the data source.

[0085] In this embodiment, in the data collection and preprocessing stage, it is crucial to determine the data sources, which need to obtain various types of data related to power grid regulation business from multiple channels to ensure the comprehensiveness and accuracy of the data. For example, historical load data, power generation data, equipment state monitoring data, weather data, and new energy prediction data. Historical load data can be used to understand the past power demand patterns, while power generation data provides the output of generating units. Equipment state monitoring data is essential for assessing the health of power grid equipment, and weather data and new energy prediction data help to predict the volatility and intermittency of new energy power generation. To achieve this goal, cooperation can be established with power grid operation and management agencies, meteorological departments, and new energy power generation enterprises to obtain high-quality raw data. At the same time, the source of the data needs to be strictly screened to ensure the reliability and timeliness of the data. By integrating these multi-source data, a solid data foundation can be laid for the subsequent scene generation method, ensuring that the generated power grid regulation business scenarios not only conform to historical laws but also adapt to future operation needs.

[0086] Step S202, data acquisition and arrangement.

[0087] In this embodiment, data acquisition and arrangement is a key step in the data collection and preprocessing stage. When acquiring data, cooperation can be established with power grid operation and management agencies, meteorological departments, etc. through data sharing agreements, API interfaces, etc. to obtain high-quality raw data. These data include historical load data, power generation data, equipment state monitoring data, weather data, and new energy prediction data. After obtaining the data, it needs to be preliminarily arranged, such as checking the completeness of the data to ensure that there is no missing important information, classifying and labeling the data for subsequent processing, and converting different formats of data into a unified format for data fusion and analysis. For example, unifying load data from different sources into the same time series format, and spatiotemporally matching weather data and new energy prediction data. Through these arrangement work, the usability and consistency of the data can be ensured, laying a good foundation for subsequent data cleaning and standardization processing.

[0088] Step S203, data cleaning.

[0089] In this embodiment, the purpose of data cleaning is to remove noise and outliers in the data, correct inconsistent data records, and fill in missing data points, thereby improving the quality of the data. In this process, first, the data is preliminarily checked to identify obvious errors and outliers. For example, some data points may be obviously beyond the reasonable range, or there may be obvious entry errors. For these problems, various methods can be used for processing. For outliers, statistical methods such as the 3σ principle or box plot analysis can be used to identify and eliminate them; or machine learning algorithms such as Isolation Forest or Local Outlier Factor can be used to detect and handle complex abnormal situations. For missing values, common processing methods include deleting records containing missing values, filling in using mean or median, or using more advanced interpolation algorithms such as linear interpolation, spline interpolation, etc. to generate more reasonable filling values. In addition, consistency checks will be performed on the data to correct inconsistent records. For example, check whether the records from different data sources at the same time point are consistent, or check whether the time series of the data is continuous. Through these cleaning operations, the accuracy and reliability of the data can be significantly improved, providing high-quality data support for subsequent modeling and analysis.

[0090] Step S204, data standardization.

[0091] In this embodiment, the purpose of data standardization is to convert data from different sources, different dimensions, and different dimension units to a unified scale range, so as to facilitate subsequent data fusion and modeling analysis. In this process, appropriate methods need to be selected for data standardization processing according to the specific data characteristics and the needs of subsequent modeling analysis. Standardization methods can include Z-score standardization, min-max standardization, and decimal scaling standardization, etc. Z-score standardization converts data points to a distribution with a mean of 0 and a standard deviation of 1, which is suitable for data that is normally distributed; min-max standardization linearly transforms data to a specified interval, such as [0, 1], which is suitable for data that is relatively uniform; decimal scaling standardization scales data to a specified range by moving the decimal point, which is suitable for cases where data is large or the data range is wide. When selecting a standardization method, factors such as the distribution characteristics of the data, the sensitivity of the model to the scale of the data, and the needs of the actual application, etc. need to be considered. For example, for models such as neural networks that are sensitive to the scale of input data, min-max standardization may be more appropriate; while for clustering algorithms based on distance calculation, Z-score standardization may be more beneficial to highlight the distribution characteristics of the data. Through data standardization, the differences in dimensions and dimension units between different data can be eliminated, allowing data to be compared and analyzed on the same scale, thereby improving the performance and stability of the model.

[0092] Step S205, data analysis.

[0093] In this embodiment, the purpose of data analysis is to reveal the basic characteristics and inherent laws of power grid operation through in-depth analysis of a large amount of collected data, and to provide a basis for subsequent feature extraction and model construction. In this process, two methods can be used, statistical analysis and clustering analysis. In terms of statistical analysis, basic statistical indicators of the data are calculated, such as mean, variance, median, extreme value, etc., to understand the concentration trend and dispersion degree of the power grid operation data. At the same time, the correlation between different variables can also be analyzed, such as load and meteorological factors, new energy power generation and time series, etc. By drawing scatter plots, calculating correlation coefficients and other methods, the main trends and periodic patterns in the data can be identified. Clustering analysis is used to divide the power grid operation data into different categories and identify operation states with similar characteristics. By selecting appropriate clustering algorithms, such as K-means clustering algorithm, Density-Based Spatial Clustering of Applications with Noise (DBSCAN), etc., the data is grouped according to its characteristics and analysis objectives. For example, clustering can be performed according to load level, new energy generation proportion, etc., to discover typical patterns and abnormal situations in power grid operation. Through these analyses, a comprehensive understanding of the distribution characteristics and internal structure of the power grid operation data can be obtained, laying a foundation for subsequent feature extraction and scenario generation.

[0094] Step S206, feature extraction.

[0095] In this embodiment, the purpose of feature extraction is to identify and extract the most representative and influential features for power grid regulation business from the analyzed data, so as to efficiently utilize these key information in subsequent model construction and scenario generation. This process generally includes two aspects of key feature selection and dimension reduction. In the aspect of key feature selection, combined with the expert knowledge of power grid operation and data analysis results, the features that have decisive influence on the operation state of power grid are determined. For example, load peak valley difference, new energy power generation volatility, equipment failure rate, etc. These features can directly reflect the operation pressure and stability of power grid. At the same time, the feature importance evaluation method such as decision tree based feature selection, correlation coefficient analysis, etc. can be used to further screen out the features that contribute most to scenario generation. Dimension reduction is to reduce the dimension of data and improve the efficiency and performance of the model. Principal component analysis, linear discriminant analysis and other dimension reduction techniques can be used to project high-dimensional data into low-dimensional space while retaining as much of the main information in the data as possible. For example, through principal component analysis, multiple correlated variables can be converted into several uncorrelated principal components, which can explain most of the variance in the data, thus reducing the dimension of the data while maintaining the key features of the data. Through feature extraction, a complex data set can be simplified into a set of representative features, which can not only effectively describe the operation state of power grid, but also significantly improve the computational efficiency and accuracy of subsequent models.

[0096] Step S207, physical model construction.

[0097] In this embodiment, the core of physical model construction is to establish a model framework that can accurately describe the operation state and characteristics of power grid according to the basic physical laws and mathematical theories of power grid operation. The key of this step is to select appropriate model types and set specific parameters of the model according to the topology structure, equipment parameters and operation constraints of the actual power grid. When constructing the physical model, first refer to the classical theories and methods in the field of power system analysis, for example, the power flow calculation model of power system is used to simulate the power flow under different operation conditions, and the short circuit fault analysis model is used to evaluate the current impact and voltage change of power grid when short circuit and other abnormal conditions occur. At the same time, considering the complexity and diversity of power grid, the model can be appropriately adjusted and extended according to the characteristics of the actual power grid to ensure that it can accurately reflect the operation characteristics of the target power grid. In order to ensure the accuracy and reliability of the physical model, it also needs to be calibrated and verified by actual power grid operation data, such as using historical data to estimate and optimize the parameters of the model, and testing the prediction ability of the model by simulating known operation scenarios. After strict calibration and verification, the physical model can provide a solid foundation for subsequent scenario generation, ensuring that the generated scenarios conform to the actual operation rules and physical constraints of the power grid.

[0098] Step S208, data-driven model construction.

[0099] In this embodiment, the data-driven model construction mainly utilizes pre-processed and feature-extracted data to train a model that can capture the operation patterns and rules of the power grid through machine learning and deep learning techniques. The key of this step is to select a suitable model architecture according to the characteristics of the data and the task requirements, and to make the model achieve the best performance through effective training strategies. When selecting the model architecture, factors such as the type and dimension of the data, and the complexity of the power grid operation can be considered. For example, for load data and new energy generation data with time series characteristics, recurrent neural network architectures such as long short-term memory network may be more suitable, as they can effectively handle the temporal dependencies in sequence data. For high-dimensional and nonlinear data, deep feedforward neural networks or convolutional neural networks may be more advantageous. In the model training process, cross-validation and other methods are used to optimize the hyperparameters of the model, prevent overfitting and underfitting phenomena, and ensure that the model has good generalization ability. At the same time, the training process can be monitored, and through observation of indicators such as the trend of the loss function, the training strategy can be adjusted in time to improve the convergence speed and accuracy of the model. The advantage of data-driven models is that they can automatically learn complex patterns and rules from a large amount of historical data, so that they can make accurate predictions and scenario generation based on data characteristics when facing new operating conditions. Data-driven models will be combined with physical models to provide support for the generation of power grid regulation business scenarios, make up for the shortcomings of single models, and improve the accuracy and flexibility of scenario generation.

[0100] Step S209, hybrid model construction.

[0101] In this embodiment, the hybrid model construction aims to organically combine the physical model and the data-driven model to form a comprehensive model that has both the constraints of physical laws and the flexibility to adapt to data characteristics. The core of this step is to design a reasonable fusion strategy to make the advantages of both models fully play while making up for their respective shortcomings. In terms of fusion strategy, it can be achieved through model series connection, parallel connection, and weight-based fusion, etc. For example, in series fusion, the data-driven model can first extract features from the input data and make a preliminary prediction, then pass the result to the physical model, which will further correct and optimize it according to the physical laws of the power grid, and finally output more accurate scenario results. In parallel fusion, both models can process the same input data simultaneously, and then the output results of the two are integrated through a certain weight distribution mechanism to obtain the final scenario prediction. In order to determine the best fusion strategy and parameters, a large number of experiments and verifications can be carried out to compare the performance of different fusion methods under various operating scenarios. In this way, a hybrid model that conforms to the physical operating laws of the power grid and can flexibly adapt to complex data characteristics can be constructed, providing strong support for subsequent scenario generation and ensuring that the generated scenarios are accurate and practical.

[0102] Step S210, data-driven scenario generation.

[0103] In this embodiment, the purpose of data-driven scenario generation is to use the trained and optimized data-driven model to predict the possible future operating state of the power grid based on the patterns and laws in historical data, and generate diverse scenarios. This helps to fully understand the potential operating conditions of the power grid under different conditions and provides data support for subsequent analysis and decision-making. In this process, first, ensure that the data-driven model has been trained and verified with a large amount of historical data and has good prediction performance. Then, according to the research objectives and time range, determine the input variables and initial conditions of the model. For example, when predicting the load scenario for the next 24 hours, the input variables include historical load data, weather forecasts, and new energy generation forecasts, etc. Next, run the model to generate a large number of possible power grid operating scenarios, and use statistical and machine learning methods to select representative and reasonable scenarios to ensure that these scenarios can cover different operating states and conform to the basic laws of power grid operation.

[0104] Step S211, physical-driven scenario generation.

[0105] In this embodiment, the core of model-driven scenario generation is based on the physical model of the power grid, simulating the state of the power grid under different operating conditions, and generating scenarios that conform to physical laws. This method makes up for the shortcomings of data-driven methods in generating extreme scenarios, ensuring the accuracy and credibility of the scenarios. In this process, first ensure that the physical model has been strictly calibrated and verified, and can accurately reflect the operating characteristics and physical laws of the target power grid. Then, set the parameters of the model according to the research objectives, such as device operating state, fault mode, etc. Next, define the operating conditions of the model, including normal operation, device failure and extreme weather, and set the corresponding boundary conditions and constraints, such as voltage range and device capacity limit. Finally, run the model to simulate the operating state of the power grid under these set conditions, generating the corresponding scenarios. To improve the rationality and accuracy of the scenarios, optimization algorithms can be used to adjust the model parameters, ensuring that the generated scenarios not only conform to the physical laws, but also meet the requirements of actual operation.

[0106] Step S212, scenario fusion.

[0107] In this embodiment, scenario fusion aims to integrate data-driven and model-driven generated scenarios, taking their strengths and weaknesses, and generating more comprehensive and accurate integrated scenarios. Through reasonable weight allocation and fusion strategy, the flexibility of data-driven and the accuracy of model-driven are balanced, and the overall quality of the scenarios is improved. In this process, first select the appropriate fusion strategy according to the characteristics and application requirements of the scenarios, such as weighted average, voting mechanism, etc. At the same time, consider the credibility and representativeness of the scenarios, determine the weight allocation method of each scenario. Then, the scenarios generated by the two methods are fused according to the selected strategy, and the fused scenarios are standardized to ensure the consistency of format and dimension, so as to facilitate subsequent application and analysis. Finally, evaluate the fused scenarios, check their rationality and completeness, and analyze whether there are omissions or unreasonable places. According to the evaluation results, adjust the fusion strategy and weight allocation, optimize the fused scenarios, and ensure that they can accurately reflect the operating state of the power grid, providing reliable decision support for power grid regulation.

[0108] Step S213, scenario application.

[0109] In this embodiment, the purpose of the scene application is to actually apply the comprehensive scene generated in the previous stage to various businesses of power grid regulation, providing decision support for the safe, stable and efficient operation of the power grid. In this process, first, the generated scene data is integrated into the business systems of power grid regulation, such as energy management systems, power grid dispatching automation systems, etc. These scene data can provide scientific basis for dispatchers to make scheduling plans, conduct load forecasting, arrange equipment maintenance, etc. For example, when making scheduling plans, dispatchers can refer to the generated scene to estimate the state of the power grid under different operating conditions, thereby making more reasonable and flexible scheduling plans. In addition, the generated scene can also be used for fault simulation and emergency drills of the power grid. By simulating various fault scenarios, regulatory personnel can familiarize themselves with response measures in advance and improve emergency handling capabilities. In terms of new energy consumption, scene data can help analyze the impact of the volatility and intermittency of new energy generation on the power grid, thereby developing more effective consumption strategies.

[0110] Step S214, feedback collection and analysis.

[0111] In this embodiment, the main purpose of feedback collection and analysis is to collect feedback information in actual application, evaluate the actual effect of the scene generation method, and provide basis for subsequent optimization. This step is crucial for continuously improving the scene generation method and improving its applicability. In the actual application process, feedback opinions (feedback information) from power grid regulatory personnel, operation and maintenance personnel, and relevant management departments can be collected. These feedback information includes evaluation of the accuracy, practicality and reliability of the generated scene, as well as problems and challenges encountered in actual application. For example, regulatory personnel may feedback that some scenes deviate from the actual operating conditions, or that the guidance of the scene is not clear enough in specific situations. At the same time, quantitative analysis of application effect is carried out, such as comparing the differences between scheduling decisions based on generated scenes and actual operating results, evaluating the influence of scene generation method on power grid operation efficiency, new energy consumption rate, fault response speed, etc. Through these feedback and analysis, the performance of the scene generation method in actual application can be deeply understood, providing a clear direction for subsequent optimization work.

[0112] Step S215, model and scene optimization.

[0113] In this embodiment, model and scenario optimization aims to optimize and improve the models and scenarios in the scene generation method according to the results of feedback collection and analysis, in order to improve their performance and applicability. According to the feedback information, firstly, the data-driven model and the physical model are optimized. For example, if it is found that the prediction accuracy of the model in some specific scenarios is insufficient, the parameters, structure or training algorithm of the model can be adjusted to improve its performance in these scenarios. For data-driven models, the model can be retrained, more relevant data can be introduced or more advanced machine learning algorithms can be used; for physical models, model parameters can be recalibrated, and boundary conditions and constraint settings of the model can be improved. At the same time, the strategy and method of scene generation can also be optimized. For example, the fusion strategy of data-driven and model-driven scenes can be adjusted, and the index system of scene screening and evaluation can be optimized to generate more comprehensive scenes that meet actual needs. In addition, according to the problems in actual application, the resolution, time scale, spatial coverage range of the scene can be adjusted and optimized. Through these optimization measures, the quality and effect of the scene generation method can be continuously improved, so that it can better serve the actual needs of power grid control business and provide more powerful support for the safe and stable operation and optimal scheduling of power grids.

[0114] Step S216, data updating and integration.

[0115] In this embodiment, data updating and integration is the basis of the continuous learning and updating stage, ensuring that the scene generation method can keep pace with the times and adapt to new changes and characteristics of power grid operation. As the power grid continues to develop and operation data continues to accumulate, the original data foundation may gradually lose its timeliness and representativeness. Therefore, it is necessary to regularly collect new power grid operation data, including load data, power generation data, equipment state monitoring data, weather data and new energy prediction data, etc. After collecting new data, integrate it with historical data to form a complete and continuously updated data set. This process not only involves the simple superposition of data, but also requires consistency checking and format unification of the data to ensure the compatibility between new and old data. For example, new data may need to be cleaned to remove noise and outliers, fill in missing values, and standardized to make it compatible with historical data in the same scale and format. Through continuous data updating and integration, the latest and most comprehensive data support can be provided for subsequent model optimization and scene generation, so that the scene generation method can better reflect the current operation status and future development trend of the power grid.

[0116] Step S217, model continuous optimization.

[0117] In this embodiment, model continuous optimization is the core of the continuous learning and updating phase, aiming to continuously improve the performance and adaptability of data-driven models and physical models according to new data and feedback information in actual applications. As the power grid operating environment changes and new technologies emerge, the original model may have problems such as decreased prediction accuracy or inability to accurately capture new operating modes. Therefore, the model needs to be evaluated and optimized regularly. In the optimization process, first analyze the characteristics of the new data and the prediction error of the model to determine where the model needs to be improved. For example, if it is found that the data-driven model has insufficient prediction accuracy when dealing with certain types of load fluctuations, the model structure can be adjusted, the number of hidden layer nodes can be increased, or a more complex neural network architecture can be used. For physical models, parameters can be recalibrated based on new device parameters or operating constraints, and boundary condition settings can be improved. At the same time, new algorithms and modeling techniques can be introduced to further improve model performance. For example, use ensemble learning methods to combine the prediction results of multiple models to improve the stability and accuracy of the prediction. Through continuous optimization and adjustment, the model can still provide high-precision and high-reliability prediction and simulation results when facing new operating conditions and data characteristics, providing solid technical support for power grid control business scenarios.

[0118] Step S218, scene dynamic updating and verification.

[0119] In this embodiment, scene dynamic updating and verification is a key step in the continuous learning and updating phase, ensuring that the generated scenarios can timely reflect the latest operating conditions and future development trends of the power grid, and always maintain their accuracy and practicality. As the power grid continues to develop and operating data continues to update, the original scenario may gradually lose its representativeness and guiding significance. Therefore, the generated scenario needs to be updated regularly, and the updated scenario needs to be strictly verified. In the scene dynamic updating process, a series of power grid operating scenarios are generated again using the optimized model and the latest data. These scenarios not only include predictions under normal operating conditions, but also cover various possible fault conditions and operating conditions under extreme weather conditions. By comparing and analyzing with actual operating data, the accuracy and reasonableness of the updated scenario are verified. For example, error indicators such as root mean square error and mean absolute error can be calculated to evaluate the prediction accuracy of the scenario. At the same time, power grid control experts can also subjectively evaluate the generated scenario to ensure that the scenario has good feasibility and guidance in actual application. Through continuous dynamic updating and verification, the generated scenario can always be close to the actual power grid operating conditions, providing timely and accurate decision support for power grid control personnel, helping them better cope with various challenges in power grid operation, and ensuring the safe and stable operation and optimal scheduling of the power grid.

[0120] In the embodiment of the present application, through the above-mentioned steps, a power grid regulation service scene generation method based on data-model double driving is formed, which solves the problems of the existing power grid regulation service scene generation method in extreme scene generation, complex condition adaptability and insufficient data and model fusion, improves the accuracy, reliability and flexibility of scene generation, provides more accurate decision support for power grid regulation, plays an important role in ensuring the safe and stable operation of the power grid, assisting new energy consumption and optimizing the operation efficiency of the power grid.

[0121] Embodiment 3

[0122] The embodiment of the present application provides a power grid operation scene information generation device, and it should be noted that the power grid operation scene information generation device of the embodiment of the present application can be used to execute Figure 1 The power grid operation scene information generation method provided in the embodiment of the present application. The power grid operation scene information generation device provided in the embodiment of the present application is introduced as follows.

[0123] Figure 3 It is a structure schematic diagram of a power grid operation scene information generation device according to the embodiment of the present application, as Figure 3 shown, the power grid operation scene information generation device 300 can include: an acquisition unit 302, a first generation unit 304, a second generation unit 306 and a fusion unit 308.

[0124] The acquisition unit 302 is configured to acquire power grid operation data of a power grid, wherein the power grid operation data is used to at least represent the operation state of the power grid in the operation process.

[0125] The first generation unit 304 is configured to analyze the power grid operation data by using a data-driven model to obtain first operation scene information, wherein the data-driven model is used to determine the operation mode and operation law of the power grid, and the first operation scene information is used to represent the operation state of the power grid under normal operation conditions.

[0126] The second generation unit 306 is configured to simulate the operation state of the power grid by using a physical model to generate second operation scene information, wherein the physical model is used to generate operation scene information that conforms to the operation law and physical constraints of the power grid, and the second operation scene information is used to represent the operation state of the power grid under abnormal operation conditions.

[0127] The fusion unit 308 is configured to fuse the first operation scene information and the second operation scene information to obtain target operation scene information, wherein the target operation scene information is used to represent the operation state of the power grid under different operation conditions.

[0128] The device provided by the embodiment of the present application comprises a first generation unit 304, a second generation unit 306 and a fusion unit.

[0129] Embodiment 4

[0130] According to the embodiment of the present application, a computer readable storage medium is also provided, and the computer readable storage medium stores a program which is executed by a processor to implement the method of the embodiment of the present application.

[0131] Embodiment 5

[0132] According to the embodiment of the present application, a processor is also provided, and the processor is used to run a program, wherein the program is executed to implement the method of the embodiment of the present application.

[0133] Embodiment 6

[0134] Figure 4 is a schematic diagram of an electronic device according to the embodiment of the present application, as shown in Figure 4 The embodiment of the present application also provides an electronic device 400, and the device comprises a processor 401, a memory 402 and a program stored in the memory and capable of running on the processor, and the processor implements the above steps when executing the program.

[0135] The device in the present application can be a server, a PC, a tablet computer (PAD), a mobile phone and the like.

[0136] The present application also provides a computer program product which, when executed on a data processing device, is adapted to execute a program initialized with the above method steps.

[0137] Those skilled in the art will appreciate that embodiments of the application can be readily used as a method, a system or a computer program product. Accordingly, the application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the application can take the form of a computer program product on one or more computer readable storage media (including, but not limited to, disk memory, CD-ROM, optical storage devices, etc.) embodying computer readable program code.

[0138] The application is described in relation to flow diagrams and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It is understood that each flow and / or block in the flow diagrams and / or block diagrams, and combinations of flows and / or blocks in the flow diagrams and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, an embedded processor or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flow diagrams and / or block diagrams block or blocks. Figure 1 Figure 1

[0139] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the flow diagrams and / or block diagrams block or blocks. Figure 1 Figure 1

[0140] These computer program instructions can also be loaded into a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the flow diagrams and / or block diagrams block or blocks. Figure 1 Figure 1

[0141] In one typical configuration, the computing device includes one or more Central Processing Units (CPU), input / output interfaces, network interfaces, and memory.

[0142] ​​​​​​The memory can include non-persistent memory and / or volatile memory, such as a random access memory (RAM) including a static RAM (SRAM) and / or dynamic RAM (DRAM), for example. The memory is an example of computer readable media.

[0143] Computer readable media includes permanent and non-permanent, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile discs (DVDs) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible to a computing device. According to the definition herein, computer readable media does not include transitory media such as modulated data signals and carrier waves.

[0144] It is also to be noted that the terms "comprising", "including", and any other variation thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can also include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the statement "comprising a" does not exclude the existence of additional identical elements in the process, method, article, or apparatus including the element.

[0145] Those skilled in the art will appreciate that embodiments of the present application can be devised for a method, a system, or a computer program product. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) embodying computer-readable program code thereon for use by or in connection with an instruction execution system. For the purposes of this description, a computer-usable or computer readable storage medium can be any apparatus that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device.

[0146] The foregoing is merely illustrative of the embodiments of this application, and is not intended to limit the application. Numerous variations and modifications can be possible to the embodiments without departing from the spirit and scope of the application. Any equivalent modifications or variations, made within the spirit and scope of the application, should be considered within the scope of the application.

Claims

1. A method for generating power grid operation scenario information, characterized in that, include: Acquire power grid operation data, wherein the power grid operation data is used to represent at least the operating status of the power grid during operation; Using a data-driven model, the power grid operation data is analyzed to obtain first operation scenario information. The data-driven model is used to determine the operation mode and operation pattern of the power grid, and the first operation scenario information is used to represent the operation status of the power grid under normal operating conditions. Using a physical model, the operating state of the power grid is simulated to generate second operating scenario information. The physical model is used to generate operating scenario information that conforms to the operating rules and physical constraints of the power grid, and the second operating scenario information is used to represent the operating state of the power grid under abnormal operating conditions. The first operating scenario information and the second operating scenario information are fused to obtain target operating scenario information, wherein the target operating scenario information is used to represent the operating status of the power grid under different operating conditions.

2. The method for generating power grid operation scenario information according to claim 1, characterized in that, The first and second running scenario information are fused to obtain target running scenario information, including: According to the fusion strategy, the first running scenario information and the second running scenario information are fused to obtain a fusion result. The fusion strategy is used to represent the rules for fusion of the first running scenario information and the second running scenario information. The fusion strategy includes serial fusion and / or parallel fusion. Based on the fusion results, the target operating scenario information is determined.

3. The method for generating power grid operation scenario information according to claim 2, characterized in that, Based on the fusion results, the target operating scenario information is determined, including: The fusion results are filtered to obtain a set of operation scenario information, wherein the set of operation scenario information includes different operation scenario information of the power grid; The different running scenario information in the running scenario information set is sorted to obtain the target running scenario information, wherein the target running scenario information is the running scenario information in the running scenario information set that meets the running scenario information threshold.

4. The method for generating power grid operation scenario information according to claim 1, characterized in that, Using a data-driven model, the power grid operation data is analyzed to obtain information about the first operation scenario, including: The power grid operation data is classified to obtain classification results; The classification results are filtered to obtain key data from the power grid operation data, wherein the key data is used to represent the operating status and operating trend of the power grid under the normal operating conditions; The key data is identified using the data-driven model to obtain the identification results; Using the identification results, the operating status of the power grid under the normal operating conditions is predicted; In response to the prediction result meeting the prediction result threshold, the first running scenario information is generated; The method further includes: The first operational scenario information is evaluated to obtain a first evaluation result; Based on the first evaluation result, the data-driven model is updated.

5. The method for generating power grid operation scenario information according to claim 1, characterized in that, Using a physical model, the operating state of the power grid is simulated to generate second operating scenario information, including: In the physical model, the operating state of the power grid is simulated based on the topology and operating constraints to obtain simulation results. The topology represents the connection rules and physical location relationships between multiple components within the power grid, and the operating constraints represent the rules that the power grid must satisfy during operation. Key information is extracted from the simulation results, wherein the key information is used to represent the operating status of the power grid under the abnormal operating conditions; Based on the aforementioned key information, the second operational scenario information is generated; The method further includes: The second operational scenario information is evaluated to obtain a second evaluation result; The physical model is updated based on the second evaluation result.

6. The method for generating power grid operation scenario information according to claim 1, characterized in that, After acquiring the power grid operation data, the method further includes: The power grid operation data is cleaned and / or checked for consistency to obtain the processed power grid operation data. The processed power grid operation data is standardized to obtain standardized power grid operation data, wherein the standardization process is used to convert the power grid operation data to a uniform scale range. The standardized power grid operation data is analyzed to determine the central tendency and dispersion of the processed power grid operation data within a unified scale range.

7. The method for generating power grid operation scenario information according to any one of claims 1 to 5, characterized in that, The method further includes: Collect feedback information, wherein the feedback information is used to indicate the deviation between the target operating scenario information and the predicted operating scenario information, wherein the predicted operating scenario information is used to indicate the expected operating scenario information of the future operating state of the power grid based on the power grid operating data before the application of the target operating scenario information; Determine the parameter adjustment value corresponding to the feedback information, and adjust the parameters in the data-driven model and the parameters in the physical model according to the parameter adjustment value; Based on the adjusted data-driven model and the physical model, the first running scenario information and the second running scenario information are generated.

8. A device for generating power grid operation scenario information, characterized in that, The device includes: An acquisition unit is used to acquire power grid operation data of the power grid, wherein the power grid operation data is used to at least represent the operating status of the power grid during operation; The first generation unit is used to analyze the power grid operation data using a data-driven model to obtain first operation scenario information, wherein the data-driven model is used to determine the operation mode and operation law of the power grid, and the first operation scenario information is used to represent the operation status of the power grid under normal operating conditions. The second generation unit is used to simulate the operating state of the power grid using a physical model and generate second operating scenario information. The physical model is used to generate operating scenario information that conforms to the operating rules and physical constraints of the power grid, and the second operating scenario information is used to represent the operating state of the power grid under abnormal operating conditions. The fusion unit is used to fuse the first operating scenario information and the second operating scenario information to obtain target operating scenario information, wherein the target operating scenario information is used to represent the operating status of the power grid under different operating conditions.

9. A processor, characterized in that, The processor is used to run a program, wherein the program is executed by the processor to perform the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein, when the program is executed, it controls the device on which the computer-readable storage medium is located to perform the method according to any one of claims 1 to 7.

11. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to run the computer program to perform the method according to any one of claims 1 to 7.

12. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method according to any one of claims 1 to 7.

Images