Virtual power plant optimization scheduling method based on multi-time scale collaborative decision

Abstract

The invention provides a virtual power plant optimal scheduling method based on multi-time scale collaborative decision, and belongs to the technical field of power supply, and the method comprises the steps: 1, obtaining a power market mechanism, a distributed energy set, power data, energy storage data, user power consumption data and historical operation data; 2, determining a power time vector, a first time scale vector, a second time scale vector, a virtual time scale vector and a scheduling target vector; 3, constructing a plurality of optimization scheduling models; and 4, determining the hierarchy of each optimization scheduling model, and realizing the coordinated optimization scheduling of the virtual power plant based on the virtual time scale vector. According to the method, an optimal scheduling model and a hierarchical system can be constructed in a layered manner, accurate collaborative decision-making under multiple time scales is realized, market price fluctuation and energy output change are quickly responded, the resource configuration efficiency of the virtual power plant is improved, the operation cost is reduced, and the competitiveness and flexibility in a complex power market are enhanced.

Description

Technical Field

[0001] The present invention relates to the field of power supply technology, and in particular to a virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making. Background Art

[0002] Early power systems were primarily based on centralized power generation and simple dispatch models. With the large-scale integration of distributed energy resources (such as photovoltaic and wind power) and the advancement of power market reforms, virtual power plants (VPPs) have emerged as a key form of integrating decentralized resources into the market. Traditional VPP dispatch often uses fixed timescales (e.g., hourly), making it difficult to adapt to high-frequency trading in the power market (such as the 15-minute cycle in the real-time market) and the short-term fluctuations of distributed energy resources. Furthermore, VPPs employ a single data model, failing to fully integrate market rules, pricing mechanisms, and energy equipment parameters. Furthermore, existing dispatch models often employ a single architecture, failing to balance long-term planning with short-term, real-time regulation needs. This results in inefficient resource dispatch and delayed market response, making it difficult to meet the flexibility and cost-effectiveness requirements of the new power system.

[0003] Therefore, the present invention provides a virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making. Summary of the Invention

[0004] The present invention provides a virtual power plant optimization scheduling method based on multi-time-scale collaborative decision-making. Based on the acquired power market mechanism and the distributed energy resource set of the virtual power plant, a power time vector is determined. A first time scale vector and a second time scale vector are determined based on historical operating data and the power time vector. Furthermore, a virtual time scale vector and a scheduling target vector are determined. Multiple optimization scheduling models are constructed, and the hierarchy of each optimization scheduling model is determined, thereby achieving coordinated optimization scheduling of the virtual power plant based on the virtual time scale vector. This method can integrate multi-dimensional power market data with the characteristics of distributed energy resources, constructing a hierarchical optimization scheduling model and hierarchical system, achieving precise collaborative decision-making at multiple time scales, rapidly responding to market price fluctuations and energy output changes, improving the resource allocation efficiency of the virtual power plant, reducing operating costs, and enhancing its competitiveness and flexibility in complex power markets.

[0005] The present invention provides a virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making, comprising: Step 1: Obtain the electricity market mechanism, obtain the distributed energy collection of the virtual power plant, obtain the power supply data and energy storage data of the virtual power plant, collect user electricity consumption data, and obtain the historical operation data of the virtual power plant; Step 2: Based on the power market mechanism, determine the power time vector, determine the first time scale vector and the second time scale vector based on the historical operation data and the power time vector, and determine the virtual time scale vector and the scheduling target vector; Step 3: Construct multiple optimization scheduling models based on power supply operation data, energy storage operation data, user power consumption data, virtual time scale vector, and scheduling target vector; Step 4: Determine the level of each optimization scheduling model based on the virtual time scale vector, and realize the coordinated optimization scheduling of the virtual power plant based on the virtual time scale vector based on all optimization scheduling models and the levels of all optimization scheduling models.

[0006] According to the virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making provided by the present invention, the power market mechanism includes power transaction rule data, power price mechanism data, market transaction data and power policy and regulation data; A distributed energy resource pool includes multiple distributed energy resources in a virtual power plant.

[0007] According to the virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making provided by the present invention, the power supply data and energy storage data of the virtual power plant are obtained, and the user electricity consumption data is collected, including: Obtaining power sub-data for each distributed energy in the distributed energy set of the virtual power plant, wherein the power sub-data includes at least the name of the distributed energy device, its geographical location, its grid access node, its power generation parameter set, and parameter data for each power generation parameter in the power generation parameter set; Determine the power supply data of the virtual power plant based on the power supply sub-data of all distributed energy resources in the distributed energy resource set of the virtual power plant; Obtaining a set of energy storage devices of the virtual power plant, and obtaining device sub-data of each energy storage device in the set of energy storage devices of the virtual power plant, wherein the device sub-data includes at least the name of the energy storage device, the geographical location, the grid access node, the energy storage parameter set, and the energy storage data of each energy storage parameter in the energy storage parameter set; Collect the electronic data of each user in real time and determine the user's electricity consumption data based on the electronic data of all users; Obtain historical operating data of each distributed energy in the distributed energy set of the virtual power plant within a specified historical time period, wherein the historical operating data includes the response delay time at each time point within the specified historical time period, the historical transaction time scale set, and the transaction volume of each historical time scale in the historical time scale set.

[0008] The virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making provided by the present invention determines the power time vector based on the power market mechanism, including: Performing a first extraction of key time parameters in power transaction rule data in a power market mechanism, and determining a power rule time vector based on all the first extracted key time parameters; performing a second extraction of key time parameters in the power price mechanism data in the power market mechanism, and determining a power price time vector based on all the second extracted key time parameters; performing a third extraction of key time parameters from market transaction data in the power market mechanism, and determining a power market time vector based on all key time parameters extracted by the third extraction; performing a fourth extraction of key time parameters from the power policy and regulation data in the power market mechanism, and determining a power policy time vector based on all key time parameters extracted from the fourth extraction; An electricity time vector is determined based on an electricity rule time vector, an electricity price time vector, an electricity market time vector, and an electricity policy time vector.

[0009] According to the virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making provided by the present invention, a first time scale vector and a second time scale vector are determined based on historical operation data and a power time vector, including: Randomly select N1 key time parameters in the power time vector as initial cluster centers, calculate N1 first cluster sets of the virtual power plant based on the power market mechanism based on the power rule time vector, the power price time vector, the power market time vector, the power policy time vector, the power time vector and all initial cluster centers, and determine the first time scale vector of the virtual power plant based on the first cluster time scales of all first cluster sets; Calculating a time-sensitive value for each distributed energy resource in the distributed energy resource set based on power supply data and historical operating data of each distributed energy resource in the distributed energy resource set; Based on the time-sensitive values ​​of all distributed energy resources in the distributed energy resource set and the number of first cluster sets, cluster analysis is performed on all distributed energy resources in the distributed energy resource set to determine N1 second cluster sets, and based on the time-sensitive values ​​of all distributed energy resources in each second cluster set, a second clustering time scale of each second cluster set is determined; A second time scale vector of the virtual power plant is determined based on the second cluster time scales of all second sets of clusters.

[0010] According to the virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making provided by the present invention, the virtual time scale vector and the scheduling target vector are determined, including: Calculating a first scale selection vector based on the first time scale vector of the virtual power plant and historical operating data within a historical specified time period, and determining the first scale selected vector; Calculating a second scale selection vector based on a second time scale vector of the virtual power plant, and determining the second scale selected vector; Determining a virtual time scale vector based on the first scale selected vector and the second scale selected vector; Based on each first cluster time scale or second cluster time scale in the virtual time scale vector, a scheduling target vector of the virtual time scale vector is determined, wherein the scheduling target vector includes a scheduling target of each first cluster time scale or second cluster time scale.

[0011] The virtual power plant optimization scheduling method based on multi-time-scale collaborative decision-making provided by the present invention constructs multiple optimization scheduling models based on power supply operation data, energy storage operation data, user electricity consumption data, virtual time-scale vectors, and scheduling target vectors, including: Based on each first cluster time scale or second cluster time scale in the virtual time scale vector and the scheduling target of each first cluster time scale or second cluster time scale in the scheduling target vector, an optimization scheduling model for each first cluster time scale or second cluster time scale in the virtual time scale vector is constructed; Based on the power supply operation data, the energy storage operation data and the user electricity consumption data, the optimization scheduling model of each first cluster time scale or second cluster time scale in the virtual time scale vector is trained.

[0012] According to the virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making provided by the present invention, the hierarchy of each optimization scheduling model is determined based on the virtual time scale vector, and based on all optimization scheduling models and the hierarchy of all optimization scheduling models, the coordinated optimization scheduling of the virtual power plant based on the virtual time scale vector is realized, including: determining a level of each optimized scheduling model based on each first cluster time scale or second cluster time scale in the virtual time scale vector; Based on the optimization scheduling models of all first cluster time scales or second cluster time scales in the virtual time scale vector and the hierarchy of all optimization scheduling models, the coordinated optimization scheduling of the virtual power plant based on the virtual time scale vector is realized.

[0013] Compared with the prior art, the present invention has the following advantages: Based on the acquired electricity market mechanism and the distributed energy resources of the virtual power plant, the power time vector is determined. Based on historical operating data and the power time vector, the first and second time scale vectors are determined. Furthermore, the virtual time scale vector and the dispatch target vector are determined. Multiple optimization dispatch models are constructed, and the hierarchy of each optimization dispatch model is determined. This enables coordinated optimization dispatch of the virtual power plant based on the virtual time scale vector. This approach integrates multi-dimensional electricity market data with the characteristics of distributed energy resources, constructing a hierarchical optimization dispatch model and hierarchical system. This enables precise collaborative decision-making across multiple time scales, rapidly responding to market price fluctuations and energy output changes, improving the resource allocation efficiency of the virtual power plant, reducing operating costs, and enhancing its competitiveness and flexibility in complex electricity markets. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0015] Figure 1 It is a flow chart of a virtual power plant optimization scheduling method based on multi-time-scale collaborative decision-making provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0016] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0017] Example 1: The embodiment of the present invention provides a virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making, such as Figure 1 Shown, including: Step 1: Obtain the electricity market mechanism, obtain the distributed energy collection of the virtual power plant, obtain the power supply data and energy storage data of the virtual power plant, collect user electricity consumption data, and obtain the historical operation data of the virtual power plant; Step 2: Based on the power market mechanism, determine the power time vector, determine the first time scale vector and the second time scale vector based on the historical operation data and the power time vector, and determine the virtual time scale vector and the scheduling target vector; Step 3: Construct multiple optimization scheduling models based on power supply operation data, energy storage operation data, user power consumption data, virtual time scale vector, and scheduling target vector; Step 4: Determine the level of each optimization scheduling model based on the virtual time scale vector, and realize the coordinated optimization scheduling of the virtual power plant based on the virtual time scale vector based on all optimization scheduling models and the levels of all optimization scheduling models.

[0018] In this embodiment, the power market mechanism, the virtual power plant's distributed energy resource pool, power supply data, energy storage data, user electricity usage data, and historical virtual power plant operation data are collected. This data forms the basis for optimizing the virtual power plant's scheduling. The power market mechanism determines the economic objectives of scheduling, the distributed energy resource pool and power supply data determine the resource basis for scheduling, energy storage data and user electricity usage data influence scheduling flexibility and demand responsiveness, and historical operation data provides an empirical reference for scheduling.

[0019] In this embodiment, the power time vector reflects multiple time scales of the power market mechanism, such as the day-ahead market (24 hours), the real-time market (5 minutes), etc.

[0020] In this embodiment, a weighted K-means algorithm is used for the first time scale vector (power market mechanism) to cluster the key time parameters of all data in the power time vector into N1 sets. For example, "24 hours (day-ahead market)", "30 days (monthly contract)", and "1 year (policy cycle)" are clustered into the "long-term market scale", while "15 minutes (real-time market)" and "1 hour (load forecast update cycle)" are clustered into the "short-term market scale".

[0021] In this embodiment, the second time scale vector (resource driven) calculates the time sensitivity value for each distributed energy source, and clusters the energy into N1 sets (consistent with the number of market clusters) according to the time sensitivity value, for example: the first sensitive set, the second sensitive set, the third sensitive set, the fourth sensitive set, and the fifth sensitive set.

[0022] In this embodiment, a dedicated goal is defined for each time scale, for example: 24 hours: maximizing the day-ahead market revenue; 10 minutes: minimizing the real-time power balance error.

[0023] In this embodiment, each model corresponds to a time scale, such as a day-ahead scheduling model, an intraday rolling optimization model, and a real-time scheduling model.

[0024] The beneficial effects of the above technical solution include: determining the power time vector based on the acquired power market mechanism and the distributed energy collection of the virtual power plant; determining the first time scale vector and the second time scale vector based on historical operating data and the power time vector; determining the virtual time scale vector and the scheduling target vector; constructing multiple optimization scheduling models and determining the hierarchy of each optimization scheduling model, and achieving coordinated optimization scheduling of the virtual power plant based on the virtual time scale vector. It can integrate multi-dimensional power market data and distributed energy characteristics, and construct a hierarchical optimization scheduling model and hierarchical system in layers, achieving precise collaborative decision-making at multiple time scales, quickly responding to market price fluctuations and energy output changes, improving the resource allocation efficiency of the virtual power plant, reducing operating costs, and enhancing competitiveness and flexibility in complex power markets.

[0025] Example 2: The embodiment of the present invention provides a virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making, wherein the power market mechanism includes power transaction rule data, power price mechanism data, market transaction data, and power policy and regulation data; A distributed energy resource pool includes multiple distributed energy resources in a virtual power plant.

[0026] In this embodiment, the power market mechanism can be obtained through the official website of the power trading center and the power market trading platform, and public trading rules documents, policy documents, market announcements, etc. can be obtained; the API interface of the power trading platform can be connected to obtain the latest electricity price information, trading data, auxiliary service declarations and bidding results and other dynamic data in real time; subscribe to the information push services of the power market information platform and industry associations to obtain unstructured data such as policy changes and market dynamics analysis in a timely manner; communicate with local energy bureaus, development and reform commissions and other departments to obtain internal information such as policy documents and subsidy application guidelines.

[0027] In this embodiment, the power transaction rule data represents the access conditions, transaction subject qualifications, transaction process, contract signing and execution specifications, etc. for power transactions.

[0028] In this embodiment, the electricity price mechanism data represents time-of-use electricity prices (peak, valley and flat electricity price period division, electricity price standards), real-time electricity price fluctuation range, ancillary service prices (frequency regulation, standby and other service fees), and capacity electricity price policies.

[0029] In this embodiment, the market transaction data represents the transaction volume, transaction price, and transaction schedule of the day-ahead market, real-time market, and medium- to long-term market; and the trading rules and market conditions of electricity futures and options.

[0030] In this embodiment, the power policy and regulation data represents power market reform policies, subsidy policies, new energy consumption policies, environmental protection constraint policies, etc. issued by the government.

[0031] In this embodiment, each distributed energy in the distributed energy set can be photovoltaic, wind power, electricity, etc.

[0032] The beneficial effects of the above technical solution are: determining the electricity market mechanism and the distributed energy collection, which can provide a data basis for determining the electricity time vector.

[0033] Example 3: The embodiment of the present invention provides a virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making, which obtains power supply data and energy storage data of the virtual power plant and collects user electricity consumption data, including: Obtaining power sub-data for each distributed energy in the distributed energy set of the virtual power plant, wherein the power sub-data includes at least the name of the distributed energy device, its geographical location, its grid access node, its power generation parameter set, and parameter data for each power generation parameter in the power generation parameter set; Determine the power supply data of the virtual power plant based on the power supply sub-data of all distributed energy resources in the distributed energy resource set of the virtual power plant; Obtaining a set of energy storage devices of the virtual power plant, and obtaining device sub-data of each energy storage device in the set of energy storage devices of the virtual power plant, wherein the device sub-data includes at least the name of the energy storage device, the geographical location, the grid access node, the energy storage parameter set, and the energy storage data of each energy storage parameter in the energy storage parameter set; Collect the electronic data of each user in real time and determine the user's electricity consumption data based on the electronic data of all users; Obtain historical operating data of each distributed energy in the distributed energy set of the virtual power plant within a specified historical time period, wherein the historical operating data includes the response delay time at each time point within the specified historical time period, the historical transaction time scale set, and the transaction volume of each historical time scale in the historical time scale set.

[0034] In this embodiment, for each distributed energy source in the virtual power plant, comprehensive basic information including device name, geographical location, grid access node, as well as a set of power generation parameters (such as photovoltaic conversion efficiency, wind power cut-in and cut-out wind speed, etc.) and corresponding parameter data are collected.

[0035] In this embodiment, the power sub-data of all distributed energy sources are integrated and analyzed, and the power data of the virtual power plant is constructed from a holistic level.

[0036] In this embodiment, for each device in the energy storage device set, information including the device name, geographical location, grid access node, energy storage parameter set, and corresponding energy storage data is collected.

[0037] In this embodiment, the electricity consumption data of each user is collected in real time, and all user data are integrated and processed to form complete user electricity consumption data.

[0038] In this embodiment, the power generation parameter set includes at least installed capacity, rated power, power generation efficiency curve, equipment model, service life, start time, stop time, power generation, power generation, equipment operating status and fault information.

[0039] In this embodiment, the parameter data can be a numerical value, text, time, or image, for example: the parameter data values ​​of installed capacity, rated power, service life, generated power, and power generation; the parameter data of equipment model, power supply operating status, and fault information are text; the parameter data of the power generation efficiency curve is an image (such as the curve of the conversion efficiency of the photovoltaic panel changing with the light intensity); and the parameter data of the start time and stop time are time.

[0040] In this embodiment, the energy storage parameter set includes at least energy storage type, total capacity, rated charge and discharge power, charging efficiency, discharge efficiency, cycle life, charge cut-off voltage, discharge cut-off voltage, charge capacity, discharge capacity and device operating status.

[0041] In this embodiment, the energy storage data can be a numerical value, text, time, or an image, for example: the energy storage type (lithium battery, flow battery, etc.) and the device operating status are text, the total capacity, rated charge and discharge power, charge cut-off voltage, discharge cut-off voltage, cycle life, charge capacity, and discharge capacity are values, and the charging efficiency and discharge efficiency are images.

[0042] In this embodiment, the electronic data at least includes the user's power consumption, power consumption time, voltage, current, power factor, daily power consumption, and monthly power consumption.

[0043] The beneficial effects of the above technical solution are: obtaining the power supply data and energy storage data of the virtual power plant, collecting user electricity consumption data, which can improve the data quality and comprehensiveness.

[0044] Example 4: The embodiment of the present invention provides a virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making, which determines the power time vector based on the power market mechanism, including: Performing a first extraction of key time parameters in power transaction rule data in a power market mechanism, and determining a power rule time vector based on all the first extracted key time parameters; performing a second extraction of key time parameters in the power price mechanism data in the power market mechanism, and determining a power price time vector based on all the second extracted key time parameters; performing a third extraction of key time parameters from market transaction data in the power market mechanism, and determining a power market time vector based on all key time parameters extracted by the third extraction; performing a fourth extraction of key time parameters from the power policy and regulation data in the power market mechanism, and determining a power policy time vector based on all key time parameters extracted from the fourth extraction; An electricity time vector is determined based on an electricity rule time vector, an electricity price time vector, an electricity market time vector, and an electricity policy time vector.

[0045] In this embodiment, the power trading rules cover numerous specifications, including market access, transaction processes, and contract signing. These include a large number of key time parameters, such as the day-ahead market submission deadline and the fulfillment period for medium- and long-term contracts. A "first extraction" of these time parameters can identify time nodes closely related to transaction execution. For example, by clarifying the day-ahead market submission deadline, virtual power plants can plan their next-day power generation plans in advance and submit them in a timely manner, ensuring compliance with trading rules, avoiding missed trading opportunities due to missed submission deadlines, and ensuring effective market participation.

[0046] In this embodiment, key time parameters within the electricity pricing mechanism data include the time period divisions for time-of-use electricity prices and the frequency of real-time electricity price updates. By obtaining these parameters through "second extraction," we can clearly understand the fluctuation patterns of electricity prices over time. For example, if real-time electricity prices are updated every 15 minutes, the virtual power plant can more flexibly adjust its power generation and consumption strategies accordingly, increasing power consumption or energy storage charging during low electricity prices and reducing power consumption or increasing power generation and grid connection during peak electricity prices, thereby reducing costs and increasing efficiency.

[0047] In this embodiment, market transaction data records information such as the time, power consumption, and price of historical transactions. By extracting key time parameters from this data, we can uncover temporal patterns and trends in market transactions. For example, if analysis reveals a surge in market transaction volume and rising prices during a certain period, the virtual power plant can proactively plan to increase power generation during that period in subsequent scheduling to maximize market returns. Furthermore, by studying parameters such as transaction time intervals, the virtual power plant can optimize its trading rhythm and enhance its market competitiveness.

[0048] In this example, electricity policies and regulations, such as the effective and expiration dates of subsidy policies and the implementation timelines of environmental protection policies, significantly impact virtual power plant operations. The "fourth extraction" of these key time parameters helps virtual power plants plan ahead. For example, before new energy subsidy policies take effect, the construction and commissioning of distributed energy equipment can be accelerated to ensure timely access to policy benefits. Energy mix and power generation methods can also be adjusted based on the implementation timelines of environmental protection policies to ensure compliance and avoid operational risks associated with policy changes.

[0049] The beneficial effects of the above technical solution are: based on the power market mechanism, the power time vector is determined, which can enhance the responsiveness of scheduling decisions to market dynamics and policy changes.

[0050] Example 5: An embodiment of the present invention provides a virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making, which determines a first time scale vector and a second time scale vector based on historical operation data and a power time vector, including: Randomly select N1 key time parameters in the power time vector as initial cluster centers, calculate N1 first cluster sets of the virtual power plant based on the power market mechanism based on the power rule time vector, the power price time vector, the power market time vector, the power policy time vector, the power time vector and all initial cluster centers, and determine the first time scale vector of the virtual power plant based on the first cluster time scales of all first cluster sets; Calculating a time-sensitive value for each distributed energy resource in the distributed energy resource set based on power supply data and historical operating data of each distributed energy resource in the distributed energy resource set; Based on the time-sensitive values ​​of all distributed energy resources in the distributed energy resource set and the number of first cluster sets, cluster analysis is performed on all distributed energy resources in the distributed energy resource set to determine N1 second cluster sets, and based on the time-sensitive values ​​of all distributed energy resources in each second cluster set, a second clustering time scale of each second cluster set is determined; A second time scale vector of the virtual power plant is determined based on the second cluster time scales of all second sets of clusters.

[0051] In this embodiment, based on the power rule time vector, the power price time vector, the power market time vector, the power policy time vector, the power time vector, and all initial cluster centers, the N1 first cluster sets of the virtual power plant based on the power market mechanism are calculated, and the first time scale vector of the virtual power plant is determined based on the first cluster time scales of all the first cluster sets. The calculation formula can be expressed as: ; ; ; ; in, represents the first time scale vector of the virtual power plant based on the electricity market mechanism, They represent the first clustering time scales of the 1st first cluster set, the kth first cluster set, and the N1th first cluster set, respectively. represents the number of key time parameters in the power time vector, represents the kth initial cluster center, represents the weighted time parameter of the i-th key time parameter in the power time vector, The first indicator function representing the i-th key time parameter in the power time vector based on the j-th data in the power market mechanism, represents the weight of the jth data in the electricity market mechanism, represents the i-th key time parameter in the power time vector, , represents the power rule time vector, represents the electricity price time vector, represents the electricity market time vector, represents the electricity policy time vector.

[0052] In this embodiment, j=1 represents the power transaction rule data in the power market mechanism, j=2 represents the power price mechanism data in the power market mechanism, j=3 represents the market transaction data in the power market mechanism, and j=4 represents the power policy and regulation data in the power market mechanism.

[0053] In this embodiment, based on the power supply data and historical operation data of each distributed energy in the distributed energy set, the time-sensitive value of each distributed energy in the distributed energy set is calculated. The calculation formula can be expressed as: ; ; ; in, Represents the time-sensitive value of the ath distributed energy in the distributed energy set, represents the time response sensitivity value based on the fitted response rate, Indicates the time rate sensitivity value based on the power regulation rate, It represents the fitting response rate of the ath distributed energy in the distributed energy set within the specified historical time period. represents the time decay factor, Indicates the length of the specified historical time period. Represents the time point of the ath distributed energy in the distributed energy set within the specified historical time period The response delay time under , represents the power regulation rate of the ath distributed energy in the distributed energy set, represents the maximum output power of the ath distributed energy in the distributed energy set, represents the minimum output power of the ath distributed energy in the distributed energy set, It represents the power ramp time of the ath distributed energy in the distributed energy set.

[0054] In this embodiment, the time rate sensitive value Indicates the first preset value for converting the fitted response rate into a time-sensitive value. The time-rate sensitive value A second preset value representing a conversion of the power regulation rate into a time-sensitive value.

[0055] In this embodiment, based on the second clustering time scales of all second cluster sets, the calculation formula for determining the second time scale vector of the virtual power plant can be expressed as: ; in, represents the second time scale vector of the virtual power plant based on the distributed energy collection, They represent the second clustering time scales of the 1st second cluster set, the kth second cluster set, and the N1th second cluster set respectively.

[0056] In this embodiment, the second clustering time scales of all second cluster sets are integrated to determine the second time scale vector of the virtual power plant.

[0057] The beneficial effects of the above technical solution are: by determining the first time scale vector and the second time scale vector based on historical operating data and the power time vector, the multi-source time parameters of the power market and the time-sensitive characteristics of distributed energy can be combined through double-layer clustering to improve the accuracy and flexibility of multi-time scale coordinated scheduling.

[0058] Example 6: The embodiment of the present invention provides a virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making, and determines a virtual time scale vector and a scheduling target vector, including: Calculating a first scale selection vector based on the first time scale vector of the virtual power plant and historical operating data within a historical specified time period, and determining the first scale selected vector; Calculating a second scale selection vector based on a second time scale vector of the virtual power plant, and determining the second scale selected vector; Determining a virtual time scale vector based on the first scale selected vector and the second scale selected vector; Based on each first cluster time scale or second cluster time scale in the virtual time scale vector, a scheduling target vector of the virtual time scale vector is determined, wherein the scheduling target vector includes a scheduling target of each first cluster time scale or second cluster time scale.

[0059] In this embodiment, based on the first time scale vector of the virtual power plant and the historical operating data within the historical specified time period, the first scale selection vector is calculated, and the calculation formula for determining the first scale selected vector can be expressed as: ; in, represents the first scale selection vector based on the first time scale vector, , Indicates the transaction volume completed based on the historical time scale corresponding to the first cluster time scale of the kth first cluster set within the specified historical time period. represents the selected value of the first cluster time scale of the kth first cluster set in the first time scale vector, represents the transaction volume weight, Indicates the maximum transaction volume completed based on the first cluster time scale of all first cluster sets within a specified historical time period.

[0060] In this embodiment, the first scale selection vector includes selection values ​​of the first cluster time scales of all first cluster sets in the first time scale vector; In this embodiment, =0, indicating that the first clustering time scale of the kth first clustering set in the first time scale vector is not selected. =1, it indicates that the first clustering time scale of the kth first clustering set in the first time scale vector is selected.

[0061] In this embodiment, for example, , indicating that the first cluster time scales of the first first cluster set and the third cluster set in the first time scale vector are selected, and the first scale selected vector includes the first cluster time scales of the first first cluster set and the third cluster set.

[0062] In this embodiment, based on the second time scale vector of the virtual power plant, the second scale selection vector is calculated, and the second scale selected vector is determined. The calculation formula can be expressed as: ; in, represents the second scale selection vector based on the second time scale vector, , represents the number of second clustering time scales allowed to be selected, represents the selected value of the second clustering time scale of the kth second cluster set in the second time scale vector, represents the kth second cluster set, Indicates that the time-sensitive value of the ath distributed energy in the distributed energy set belongs to the kth second cluster set, Di represents the sensitive value weight, represents the comprehensive sensitivity value of all distributed energy in the kth second cluster set, Represents the comprehensive sensitivity value of all distributed energy resources in all second cluster sets.

[0063] In this embodiment, the second scale selection vector includes selection values ​​of the second cluster time scales of all second cluster sets in the second time scale vector; In this embodiment, = 0, indicating that the second clustering time scale of the kth second clustering set in the second time scale vector is not selected. =1, it indicates that the second clustering time scale of the kth second clustering set in the second time scale vector is selected.

[0064] For example, , indicating that the first cluster time scales of the first cluster set, the third cluster set, and the fourth cluster set in the first time scale vector are selected, and the first scale selected vector includes the first cluster time scales of the first cluster set, the third cluster set, and the fourth cluster set.

[0065] In this embodiment, vector addition is performed on the first scale selected vector and the second scale selected vector to determine the virtual time scale vector.

[0066] In this embodiment, for each time scale in the virtual time scale vector, a dedicated scheduling target is defined based on its attributes (market or resource). For example, for 24h, the target is "maximizing the day-ahead market revenue", and the associated parameters include time-of-use electricity prices, contract electricity deviation penalties, etc.; for 5min, the target is "minimizing the real-time power balance error", and the associated parameters include renewable energy prediction errors, energy storage SOC dynamic constraints, etc.

[0067] The beneficial effects of the above technical solution are: determining the virtual time scale vector and the scheduling target vector can accurately match the market cycle and resource characteristics, dynamically adjust the scheduling granularity and targets in a complex operating environment, improve the flexibility and optimization efficiency of multi-time scale collaboration, and provide an innovative solution for virtual power plants to maximize their global benefits in a volatile market.

[0068] Example 7: The embodiment of the present invention provides a virtual power plant optimization scheduling method based on multi-timescale collaborative decision-making. Based on power supply operation data, energy storage operation data, user electricity consumption data, virtual timescale vectors, and scheduling target vectors, multiple optimization scheduling models are constructed, including: Based on each first cluster time scale or second cluster time scale in the virtual time scale vector and the scheduling target of each first cluster time scale or second cluster time scale in the scheduling target vector, an optimization scheduling model for each first cluster time scale or second cluster time scale in the virtual time scale vector is constructed; Based on the power supply operation data, the energy storage operation data and the user electricity consumption data, the optimization scheduling model of each first cluster time scale or second cluster time scale in the virtual time scale vector is trained.

[0069] In this embodiment, an adaptive algorithm framework is selected according to the characteristics of the time scale. For example, for long time scales (such as 24 hours), a mixed integer programming (MIP) model is used to handle discrete decision-making problems such as unit start-up and shutdown, and equipment maintenance. For short time scales (such as 5 minutes), a model predictive control (MPC) or stochastic optimization model is used to cope with short-term fluctuations in renewable energy and high-frequency changes in market prices.

[0070] The beneficial effects of the above technical solution are: based on power supply operation data, energy storage operation data, user electricity consumption data, virtual time scale vector and scheduling target vector, multiple optimization scheduling models are constructed, so that each optimization scheduling model can be deeply adapted to the market rules, resource characteristics and scheduling goals under a specific time scale, thereby improving the model's prediction accuracy and decision-making efficiency, and avoiding parameter coupling and optimization deviation caused by multi-scale hybrid modeling.

[0071] Example 8: An embodiment of the present invention provides a virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making. The method determines the level of each optimization scheduling model based on a virtual time scale vector, and implements coordinated optimization scheduling of the virtual power plant based on the virtual time scale vector based on all optimization scheduling models and the levels of all optimization scheduling models. The method includes: determining a level of each optimized scheduling model based on each first cluster time scale or second cluster time scale in the virtual time scale vector; Based on the optimization scheduling models of all first cluster time scales or second cluster time scales in the virtual time scale vector and the hierarchy of all optimization scheduling models, the coordinated optimization scheduling of the virtual power plant based on the virtual time scale vector is realized.

[0072] In this embodiment, the optimization scheduling model is divided into different levels based on the granularity of each first or second cluster time scale in the virtual time scale vector. For example, the long-term level (e.g., T = 24 hours) is responsible for strategic decision-making, such as distributed energy investment planning and medium- and long-term contract power allocation, with the goal of maximizing annual revenue and meeting carbon emission standards. The medium-term level (e.g., T = 1 hour) is responsible for tactical scheduling, such as intraday rolling optimization and energy storage charging and discharging planning, with the goal of balancing market revenue and equipment operating costs. The short-term level (e.g., T = 15 minutes / 5 minutes) is responsible for operational control, such as real-time power balancing and frequency regulation, with the goal of minimizing prediction error and regulation costs. The hierarchical division follows the principle of "the coarser the time granularity, the higher the level; the more far-reaching the decision impact, the higher the level."

[0073] In this embodiment, by adopting a rolling optimization and feedback correction mechanism, the constraint boundaries are transferred step by step from the output results of the high-level optimization scheduling model to the low-level optimization scheduling model, and the output results of the low-level optimization scheduling model are transferred step by step to the high-level optimization scheduling model to feedback correction parameters, thereby realizing coordinated optimization scheduling of the virtual power plant based on the virtual time scale vector.

[0074] In this embodiment, the inter-layer constraint transmission mechanism is as follows: The upper layer guides the lower layer: the output of the long-term layer model serves as the boundary condition of the medium-term layer model. For example, if the long-term layer determines that the photovoltaic output ratio in a certain month must be no less than 40%, the medium-term layer model must ensure that this constraint is met cumulatively on a monthly basis when formulating the daily power generation plan. The lower layer provides feedback to the upper layer: real-time operating data from the short-term layer model (such as the actual charging and discharging efficiency of energy storage and load regulation deviation) is fed back to the medium- and long-term layer models to correct the prediction parameters. For example, if the short-term layer discovers that the charging and discharging efficiency of a certain energy storage device has decreased by 10% due to aging, the medium-term layer model will automatically adjust the available capacity forecast for that device to prevent the upper-layer decision from being out of sync with actual capacity.

[0075] The beneficial effects of the above technical solution are: determining the level of each optimization scheduling model based on the virtual time scale vector, and realizing coordinated optimization scheduling of virtual power plants based on virtual time scale vectors based on all optimization scheduling models and the levels of all optimization scheduling models. This can achieve the organic unity of global goals and local optimization, avoid parameter coupling and optimization deviation caused by multi-scale hybrid modeling, and provide highly robust technical support for the dynamic optimization of virtual power plants in complex scenarios, thereby reducing scheduling errors and operation risks.

[0076] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0077] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0078] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making, characterized by: include: Step 1: Obtain the electricity market mechanism, obtain the distributed energy collection of the virtual power plant, obtain the power supply data and energy storage data of the virtual power plant, collect user electricity consumption data, and obtain the historical operation data of the virtual power plant; Step 2: Based on the power market mechanism, determine the power time vector, determine the first time scale vector and the second time scale vector based on the historical operation data and the power time vector, and determine the virtual time scale vector and the scheduling target vector; Step 3: Construct multiple optimization scheduling models based on power supply operation data, energy storage operation data, user power consumption data, virtual time scale vector, and scheduling target vector; Step 4: Determine the level of each optimization scheduling model based on the virtual time scale vector, and realize the coordinated optimization scheduling of the virtual power plant based on the virtual time scale vector based on all optimization scheduling models and the levels of all optimization scheduling models.

2. According to the virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making according to claim 1, the power market mechanism includes power transaction rule data, power price mechanism data, market transaction data and power policy and regulation data; A distributed energy resource pool includes multiple distributed energy resources in a virtual power plant.

3. The virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making according to claim 2 is characterized in that: Obtain power supply data and energy storage data of the virtual power plant, and collect user electricity consumption data, including: Obtaining power sub-data for each distributed energy in the distributed energy set of the virtual power plant, wherein the power sub-data includes at least the name of the distributed energy device, its geographical location, its grid access node, its power generation parameter set, and parameter data for each power generation parameter in the power generation parameter set; Determine the power supply data of the virtual power plant based on the power supply sub-data of all distributed energy resources in the distributed energy resource set of the virtual power plant; Obtaining a set of energy storage devices of the virtual power plant, and obtaining device sub-data of each energy storage device in the set of energy storage devices of the virtual power plant, wherein the device sub-data includes at least the name of the energy storage device, the geographical location, the grid access node, the energy storage parameter set, and the energy storage data of each energy storage parameter in the energy storage parameter set; Collect the electronic data of each user in real time and determine the user's electricity consumption data based on the electronic data of all users; Obtain historical operating data of each distributed energy in the distributed energy set of the virtual power plant within a specified historical time period, wherein the historical operating data includes the response delay time at each time point within the specified historical time period, the historical transaction time scale set, and the transaction volume of each historical time scale in the historical time scale set.

4. The virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making according to claim 2 is characterized in that: Based on the electricity market mechanism, determine the electricity time vector, including: Performing a first extraction of key time parameters in power transaction rule data in a power market mechanism, and determining a power rule time vector based on all the first extracted key time parameters; performing a second extraction of key time parameters in the power price mechanism data in the power market mechanism, and determining a power price time vector based on all the second extracted key time parameters; performing a third extraction of key time parameters from market transaction data in the power market mechanism, and determining a power market time vector based on all key time parameters extracted by the third extraction; performing a fourth extraction of key time parameters from the power policy and regulation data in the power market mechanism, and determining a power policy time vector based on all key time parameters extracted from the fourth extraction; An electricity time vector is determined based on an electricity rule time vector, an electricity price time vector, an electricity market time vector, and an electricity policy time vector.

5. The virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making according to claim 4 is characterized in that: Determining a first time scale vector and a second time scale vector based on historical operating data and a power time vector includes: Randomly select N1 key time parameters in the power time vector as initial cluster centers, calculate N1 first cluster sets of the virtual power plant based on the power market mechanism based on the power rule time vector, the power price time vector, the power market time vector, the power policy time vector, the power time vector and all initial cluster centers, and determine the first time scale vector of the virtual power plant based on the first cluster time scales of all first cluster sets; Calculating a time-sensitive value for each distributed energy resource in the distributed energy resource set based on power supply data and historical operating data of each distributed energy resource in the distributed energy resource set; Based on the time-sensitive values ​​of all distributed energy resources in the distributed energy resource set and the number of first cluster sets, cluster analysis is performed on all distributed energy resources in the distributed energy resource set to determine N1 second cluster sets, and based on the time-sensitive values ​​of all distributed energy resources in each second cluster set, a second clustering time scale of each second cluster set is determined; A second time scale vector of the virtual power plant is determined based on the second cluster time scales of all second sets of clusters.

6. The virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making according to claim 5 is characterized in that: And determine the virtual time scale vector and the scheduling target vector, including: Calculating a first scale selection vector based on the first time scale vector of the virtual power plant and historical operating data within a historical specified time period, and determining the first scale selected vector; Calculating a second scale selection vector based on a second time scale vector of the virtual power plant, and determining the second scale selected vector; Determining a virtual time scale vector based on the first scale selected vector and the second scale selected vector; Based on each first cluster time scale or second cluster time scale in the virtual time scale vector, a scheduling target vector of the virtual time scale vector is determined, wherein the scheduling target vector includes a scheduling target of each first cluster time scale or second cluster time scale.

7. The virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making according to claim 6 is characterized in that: Based on power supply operation data, energy storage operation data, user power consumption data, virtual time scale vector, and dispatch target vector, multiple optimization dispatch models are constructed, including: Based on each first cluster time scale or second cluster time scale in the virtual time scale vector and the scheduling target of each first cluster time scale or second cluster time scale in the scheduling target vector, an optimization scheduling model for each first cluster time scale or second cluster time scale in the virtual time scale vector is constructed; Based on the power supply operation data, the energy storage operation data and the user electricity consumption data, the optimization scheduling model of each first cluster time scale or second cluster time scale in the virtual time scale vector is trained.

8. The virtual power plant optimization scheduling method based on multi-time scale collaborative decision-making according to claim 6 is characterized in that: The level of each optimization scheduling model is determined based on the virtual time scale vector. Based on all optimization scheduling models and the levels of all optimization scheduling models, the coordinated optimization scheduling of the virtual power plant based on the virtual time scale vector is realized, including: determining a level of each optimized scheduling model based on each first cluster time scale or second cluster time scale in the virtual time scale vector; Based on the optimization scheduling models of all first cluster time scales or second cluster time scales in the virtual time scale vector and the hierarchy of all optimization scheduling models, the coordinated optimization scheduling of the virtual power plant based on the virtual time scale vector is realized.

Images