Source network load storage cooperative scheduling control method

Abstract

The invention belongs to the technical field of power system scheduling and management, and particularly relates to a source-network-load-storage cooperative scheduling control method, which comprises the following steps: step 1, data acquisition and local prediction; step 2, solving a distributed game; step 3, global coordination and verification; and 4, executing and feeding back a scheduling strategy. In the step 1, local data is firstly collected, the local data comprises collecting local load data, then collecting output data, collecting energy storage system data and collecting power grid state data, and after the data collection is completed, a local data set is output. Through distributed decision-making, edge calculation, lightweight global coordination and block chain technologies, the method can adapt to the dispersibility and autonomy of a distributed energy system, significantly reduces the communication burden and calculation complexity, achieves the data sharing between the load demands and energy storage states of all subjects through global coordination and verification, and improves the reliability of the distributed energy system. The overall operation efficiency of the system is improved, and meanwhile the privacy data of all the main bodies are protected.

Description

Technical Field

[0001] The present invention belongs to the technical field of power system dispatching and management, and specifically is a source-grid-load-storage coordinated dispatching control method. Background Art

[0002] The traditional power grid dispatching mode relies on the regulation of stable power sources such as thermal power, which is difficult to adapt to the large-scale grid connection of new energy. It is urgent to achieve dynamic matching of multiple resources through coordinated control of source, grid, load and storage. With the rapid growth of distributed energy resources (DERs), traditional centralized scheduling methods face the problems of high computational complexity, heavy communication burden, and difficult privacy protection. Specifically: Centralized optimization has high computational complexity and is difficult to handle for large-scale distributed systems. Traditional centralized scheduling methods are difficult to adapt to the decentralization and autonomy of distributed energy systems. Each entity is unwilling to share private data on load demand and energy storage status, and fully distributed scheduling will lead to a decrease in the overall performance of the system and be inconvenient to use. Therefore, a source-grid-load-storage coordinated scheduling control method is proposed to address the above problems. Summary of the invention

[0003] The purpose of the present invention is to solve the technical problems raised in the background technology by setting up a source-grid-load-storage coordinated scheduling control method in view of the shortcomings of the prior art.

[0004] In view of the above technical problems, the following technical solutions are adopted: a source-grid-load-storage coordinated dispatching control method, comprising the following steps: Step 1: Data collection and local prediction; Step 2: Distributed game solution; Step 3: Global coordination and verification; Step 4: Scheduling strategy execution and feedback.

[0005] Preferably, in the step 1, it includes collecting local load data, then collecting output data, collecting energy storage system data, collecting grid status data, and after the above data collection is completed, outputting the local data set; Then, the data is preprocessed, specifically, outliers and noise are removed, the data is standardized to a uniform range, key features are extracted, where the key features include load trends and weather correlation, and the preprocessed data set is output; Finally, the LSTM model is used to predict local load, the photovoltaic output model is used to predict renewable energy output, and the charging and discharging strategy of the energy storage system is calculated to output local prediction results and preliminary scheduling decisions.

[0006] Preferably, an LSTM model is used for local load forecasting, and the load forecasting formula of the LSTM model is: ; in: is the load forecast value for the next moment; is the historical load data; For historical weather data.

[0007] Preferably, the renewable energy output is predicted by a photovoltaic output model, and the photovoltaic output model is: ; in: Output power for photovoltaic modules; is the electrical efficiency of photovoltaic cells under standard conditions; is the effective light-receiving area of ​​the photovoltaic module; is the solar irradiance at time t; is the operating temperature of the photovoltaic cell; is the standard test temperature; is the temperature correction factor.

[0008] Preferably, the load data includes residential load and industrial load; The output data includes photovoltaic energy; The energy storage system data includes SOC and charging and discharging power; The grid status data includes voltage and frequency.

[0009] Preferably, in step 2, each subject, specifically power generation, load and energy storage, defines its own objective function, defines game rules, specifically power balance constraints, initializes the power generation plan of each subject, and outputs an initial game model; Distributed iterative solution: Use the alternating direction multiplier method for distributed iteration. Each agent updates its own strategy based on local information, exchanges information on power demand and available capacity, and determines whether the convergence condition is met. If yes, proceed to the next step. If no, continue to iterate and output the Nash equilibrium solution. Generate collaborative scheduling strategy: Each subject generates the final scheduling strategy based on the Nash equilibrium solution, checks whether the strategy meets the global constraints of system frequency and node voltage, and enters the global coordination layer if it does. If not, adjust the game model parameters, solve again, and output the collaborative scheduling strategy.

[0010] Preferably, the convergence condition formula is: ; in: is the value before the strategy change; is the value before the next strategy change; is the convergence threshold. To determine whether the iteration reaches the Nash equilibrium, more preferably, the strategy change is less than the threshold.

[0011] Preferably, in step three, lightweight global coordination is performed, and a consistency algorithm is used to share the global information of system frequency and node voltage, and whether the global information of system frequency and node voltage is consistent is determined. If yes, the next step is entered; if no, re-coordination is performed to output a globally consistent scheduling strategy; Blockchain recording and verification: record the decision and transaction information of each subject in the blockchain, use smart contracts to verify the legitimacy and consistency of the decision, and determine whether the blockchain verification is passed. If yes, enter the application layer. If not, regenerate the decision and output the verified scheduling strategy.

[0012] As a preferred method, a consistency algorithm is used to share global information, and data coordination is performed using the following formula. The consistency coordination formula is: ; in: k is the number of iterations; j represents the neighbor node number of node i; is a local state variable; is the coordinated step length; is the neighbor set of node i.

[0013] Preferably, in step 4, the scheduling strategy is executed: each subject executes the final coordinated scheduling strategy, specifically, power generation scheduling, energy storage charging and discharging, and demand response, monitors the execution effect of power balance and system frequency in real time, and outputs the execution result; Real-time monitoring and feedback: Collect load, power generation and energy storage system status data in real time to determine whether the system status is normal. If yes, end the current cycle and enter the next cycle. If not, trigger the exception handling mechanism, restart the process, and output feedback information.

[0014] Beneficial effects of the present invention: (1) Through distributed decision-making, edge computing, lightweight global coordination and blockchain technology, it can adapt to the decentralization and autonomy of distributed energy systems, significantly reducing the communication burden and computational complexity; (2) Through global coordination and verification, data sharing between load demands and energy storage status of each entity is achieved, which improves the overall operating efficiency of the system while protecting the privacy data of each entity. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] In the attached picture: Figure 1 It is the overall flow chart of the collaborative scheduling control method of the present invention; Figure 2 It is a partial flow chart of the collaborative scheduling control method of the present invention. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0018] Specific examples are given below.

[0019] Example See also Figure 1-Figure 2 The present invention provides a source-grid-load-storage coordinated dispatching control method, comprising the following steps: Step 1: Data collection and local prediction; Step 2: Distributed game solution; Step 3: Global coordination and verification; Step 4: Scheduling strategy execution and feedback.

[0020] Furthermore, in the step 1, local data is first collected, including collecting local load data, including residential load and industrial load; Then collect renewable energy output data, including wind power and photovoltaic energy, and collect energy storage system data, including SOC and charging and discharging power; Collect grid status data, including voltage and frequency. After the above data collection is completed, output the local data set; Then, the data is preprocessed, specifically, outliers and noise are removed, the data is standardized to a uniform range, key features are extracted, where the key features include load trends and weather correlation, and the preprocessed data set is output; In the process of data preprocessing, it is necessary to The outlier elimination formula eliminates abnormal data to ensure the reliability of the data. The outlier removal formula is as follows: ; in: is the original data point.

[0021] is the data mean.

[0022] is the standard deviation of the data.

[0023] By using statistical methods to propose data that deviates from the mean by more than three times the standard deviation, the data can be made closer, thus ensuring the reliability of the data; At the same time, the data also needs to be standardized. In the process of data standardization, The standardization formula is used to standardize the data. The formula is as follows: ; in: For standardized data, through data standardization, the data can be scaled to a distribution with a mean of 0 and a variance of 1, eliminating dimensional differences.

[0024] Finally, the LSTM model is used to predict local load, the photovoltaic output model is used to predict renewable energy output, and the charging and discharging strategy of the energy storage system is calculated to output local prediction results and preliminary scheduling decisions.

[0025] Furthermore, the LSTM model is used for local load forecasting, and the load forecasting formula of the LSTM model is: ; in: is the load forecast value for the next moment; is the historical load data; For historical weather data; The long short-term memory network, or LSTM model, is used to capture the temporal correlation between load and weather.

[0026] Furthermore, in this embodiment, the renewable energy output can also be predicted by a photovoltaic output model, and the photovoltaic output model is: ; in: is the output power of the photovoltaic module; the power is W.

[0027] It is the electrical efficiency of photovoltaic cells under standard conditions; generally, the electrical efficiency has been determined when the product leaves the factory.

[0028] It is the effective light-receiving area of ​​the photovoltaic module, and its unit is square meter.

[0029] is the solar irradiance at time t, with the unit of W per square meter.

[0030] is the operating temperature of the photovoltaic cell, in degrees Celsius.

[0031] This is the standard test temperature, usually 25 degrees Celsius.

[0032] It is the temperature correction coefficient, which reflects the effect of temperature on efficiency.

[0033] Furthermore, the load data includes residential load and industrial load; The output data includes wind power energy and photovoltaic energy; The energy storage system data includes SOC and charging and discharging power; The grid status data includes voltage and frequency.

[0034] Furthermore, in step 2, each subject, specifically power generation, load and energy storage, defines its own objective function. This process minimizes the power generation cost through a quadratic function model. The objective function on the power generation side is: ; in: t is the time variable; is the power generation; , , is the power generation cost coefficient; the minimized power generation cost value is obtained through the quadratic function model.

[0035] Define the power balance constraint game rules, initialize the power generation plan strategy of each subject, and output the initial game model; Distributed iterative solution: Use the alternating direction multiplier method (ADMM) for distributed iteration. The ADMM iteration formula is as follows: ; in: It is a mark of the number of iterations. During the distributed optimization process, the algorithm will perform multiple iterations, and each iteration will update the node strategy until the convergence condition is reached. is the penalty parameter; is the current strategy on the i-th node. During the iteration process, these strategies will be updated according to the formula; is the local loss function of the i-th node, indicating that the node is losses or costs arising from is the value of the Lagrange multiplier at the kth iteration; is the global average power.

[0036] This process can optimize the strategies of each subject in a distributed manner and approach the global optimal solution through penalty terms.

[0037] Each subject updates its own strategy based on local information, exchanges necessary information on power demand and available capacity, and determines whether the convergence condition is met and the strategy change is less than the threshold. If yes, it proceeds to the next step, otherwise it continues to iterate and outputs the Nash equilibrium solution. Generate collaborative scheduling strategy: Each subject generates the final scheduling strategy based on the Nash equilibrium solution, and checks whether the strategy meets the global constraints of system frequency and node voltage. If yes, it enters the global coordination layer. If not, it adjusts the game model parameters, solves the problem again, and outputs the collaborative scheduling strategy.

[0038] Furthermore, the convergence condition formula is: ; in: is the value before the strategy change; is the value before the next strategy change; is the convergence threshold, which determines whether the iteration reaches the Nash equilibrium. More specifically, the strategy change is less than the threshold.

[0039] Furthermore, in the step 3, lightweight global coordination is performed, and the global information of the system frequency and the node voltage is shared using a consistency algorithm to determine whether the global information of the system frequency and the node voltage is consistent. If yes, the next step is entered; if no, the next step is performed again to output a globally consistent scheduling strategy. Blockchain recording and verification: record the decision and transaction information of each subject in the blockchain, use smart contracts to verify the legitimacy and consistency of the decision, and determine whether the blockchain verification is passed. If yes, enter the application layer. If not, regenerate the decision and output the verified scheduling strategy.

[0040] Furthermore, the consistency algorithm is used to share the global information of system frequency and node voltage, and the data is coordinated through the following formula. The consistency coordination formula is: ; in: k is the number of iterations; j represents the neighbor node number of node i; is the local state variable (voltage); is the coordinated step length; is the neighbor set of node i. The distributed average algorithm is used to achieve the consistency of the whole network state; The blockchain verification formula is as follows: ; in: It is a data concatenation symbol. The hash chain ensures that the scheduling decision cannot be tampered with.

[0041] Furthermore, in step 4, the scheduling strategy is executed: each subject executes the final coordinated scheduling strategy, specifically, power generation scheduling, energy storage charging and discharging, and demand response, monitors the execution effect power balance or system frequency in real time, and outputs the execution result. In this step, the energy storage charging and discharging control formula is: ; in: is the actual photovoltaic output at time t, in KW; is the PV output reference value at time t, which is based on the needs of prediction or scheduling; is the charging or discharging power of the energy storage at time t, in KW, with charging being negative and discharging being positive; , are weight coefficients, representing the priorities of fluctuation suppression and equipment loss respectively; is the energy storage loss cost function, which is related to the charge and discharge power depth and the number of cycles; By minimizing the PV output deviation , can reduce the impact on the power grid, which is conducive to prolonging the normal operation of the power grid; Through loss cost Limiting overcharging or over-discharging can effectively extend the service life of the energy storage body.

[0042] Real-time monitoring and feedback: Collect system status data, load, power generation and energy storage status in real time to determine whether the system status is normal. If yes, end the current cycle and enter the next cycle. If not, trigger the exception handling mechanism, restart the process and output feedback information.

[0043] In summary, this application can adapt to the decentralization and autonomy of distributed energy systems through distributed decision-making, edge computing, lightweight global coordination and blockchain technology, and significantly reduce the communication burden and computational complexity; And through global coordination and verification, data sharing between load demands and energy storage status of each entity is achieved, which improves the overall operating efficiency of the system while protecting the privacy data of each entity.

[0044] In the description of the present invention, it is necessary to understand that the directions or positional relationships indicated by the terms "front and back", "left and right", etc. are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or component referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as a limitation on the present invention.

[0045] Of course, in this technical solution, those skilled in the art should understand that the term "one" should be understood as "at least one" or "one or more", that is, in one embodiment, the number of an element can be one, while in another embodiment, the number of the element can be multiple. The term "one" should not be understood as a limitation on quantity.

[0046] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art under the technical guidance of the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A source-grid-load-storage coordinated dispatching control method, characterized in that: The following steps are involved: Step 1: Data collection and local prediction; Step 2: Distributed game solution; Step 3: Global coordination and verification; Step 4: Scheduling strategy execution and feedback.

2. A source-grid-load-storage coordinated dispatching control method according to claim 1, characterized in that: In the step 1, it includes collecting local load data, then collecting output data, collecting energy storage system data, collecting grid status data, and after the above data collection is completed, outputting the local data set; Then, the data is preprocessed, specifically, outliers and noise are removed, the data is standardized to a uniform range, key features are extracted, where the key features include load trends and weather correlation, and the preprocessed data set is output; Finally, the LSTM model is used to predict local load, the photovoltaic output model is used to predict renewable energy output, and the charging and discharging strategy of the energy storage system is calculated to output local prediction results and preliminary scheduling decisions.

3. A source-grid-load-storage coordinated dispatching control method according to claim 2, characterized in that: The LSTM model is used for local load forecasting. The load forecasting formula of the LSTM model is: ； in: is the load forecast value for the next moment; is the historical load data; For historical weather data.

4. A source-grid-load-storage coordinated dispatching control method according to claim 2, characterized in that: Renewable energy output is predicted through a photovoltaic output model, and the photovoltaic output model is: ； in: Output power for photovoltaic modules; is the electrical efficiency of photovoltaic cells under standard conditions; is the effective light-receiving area of ​​the photovoltaic module; is the solar irradiance at time t; is the operating temperature of the photovoltaic cell; is the standard test temperature; is the temperature correction factor.

5. The source-grid-load-storage coordinated dispatching control method according to claim 2 is characterized in that: The load data includes residential load and industrial load; The output data includes photovoltaic energy; The energy storage system data includes SOC and charging and discharging power; The grid status data includes voltage and frequency.

6. A source-grid-load-storage coordinated dispatching control method according to claim 1, characterized in that: In the step 2, the objective functions of the power generation, load and energy storage entities are defined, the game rules are defined, specifically, the power balance constraints, the power generation plans of the entities are initialized, and the initial game model is output; Distributed iterative solution: Use the alternating direction multiplier method for distributed iteration. Each agent updates its own strategy based on local information, exchanges information on power demand and available capacity, and determines whether the convergence condition is met. If yes, proceed to the next step. If no, continue to iterate and output the Nash equilibrium solution. Generate collaborative scheduling strategy: Each subject generates the final scheduling strategy based on the Nash equilibrium solution, checks whether the strategy meets the global constraints of system frequency and node voltage, and enters the global coordination layer if it does. If not, adjust the game model parameters, solve again, and output the collaborative scheduling strategy.

7. A source-grid-load-storage coordinated dispatching control method according to claim 6, characterized in that: The convergence condition formula is: ； in: is the value before the strategy change; is the value before the next strategy change; is the convergence threshold.

8. The source-grid-load-storage coordinated dispatching control method according to claim 1 is characterized by: In the step 3, lightweight global coordination is performed, and a consistency algorithm is used to share the global information of system frequency and node voltage, and to determine whether the global information of system frequency and node voltage is consistent. If yes, the next step is entered, and if no, the next step is performed again to output a globally consistent scheduling strategy. Blockchain recording and verification: record the decision and transaction information of each subject in the blockchain, use smart contracts to verify the legitimacy and consistency of the decision, and determine whether the blockchain verification is passed. If yes, enter the application layer. If not, regenerate the decision and output the verified scheduling strategy.

9. A source-grid-load-storage coordinated dispatching control method according to claim 8, characterized in that: Use the consistency algorithm to share global information and coordinate data using the following formula. The consistency coordination formula is: ; in: k is the number of iterations; j represents the neighbor node number of node i; is a local state variable; is the coordinated step length; is the neighbor set of node i.

10. The source-grid-load-storage coordinated dispatching control method according to claim 1, characterized in that: In step 4, the scheduling strategy is executed: each subject executes the final coordinated scheduling strategy, specifically, power generation scheduling, energy storage charging and discharging, and demand response, monitors the execution effect of power balance and system frequency in real time, and outputs the execution result; Real-time monitoring and feedback: Collect load, power generation and energy storage system status data in real time to determine whether the system status is normal. If yes, end the current cycle and enter the next cycle. If not, trigger the exception handling mechanism, restart the process, and output feedback information.

Images