A method, device and computer equipment for training a power grid section control model

Abstract

The present application relates to a training method, device and computer equipment for a power grid section control model. The method includes: obtaining section control model training sample data corresponding to the target power grid, the power grid section control model to be trained and the power grid section optimization network; inputting the section control model training sample data into the power grid section control model to be trained to obtain the initial power grid section control information corresponding to the target power grid; inputting the initial power grid section control information into the power grid section optimization network to obtain the optimized power grid section control information; using the optimized power grid section control information to train the power grid section control model to be trained, until the disturbance of the initial power grid section control information is less than the disturbance threshold, and the trained power grid section control model is obtained. The use of this method can achieve more intelligent and flexible power grid management, thereby improving the response speed and adaptability of the power system, and further improving the stability of the power system.

Description

Technical Field

[0001] The present application relates to the field of smart grid technology, and in particular to a method, device, computer equipment, storage medium and computer program product for training a power grid section control model. Background Art

[0002] With the development of computer technology, smart grid technology has emerged. The construction of a grid section control model is an advanced technology used in power systems. It can predict and optimize the control strategy of the grid section based on the real-time data and scenario information of the power system to improve the stability, efficiency and reliability of the power system. The grid control technology in traditional technology mainly relies on empirical rules and static models. It is difficult to adapt to complex changes in the power system due to the limitations of model accuracy and real-time performance. This leads to poor stability of the power system. Summary of the invention

[0003] Based on this, it is necessary to provide a grid section control model training method, device, computer equipment, computer readable storage medium and computer program product that can improve the stability of the power system in response to the above technical problems.

[0004] In a first aspect, the present application provides a method for training a power grid section control model, comprising:

[0005] Obtaining training sample data of the section control model corresponding to the target power grid, the section control model of the power grid to be trained, and the power grid section optimization network;

[0006] Inputting the section control model training sample data into the power grid section control model to be trained to obtain initial power grid section control information corresponding to the target power grid;

[0007] Inputting the initial power grid section control information into the power grid section optimization network to obtain optimized power grid section control information;

[0008] The grid section regulation model to be trained is trained using the optimized grid section regulation information until the disturbance of the initial grid section regulation information is less than a disturbance threshold, thereby obtaining a trained grid section regulation model.

[0009] In a second aspect, the present application also provides a power grid section control model training device, comprising:

[0010] A power grid data acquisition module is used to obtain the section control model training sample data corresponding to the target power grid, the power grid section control model to be trained, and the power grid section optimization network;

[0011] A control information calculation module, used for inputting the section control model training sample data into the power grid section control model to be trained to obtain the initial power grid section control information corresponding to the target power grid;

[0012] A control information optimization module, used for inputting the initial power grid section control information into the power grid section optimization network to obtain optimized power grid section control information;

[0013] The control model training module is used to use the optimized power grid section control information to train the power grid section control model to be trained until the disturbance of the initial power grid section control information is less than a disturbance threshold, thereby obtaining a trained power grid section control model.

[0014] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0015] Obtaining training sample data of the section control model corresponding to the target power grid, the section control model of the power grid to be trained, and the power grid section optimization network;

[0016] Inputting the section control model training sample data into the power grid section control model to be trained to obtain initial power grid section control information corresponding to the target power grid;

[0017] Inputting the initial power grid section control information into the power grid section optimization network to obtain optimized power grid section control information;

[0018] The grid section regulation model to be trained is trained using the optimized grid section regulation information until the disturbance of the initial grid section regulation information is less than a disturbance threshold, thereby obtaining a trained grid section regulation model.

[0019] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0020] Obtaining training sample data of the section control model corresponding to the target power grid, the section control model of the power grid to be trained, and the power grid section optimization network;

[0021] Inputting the section control model training sample data into the power grid section control model to be trained to obtain initial power grid section control information corresponding to the target power grid;

[0022] Inputting the initial power grid section control information into the power grid section optimization network to obtain optimized power grid section control information;

[0023] The grid section regulation model to be trained is trained using the optimized grid section regulation information until the disturbance of the initial grid section regulation information is less than a disturbance threshold, thereby obtaining a trained grid section regulation model.

[0024] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:

[0025] Obtaining training sample data of the section control model corresponding to the target power grid, the section control model of the power grid to be trained, and the power grid section optimization network;

[0026] Inputting the section control model training sample data into the power grid section control model to be trained to obtain initial power grid section control information corresponding to the target power grid;

[0027] Inputting the initial power grid section control information into the power grid section optimization network to obtain optimized power grid section control information;

[0028] The grid section regulation model to be trained is trained using the optimized grid section regulation information until the disturbance of the initial grid section regulation information is less than a disturbance threshold, thereby obtaining a trained grid section regulation model.

[0029] The above-mentioned power grid section control model training method, device, computer equipment, storage medium and computer program product obtain the section control model training sample data corresponding to the target power grid, the power grid section control model to be trained and the power grid section optimization network; input the section control model training sample data into the power grid section control model to be trained to obtain the initial power grid section control information corresponding to the target power grid; input the initial power grid section control information into the power grid section optimization network to obtain the optimized power grid section control information; use the optimized power grid section control information to train the power grid section control model to be trained until the disturbance of the initial power grid section control information is less than the disturbance threshold, thereby obtaining the trained power grid section control model.

[0030] By using the obtained section control model training sample data corresponding to the target power grid, the grid section control model to be trained is trained, and the initial grid section control information is obtained, and the initial grid section control information is input into the grid section optimization network to obtain the optimized grid section control information. Then, the grid section control model to be trained is trained using the optimized grid section control information until the disturbance of the grid section control information is less than the set disturbance threshold, thereby obtaining the trained grid section control model. By training and optimizing the grid section control model, the system can achieve more intelligent and flexible grid management, thereby improving the response speed and adaptability of the power system, helping to cope with various external and internal changes, and further improving the stability, reliability and economy of the power system. Secondly, intelligent section control can optimize the structure and power distribution of the power grid, reduce energy loss and transmission loss, reduce the operating cost of the system, and improve energy utilization efficiency. In addition, through continuous optimization and learning, the grid section control model can continuously adapt to the changing operating environment and market demand, maintain the best control strategy, and make important contributions to the sustainable development and optimized operation of the power industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0032] Figure 1 An application environment diagram of a power grid section control model training method in an embodiment;

[0033] Figure 2 A schematic diagram of a flow chart of a power grid section control model training method in one embodiment;

[0034] Figure 3 It is a flowchart of a method for obtaining initial power grid section control information in one embodiment;

[0035] Figure 4 A schematic diagram of a flow chart of a method for obtaining power grid section regulation information through training in one embodiment;

[0036] Figure 5 A schematic flow chart of a method for obtaining optimized power grid section control information in one embodiment;

[0037] Figure 6 A schematic flow chart of a method for obtaining optimized power grid section control information in another embodiment;

[0038] Figure 7 A schematic diagram of a flow chart of a method for obtaining a trained power grid section control model in one embodiment;

[0039] Figure 8 It is a structural block diagram of a power grid section control model training device in one embodiment;

[0040] Fig. 9 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0042] The present invention provides a method for training a power grid section control model, which can be applied to Figure 1 In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. The server 104 obtains the section control model training sample data corresponding to the target power grid, the power grid section control model to be trained, and the power grid section optimization network through the terminal 102; the section control model training sample data is input into the power grid section control model to be trained to obtain the initial power grid section control information corresponding to the target power grid; the initial power grid section control information is input into the power grid section optimization network to obtain the optimized power grid section control information; the optimized power grid section control information is used to train the power grid section control model to be trained until the disturbance of the initial power grid section control information is less than the disturbance threshold, and the trained power grid section control model is obtained. Among them, the terminal 102 can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, Internet of Things devices and portable wearable devices, and the Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart car-mounted devices, etc. The portable wearable device may be a smart watch, a smart bracelet, a head-mounted device, etc. The server 104 may be implemented as an independent server or a server cluster consisting of multiple servers.

[0043] In an exemplary embodiment, Figure 2 As shown in the figure, a training method for a power grid section control model is provided, and the method is applied to Figure 1 The server in the example is used to illustrate, including the following steps 202 to 208. Among them:

[0044] Step 202, obtaining the section control model training sample data corresponding to the target power grid, the power grid section control model to be trained, and the power grid section optimization network.

[0045] The target power grid may be a power grid that requires power grid section regulation.

[0046] The section control model training sample data may be training sample data used to train the power grid section control model.

[0047] The power grid section control model to be trained may be a power grid section control model that has not been trained yet.

[0048] Among them, the power grid section optimization network can be a model used to optimize the output data of the power grid section control model.

[0049] Specifically, the historical operation data of the target power grid is obtained, including load data, generator data, line data, etc. According to the collected data, a model of the power system is constructed, including node model, line model, etc., and a suitable section control scheme is designed, including the power adjustment range of each node, the transmission capacity limit of the line, etc. The historical data and the designed control scheme are used to generate the section control model training sample data, including the input state and the expected output control action under different operating conditions. A suitable neural network structure is designed as the training model of the power grid section control model. A deep neural network structure suitable for the power system control task can be selected, such as multi-layer perceptron (MLP), convolutional neural network (CNN) or recurrent neural network (RNN), etc. The model parameters are initialized and the training process is prepared. The structure of the design optimization network is usually an optimization algorithm for optimizing section control, such as genetic algorithm, particle swarm optimization algorithm, reinforcement learning algorithm, etc. The parameters of the optimization network are initialized, and the section optimization process is prepared to obtain the power grid training section optimization network.

[0050] Step 204: input the section control model training sample data into the power grid section control model to be trained to obtain initial power grid section control information corresponding to the target power grid.

[0051] The initial power grid section control information may be control information that has been preliminarily calculated by a power grid section control model, but has not been verified and modified for efficiency and safety.

[0052] Specifically, the section control model training sample data is input into the section control model of the power grid to be trained in batches, ensuring that the order of data input is consistent with the features corresponding to the input layer of the model, and that the data format is consistent with the model input requirements. Forward propagation is performed on each section control model training sample data, and the input data passes through each layer of the neural network to finally obtain the initial power grid section control information corresponding to the target power grid. During the forward propagation process, the input data is multiplied by the weights and biases of the model, and the output of each layer is calculated through the activation function. The predicted output of the model is compared with the actual label of the sample data, and the value of the loss function is calculated. Commonly used loss functions include mean square error (MSE), cross entropy, etc., which are used to measure the difference between the model prediction results and the true labels.

[0053] Step 206, inputting the initial power grid section regulation information into the power grid section optimization network to obtain the optimized power grid section regulation information.

[0054] The optimized power grid section control information may be the optimized initial power grid section control information.

[0055] Specifically, the initial grid section control information is input into the grid section optimization network, and the grid section optimization network performs optimization calculations based on the input initial grid section control information to improve the efficiency and performance of grid operation. The optimized grid section control information is calculated through the optimization algorithm and the objective function. The optimized grid section control information may include adjusted node power distribution, line load distribution, generator output adjustment, etc., to achieve balance and stability of the grid.

[0056] Step 208, using the optimized power grid section regulation information, the power grid section regulation model to be trained is trained until the disturbance of the initial power grid section regulation information is less than the disturbance threshold, thereby obtaining a trained power grid section regulation model.

[0057] The trained power grid section control model may be a trained power grid section control model.

[0058] Specifically, the power grid section control information is optimized, or the training sample data of the section control model is combined to train the power grid section control model to be trained, and the weight and bias of the model are updated through the back propagation algorithm to reduce the prediction error of the model. After each round of calculation, the size between the disturbance of the initial power grid section control information corresponding to the model and the disturbance threshold set by the target power grid is calculated. If the disturbance of the initial power grid section control information is less than the disturbance threshold, the training is stopped, and the model has converged to a better state, and the trained power grid section control model is obtained; otherwise, the next round of training is continued until the disturbance of the initial power grid section control information is less than the disturbance threshold, and the trained power grid section control model is obtained. The model in the training process is evaluated at the same time to check the performance and generalization ability of the model on the verification set. The prediction accuracy and accuracy of the model can be evaluated by comparing with the actual operation data and using the verification set.

[0059] In the above-mentioned power grid section control model training method, the section control model training sample data corresponding to the target power grid, the power grid section control model to be trained and the power grid section optimization network are obtained; the section control model training sample data is input into the power grid section control model to be trained to obtain the initial power grid section control information corresponding to the target power grid; the initial power grid section control information is input into the power grid section optimization network to obtain the optimized power grid section control information; the optimized power grid section control information is used to train the power grid section control model to be trained until the disturbance of the initial power grid section control information is less than the disturbance threshold, so as to obtain the trained power grid section control model.

[0060] By using the obtained section control model training sample data corresponding to the target power grid, the grid section control model to be trained is trained, and the initial grid section control information is obtained, and the initial grid section control information is input into the grid section optimization network to obtain the optimized grid section control information. Then, the grid section control model to be trained is trained using the optimized grid section control information until the disturbance of the grid section control information is less than the set disturbance threshold, thereby obtaining the trained grid section control model. By training and optimizing the grid section control model, the system can achieve more intelligent and flexible grid management, thereby improving the response speed and adaptability of the power system, helping to cope with various external and internal changes, and further improving the stability, reliability and economy of the power system. Secondly, intelligent section control can optimize the structure and power distribution of the power grid, reduce energy loss and transmission loss, reduce the operating cost of the system, and improve energy utilization efficiency. In addition, through continuous optimization and learning, the grid section control model can continuously adapt to the changing operating environment and market demand, maintain the best control strategy, and make important contributions to the sustainable development and optimized operation of the power industry.

[0061] In an exemplary embodiment, Figure 3 As shown, the section control model training sample data is input into the power grid section control model to be trained to obtain the initial power grid section control information corresponding to the target power grid, including steps 302 to 306. Among them:

[0062] Step 302: Input the section control model training sample data into the power grid section control model to be trained to obtain the training power grid section control information.

[0063] The training grid section control information may be a calculation result of a grid section control model, but the calculation result has not been verified and optimized by simulation.

[0064] Specifically, the section control model training sample data is input into the section control model of the power grid to be trained in batches, ensuring that the order of data input is consistent with the features corresponding to the input layer of the model, and that the data format is consistent with the model input requirements. Forward propagation is performed on each section control model training sample data, and the input data passes through each layer of the neural network to finally obtain the training power grid section control information corresponding to the target power grid. During the forward propagation process, the input data is multiplied by the weight and bias of the model, and the output of each layer is calculated through the activation function. The predicted output of the model is compared with the actual label of the sample data, and the value of the loss function is calculated. Commonly used loss functions include mean square error (MSE), cross entropy, etc., which are used to measure the difference between the model prediction result and the true label.

[0065] Step 304, importing the training power grid section control information into the power grid system simulation network to obtain power grid operation simulation data.

[0066] Among them, the power grid system simulation network can be a model used to test whether the training power grid section regulation information meets the standards of the target power grid.

[0067] The power grid operation simulation data may be the result obtained after testing the training power grid section control information.

[0068] Specifically, the training power grid section regulation information is input into the power grid system simulation network for simulation operation. The power grid system simulation network simulates the operation of the power grid under different loads, power generation, line status, etc. according to the input regulation information. The randomness and uncertainty in the simulation environment of the power grid system simulation network are considered to simulate the actual power grid operation as realistically as possible. During the power grid system simulation operation, various parameters, node status, line load and other information of the simulation network are recorded, and finally the power grid operation simulation data is output.

[0069] Step 306, training the grid section regulation model to be trained according to the grid operation simulation data until the grid operation simulation data meets the power operation standard, and using the trained grid section regulation information as the initial grid section regulation information.

[0070] The power operation standard may be an operation standard specified by the target power grid.

[0071] Specifically, the power grid operation simulation data is input into the power grid section control model to be trained, and the parameters of the model are continuously adjusted through the back propagation algorithm, so that the model gradually approaches the mode and standard of actual power grid operation. After each round of calculation, the performance of the model is evaluated using methods such as validation set or cross-validation, and the accuracy, loss value and other indicators of the model between the power grid operation simulation data and the power operation standard are checked, and the hyperparameters and structure of the model are adjusted to improve the generalization ability and prediction accuracy of the model. If the power grid operation simulation data meets the power operation standard, the training is stopped, and the training power grid section control information just obtained is used as the initial power grid section control information. Otherwise, the training is repeated until the power grid operation simulation data meets the power operation standard, and the training power grid section control information is used as the initial power grid section control information. Among them, the training process can set the number of training rounds, learning rate attenuation strategy, etc. as needed to control the convergence speed and stability of the training process.

[0072] In this embodiment, by inputting the training sample data of the section control model into the power grid section control model to be trained, the training power grid section control information is obtained, and then the information is imported into the power grid system simulation network to obtain the power grid operation simulation data. The power grid section control model to be trained is trained using these simulation data until the simulation data meets the power operation standard. The main advantage of this process is that through the training of simulation data, the power grid control strategy can be effectively optimized and the operation efficiency and stability of the power grid can be improved. At the same time, by continuously adjusting the model parameters, the model can better adapt to different power operation conditions, thereby providing more reliable support and guidance for the actual power grid operation.

[0073] In an exemplary embodiment, Figure 4 As shown, the grid section control model to be trained includes a grid control strategy network and a grid control value network; the section control model training sample data is input into the grid section control model to be trained to obtain the training grid section control information, including steps 402 to 410. Among them:

[0074] Step 402: Input the section control model training sample data into the power grid control strategy network to obtain power grid control decision information.

[0075] The power grid control strategy network may be a model for generating a control strategy based on data samples.

[0076] The power grid control decision information may be strategy information for controlling the power grid section control model to be trained.

[0077] Specifically, the section control model training sample data is input into the power grid control strategy network. It is necessary to ensure that the dimension and format of the input data match the input layer of the network. Through the forward propagation algorithm, the section control model training sample data is passed from the input layer to the output layer to calculate the output result of the network. After each node receives the input, it is processed by the activation function and the result is passed to the next layer of nodes until the output layer. According to the output result of the network, the control decision information of the power grid is calculated.

[0078] Step 404: input the section control model training sample data into the power grid control value network to obtain the comprehensive reward value of the power grid.

[0079] Among them, the power grid regulation value network can be a model for calculating the operating cost after regulation based on data samples.

[0080] The comprehensive reward value of the power grid may be the operating cost information after the power grid section control model to be trained is regulated.

[0081] Specifically, the training sample data of the section control model is input into the power grid control value network. It is necessary to ensure that the dimension and format of the input data match the input layer of the network. Through the forward propagation algorithm, the training sample data of the section control model is passed from the input layer to the output layer to calculate the output result of the network. After each node receives the input, it is processed by the activation function and the result is passed to the next layer of nodes until the output layer. According to the output result of the network, the comprehensive reward value of the power grid is calculated.

[0082] Step 406, using the power grid stability simulation model, adjusts the power grid control decision information and the power grid comprehensive reward value to obtain adjusted control decision information and adjusted comprehensive reward value.

[0083] The power grid stability simulation model may be used to simulate the stability of input parameters applied to a target point network.

[0084] The adjusted control decision information and the adjusted comprehensive reward value may be adjustment values ​​obtained from the power grid control decision information and the power grid comprehensive reward value respectively.

[0085] Specifically, a power grid stability simulation model is retrieved. The power grid stability simulation model can be a model based on physical simulation or a model based on machine learning, which is used to evaluate the stability of the power grid. The power grid stability simulation model is trained or parameterized to ensure that it can accurately simulate the operation of the power grid. The power grid stability simulation model is used to simulate the power grid regulation decision information and the comprehensive reward value of the power grid, and the stability performance of the target power grid under different decision-making schemes is analyzed. According to the actual operation of the power grid and the constraints, the feasibility and rationality of each decision-making scheme and its impact on the stability of the power grid are evaluated. According to the simulation results, the adjustment regulation decision information and the adjustment comprehensive reward value are generated.

[0086] Step 408, adjusting the model parameters of the power grid control strategy network and the power grid control value network according to the adjustment control decision information and the adjustment comprehensive reward value, to obtain a parameter-adjusted power grid section control model.

[0087] The parameter-adjusted power grid section control model may be a power grid section control model to be trained that has undergone parameter adjustment but fails to meet usage requirements.

[0088] Specifically, according to the adjustment of the control decision information and the adjustment of the comprehensive reward value, the model parameters of the power grid control strategy network and the power grid control value network are adjusted respectively. This can be achieved through optimization algorithms, such as gradient descent algorithms, genetic algorithms, etc., to obtain a parameter-adjusted power grid section control model; at the same time, according to the adjustment of the comprehensive reward value, the gradient of the loss function or the objective function is calculated, and the parameters of the network are updated according to the direction of the gradient descent.

[0089] Step 410: Input the section control model training sample data into the parameter-adjusted power grid section control model to obtain training power grid section control information.

[0090] Specifically, the section control model training sample data is input into the parameter adjustment power grid section control model, and the input data is passed from the input layer to the output layer through the forward propagation algorithm to calculate the output result of the network. After each node receives the input, it is processed by the activation function and the result is passed to the next layer of nodes until the output layer. According to the output result of the network, the section control information of the training power grid is calculated.

[0091] In this embodiment, by inputting the training sample data of the section control model into the power grid control strategy network and the power grid control value network, the power grid control decision information and the comprehensive reward value of the power grid are obtained, and the decision information and reward value are adjusted by using the power grid stability simulation model to obtain the adjusted decision information and reward value. According to the adjusted information, the model parameters of the power grid control strategy network and the power grid control value network are adjusted to obtain the power grid section control model after parameter adjustment, and the training sample data is input into the power grid section control model after parameter adjustment to obtain the training power grid section control information. The main advantage of this process is that by dynamically adjusting the decision information and reward value, and optimizing the model parameters, the accuracy and adaptability of the power grid control strategy are improved, thereby enhancing the stability and efficiency of the power grid. At the same time, by continuously optimizing the model parameters, the model can better adapt to the changes and needs of the power grid, providing strong support and guarantee for the long-term stable operation of the power grid.

[0092] In an exemplary embodiment, Figure 5 As shown, the initial power grid section control information is input into the power grid section optimization network to obtain the optimized power grid section control information, including steps 502 to 506. Among them:

[0093] Step 502: randomly select a number of power safety consideration factors from the power grid system simulation network according to the initial power grid section control information.

[0094] Among them, power safety considerations may be safety factors that need to be considered by the target power grid.

[0095] Specifically, according to the requirements and objectives of power grid regulation, determine the power safety factors that need to be considered. These factors may include node voltage stability, line flow restrictions, generator output restrictions, etc. For each selected power safety consideration, according to the initial power grid section regulation information, randomly select the corresponding nodes, lines or generators from the power grid system simulation network. The selection process can be achieved through a random number generator to ensure randomness and representativeness. For each selected safety consideration, determine the specific parameters or indicators that need to be considered. These parameters may include node voltage values, line flow values, generator output values, etc.

[0096] Step 504: randomly generate a number of power efficiency calculation functions from the power grid system simulation network according to the initial power grid section control information.

[0097] The power efficiency calculation function may be a function used to calculate the efficiency of the target power grid.

[0098] Specifically, according to the objectives and requirements of power grid regulation, the types and categories of power efficiency calculation functions to be generated are determined. These calculation functions may include node voltage stability indicators, line flow loss indicators, generator efficiency indicators, etc. For each type of power efficiency indicator, the corresponding function form is randomly selected from the predefined function library according to the initial power grid section regulation information. For example, polynomial functions, exponential functions, logarithmic functions, etc. can be selected as the form of power efficiency calculation functions. For the selected calculation function, the parameters in the function are determined, including coefficients, exponents, constant terms, etc. These parameters can be generated within a certain range by a random number generator to ensure the diversity and representativeness of the function. According to the selected function form and parameters, the corresponding power efficiency calculation function is generated.

[0099] Step 506, optimizing the initial power grid section control information according to various power safety considerations and various power efficiency calculation functions to obtain optimized power grid section control information.

[0100] Specifically, according to the goals and needs of power grid regulation, the goals and indicators that need to be optimized are determined. These indicators may include safety indicators, efficiency indicators, or a combination of the two. For each selected power safety consideration factor and each power efficiency calculation function, the corresponding safety and efficiency index values ​​are calculated in combination with the initial power grid section regulation information. These index values ​​will be used to evaluate the quality of the current power grid section regulation information. According to the obtained safety and efficiency index values, the initial power grid section regulation information is optimized by comprehensively considering the importance and priority of each indicator. The optimization method can adopt heuristic algorithms, genetic algorithms, simulated annealing algorithms, etc., with the goal of maximizing or minimizing the objective function. After being processed by the optimization algorithm, the optimized power grid section regulation information is obtained, including the node adjustment scheme, the transmission capacity of the line, etc.

[0101] In this embodiment, by randomly selecting power safety considerations and generating power efficiency calculation functions from the power grid system simulation network, the safety and efficiency of the power grid are comprehensively considered. According to these safety considerations and efficiency calculation functions, the initial power grid section control information is optimized to obtain the optimized power grid section control information. The main advantage of this process is that by comprehensively considering safety and efficiency factors, the control strategy of the power grid is optimized, and the safety and operation efficiency of the power grid are improved. At the same time, through random selection and generation, the optimization process is more diverse and comprehensive, which can better cope with the complexity and uncertainty of the power grid and provide a more reliable guarantee for the stable operation of the power grid.

[0102] In an exemplary embodiment, Figure 6As shown, according to various power safety considerations and various power efficiency calculation functions, the initial power grid section control information is optimized to obtain the optimized power grid section control information, including steps 602 to 606. Wherein:

[0103] Step 602, based on power safety considerations and initial power grid section control information, calculate the load fluctuation information and new energy fluctuation information of the target power grid.

[0104] The load fluctuation information may be the fluctuation of the load of the non-new energy generating units of the target power grid.

[0105] The new energy fluctuation information may be the fluctuation of the load of the new energy power generation unit of the target power grid.

[0106] Specifically, based on power safety considerations and initial grid section control information, combined with load forecasting models or historical load data, the load fluctuation information of the target grid is calculated. The load fluctuation information may include the fluctuation range, frequency, and time period of the load curve. Similarly, based on power safety considerations and initial grid section control information, combined with renewable energy power generation models or historical renewable energy power generation data, the renewable energy fluctuation information of the target grid is calculated. The renewable energy fluctuation information may include the fluctuation range, frequency, and time period of renewable energy power generation such as wind power and photovoltaic power. The load fluctuation information and renewable energy fluctuation information are comprehensively considered and analyzed to evaluate the stability and reliability of the grid. Considering the dynamic characteristics and complexity of the power system, simulation software can be used for simulation analysis, or actual data can be used for statistical analysis.

[0107] Step 604, calculating the efficiency balance fluctuation information of the target power grid according to the power efficiency calculation function, the initial power grid section control information, the load fluctuation information and the new energy fluctuation information.

[0108] The efficiency balance fluctuation information may be the fluctuation situation when the loads of all the generator sets of the target power grid are balanced.

[0109] Specifically, the efficiency and balance of the power grid are comprehensively considered according to the power efficiency calculation function, the initial power grid section control information, the load fluctuation information and the new energy fluctuation information. According to the index value of the calculation function, combined with the initial power grid section control information, the load fluctuation and the new energy fluctuation, the efficiency balance of the power grid can be evaluated and analyzed to obtain the efficiency balance fluctuation information. The calculated efficiency balance fluctuation information is comprehensively analyzed to evaluate the operation of the power grid. If there is a large or unstable efficiency balance fluctuation, the power grid control strategy or optimization measures can be considered to improve the efficiency and balance of the power grid.

[0110] Step 606, optimizing the initial power grid section control information according to the load fluctuation information, the new energy fluctuation information and the efficiency balance fluctuation information to obtain the optimized power grid section control information.

[0111] Specifically, the impact of load fluctuation information, new energy fluctuation information and efficiency balance fluctuation information on the operation of the power grid is comprehensively considered, including the stability, efficiency and balance of the power grid, and the initial power grid section control information is optimized and adjusted so that the power grid can better adapt to load fluctuations, new energy fluctuations and maintain efficiency balance. The optimization may involve the node adjustment scheme, the transmission capacity adjustment of the line, the operation strategy of the energy storage equipment, etc., and the optimized power grid section control information is generated according to the adjusted power grid section control information.

[0112] In this embodiment, by calculating the load fluctuation information and new energy fluctuation information of the target power grid, as well as the efficiency balance fluctuation information, the load balance and energy utilization efficiency of the power grid are comprehensively considered. According to these fluctuation information, the initial power grid section control information is optimized to obtain the optimized power grid section control information. The main advantage of this process is that by comprehensively considering factors such as load fluctuation, new energy fluctuation and efficiency balance fluctuation, the section control strategy of the power grid is optimized, so that the power grid can operate more stably and efficiently when facing different load and energy fluctuations. At the same time, by calculating and analyzing the fluctuation information, the dynamic changes of the power grid operation can be better grasped, providing more accurate and reliable guidance for power grid control, and improving the overall performance and operation quality of the power grid.

[0113] In an exemplary embodiment, Figure 7 As shown, the optimized power grid section control information is used to train the power grid section control model to be trained until the disturbance of the initial power grid section control information is less than the disturbance threshold, thereby obtaining the trained power grid section control model, including steps 702 to 704. Wherein:

[0114] Step 702, when the disturbance of the initial grid section control information is greater than the disturbance threshold, the grid scenario parameters, control safety parameters, control efficiency parameters and scenario generalization parameters of the grid section control model to be trained are adjusted according to the optimized grid section control information to obtain the adjusted grid section control model.

[0115] Specifically, when the disturbance of the initial grid section control information is greater than the disturbance threshold, the grid scene parameters in the grid section control model to be trained are adjusted according to the optimized grid section control information, including the topological structure of the grid, load distribution, new energy distribution, etc., and it may be necessary to update the connection relationship between nodes, line parameters, generator capacity and other grid scene related parameters. According to the optimized grid section control information, the control safety parameters in the grid section control model to be trained are adjusted, including the rated power, flow limit, voltage limit, etc. of the equipment, and it may be necessary to update the rated power of the equipment, the transmission capacity limit of the line, the voltage limit of the node and other control safety related parameters. According to the optimized grid section control information, the control efficiency parameters in the grid section control model to be trained are adjusted, including the adjustment speed and response time of the equipment, and it may be necessary to update the start-stop response time of the generator, the charge and discharge rate of the energy storage device and other control efficiency related parameters. According to the optimized grid section control information, the scene generalization parameters in the grid section control model to be trained are adjusted to adapt to different grid scenes, and it may be necessary to update the learning rate, regularization parameter and other parameters in the model for generalization to different scenes. By adjusting the power grid scenario parameters, regulation safety parameters, regulation efficiency parameters and scenario generalization parameters, the adjusted power grid section regulation model is obtained.

[0116] Step 704, use the adjusted grid section control model as the grid section control model to be trained, return to execute the step of inputting the section control model training sample data into the grid section control model to be trained, and obtain the initial grid section control information corresponding to the target grid, until the disturbance of the initial grid section control information is less than the disturbance threshold, and obtain the trained grid section control model.

[0117] Specifically, the adjusted power grid section control model is used as the power grid section control model to be trained, and the step of "inputting the section control model training sample data into the power grid section control model to be trained to obtain the initial power grid section control information corresponding to the target power grid" is returned to execute; if the disturbance of the initial power grid section control information is less than the disturbance threshold, the trained power grid section control model is obtained; if the disturbance of the initial power grid section control information is still greater than the disturbance threshold, the training is continued until the disturbance of the initial power grid section control information is less than the disturbance threshold, and the trained power grid section control model is obtained.

[0118] In this embodiment, the parameters of the power grid section control model are dynamically adjusted to adapt to changes in the power grid operation state. When the disturbance of the initial power grid section control information is greater than the set disturbance threshold, the system will adjust the power grid scene parameters, control safety parameters, control efficiency parameters and scene generalization parameters of the power grid section control model to be trained according to the optimized power grid section control information to obtain the adjusted power grid section control model. Such optimization and adjustment can make the power grid section control model more adaptable to the current power grid operation state, and improve the flexibility and robustness of power grid control. Subsequently, the adjusted power grid section control model is used as the model to be trained, and the model is retrained until the disturbance of the initial power grid section control information is less than the set disturbance threshold, thereby obtaining the trained power grid section control model. This process ensures that the power grid section control model is highly adapted to the actual power grid operation state, and effectively improves the stability and efficiency of the power grid.

[0119] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0120] Based on the same inventive concept, the embodiment of the present application also provides a power grid section control model training device for implementing the power grid section control model training method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more power grid section control model training device embodiments provided below can refer to the limitations of a power grid section control model training method above, and will not be repeated here.

[0121] In an exemplary embodiment, Figure 8 As shown, a power grid section control model training device is provided, including: a power grid data acquisition module 802, a control information calculation module 804, a control information optimization module 806 and a control model training module 808, wherein:

[0122] The power grid data acquisition module 802 is used to acquire the section control model training sample data corresponding to the target power grid, the power grid section control model to be trained, and the power grid section optimization network;

[0123] The control information calculation module 804 is used to input the section control model training sample data into the power grid section control model to be trained to obtain the initial power grid section control information corresponding to the target power grid;

[0124] The control information optimization module 806 is used to input the initial power grid section control information into the power grid section optimization network to obtain the optimized power grid section control information;

[0125] The control model training module 808 is used to use the optimized power grid section control information to train the power grid section control model to be trained until the disturbance of the initial power grid section control information is less than the disturbance threshold, thereby obtaining a trained power grid section control model.

[0126] In one embodiment, the control information calculation module 804 is also used to input the section control model training sample data into the section control model of the power grid to be trained to obtain the trained power grid section control information; import the trained power grid section control information into the power grid system simulation network to obtain power grid operation simulation data; train the power grid section control model to be trained according to the power grid operation simulation data until the power grid operation simulation data meets the power operation standards, and use the trained power grid section control information as the initial power grid section control information.

[0127] In one embodiment, the control information calculation module 804 is also used to input the section control model training sample data into the power grid control strategy network to obtain power grid control decision information; input the section control model training sample data into the power grid control value network to obtain the power grid comprehensive reward value; use the power grid stability simulation model to adjust the power grid control decision information and the power grid comprehensive reward value to obtain adjusted control decision information and adjusted comprehensive reward value; adjust the model parameters of the power grid control strategy network and the power grid control value network according to the adjusted control decision information and the adjusted comprehensive reward value to obtain a parameter-adjusted power grid section control model; input the section control model training sample data into the parameter-adjusted power grid section control model to obtain trained power grid section control information.

[0128] In one embodiment, the control information optimization module 806 is also used to randomly select a number of power safety considerations from the power grid system simulation network based on the initial power grid section control information; randomly generate a number of power efficiency calculation functions from the power grid system simulation network based on the initial power grid section control information; optimize the initial power grid section control information based on each power safety consideration and each power efficiency calculation function to obtain optimized power grid section control information.

[0129] In one embodiment, the control information optimization module 806 is also used to calculate the load fluctuation information and new energy fluctuation information of the target power grid based on power safety considerations and the initial power grid section control information; calculate the efficiency balance fluctuation information of the target power grid based on the power efficiency calculation function, the initial power grid section control information, the load fluctuation information and the new energy fluctuation information; optimize the initial power grid section control information based on the load fluctuation information, the new energy fluctuation information and the efficiency balance fluctuation information to obtain the optimized power grid section control information.

[0130] In one embodiment, the control model training module 808 is also used to adjust the grid scenario parameters, control safety parameters, control efficiency parameters and scenario generalization parameters of the grid section control model to be trained according to the optimized grid section control information when the disturbance of the initial grid section control information is greater than the disturbance threshold, so as to obtain the adjusted grid section control model; use the adjusted grid section control model as the grid section control model to be trained, return to execute the step of inputting the section control model training sample data into the grid section control model to be trained, and obtain the initial grid section control information corresponding to the target grid, until the disturbance of the initial grid section control information is less than the disturbance threshold, so as to obtain the trained grid section control model.

[0131] Each module in the above-mentioned power grid section control model training device can be implemented in whole or in part by software, hardware and their combination. Each of the above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0132] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Fig. 9 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store server data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for training a power grid section control model is implemented.

[0133] Those skilled in the art will understand that Fig. 9 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0134] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.

[0135] In one embodiment, a computer-readable storage medium is provided, storing a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0136] In one embodiment, a computer program product or computer program is provided, the computer program product or computer program includes computer instructions, the computer instructions are stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the steps in the above-mentioned method embodiments.

[0137] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0138] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0139] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0140] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A power grid section control model training method, characterized in that: The method comprises: Obtaining training sample data of the section control model corresponding to the target power grid, the section control model of the power grid to be trained, and the power grid section optimization network; Input the section control model training sample data into the section control model of the power grid to be trained to obtain the initial power grid section control information corresponding to the target power grid, specifically including: inputting the section control model training sample data into the power grid control strategy network in the power grid section control model to be trained to obtain power grid control decision information; inputting the section control model training sample data into the power grid control value network in the power grid section control model to be trained to obtain a comprehensive power grid reward value; using a power grid stability simulation model to adjust the power grid control decision information and the comprehensive power grid reward value to obtain adjusted control decision information and adjusted comprehensive reward value; according to the The control decision information and the adjusted comprehensive reward value are adjusted to adjust the model parameters of the power grid control strategy network and the power grid control value network to obtain a parameter-adjusted power grid section control model; the section control model training sample data is input into the parameter-adjusted power grid section control model to obtain the trained power grid section control information; the trained power grid section control information is imported into the power grid system simulation network to obtain power grid operation simulation data; according to the power grid operation simulation data, the power grid section control model to be trained is trained until the power grid operation simulation data meets the power operation standard, and the trained power grid section control information is used as the initial power grid section control information; Inputting the initial grid section control information into the grid section optimization network to obtain optimized grid section control information, specifically including: randomly selecting a number of power safety considerations from the grid system simulation network according to the initial grid section control information; randomly generating a number of power efficiency calculation functions from the grid system simulation network according to the initial grid section control information; calculating the load fluctuation information and the new energy fluctuation information of the target grid according to the power safety considerations and the initial grid section control information; calculating the efficiency balance fluctuation information of the target grid according to the power efficiency calculation function, the initial grid section control information, the load fluctuation information and the new energy fluctuation information; optimizing the initial grid section control information according to the load fluctuation information, the new energy fluctuation information and the efficiency balance fluctuation information to obtain the optimized grid section control information; The grid section regulation model to be trained is trained using the optimized grid section regulation information until the disturbance of the initial grid section regulation information is less than a disturbance threshold, thereby obtaining a trained grid section regulation model.

2. The method according to claim 1, characterized in that The using the optimized power grid section regulation information to train the power grid section regulation model to be trained until the disturbance of the initial power grid section regulation information is less than a disturbance threshold, thereby obtaining a trained power grid section regulation model, comprises: When the disturbance of the initial power grid section control information is greater than the disturbance threshold, the power grid scene parameters, control safety parameters, control efficiency parameters and scene generalization parameters of the power grid section control model to be trained are adjusted according to the optimized power grid section control information to obtain an adjusted power grid section control model; The adjusted power grid section control model is used as the power grid section control model to be trained, and the step of inputting the section control model training sample data into the power grid section control model to be trained to obtain the initial power grid section control information corresponding to the target power grid is returned to, until the disturbance of the initial power grid section control information is less than the disturbance threshold, so as to obtain the trained power grid section control model.

3. A power grid section control model training device, characterized in that: The device comprises: A power grid data acquisition module is used to obtain the section control model training sample data corresponding to the target power grid, the power grid section control model to be trained, and the power grid section optimization network; The control information calculation module is used to input the section control model training sample data into the section control model of the power grid to be trained, and obtain the initial power grid section control information corresponding to the target power grid, specifically including: inputting the section control model training sample data into the power grid control strategy network in the power grid section control model to be trained, and obtaining power grid control decision information; inputting the section control model training sample data into the power grid control value network in the power grid section control model to be trained, and obtaining a comprehensive power grid reward value; using a power grid stability simulation model to adjust the power grid control decision information and the comprehensive power grid reward value, and obtain adjusted control decision information and adjusted comprehensive reward value ; According to the adjustment and control decision information and the adjustment comprehensive reward value, the model parameters of the power grid control strategy network and the power grid control value network are adjusted to obtain a parameter-adjusted power grid section control model; the section control model training sample data is input into the parameter-adjusted power grid section control model to obtain the training power grid section control information; the training power grid section control information is imported into the power grid system simulation network to obtain power grid operation simulation data; according to the power grid operation simulation data, the power grid section control model to be trained is trained until the power grid operation simulation data meets the power operation standard, and the training power grid section control information is used as the initial power grid section control information; A control information optimization module is used to input the initial grid section control information into the grid section optimization network to obtain optimized grid section control information, specifically including: randomly selecting a number of power safety considerations from the grid system simulation network according to the initial grid section control information; randomly generating a number of power efficiency calculation functions from the grid system simulation network according to the initial grid section control information; calculating the load fluctuation information and new energy fluctuation information of the target grid according to the power safety considerations and the initial grid section control information; calculating the efficiency balance fluctuation information of the target grid according to the power efficiency calculation function, the initial grid section control information, the load fluctuation information and the new energy fluctuation information; optimizing the initial grid section control information according to the load fluctuation information, the new energy fluctuation information and the efficiency balance fluctuation information to obtain the optimized grid section control information; The control model training module is used to use the optimized power grid section control information to train the power grid section control model to be trained until the disturbance of the initial power grid section control information is less than a disturbance threshold, thereby obtaining a trained power grid section control model.

4. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 2 are implemented.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 2 are implemented.

6. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 2 are implemented.

Images