Automatic matching regulation and control method and device for source-load balance of power system and electronic equipment

Abstract

The invention relates to the technical field of power system regulation and control, in particular to an automatic matching regulation and control method and device for power system source load balance and electronic equipment. Comprising the steps of obtaining current scene operation data of a power system; the current scene operation data are input to a preset neural network model, a source load balance automatic matching regulation and control strategy of the current scene is obtained, and the preset neural network model is obtained through historical scene operation data set training and a preset historical scene-source load balance regulation and control strategy mapping relation library; and performing automatic matching regulation and control on the source load balance process of the current scene based on the source load balance automatic matching regulation and control strategy. Therefore, flexible load resources based on the deep neural network participate in automatic matching regulation and control of power system source load balance, so that the problems of high source load volatility and lack of flexible regulation and control resources when high-proportion fluctuation renewable energy sources are accessed to the power system are solved; and an effective way is provided for implementation of flexible load resources participating in quick response of the power system.

Description

Technical Field

[0001] The present application relates to the technical field of power system control, and in particular to an automatic matching control method, device and electronic equipment for source-load balance of a power system. Background Art

[0002] As the global energy landscape accelerates its transition to green and low-carbon, the "dual carbon" goal has become the core orientation of the development of the power system, and the proportion of renewable energy in power supply continues to rise, which has brought unprecedented challenges to the source-load balance of the power system. On the one hand, renewable energy power generation is constrained by natural conditions, and its output fluctuates violently and is difficult to accurately predict; on the other hand, new power loads present diversified and dynamic characteristics, such as disordered charging of electric vehicles, intermittent loads of industrial production, and uncertainty in residents' electricity consumption behavior, which further increases the complexity and difficulty of source-load matching.

[0003] In a power system with a high proportion of renewable energy, the participation of flexible load resources is crucial to achieve source-load balance regulation. However, there are several problems at present: first, the research focuses on a single goal and lacks a comprehensive optimization strategy; second, the traditional optimization algorithm is complex to calculate and difficult to meet the needs of real-time regulation; third, the application of artificial intelligence and machine learning technology in flexible load regulation is insufficient, resulting in delayed regulation decisions and insufficient accuracy. These problems limit the improvement of system operation efficiency and the reliability of power supply, and cannot effectively solve the problems of strong source-load volatility and lack of flexible regulation resources when a high proportion of fluctuating renewable energy is connected to the power system, and they are in urgent need of improvement. Summary of the invention

[0004] The present application provides an automatic matching and control method, device and electronic equipment for source-load balance in an electric power system, so as to solve the problems of strong source-load volatility and lack of flexible control resources when a high proportion of fluctuating renewable energy is connected to the electric power system, and provides an effective way for flexible load resources to participate in the realization of rapid response of the electric power system.

[0005] To achieve the above objectives, the first embodiment of the present application proposes an automatic matching and control method for source-load balance of a power system, comprising the following steps:

[0006] Obtain current scenario operation data of the power system;

[0007] Inputting the current scenario operation data into a preset neural network model to obtain a source-load balance automatic matching control strategy for the current scenario, wherein the preset neural network model is obtained by training a historical scenario operation data set and a preset historical scenario-source-load balance control strategy mapping relationship library;

[0008] Based on the source-load balance automatic matching and regulation strategy, the source-load balance process of the current scene is automatically matched and regulated.

[0009] According to an embodiment of the present application, before the current scenario operation data is input into the preset neural network model to obtain the source-load balance automatic matching control strategy when the current scenario is running, it also includes:

[0010] Obtaining a historical scenario operation data set of the power system;

[0011] Preprocessing the historical scenario operation data set, and dividing the preprocessed historical scenario operation data set into a training set and a verification set;

[0012] Based on the preset historical scenario-source-load balance control strategy mapping relationship library, the preset neural network is trained using the training set to obtain an initial neural network model, and the initial neural network model is verified using the verification set until the initial neural network model meets the preset standard, and the iterative training of the preset neural network is terminated to obtain the preset neural network model, otherwise the training parameters of the preset neural network are adjusted and the iterative training continues.

[0013] According to an embodiment of the present application, before the preset neural network is trained using the training set to obtain the initial neural network model based on the preset historical scenario-source-load balance control strategy mapping relationship library, it also includes:

[0014] Obtain source-load balance to match regulation needs;

[0015] Based on the source-load balance matching regulation demand, the source-load imbalance scenario in the power system is divided into multiple typical scenarios, and the quantitative characterization index corresponding to each typical scenario is determined;

[0016] Obtaining the control requirements corresponding to each typical scenario;

[0017] Based on the quantitative characterization indicators corresponding to each typical scenario and the control requirements corresponding to each typical scenario, a source-load balance control model corresponding to each typical scenario is constructed;

[0018] Based on the historical scenario operation data set and the source-load balance control model corresponding to each typical scenario, the preset historical scenario-source-load balance control strategy mapping relationship library is established.

[0019] According to an embodiment of the present application, the typical scenario includes at least one of a renewable energy consumption scenario, a peak shaving scenario, a valley filling scenario, a frequency fluctuation scenario, and a voltage fluctuation scenario.

[0020] According to an embodiment of the present application, constructing a source-load balance control model corresponding to each typical scenario based on the quantitative characterization index corresponding to each typical scenario and the control demand corresponding to each typical scenario includes:

[0021] Determine the flexible load resource optimization control target corresponding to each typical scenario based on the quantitative characterization index corresponding to each typical scenario and the control demand corresponding to each typical scenario;

[0022] According to the flexible load resource optimization control target corresponding to each typical scenario, a source-load balance control model corresponding to each typical scenario is constructed.

[0023] According to an embodiment of the present application, the preset historical scenario-source-load balance control strategy mapping relationship library is established based on the historical scenario operation data and the source-load balance control model corresponding to each typical scenario, including:

[0024] Determine the operation scenario vector corresponding to each moment in the historical scenario operation data set;

[0025] Based on the source-load balance control model corresponding to each typical scenario, a multi-objective particle swarm algorithm is used to obtain the source-load balance control strategies for multiple historical scenarios;

[0026] Based on the source-load balance control strategies of the multiple historical scenarios, a preset historical scenario-source-load balance control strategy mapping relationship library is established.

[0027] According to the automatic matching and control method for source-load balance of the power system proposed in the embodiment of the present application, by obtaining the current scenario operation data of the power system and inputting the current scenario operation data into a preset neural network model, the automatic matching and control strategy for source-load balance of the current scenario can be obtained; based on the automatic matching and control strategy for source-load balance, the source-load balance process of the current scenario is automatically matched and controlled. Thus, by using flexible load resources based on deep neural networks to participate in the automatic matching and control of source-load balance of the power system, the problem of strong source-load volatility and lack of flexible control resources when a high proportion of fluctuating renewable energy sources are connected to the power system is solved, and an effective way is provided for the realization of flexible load resources participating in the rapid response of the power system.

[0028] To achieve the above-mentioned purpose, the second embodiment of the present application proposes an automatic matching and control device for source-load balance of a power system, comprising:

[0029] An acquisition module is used to obtain the current scenario operation data of the power system;

[0030] An acquisition module is used to input the current scene operation data into a preset neural network model to obtain a source-load balance automatic matching control strategy for the current scene, wherein the preset neural network model is obtained by training a historical scene operation data set and a preset historical scene-source-load balance control strategy mapping relationship library;

[0031] The control module is used to automatically match and control the source-load balancing process of the current scene based on the source-load balancing automatic matching control strategy.

[0032] According to an embodiment of the present application, before inputting the current scenario operation data into the preset neural network model to obtain the source-load balance automatic matching control strategy when the current scenario is running, the obtaining module further includes:

[0033] An acquisition unit, used for acquiring a historical scenario operation data set of the power system;

[0034] A preprocessing unit, used to preprocess the historical scenario operation data set, and divide the preprocessed historical scenario operation data set into a training set and a verification set;

[0035] A training unit is used to train a preset neural network based on the preset historical scenario-source-load balance control strategy mapping relationship library using the training set to obtain an initial neural network model, and to verify the initial neural network model using the verification set until the initial neural network model meets the preset standard, and then terminate the iterative training of the preset neural network to obtain the preset neural network model, otherwise, continue the iterative training after adjusting the training parameters of the preset neural network.

[0036] According to an embodiment of the present application, before the preset neural network is trained using the training set based on the preset historical scenario-source-load balance control strategy mapping relationship library to obtain the initial neural network model, the training unit further includes:

[0037] The first acquisition subunit is used to obtain the source-load balance matching regulation demand;

[0038] A determination subunit is used to divide the source-load imbalance scenario in the power system into multiple typical scenarios based on the source-load balance matching regulation demand, and determine the quantitative characterization index corresponding to each typical scenario;

[0039] A second acquisition subunit is used to acquire the control requirements corresponding to each typical scenario;

[0040] A construction subunit is used to construct a source-load balance control model corresponding to each typical scenario based on the quantitative characterization index corresponding to each typical scenario and the control demand corresponding to each typical scenario;

[0041] A subunit is established to establish the preset historical scenario-source-load balance control strategy mapping relationship library based on the historical scenario operation data set and the source-load balance control model corresponding to each typical scenario.

[0042] According to an embodiment of the present application, the typical scenario includes at least one of a renewable energy consumption scenario, a peak shaving scenario, a valley filling scenario, a frequency fluctuation scenario, and a voltage fluctuation scenario.

[0043] According to one embodiment of the present application, the construction subunit is specifically used for:

[0044] Determine the flexible load resource optimization control target corresponding to each typical scenario based on the quantitative characterization index corresponding to each typical scenario and the control demand corresponding to each typical scenario;

[0045] According to the flexible load resource optimization control target corresponding to each typical scenario, a source-load balance control model corresponding to each typical scenario is constructed.

[0046] According to one embodiment of the present application, the establishing subunit is specifically used for:

[0047] Determine the operation scenario vector corresponding to each moment in the historical scenario operation data set;

[0048] Based on the source-load balance control model corresponding to each typical scenario, a multi-objective particle swarm algorithm is used to obtain the source-load balance control strategies for multiple historical scenarios;

[0049] Based on the source-load balance control strategies of the multiple historical scenarios, a preset historical scenario-source-load balance control strategy mapping relationship library is established.

[0050] According to the automatic matching and control device for source-load balance of the power system proposed in the embodiment of the present application, by acquiring the current scenario operation data of the power system and inputting the current scenario operation data into a preset neural network model, the automatic matching and control strategy for source-load balance of the current scenario can be obtained; based on the automatic matching and control strategy for source-load balance, the source-load balance process of the current scenario is automatically matched and controlled. Thus, by using the flexible load resources based on deep neural networks to participate in the automatic matching and control of source-load balance of the power system, the problem of strong source-load volatility and lack of flexible control resources when a high proportion of fluctuating renewable energy sources are connected to the power system is solved, and an effective way is provided for the realization of flexible load resources participating in the rapid response of the power system.

[0051] To achieve the above-mentioned objectives, the third aspect embodiment of the present application proposes an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement an automatic matching and control method for source-load balance of an electric power system as in the above-mentioned embodiment.

[0052] To achieve the above objectives, the fourth aspect of the present application proposes a computer-readable storage medium on which a computer program is stored, and the program is executed by a processor to implement an automatic matching and control method for source-load balance of an electric power system as in the above embodiment.

[0053] To achieve the above objectives, the fourth aspect of the present application proposes a computer program product, which includes a computer program. When the computer program is executed by a processor, it can implement the automatic matching and control method for source-load balance of the power system as in the above embodiment.

[0054] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0056] Figure 1 A flowchart of an automatic matching and control method for source-load balance of a power system provided according to an embodiment of the present application;

[0057] Figure 2 A schematic diagram of a framework of an automatic matching and control method for source-load balance of a power system according to an embodiment of the present application;

[0058] Figure 3 A schematic diagram of a preset historical scenario-source-load balance control strategy mapping relationship library according to an embodiment of the present application;

[0059] Figure 4 A flowchart of an automatic matching and control method for source-load balance of a power system according to another embodiment of the present application;

[0060] Figure 5 A schematic diagram of an improved IEEE 30-node system topology according to an embodiment of the present application;

[0061] Figure 6 A curve diagram of a training process of an automatic matching and control method for source-load balance in a power system according to an embodiment of the present application;

[0062] Figure 7A matching schematic diagram of an automatic matching and control method for source-load balance of a power system according to an embodiment of the present application;

[0063] Figure 8 A schematic diagram of a matching result (I) of a real-time scene to be optimized according to an embodiment of the present application;

[0064] Fig. 9 A schematic diagram of a matching result (II) of a real-time scene to be optimized according to an embodiment of the present application;

[0065] Fig.10 A schematic diagram of a broken line showing the regulation effect of a renewable energy consumption scenario according to an embodiment of the present application;

[0066] Fig.11 A schematic diagram of a peak shaving scenario control effect according to an embodiment of the present application;

[0067] Fig.12 A schematic diagram of a broken line showing the control effect of a valley-filling scene according to an embodiment of the present application;

[0068] Fig.13 A schematic diagram of a frequency fluctuation scene control effect according to an embodiment of the present application;

[0069] Fig.14 A schematic diagram of a broken line showing the voltage fluctuation scenario control effect according to an embodiment of the present application;

[0070] Fig.15 A block diagram of an automatic matching and control device for source-load balance of a power system provided according to an embodiment of the present application;

[0071] Fig.16 It is a schematic diagram of the structure of an electronic device provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0072] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0073] The following describes the automatic matching and control method, device and electronic device for source-load balance of the power system proposed in the embodiments of the present application with reference to the accompanying drawings. First, the automatic matching and control method for source-load balance of the power system proposed in the embodiments of the present application will be described with reference to the accompanying drawings.

[0074] Figure 1 It is a flow chart of an automatic matching and control method for source-load balance of an electric power system according to an embodiment of the present application.

[0075] Before introducing the automatic matching and control method for source-load balance of an electric power system proposed in the embodiment of the present application, the relevant technical background is first introduced.

[0076] At present, the power system faces a significant scarcity problem in terms of flexible regulation resources. The capacity of traditional dispatchable power sources is shrinking, resulting in a decrease in the system's inertia support and frequency and voltage regulation capabilities, making it difficult to effectively cope with the uncertainty shocks on both the source and load sides. In this context, tapping and utilizing the potential of flexible load resources to participate in the rapid response regulation of the system has become a key way to improve the real-time source-load balance regulation capabilities of power systems with a high proportion of renewable energy. However, looking at the current technical landscape, there are still many problems and shortcomings that need to be solved:

[0077] (1) At present, the research on the participation of flexible loads in the system source-load balance regulation is mainly focused on a single target scenario. In the context of renewable energy consumption, some optimization models only focus on maximizing the amount of consumption, but ignore the need for voltage stability and frequency regulation during load fluctuations; most of the research on peak shaving and valley filling starts from the adjustment of load form, and insufficient consideration is given to the dynamic balance under the fluctuation of renewable energy power generation; and in terms of system frequency response and voltage regulation, there is also a lack of comprehensive optimization control strategies for various working conditions. This single-target-oriented research method leads to insufficient adaptability of the control scheme when facing complex and changeable multi-scenario switching in the actual system, and cannot meet the system's all-round and multi-level rapid response requirements, which greatly limits the full realization of the value of flexible load resources and the improvement of the overall system operation efficiency.

[0078] (2) The generation of current control strategies mainly relies on traditional optimization algorithms, such as multi-objective particle swarm optimization and genetic algorithms. Although these algorithms can achieve precise optimization in theory, their computational complexity increases significantly with the increase of system scale and variable dimension, resulting in a lengthy iterative optimization process. In the scenario of real-time source-load balance control, facing the ever-changing system conditions, the solution speed of these algorithms cannot meet the control requirements, resulting in delays in the issuance of control instructions, making it impossible for the system to capture and respond to changes in operating status in a timely manner and to intervene accurately, thereby seriously weakening the system's ability to respond to source-load fluctuations, causing power quality to decline and operational stability to be damaged, and laying hidden dangers for reliable power supply.

[0079] (3) Although artificial intelligence and machine learning technologies have emerged in many areas of power systems, progress is still slow in deeply integrating flexible load resources into system regulation and innovative applications. Flexible load resources are diverse in type, complex in characteristics, and large in scale. Their regulation process involves the collection and processing of large amounts of data, the construction of complex models, and the formulation of real-time decisions. Traditional technical methods are unable to cope with such complex tasks, and advanced technologies such as deep neural networks (DNN) have not yet been fully explored and effectively applied. This makes it difficult for the system to efficiently mine data value and accurately extract load characteristics and operating laws when faced with large amounts of data and complex working conditions, and it is impossible to achieve intelligent generation and adaptive adjustment of regulation strategies, resulting in delayed regulation decisions and insufficient accuracy, which has become a key bottleneck restricting the intelligent upgrade and efficient operation of the power system.

[0080] Based on the above problems, the embodiment of the present application proposes an automatic matching and control method for the source-load balance of the power system. By involving flexible load resources based on deep neural networks in the automatic matching and control of the source-load balance of the power system, the problem of strong source-load volatility and lack of flexible control resources when a high proportion of fluctuating renewable energy is connected to the power system is solved, and an effective way is provided for flexible load resources to participate in the realization of rapid response of the power system.

[0081] For example, Figure 1 As shown, the automatic matching and control method for source-load balance of the power system includes the following steps:

[0082] In step S101, current scenario operation data of the power system is acquired.

[0083] Specifically, using sensors, measuring equipment or remote monitoring systems, the operating data of the power system can be collected in real time, which may include data on the power generation side (such as the output power of each power plant, fuel consumption, etc.) and the load side (such as load distribution, etc.).

[0084] In step S102, the current scene operation data is input into a preset neural network model to obtain the source-load balance automatic matching control strategy of the current scene, wherein the preset neural network model is obtained by training the historical scene operation data set and a preset historical scene-source-load balance control strategy mapping relationship library.

[0085] It can be understood that source-load balance refers to the balance between power generation (source) and power consumption (load); the automatic matching control strategy is an automated adjustment method that aims to achieve the best matching state between power generation and power consumption.

[0086] That is to say, the current scenario operation data obtained in step S101 will be input into a neural network model (i.e., a preset neural network model) trained with historical data. The neural network model is established by training with the operation data set of the historical scenario. In addition, the model also combines a mapping relationship library of pre-set historical scenarios and source-load balance control strategies. In this way, the preset neural network model can output the source-load balance automatic matching control strategy for the current power system scenario based on these input data.

[0087] In step S103, based on the source-load balance automatic matching and regulation strategy, the source-load balance process of the current scene is automatically matched and regulated.

[0088] That is to say, according to the source-load balance automatic matching control strategy output in step S102, the system will automatically adjust the source-load balance process of the current scene to achieve more efficient and stable power supply and ensure the stable operation of the power system.

[0089] To facilitate understanding, the automatic matching and control method for source-load balance of the power system proposed in the embodiment of the present application will be further elaborated in detail.

[0090] First, it is explained in detail how to obtain the preset neural network model.

[0091] As a possible implementation method, in some embodiments, before inputting the current scenario operation data into a preset neural network model to obtain the source-load balance automatic matching control strategy when the current scenario is running, it also includes: obtaining a historical scenario operation data set of the power system; preprocessing the historical scenario operation data set, and dividing the preprocessed historical scenario operation data set into a training set and a verification set; based on a preset historical scenario-source-load balance control strategy mapping relationship library, using the training set to train the preset neural network to obtain an initial neural network model, and using the verification set to verify the initial neural network model until the initial neural network model meets the preset standards, and ending the iterative training of the preset neural network to obtain the preset neural network model, otherwise adjusting the training parameters of the preset neural network and continuing the iterative training.

[0092] That is to say, in an embodiment of the present application, the historical scenario operation data set of the power system is used as input and the control strategy is used as output, and a preset neural network (i.e., a deep neural network) is used for training to form a preset neural network model. The preset neural network model reflects the nonlinear input-output "mapping" relationship between the operation scenario and the control strategy. By inputting the real-time scenario operation data of the power system into the preset neural network model, a control strategy that matches the current scenario operation can be directly obtained.

[0093] Specifically, firstly, the historical scenario operation data sets of the power system are obtained (these data sets contain the actual operation conditions of the power system in different historical periods), and then these historical scenario operation data sets are preprocessed (such as data cleaning, data normalization, outlier processing, etc.) to ensure the quality and consistency of the data. Next, the preprocessed historical scenario operation data sets can be divided into a training set and a validation set. The training set is used to train the neural network model, while the validation set is used to evaluate the performance and generalization ability of the model. In this way, it can be ensured that the trained model can also perform well on unseen data. Based on the preset historical scenario-source-load balance control strategy mapping relationship library (which defines the corresponding relationship between source-load balance control strategies in different scenarios), the training set can be used to train the deep neural network to obtain an initial neural network model. After obtaining the initial neural network model, it can be verified using the validation set. If the performance of the model meets the preset standards, the iterative training of the deep neural network can be ended to obtain the final preset neural network model. If the performance of the model does not meet the preset standards, the training parameters of the preset neural network, such as the learning rate, batch size, optimization algorithm, etc., can be adjusted, and then the iterative training can be continued. Through continuous iterative training and adjustment, we can eventually obtain a preset neural network model that meets the preset standards. This model can accurately predict the source-load balance control strategy based on historical scenarios, thereby providing strong decision-making support for the operation of the power system.

[0094] Among them, deep neural network is a machine learning model that simulates the structure and function of the human brain and is designed to solve complex problems. The model consists of multiple layers of artificial neurons, which are distributed in multiple levels, including input layer, several hidden layers and output layer. The core operating mechanism of deep neural network involves two key steps: forward propagation and back propagation.

[0095] Forward propagation refers to the process of starting from the input layer, calculating layer by layer and passing it to the output layer layer by layer. This process is divided into the following steps:

[0096] First, define the input layer:

[0097]

[0098] Among them, n is the number of historical scenes, m is the dimension of the scene vector, and each row vector x in the matrix 11 ,x 12 ,…,x 1m is the scene vector.

[0099] The data from the input layer flows to a series of hidden layers, each of which performs a weighted sum on the data, applies a bias, and performs a nonlinear transformation through an activation function. (l) The output expression is as follows:

[0100] h (l) =f(W (l) h (l-1) +b (l) ); (2)

[0101] Among them, W (l) is the weight of the hidden layer, b (l) is the bias of the hidden layer, f is the activation function, and the nonlinear activation function can help the network learn nonlinear complex functions, thereby solving the nonlinear mapping problem. The ReLU function is efficient and can avoid the gradient disappearance problem. Its function expression is as follows:

[0102]

[0103] Next, the output of the hidden layer is passed to the output layer, which is usually processed by the Softmax function to obtain the probability distribution of the classification:

[0104] Y = Softmax(W (out) h (L) +b (out) ); (4)

[0105] Among them, L is the last hidden layer, the Softmax function is shown in formula (6), and Y is the output layer, which is defined as follows:

[0106]

[0107] Among them, y ij is the probability of scene i being classified as j, each row of y ij The class number corresponding to the maximum value of is the matching result y of scene i.

[0108] Backpropagation is the key mechanism for optimizing the weights and biases of a neural network to reduce prediction errors. It can be achieved by following the steps below:

[0109] First, the loss function is calculated, that is, the cross entropy loss between the matching result y and the true label t is calculated:

[0110]

[0111] Then, for each layer in the network, the loss function is calculated with respect to the weight W (l) and bias b (l) The partial derivatives of , and the gradient descent method is used to update the weights and biases of each layer:

[0112]

[0113] Among them, η is the learning rate, which determines the step size of parameter update in gradient descent.

[0114] The mapping process of the deep neural network is to learn the complex relationship between input features and output labels through the input layer, hidden layer and output layer. In the embodiment of the present application, the schematic diagram of the automatic matching method of the flexible load "scenario-control scheme" considering the real-time source-load balance regulation demand of the system is as follows Figure 2 As shown in the figure, the input and output of the deep neural network are scene vectors and control strategies respectively. The deep neural network learns a large number of nonlinear mapping relationships between scenes and control strategies through offline training, realizes automatic recognition of the features of scene vectors, and matches them to the optimal control strategy, avoiding real-time optimization calculations.

[0115] Furthermore, in order to ensure that the method proposed in the embodiment of the present application can maintain a high matching accuracy in various scenarios, it is necessary to evaluate the matching and control effects. The embodiment of the present application proposes a matching accuracy evaluation index A to evaluate the automatic matching effect, and its calculation formula is as follows:

[0116]

[0117] Among them, z k is the scene vector of the real-time scene, x k is the scene vector in the matching historical scene library, m is the length of the vector, and k is a natural number.

[0118] Next, how to obtain the preset historical scenario-source-load balance control strategy mapping relationship library is explained in detail.

[0119] As a possible implementation method, in some embodiments, before using a training set to train a preset neural network based on a preset historical scenario-source-load balance control strategy mapping relationship library to obtain an initial neural network model, it also includes: obtaining source-load balance matching control requirements; dividing the source-load imbalance scenarios in the power system into multiple typical scenarios based on the source-load balance matching control requirements, and determining the quantitative characterization indicators corresponding to each typical scenario; obtaining the control requirements corresponding to each typical scenario; constructing a source-load balance control model corresponding to each typical scenario based on the quantitative characterization indicators corresponding to each typical scenario and the control requirements corresponding to each typical scenario; establishing a preset historical scenario-source-load balance control strategy mapping relationship library based on the historical scenario operation data set and the source-load balance control model corresponding to each typical scenario.

[0120] This step involves a detailed analysis and classification of the scenarios of source-load imbalance in the power system, dividing them into a variety of representative typical scenarios, and determining the corresponding quantitative characterization indicators for each typical scenario, which can accurately reflect the characteristics of source-load imbalance in the scenario. Next, for each typical scenario, obtain its corresponding control requirements, which can guide the effective control of the power system. Based on the quantitative characterization indicators and control requirements of each typical scenario, a corresponding source-load balance control model can be constructed (different source-load balance control models have different optimization objective functions), which will provide scientific control strategies for each typical scenario. Finally, using the historical scenario operation data set and combining the source-load balance control model corresponding to each typical scenario, the mapping relationship library between the historical scenario and the source-load balance control strategy (i.e., the preset historical scenario-source-load balance control strategy mapping relationship library) can be obtained to deal with various possible source-load imbalance scenarios.

[0121] Optionally, in some embodiments, the typical scenario includes at least one of a renewable energy consumption scenario, a peak shaving scenario, a valley filling scenario, a frequency fluctuation scenario, and a voltage fluctuation scenario.

[0122] It can be understood that the renewable energy consumption scenario refers to the imbalance between the source and load of the power system due to the uncertainty of the power generation of renewable energy such as solar and wind energy. The power generation of solar and wind energy is significantly affected by multiple factors such as weather, climate, day and night changes, and seasonal changes, which in turn causes the instability of power supply. When the power system cannot fully absorb renewable energy, wind and solar power will be abandoned. The peak load scenario refers to the situation where the load demand of the power grid reaches a peak during a specific period of time, which usually occurs at certain specific times of the day, such as the morning and evening peak hours on weekdays. The causes of the peak load scenario include factors such as synchronous high demand caused by life and work patterns and seasonal effects. The valley filling scenario refers to the period when the power demand in the power grid is at its lowest point, which usually occurs at night or during non-peak hours on weekdays. The causes of the valley load scenario include time factors, seasonal changes, holiday effects, etc. The frequency fluctuation scenario refers to the situation where the power grid frequency fluctuates up and down from its standard value of 50Hz. This fluctuation affects the power quality and may cause power grid stability problems. The causes of the frequency fluctuation scenario include supply and demand imbalance or emergencies. For example, sudden events such as the sudden start-up of large-scale industrial equipment or the sudden shutdown of large-scale power generation equipment, failure of power stations connected to the grid, or failure of transmission lines can instantly change the relationship between power supply and demand, causing frequency fluctuations. Voltage fluctuation scenarios refer to short-term or continuous deviations of voltage levels in the grid from their standard values. Such fluctuations may affect the stability of the grid and the operating efficiency of power equipment, affecting the energy efficiency and safe operation of the grid. The causes of voltage fluctuation scenarios include the start-up and shutdown of high-power equipment, rapid changes in loads, failures of transmission lines, fluctuations in the output of renewable energy, etc.

[0123] In view of the fact that the power system may face a variety of scenarios during actual operation, the embodiment of the present application introduces a scenario priority mechanism to resolve and classify conflicts, ensuring that the more important regulation needs of the power system are met first. The setting of scenario priority comprehensively considers factors such as the safety, stability and economy of power grid operation. The order from high to low is frequency fluctuation scenario, voltage fluctuation scenario, peak shaving scenario, valley filling scenario and renewable energy consumption scenario. The setting of this priority refers to relevant national standards and theoretical research literature on load regulation. For example, if the current scenario operation data meets the indicator conditions of both the frequency fluctuation scenario and the peak shaving scenario (that is, the frequency deviation Δf>50% and the load factor r load >85%), it is preferentially classified as a frequency fluctuation scenario.

[0124] That is to say, when studying and analyzing the source-load imbalance problem that occurs in the operation of the power system, the embodiments of the present application can summarize these complex scenarios into five typical scenarios, including: renewable energy consumption scenario, peak shaving scenario, valley filling scenario, frequency fluctuation scenario and voltage fluctuation scenario. Through this classification, the most common source-load imbalance problems in power grid operation can be understood and dealt with more systematically, including but not limited to the problem of efficient consumption of renewable energy, load fluctuations caused by various factors, and the demand for power grid frequency stability and voltage stability.

[0125] By constructing corresponding quantitative characterization indicators for typical scenarios and a source-load balance control model with corresponding optimization control objective function, an optimization control strategy for different typical scenarios can be formed.

[0126] Among them, the quantitative characterization indicators corresponding to each typical scenario may include:

[0127] (1) Renewable energy absorption rate: The renewable energy absorption rate measures the actual utilization of renewable energy by the power system within a certain period of time, reflecting the power system's ability to absorb renewable energy.

[0128] Its expression is as follows:

[0129]

[0130] Among them, r renew is the renewable energy consumption rate, P renew_creat is the actual power generation of renewable energy, P renew_max It is the maximum power generation under the current wind speed or irradiation conditions.

[0131] (2) Load factor: The load factor measures the ratio of the actual load of the power system at a certain moment to the maximum load in the past year. It is used to quantify the flexible load control requirements in peak shaving and valley filling scenarios.

[0132] Its expression is as follows:

[0133]

[0134] Among them, r load is the load factor, P L is the current actual load, P Lmax It is the maximum load value of the load curve in the past year.

[0135] (3) Frequency deviation: Frequency deviation is an important indicator to measure the degree of deviation between the actual operating frequency of the power system and the rated frequency.

[0136] Its expression is as follows:

[0137]

[0138] Among them, Δf is the frequency deviation, f is the actual frequency of the power grid, and f N The rated frequency is 50Hz, and 0.2 is the maximum allowable range of frequency deviation in my country's frequency standard.

[0139] (4) System voltage deviation rate: The system voltage deviation rate is an important indicator for measuring the voltage stability of the power system, and is used to measure the comprehensive degree of voltage deviation of all nodes.

[0140] Its expression is as follows:

[0141]

[0142] Among them, ΔU is the system voltage deviation rate, N is the total number of nodes, and U i is the actual voltage of node i, which is the actual operation monitoring data of the SCADA (Supervisory Control And Data Acquisition) system, U Ni is the rated voltage of node i.

[0143] In order to improve the renewable energy consumption of the power system, the renewable energy consumption rate is set to be less than 95% as the renewable energy consumption scenario; based on the power system operation stability, safety margin and actual operation experience, the load rate thresholds of the power system peak shaving scenario and valley filling scenario are set to 85% and 35% respectively; according to the relevant provisions of the national standard, the frequency fluctuation and voltage fluctuation scenario ranges are set to Δf>50% and ΔU>50% respectively. Among them, the corresponding relationship between each typical scenario and the quantitative index is shown in Table 1:

[0144] Table 1

[0145] Typical scenarios Quantitative indicators Scene classification label Renewable energy consumption scenarios <![CDATA[r renew <95%]]> 1 System peak shaving scenario <![CDATA[r load >85%]]> 2 System valley filling scenario <![CDATA[r load <35%]]> 3 System frequency fluctuation scenario Δf>50% 4 System voltage fluctuation scenario ΔU>50% 5

[0146] Furthermore, in some embodiments, based on the historical scenario operation data and the source-load balance control model corresponding to each typical scenario, a preset historical scenario-source-load balance control strategy mapping relationship library is established, including: determining the operation scenario vector corresponding to each moment in the historical scenario operation data set; based on the source-load balance control model corresponding to each typical scenario, using a multi-objective particle swarm algorithm to obtain the source-load balance control strategies for multiple historical scenarios; based on the source-load balance control strategies for multiple historical scenarios, a preset historical scenario-source-load balance control strategy mapping relationship library is established.

[0147] Specifically, since the operation data of the power system at a certain moment can reflect a typical scenario in which the power system is currently located, and the optimization model and algorithm can be used to obtain the optimal source-load balance control strategy based on the operation data, the embodiment of the present application can use the operation data of a single time section of the power system as the basic data for scene construction, including renewable energy power, load power, and interconnection line power, thereby forming a scene vector to reflect the source-load power situation of the current power system. Among them, the expression form of the operation scene vector is as follows:

[0148]

[0149] Among them, |N R | is the number of renewable energy units, |N L | is the number of temperature control loads, |N V | is the number of electric vehicle loads, |N S | is the number of distributed energy storage, |N NL | is the number of normal loads, P R is the active power of renewable energy, P FL is the active power of the flexible load, Q FL is the reactive power of flexible load, P NL is the normal load active power, Q NL is the reactive power of normal load, P T is the active power of the tie line, Q T is the reactive power of the interconnection line.

[0150] Since the scenario vector P reflects the source-load power situation of the current power system, when the scenario vectors P at two moments are the same, the power system can adopt the same regulation strategy.

[0151] Furthermore, based on the source-load balance control model under different typical scenarios, the traditional multi-objective particle swarm algorithm is used to accurately solve each specific scenario, thereby obtaining a series of control strategies for historical scenarios, thereby obtaining a preset historical scenario-source-load balance control strategy mapping relationship library.

[0152] An example of the preset historical scenario-source-load balance control strategy mapping relationship library can be as follows: Figure 3 As shown, Figure 3The 17 columns represent 17 historical scenes. The first 4 rows are scene characterization indicators, which are classified into typical scenes according to the indicators. The middle 28 rows are historical scene library data, which are the transpose of the scene vector P defined by formula (14). The last 7 rows are source-load balance control library data, which are the adjustment powers of various flexible loads obtained by combining historical scene data with typical scene control models and traditional multi-objective particle swarm algorithms. The automatic matching control method of the embodiment of the present application will match the historical scene closest to the current scene vector through a deep neural network, so as to use the control strategy of the historical scene as the control strategy of the current real-time scene.

[0153] The following describes in detail how to construct the source-load balance control model corresponding to each typical scenario.

[0154] Optionally, in some embodiments, based on the quantitative characterization indicators corresponding to each typical scenario and the control requirements corresponding to each typical scenario, a source-load balance control model corresponding to each typical scenario is constructed, including: based on the quantitative characterization indicators corresponding to each typical scenario and the control requirements corresponding to each typical scenario, determining the flexible load resource optimization control target corresponding to each typical scenario; according to the flexible load resource optimization control target corresponding to each typical scenario, constructing a source-load balance control model corresponding to each typical scenario.

[0155] That is to say, according to the quantitative characterization indicators corresponding to different typical scenarios and the control requirements corresponding to each typical scenario, the optimization control objective function of the flexible load resources corresponding to each typical scenario can be further set. Flexible load resources refer to those power resources that can be adjusted according to the needs of the power grid, such as energy storage equipment, adjustable industrial loads, etc. According to the optimization control objectives of the flexible load resources corresponding to each typical scenario, a source-load balance control model corresponding to each typical scenario can be constructed. The source-load balance control model is a mathematical model that can help us optimize the allocation of power resources and ensure the balance between power supply and demand in different scenarios, thereby improving the operating efficiency and reliability of the entire power system.

[0156] Specifically, the embodiment of the present application takes the lowest control cost as the comprehensive objective function of various typical scenarios:

[0157] minF=F tl +F ev +F es ; (15)

[0158] Among them, F tl is the control cost function of air conditioning load, F ev is the regulation cost function of electric vehicle load, F esis the regulation cost function of distributed energy storage. Since the flexible load resources with large-scale regulation capability in the power system include air conditioning load, electric vehicle load and distributed energy storage, these three are selected as the regulation objects of the embodiment of the present application, and the formulas are shown in formulas (16)-(18):

[0159]

[0160] Among them, N L is the set of air conditioning loads, N V is the collection of electric vehicle loads, N S It is a collection of distributed energy storage. is the control cost coefficient of air conditioning load, is the control cost coefficient of electric vehicle load, is the control cost coefficient of distributed energy storage, is the regulating power of the air conditioning load, is the regulated power of the electric vehicle load, and its positive and negative values ​​represent upward and downward regulation respectively. for The maximum value of It is the charging and discharging power of distributed energy storage. Its positive value indicates charging, and its negative value indicates discharging. It is the reactive power of distributed energy storage. Since energy storage equipment is usually equipped with advanced inverters, its reactive power can be adjusted independently of the active power.

[0161] Since each typical scenario has its own regulation requirements, in order to meet the regulation requirements of different typical scenarios, the embodiment of the present application sets a characteristic objective function corresponding to each typical scenario.

[0162] (1) Renewable energy consumption scenario, setting maximum renewable energy consumption as the characteristic goal:

[0163]

[0164] in, is the power of a single renewable energy source, N R A collection of renewable energy sources.

[0165] (2) System peak shaving scenario, set the minimum load as the characteristic target:

[0166]

[0167] (3) System valley filling scenario, set maximum load as the characteristic target:

[0168]

[0169] (4) In the system frequency fluctuation scenario, the frequency deviation is characterized by the power deviation of the tie line between the distribution network and the large power grid. Therefore, the characteristic optimization goal is to minimize the power deviation of the tie line:

[0170]

[0171] Among them, P T is the tie line power, is the normal load power, N NL It is the collection of all common loads.

[0172] (5) System voltage fluctuation scenario, with minimum voltage deviation as the characteristic objective function:

[0173]

[0174] The constraints are:

[0175] (1) Flexible load power upper and lower limit constraints:

[0176]

[0177] in, is the actual adjustment power of the air conditioning load, is the maximum adjustable power of the air conditioning load, is the maximum adjustable power value, is the maximum allowable charging power of the distributed energy storage load, is the maximum allowable discharge power of the distributed energy storage load, is the maximum reactive power that can be absorbed by the distributed energy storage load, It is the maximum reactive power that can be generated by the distributed energy storage load.

[0178] (2) Power balance constraints:

[0179]

[0180] Among them, P i is the active power injected into node i, Q i is the reactive power injected into node i, V i is the voltage amplitude of node i, V j is the voltage amplitude at node j, G ij is the real part of the admittance matrix, B ij is the imaginary part of the admittance matrix, θ ij is the phase angle difference between i and j, is the active power of renewable energy, which is used as a random variable in the embodiment of the present application. is the active power of ordinary load, is a random variable, is the reactive power of ordinary load and is a random variable.

[0181] (3) Node voltage upper and lower limit constraints:

[0182]

[0183] in, is the maximum safe lower limit of the voltage amplitude at node i, is the maximum safe upper limit of the voltage amplitude at node i, and N is the set of system nodes.

[0184] (4) Line transmission capacity constraints:

[0185]

[0186] Among them, P ij is the active power of the line between nodes i and j, Q ij is the reactive power of the line between nodes i and j, S ij is the transmission capacity of the line.

[0187] To facilitate those skilled in the art to further understand the automatic matching and control method for power system source-load balance proposed in the embodiment of the present application, the following is combined with Figure 4 and Fig.14 For further explanation.

[0188] like Figure 4 As shown, the automatic matching and control method for source-load balance of the power system may include the following steps:

[0189] Step S401, obtaining a large amount of historical scene operation data.

[0190] Step S402, generating a historical scene database based on a large amount of historical scene operation data (scene vectors).

[0191] Step S403, forming five typical scenes based on the scene quantitative characterization index classification.

[0192] Step S404, based on the source-load balance control models of different typical scenarios and a large amount of scenario operation data, a traditional multi-objective particle swarm algorithm is used to optimize and solve to build a source-load balance control strategy library.

[0193] Step S405, training the DNN model using the historical scenario operation data and the source-load balance control strategy as input and output.

[0194] Step S406, inputting real-time scene operation data into the trained DNN model.

[0195] Step S407, calculate the matching accuracy of the real-time scene, and determine whether the matching accuracy is greater than 90%. If so, execute step S408, otherwise execute step S402 to re-solve and calculate, and update the scene and the calculated control strategy to the historical scene library and strategy library to participate in the next DNN training. The DNN training interval can be three months, while maintaining the accuracy of the model, avoiding the computational burden caused by frequent retraining.

[0196] Step S408: obtaining a corresponding control strategy from a control strategy library.

[0197] The following is a detailed explanation based on specific application examples.

[0198] The example in this application is based on the improved IEEE-30 node system (including 30 nodes), such as Figure 5 As shown in the figure, PV (photovoltaic) is connected to nodes 1, 17, and 21 as a power fluctuation source, air conditioning load is connected to nodes 10, 11, and 18, electric vehicle load is connected to nodes 4 and 16, and distributed battery energy storage load and 7 ordinary loads are connected to nodes 3 and 13. The two-year operation data of a distribution network system is extracted with a time section of 15 minutes. After screening the quantitative indicators of typical scenarios, a historical scenario library consisting of 5086 historical typical scenarios is formed, which are classified into five typical scenarios. Combined with the optimization objective function and control model of each scenario, a source-load balance control strategy library (such as Figure 3 shown), Figure 3 In the above formula, the 28-dimensional scene operation vector P is calculated by formula (14).

[0199] Taking 15 minutes as the step length, we take 12 real-time scene points to be matched at each moment, and the classification results are shown in Table 2:

[0200] Table 2

[0201]

[0202]

[0203] In Table 2, the scene classification labels 1-5 represent renewable energy consumption scenes, peak shaving scenes, valley filling scenes, frequency fluctuation scenes, and voltage fluctuation scenes. After the 12 real-time scenes were classified, real-time online "scenario-control strategy" matching was performed based on DNN on the python3.9 platform. 5086 historical scenes were used as training data, and the ratio of training set to validation set was 8:2. The hyperparameter settings of DNN training can be shown in Table 3.

[0204] Table 3

[0205] Hyperparameters Value Hyperparameters Value Number of hidden units 1000 num_classes 5086 Learning Rate 1e-3 num_nue_hidden1 512 Number of training iterations 600 num_nue_hidden2 256 batch_size 15 num_nue_hidden3 128 input_size 24

[0206] Figure 6 The figure shows the change of the cross entropy loss function of the model in 600 training rounds. It can be seen from the figure that with the increase of training rounds, the loss function value gradually decreases and tends to be stable, which indicates that the training process of the model is effective, the loss function gradually converges, and the fitting effect of the model gradually improves.

[0207] The embodiments of the present application have achieved some positive effects during actual use and have obvious advantages over the prior art. The following will describe them in conjunction with data, charts, etc. from the test process.

[0208] The matching diagram of 12 real-time optimization scenarios is as follows: Figure 7 As shown, the results are Figure 8 (Scenarios 1-6 are shown) and Fig. 9 (Showing scenes 7-12) as shown by Figure 8 and Fig. 9 It can be seen that the scene vectors P of the real-time scene and the matched scene are roughly the same, so the control strategy of the matched scene can be used as the control strategy of the real-time scene. The matching accuracy evaluation index A is calculated and the control strategy of the matched scene is extracted as shown in Table 4.

[0209] Table 4

[0210]

[0211]

[0212] The comparison between the matching time and the traditional optimization time is shown in Table 5:

[0213] Table 5

[0214] Real-time scenario Matching time / s Traditional optimization time / s 1 0.31 72.45 2 0.36 75.21 3 0.35 78.76 4 0.47 81.98 5 0.44 82.44 6 0.31 81.65 7 0.48 76.69 8 0.44 71.76 9 0.38 72.31 10 0.34 73.66 11 0.41 73.17 12 0.35 69.26

[0215] In Table 5, "traditional optimization time" refers to the calculation time of using the multi-objective particle swarm algorithm to calculate the control strategy through the control model, including model initialization, iterative solution and other processes. In contrast, "matching time" refers to the time required to directly infer the control strategy based on the DNN model. The matching time is much shorter than the traditional optimization time because the DNN-based matching method learns the optimization results in advance, and only needs to infer in the real-time stage, which greatly shortens the generation time of the control strategy.

[0216] It can be seen from Tables 4 and 5 that, except for scenes 2 and 9, other real-time scenes can be matched to corresponding scenes and strategies in the historical scene library with a high matching accuracy. For scenes 2 and 9 to be matched, it is necessary to combine the source-load balance control model of typical scenes, solve them in real time through the multi-objective particle swarm algorithm, and update the scene and the calculated control strategy to the preset historical scene-source-load balance control strategy mapping relationship library.

[0217] In summary, it can be seen that the automatic matching and control method for source-load balance of the power system proposed in the embodiment of the present application has high matching accuracy and greatly reduces the strategy generation time compared with the traditional optimization method, providing an effective way for flexible load resources to participate in the realization of rapid response of the system.

[0218] After applying the control scheme obtained by matching in Table 4, the results of the quantitative indicators corresponding to each scenario are calculated and compared with the original indicator values. The results are as follows:

[0219] Depend on Figure 10-13 It can be seen that the automatic matching and control method for source-load balance of the power system proposed in the embodiment of the present application can improve the new energy and consumption rate in the renewable energy consumption scenario, reduce the load in the peak load shaving scenario, increase the load in the valley filling scenario, and reduce the frequency fluctuation in the frequency fluctuation scenario, and the control results are highly consistent with the control results of the traditional method.

[0220] Since only scenario 4 is a voltage fluctuation scenario among the 12 real-time scenarios, the voltage values ​​of each node before and after regulation in scenario 4 are obtained as follows: Fig.14 As shown. Fig.14 It can be seen that the automatic matching and control method for source-load balance of the power system proposed in the embodiment of the present application can significantly reduce the voltage deviation in the voltage fluctuation scenario, and is also highly consistent with the control results of the traditional method.

[0221] In summary, the automatic matching and control method for source-load balance of the power system proposed in the embodiment of the present application has the following beneficial effects:

[0222] (1) This application divides typical scenarios and constructs control models with different objective functions under different typical scenarios, so that flexible load resources can provide source-load balance control strategies that meet the control requirements of typical scenarios based on the current scenario operation data; through deep neural networks (DNNs), the current operation scenario and the control strategy can be automatically matched, and a high matching accuracy can be achieved, which greatly improves the control strategy generation speed while achieving similar control effects as traditional methods; by introducing a scenario library update mechanism, the historical scenario library and strategy library can be dynamically optimized and expanded when new operation data and scenario information are continuously pouring in, thereby greatly improving the power system's rapid control response capability and accuracy under the condition of source-load strength uncertainty.

[0223] (2) The flexible load "scenario-control scheme" automatic matching method proposed in this application can provide a real-time flexible load source-load balance control strategy that meets the control requirements of typical scenarios based on different typical scenarios. Traditional methods (for example, table lookup method) cannot handle a large number of complex scenarios and cannot meet the system's high requirements for real-time performance. The method of this application greatly shortens the control strategy generation time while being highly consistent with the control effect of traditional methods.

[0224] According to the automatic matching and control method for source-load balance of the power system proposed in the embodiment of the present application, by obtaining the current scenario operation data of the power system and inputting the current scenario operation data into a preset neural network model, the automatic matching and control strategy for source-load balance of the current scenario can be obtained; based on the automatic matching and control strategy for source-load balance, the source-load balance process of the current scenario is automatically matched and controlled. Thus, by using flexible load resources based on deep neural networks to participate in the automatic matching and control of source-load balance of the power system, the problem of strong source-load volatility and lack of flexible control resources when a high proportion of fluctuating renewable energy sources are connected to the power system is solved, and an effective way is provided for the realization of flexible load resources participating in the rapid response of the power system.

[0225] Next, the automatic matching and control device for source-load balance of the power system proposed in the embodiment of the present application is described with reference to the accompanying drawings.

[0226] Fig.15 It is a block diagram of an automatic matching and control device for source-load balance of an electric power system according to an embodiment of the present application.

[0227] like Fig.15 As shown, the automatic matching and control device 10 for source-load balance of the power system includes: an acquisition module 100 , an acquisition module 200 and a control module 300 .

[0228] Wherein, the acquisition module 100 is used to acquire the current scenario operation data of the power system;

[0229] An acquisition module 200 is used to input the current scene operation data into a preset neural network model to obtain the source-load balance automatic matching control strategy of the current scene, wherein the preset neural network model is obtained by training the historical scene operation data set and a preset historical scene-source-load balance control strategy mapping relationship library;

[0230] The control module 300 is used to automatically match and control the source-load balancing process of the current scene based on the source-load balancing automatic matching control strategy.

[0231] Optionally, in some embodiments, before inputting the current scenario operation data into a preset neural network model to obtain the source-load balance automatic matching control strategy when the current scenario is running, obtaining module 200 further includes:

[0232] An acquisition unit, used for acquiring a historical scenario operation data set of a power system;

[0233] A preprocessing unit, used for preprocessing the historical scenario operation data set, and dividing the preprocessed historical scenario operation data set into a training set and a verification set;

[0234] The training unit is used to train the preset neural network based on the preset historical scenario-source-load balance control strategy mapping relationship library using the training set to obtain the initial neural network model, and to verify the initial neural network model using the verification set until the initial neural network model meets the preset standard, and then terminate the iterative training of the preset neural network to obtain the preset neural network model, otherwise, continue the iterative training after adjusting the training parameters of the preset neural network.

[0235] Optionally, in some embodiments, before training a preset neural network using a training set to obtain an initial neural network model based on a preset historical scenario-source-load balance control strategy mapping relationship library, the training unit further includes:

[0236] The first acquisition subunit is used to obtain the source-load balance matching regulation demand;

[0237] Determine a subunit for dividing the source-load imbalance scenario in the power system into multiple typical scenarios based on the source-load balance matching regulation demand, and determine the quantitative characterization index corresponding to each typical scenario;

[0238] The second acquisition subunit is used to obtain the control requirements corresponding to each typical scenario;

[0239] Constructing a subunit for constructing a source-load balance control model corresponding to each typical scenario based on the quantitative characterization index corresponding to each typical scenario and the control demand corresponding to each typical scenario;

[0240] A subunit is established to run the data set based on historical scenarios and the source-load balance control model corresponding to each typical scenario, and to establish a preset historical scenario-source-load balance control strategy mapping relationship library.

[0241] Optionally, in some embodiments, the typical scenario includes at least one of a renewable energy consumption scenario, a peak shaving scenario, a valley filling scenario, a frequency fluctuation scenario, and a voltage fluctuation scenario.

[0242] Optionally, in some embodiments, a subunit is constructed, specifically for:

[0243] Based on the quantitative characterization indicators corresponding to each typical scenario and the control requirements corresponding to each typical scenario, determine the flexible load resource optimization control target corresponding to each typical scenario;

[0244] According to the flexible load resource optimization control target corresponding to each typical scenario, a source-load balance control model corresponding to each typical scenario is constructed.

[0245] Optionally, in some embodiments, a subunit is established, specifically for:

[0246] Determine the operation scenario vector corresponding to each moment in the historical scenario operation data set;

[0247] Based on the source-load balance control model corresponding to each typical scenario, the multi-objective particle swarm algorithm is used to obtain the source-load balance control strategies for multiple historical scenarios;

[0248] Based on the source-load balance control strategies of multiple historical scenarios, a preset historical scenario-source-load balance control strategy mapping relationship library is established.

[0249] It should be noted that the aforementioned explanation of the embodiment of the method for automatic matching and control of source-load balance in an electric power system is also applicable to the automatic matching and control device for automatic matching and control of source-load balance in an electric power system of this embodiment, and will not be repeated here.

[0250] According to the automatic matching and control device for source-load balance of the power system proposed in the embodiment of the present application, by acquiring the current scenario operation data of the power system and inputting the current scenario operation data into a preset neural network model, the automatic matching and control strategy for source-load balance of the current scenario can be obtained; based on the automatic matching and control strategy for source-load balance, the source-load balance process of the current scenario is automatically matched and controlled. Thus, by using the flexible load resources based on deep neural networks to participate in the automatic matching and control of source-load balance of the power system, the problem of strong source-load volatility and lack of flexible control resources when a high proportion of fluctuating renewable energy sources are connected to the power system is solved, and an effective way is provided for the realization of flexible load resources participating in the rapid response of the power system.

[0251] Fig.16 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include:

[0252] A memory 1601 , a processor 1602 , and a computer program stored in the memory 1601 and executable on the processor 1602 .

[0253] When the processor 1602 executes the program, the automatic matching and control method for source-load balance of the power system provided in the above embodiment is implemented.

[0254] Furthermore, the electronic device further comprises:

[0255] The communication interface 1603 is used for communication between the memory 1601 and the processor 1602 .

[0256] The memory 1601 is used to store computer programs that can be executed on the processor 1602 .

[0257] The memory 1601 may include a high-speed RAM (Random Access Memory) memory, and may also include a non-volatile memory, such as at least one disk memory.

[0258] If the memory 1601, the processor 1602 and the communication interface 1603 are implemented independently, the communication interface 1603, the memory 1601 and the processor 1602 can be connected to each other through a bus and communicate with each other. The bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Fig.16 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0259] Optionally, in a specific implementation, if the memory 1601, the processor 1602 and the communication interface 1603 are integrated on a chip, the memory 1601, the processor 1602 and the communication interface 1603 can communicate with each other through an internal interface.

[0260] The processor 1602 may be a CPU (Central Processing Unit), or an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present application.

[0261] An embodiment of the present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned automatic matching and control method for source-load balance of the power system.

[0262] An embodiment of the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, it implements the above-mentioned automatic matching and control method for source-load balance of the power system.

[0263] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of this application, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0264] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms are not necessarily directed to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0265] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. An automatic matching and control method for source-load balance of a power system, characterized in that: The following steps are involved: Obtain current scenario operation data of the power system; Inputting the current scenario operation data into a preset neural network model to obtain a source-load balance automatic matching control strategy for the current scenario, wherein the preset neural network model is obtained by training a historical scenario operation data set and a preset historical scenario-source-load balance control strategy mapping relationship library; Based on the source-load balance automatic matching and regulation strategy, the source-load balance process of the current scene is automatically matched and regulated.

2. The method according to claim 1, characterized in that Before inputting the current scenario operation data into the preset neural network model to obtain the source-load balance automatic matching control strategy when the current scenario is running, it also includes: Obtaining a historical scenario operation data set of the power system; Preprocessing the historical scenario operation data set, and dividing the preprocessed historical scenario operation data set into a training set and a verification set; Based on the preset historical scenario-source-load balance control strategy mapping relationship library, the preset neural network is trained using the training set to obtain an initial neural network model, and the initial neural network model is verified using the verification set until the initial neural network model meets the preset standard, and the iterative training of the preset neural network is terminated to obtain the preset neural network model, otherwise the training parameters of the preset neural network are adjusted and the iterative training continues.

3. The method according to claim 2, characterized in that Before the preset neural network is trained by using the training set based on the preset historical scenario-source-load balance control strategy mapping relationship library to obtain the initial neural network model, the method further includes: Obtain source-load balance to match regulation needs; Based on the source-load balance matching regulation demand, the source-load imbalance scenario in the power system is divided into multiple typical scenarios, and the quantitative characterization index corresponding to each typical scenario is determined; Obtaining the control requirements corresponding to each typical scenario; Based on the quantitative characterization indicators corresponding to each typical scenario and the control requirements corresponding to each typical scenario, a source-load balance control model corresponding to each typical scenario is constructed; Based on the historical scenario operation data set and the source-load balance control model corresponding to each typical scenario, the preset historical scenario-source-load balance control strategy mapping relationship library is established.

4. The method according to claim 3, characterized in that The typical scenarios include at least one of renewable energy consumption scenarios, peak shaving scenarios, valley filling scenarios, frequency fluctuation scenarios and voltage fluctuation scenarios.

5. The method according to claim 3, characterized in that: The constructing of a source-load balance control model corresponding to each typical scenario based on the quantitative characterization index corresponding to each typical scenario and the control demand corresponding to each typical scenario includes: Determine the flexible load resource optimization control target corresponding to each typical scenario based on the quantitative characterization index corresponding to each typical scenario and the control demand corresponding to each typical scenario; According to the flexible load resource optimization control target corresponding to each typical scenario, a source-load balance control model corresponding to each typical scenario is constructed.

6. The method according to claim 3, characterized in that The establishing of the preset historical scenario-source-load balance control strategy mapping relationship library based on the historical scenario operation data and the source-load balance control model corresponding to each typical scenario includes: Determine the operation scenario vector corresponding to each moment in the historical scenario operation data set; Based on the source-load balance control model corresponding to each typical scenario, a multi-objective particle swarm algorithm is used to obtain the source-load balance control strategies for multiple historical scenarios; Based on the source-load balance control strategies of the multiple historical scenarios, a preset historical scenario-source-load balance control strategy mapping relationship library is established.

7. An automatic matching and control device for source-load balance of a power system, characterized in that: include: An acquisition module is used to obtain the current scenario operation data of the power system; An acquisition module is used to input the current scene operation data into a preset neural network model to obtain a source-load balance automatic matching control strategy for the current scene, wherein the preset neural network model is obtained by training a historical scene operation data set and a preset historical scene-source-load balance control strategy mapping relationship library; The control module is used to automatically match and control the source-load balancing process of the current scene based on the source-load balancing automatic matching control strategy.

8. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the automatic matching and control method for source-load balance of an electric power system as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the automatic matching and control method for source-load balance of a power system as described in any one of claims 1 to 6.

10. A computer program product, characterized in that It comprises a computer program, which, when executed by a processor, is used to implement the automatic matching and control method for source-load balance of the power system as described in any one of claims 1-6.

Images