Static voltage stability prevention and control method and device for power system, terminal equipment and storage medium

Abstract

The invention discloses an electric power system static voltage stability prevention and control method and device, terminal equipment and a storage medium, and belongs to the technical field of electric power system stability control, and the method comprises the steps: obtaining the current generator active output, terminal voltage, reactive compensation, line load, node voltage and node power of an electric power system; inputting the active output, the terminal voltage, the reactive compensation and the line load of the generator into a preset load margin prediction model constructed based on a convex neural network to obtain a first load margin prediction value; inputting the first load margin predicted value, the node voltage and the node power into a preset voltage stability control model to obtain a current control decision parameter of the power system; and finally performing voltage stability control on the power system according to the control decision parameter. By implementing the method, the problems of complex calculation and low solving efficiency of a method for formulating a control strategy by iteratively solving non-convex nonlinear constraints in the prior art can be solved.

Description

Technical Field

[0001] This invention relates to the field of power system stability control technology, and in particular to a method, device, terminal equipment and storage medium for preventing and controlling static voltage stability in power systems. Background Technology

[0002] With the high penetration rate of new energy sources in the new power system and the increase of uncertain loads, the output of new energy power generation is difficult to predict accurately and fluctuates greatly due to its inherent intermittency, volatility and randomness (such as wind power depending on wind speed changes and photovoltaic power being constrained by sunlight conditions). The spatiotemporal distribution of uncertain loads and their random start-up and shutdown behavior further exacerbate the dynamic fluctuations in the power supply and demand balance of the system, causing the power system to operate closer to its safe and stable limits and significantly increasing the risk of static voltage instability. Therefore, it is particularly important to assess the voltage stability margin of the power system and formulate control strategies.

[0003] In existing technologies, traditional model-driven preventive control methods are commonly used for stability control, such as the Continuous Power Flow (CPF) method and nonlinear programming. These methods mainly rely on accurate physical modeling and iteratively solve non-convex nonlinear constraints to evaluate voltage stability margin and formulate control strategies. However, because the power flow equations and constraints they construct are themselves non-convex and highly nonlinear, and because the voltage magnitude, phase angle, and power of each node are strongly coupled, they suffer from computational complexity and low solution efficiency. Summary of the Invention

[0004] This invention provides a method, device, terminal equipment, and storage medium for preventing and controlling static voltage stability in power systems. It can solve the problems of computational complexity and low solution efficiency in existing technologies that rely on accurate physical modeling and iteratively solving non-convex nonlinear constraints to formulate control strategies.

[0005] An embodiment of the present invention provides a method for preventing and controlling static voltage stability in a power system, comprising: Obtain current power data of the power system; wherein, the above power data includes: generator active power output, generator terminal voltage, reactive power compensation, line load, node voltage and node power; The active power output of the generator, the terminal voltage, the reactive power compensation and the line load are input into the preset load margin prediction model based on the convex neural network, so that the preset load margin prediction model can predict the first load margin prediction value of the power system. The first load margin prediction value, node voltage and node power are input into the preset voltage stability control model so that the encoder built into the preset voltage stability control model maps the first load margin prediction value, node voltage and node power to obtain the mapping result. The mapping result is then input into the decoder built into the preset voltage stability control model to obtain the current control decision parameters of the power system. Voltage stability control of the power system is performed based on the above control decision parameters.

[0006] Furthermore, the training of the aforementioned pre-defined load margin prediction model includes: Acquire several sets of power sample data with a first true label; wherein, the power sample data includes: generator active power output sample data, generator terminal voltage sample data, reactive power compensation sample data, and line load sample data; the first true label is used to represent the actual load margin of each set of power sample data; Several sets of power sample data are input into the load margin prediction model to be trained for iterative training until the first loss function converges, and the trained preset load margin prediction model is obtained. In each iteration of training, the current load margin prediction sample value is obtained based on the current group of power sample data; the current first loss function is calculated based on the current load margin prediction sample value and the corresponding first true label; it is determined whether the current first loss function has converged. If it has converged, the current load margin prediction model is used as the preset load margin prediction model; otherwise, the model parameters of the current load margin prediction model are adjusted and training continues.

[0007] Furthermore, the above-mentioned acquisition of several sets of electricity sample data with a first real label includes: Based on the preset value range of each power sample data, determine the limit value of each power sample data; For each power sample data, the width of the sub-interval used for interval division is calculated based on the limit value of the current power sample data; Based on the aforementioned width, the corresponding preset value range is divided into several sub-intervals; Uniform sampling is performed across all sub-intervals to obtain several sets of power sample data; For each set of power sample data, power flow calculation is performed to obtain the corresponding actual load margin, and the above actual load margin is used as the first true label of the corresponding set of power sample data.

[0008] Furthermore, the above adjustments to the model parameters of the current load margin forecasting model include: Obtain the current model parameters of the current load margin prediction model, as well as the current value of the first loss function; Based on the current first loss function value, the current model parameters of the current load margin prediction model are updated by projecting gradients to obtain the adjusted model parameters of the load margin prediction model.

[0009] Furthermore, the training of the aforementioned preset voltage stability control model includes: Acquire several sets of historical power data; among which, the aforementioned historical power data includes: historical generator active power output, historical generator terminal voltage, historical reactive power compensation, historical line load, historical node voltage, and historical node power; For each set of historical power data, the historical generator active power output, historical generator terminal voltage, historical reactive power compensation, and historical line load in the current set of historical power data are input into the preset load margin prediction model to obtain the second load margin prediction value corresponding to the current set of historical power data. Based on all the predicted values ​​of the second load margin, historical node voltages and historical node powers, several sets of training sample data are calculated for training the preset voltage stability control model; each set of training sample data includes: load margin samples, node voltage sample data, node power sample data and actual optimal control decision parameters. All training sample data are input into the voltage stability control model to be trained for iterative training until the second loss function converges, and the trained preset voltage stability control model is obtained. In each iteration, the current predicted optimal control decision parameters are obtained based on the current load margin sample, node voltage sample data, and node power sample data. The current second loss function is calculated based on the current predicted optimal control decision parameters and the corresponding actual optimal control decision parameters. It is then determined whether the current second loss function has converged. If it has converged, the current voltage stability control model is used as the pre-trained preset voltage stability control model. Otherwise, the model parameters in the current voltage stability control model are adjusted, and training continues.

[0010] Furthermore, based on all the predicted values ​​of the second load margin, historical node voltages, and historical node power, several sets of training sample data are calculated for training the preset voltage stability control model, including: Obtain the single switching cost of the reactive power compensation equipment and the base power of the generator. Based on the preset minimum load margin threshold, preset generator active power output cost coefficient, single switching cost and benchmark power, a control strategy optimization model and corresponding constraints are constructed with the goal of minimizing operating costs. Based on all the predicted values ​​of the second load margin, historical node voltage and historical node power, under the above constraints, the control strategy optimization model is solved to obtain several optimal control decision parameters, as well as the predicted values ​​of the second load margin, historical node voltage and historical node power corresponding to each optimal control decision parameter. Each optimal control decision parameter is used as the actual optimal control decision parameter; the second load margin prediction value corresponding to each optimal control decision parameter is used as the corresponding load margin sample; the historical node voltage corresponding to each optimal control decision parameter is used as the corresponding node voltage sample data; and the historical node power corresponding to each optimal control decision parameter is used as the corresponding node power sample data, thereby obtaining several sets of training sample data.

[0011] Furthermore, the aforementioned adjustments to the model parameters in the current voltage stability control model include: Obtain the current model parameters of the current voltage stability control model, as well as the current value of the second loss function; Based on the current value of the second loss function, backpropagation optimization is performed on the current model parameters of the current voltage stability control model to obtain the adjusted model parameters of the voltage stability control model.

[0012] Based on the above method embodiments, the present invention provides corresponding apparatus embodiments; This invention provides a power system static voltage stability prevention and control device, comprising: The module includes a data acquisition module, a load margin prediction module, a control decision parameter output module, and a voltage stability control module. The aforementioned data acquisition module is used to acquire the current power data of the power system; wherein, the aforementioned power data includes: generator active power output, generator terminal voltage, reactive power compensation, line load, node voltage and node power; The aforementioned load margin prediction module is used to input the active power output of the generator, the terminal voltage, the reactive power compensation, and the line load into a preset load margin prediction model constructed based on a convex neural network, so that the preset load margin prediction model can predict the first load margin prediction value of the power system. The aforementioned control decision parameter output module is used to input the aforementioned first load margin prediction value, node voltage and node power into the preset voltage stability control model, so that the encoder built into the preset voltage stability control model maps the aforementioned first load margin prediction value, node voltage and node power to obtain the mapping result, and inputs the mapping result into the decoder built into the preset voltage stability control model to obtain the current control decision parameters of the power system. The voltage stability control module described above is used to control the voltage stability of the power system based on the aforementioned control decision parameters.

[0013] Based on the above method embodiments, the present invention provides a corresponding terminal device embodiment; The present invention provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the power system static voltage stability prevention and control method described in any embodiment of the present invention.

[0014] Based on the above method embodiments, the present invention provides a corresponding storage medium embodiment; The present invention provides a storage medium including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the power system static voltage stability prevention and control method described in any embodiment of the present invention.

[0015] The embodiments of the present invention have the following beneficial effects: This invention provides a method, apparatus, terminal equipment, and storage medium for preventing and controlling static voltage stability in a power system. The method includes: acquiring the current generator active power output, terminal voltage, reactive power compensation, line load, node voltage, and node power of the power system; then inputting the generator active power output, terminal voltage, reactive power compensation, and line load into a preset load margin prediction model constructed based on a convex neural network, so that the preset load margin prediction model predicts a first load margin prediction value for the power system; then inputting the first load margin prediction value, node voltage, and node power into a preset voltage stability control model, so that the encoder built into the preset voltage stability control model maps the first load margin prediction value, node voltage, and node power to obtain a mapping result, and inputting the mapping result into the decoder built into the preset voltage stability control model to obtain the current control decision parameters of the power system; finally, performing voltage stability control on the power system based on the control decision parameters. Therefore, in this invention, a pre-defined load margin prediction model based on a convex neural network is constructed to achieve a direct mapping from input data to load margin, eliminating non-convexity at the source and replacing the iteration of non-convex nonlinear equations in traditional solution methods. This allows for faster solution of load margin. Furthermore, the pre-defined voltage stability control model only requires one forward propagation to obtain control decisions, eliminating the need for iteratively solving the constraint optimization problem for each operating scenario as in nonlinear programming, thus greatly improving computational efficiency. Attached Figure Description

[0016] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating a static voltage stability prevention and control method for a power system according to an embodiment of the present invention.

[0018] Figure 2 This is a schematic diagram of the structure of a power system static voltage stability prevention and control device according to an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0021] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0022] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0023] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0024] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).

[0025] In the description of the embodiments of this application, unless otherwise expressly specified and limited, technical terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. For those skilled in the art, the specific meaning of the above terms in the embodiments of this application can be understood according to the specific circumstances.

[0026] See Figure 1 To address the shortcomings of existing technologies that rely on precise physical modeling and iteratively solving non-convex nonlinear constraints to formulate control strategies, which suffer from computational complexity and low efficiency, this invention provides a method for preventing and controlling static voltage stability in power systems, comprising: Step S101: Obtain the current power data of the power system; wherein, the power data includes: generator active power output, generator terminal voltage, reactive power compensation, line load, node voltage and node power; Specifically, the aforementioned power data includes all relevant power data from generators, reactive power compensation equipment, lines, and nodes involved in the power system. This data can be obtained from the relevant monitoring equipment deployed within the power system. The node power figures include both active and reactive power.

[0027] Step S102: Input the above generator active power output, terminal voltage, reactive power compensation and line load into the preset load margin prediction model based on convex neural network, so that the preset load margin prediction model can predict the first load margin prediction value of the above power system. Specifically, the aforementioned pre-defined load margin prediction model is a convex surrogate model built on an Intercontinental Neural Network (ICNN), which contains several hidden layers. From input to output, the following inter-layer relationship formula is used: In the formula, This represents the output vector of the l-th layer. This represents the weight matrix of the l-th layer. This represents the output vector of the (l-1)th layer. This represents the weight matrix of the (l-1)th layer, which directly connects the input vector X to the lth layer, where X represents the input vector of the preset load margin prediction model. This represents the bias vector of the l-th layer. denoted by , where k' represents the activation function and k' represents the number of hidden layers.

[0028] Specifically, the input to the aforementioned pre-defined load margin prediction model can be expressed as: In the formula, x represents the input vector of the preset load margin prediction model. This indicates the generator's active power output. Indicates the terminal voltage. Indicates reactive power compensation. This indicates the line load.

[0029] In a preferred embodiment, training the aforementioned preset load margin prediction model includes: Acquire several sets of power sample data with a first true label; wherein, the power sample data includes: generator active power output sample data, generator terminal voltage sample data, reactive power compensation sample data, and line load sample data; the first true label is used to represent the actual load margin of each set of power sample data; Several sets of power sample data are input into the load margin prediction model to be trained for iterative training until the first loss function converges, and the trained preset load margin prediction model is obtained. In each iteration of training, the current load margin prediction sample value is obtained based on the current group of power sample data; the current first loss function is calculated based on the current load margin prediction sample value and the corresponding first true label; it is determined whether the current first loss function has converged. If it has converged, the current load margin prediction model is used as the preset load margin prediction model; otherwise, the model parameters of the current load margin prediction model are adjusted and training continues.

[0030] Specifically, the formula for calculating the first loss function is: In the formula, Let N represent the first loss function, and N represent the number of sets of electricity sample data. This represents the load margin prediction sample value corresponding to the j'th group of power sample data. This represents the actual load margin corresponding to the j'th group of power sample data. Represents the regularization parameter. This represents the L2 norm of the weight matrix W.

[0031] Preferably, the preset load margin prediction model based on the input convex neural network realizes the convex parameterization expression of the nonlinear voltage stability boundary, avoids the convergence problem of traditional power flow iterative calculation, and significantly improves the computational efficiency and real-time performance of the model.

[0032] In this preferred embodiment, the load margin prediction model was iteratively trained based on several sets of power sample data with first real labels.

[0033] In another preferred embodiment, the acquisition of several sets of power sample data with a first real label includes: Based on the preset value range of each power sample data, determine the limit value of each power sample data; Specifically, the preset value ranges for each power sample data are determined based on the technical parameters and operating limits of the power grid equipment. These parameters are usually derived from the equipment manufacturer's technical manual or power grid operation specifications.

[0034] Preferably, the preset value range of the generator active power output sample data is set between the minimum and maximum active power output of the corresponding generator; the preset value range of the generator terminal voltage sample data, referring to the grid voltage operation standard, can be set to [0.94, 1.06] pu to ensure that it is within the normal operating range. The preset value range for reactive power compensation sample data can be set to [0, 10] Mvar, depending on the capacity of the parallel capacitor or static var compensator. The specific upper limit is determined by the equipment configuration. The preset value range for line load sample data can be set based on typical load levels. This reflects ±30% uncertainty in load changes, among which... This indicates the typical load level of the line.

[0035] For each power sample data, the width of the sub-interval used for interval division is calculated based on the limit value of the current power sample data; Specifically, assuming a total of N sets of power sample data are needed, the width of its sub-interval can be calculated using the following formula: In the formula, This represents the lower limit value of the i'th power sample data. This represents the upper limit value of the i'th electricity sample data. This represents the width of the sub-interval of the i'th power sample data.

[0036] Based on the aforementioned width, the corresponding preset value range is divided into several sub-intervals; Specifically, by dividing the preset value range according to these widths, N equally spaced sub-intervals can be obtained.

[0037] Uniform sampling is performed across all sub-intervals to obtain several sets of power sample data; Specifically, a randomly arranged index is generated for each power sample data to ensure that each sub-interval is sampled uniformly. Then, a sample point is randomly selected from each sub-interval using the midpoint adjustment formula. Each iteration of the sampling yields a set of power sample data. By repeating the sampling process, several sets of power sample data can be obtained. The midpoint adjustment formula is as follows: In the formula, This represents the j'-th sample point extracted from the i'-th electricity sample data. This represents the normalized sampling factor randomly generated for the j'th sample point within its corresponding sub-interval, with a value range of [0,1].

[0038] Preferred, This allows the extracted sample points to be centered within the corresponding sub-interval.

[0039] For each set of power sample data, power flow calculation is performed to obtain the corresponding actual load margin, and the above actual load margin is used as the first true label of the corresponding set of power sample data.

[0040] Specifically, for each set of power sample data, its actual load margin can be calculated using the continuous power flow method. This is existing technology and will not be elaborated here.

[0041] In this preferred embodiment, after dividing the preset value range of each power sample data, uniform sampling is performed on each sub-interval to obtain several sets of power sample data. Then, power flow calculation is performed to determine the first true label of these sample data.

[0042] In another preferred embodiment, the above-mentioned adjustment of the model parameters of the current load margin forecasting model includes: Obtain the current model parameters of the current load margin prediction model, as well as the current value of the first loss function; Based on the current first loss function value, the current model parameters of the current load margin prediction model are updated by projecting gradients to obtain the adjusted model parameters of the load margin prediction model.

[0043] Specifically, to optimize the model parameters of the load margin prediction model, the projected gradient descent method is used for updating and optimization, as shown in the following formula: In the formula, This represents the model parameters of the load margin prediction model at the (t+1)th iteration, including all weight matrices. and bias vector , This represents the projection operator, which projects negative values ​​to zero to ensure that the weights and "through-layer" weights are non-negative, thus maintaining the convexity of the model. This represents the model parameters of the load margin prediction model at the t-th iteration. This represents the learning rate, typically set to 0.01.

[0044] In this preferred embodiment, the model parameters are updated and adjusted by using the projected gradient descent method, combined with the current first loss function value of the load margin prediction model.

[0045] Step S103: Input the first load margin prediction value, node voltage and node power into the preset voltage stability control model so that the encoder built into the preset voltage stability control model maps the first load margin prediction value, node voltage and node power to obtain the mapping result, and input the mapping result into the decoder built into the preset voltage stability control model to obtain the current control decision parameters of the power system. Specifically, the preset voltage stability control model is a Physically Guided Neural Network (PGNN) employing a quasi-autoencoder structure with an encoder and decoder. This model's control framework is a key component in connecting the ICNN proxy model, primarily addressing the issue that while ICNN is efficient, its physical constraints are not stringent enough. By introducing PGNN to extend the ICNN proxy, an end-to-end parameterized preventative control framework is constructed, enabling online calculation of static voltage stability.

[0046] Specifically, the encoder maps the input first load margin prediction, node voltage, and node power into a latent representation, obtaining the mapping result: In the formula, Z represents the latent representation. This refers to the encoding process of input data under the encoder. This represents the input data consisting of "predicted first load margin, node voltage, and node power". This indicates the encoder.

[0047] Subsequently, the decoder converts the mapping results into final control decision parameters, including generator output adjustment, terminal voltage regulation, and reactive power compensation. The decoder's conversion process is represented by the following equation: In the formula, Indicates control decision parameters, This represents the decoding process of the latent representation under the decoder. This indicates the decoder.

[0048] In a preferred embodiment, the training of the aforementioned preset voltage stability control model includes: Acquire several sets of historical power data; among which, the aforementioned historical power data includes: historical generator active power output, historical generator terminal voltage, historical reactive power compensation, historical line load, historical node voltage, and historical node power; For each set of historical power data, the historical generator active power output, historical generator terminal voltage, historical reactive power compensation, and historical line load in the current set of historical power data are input into the preset load margin prediction model to obtain the second load margin prediction value corresponding to the current set of historical power data. Preferably, the aforementioned historical power data can also be replaced with power data obtained through simulation.

[0049] Based on all the predicted values ​​of the second load margin, historical node voltages and historical node powers, several sets of training sample data are calculated for training the preset voltage stability control model; each set of training sample data includes: load margin samples, node voltage sample data, node power sample data and actual optimal control decision parameters. All training sample data are input into the voltage stability control model to be trained for iterative training until the second loss function converges, and the trained preset voltage stability control model is obtained. In each iteration, the current predicted optimal control decision parameters are obtained based on the current load margin sample, node voltage sample data, and node power sample data. The current second loss function is calculated based on the current predicted optimal control decision parameters and the corresponding actual optimal control decision parameters. It is then determined whether the current second loss function has converged. If it has converged, the current voltage stability control model is used as the pre-trained preset voltage stability control model. Otherwise, the model parameters in the current voltage stability control model are adjusted, and training continues.

[0050] Specifically, to ensure that the model's output conforms to physical laws, a second loss function is designed as a convex physical information loss function. The formula for calculating the second loss function is as follows: In the formula, This represents the second loss function. This represents the weighting coefficient corresponding to the power balance loss. Indicates power balance loss, This represents the weighting coefficient corresponding to the generator constraint loss. Indicates generator constraint loss. This represents the weighting coefficient corresponding to reactive power compensation loss. This indicates compensation for reactive power loss. This represents the weighting coefficient corresponding to the load margin loss. Indicates load margin loss. This represents the weighting coefficient corresponding to the total power imbalance loss. This indicates the total power imbalance loss. This represents the active power injection amount of the node. This represents the reactive power injection amount at the node. Indicates the actual active power of the node. Indicates the actual reactive power of the node. This represents the generator output adjustment of the i-th generator obtained from the voltage stability control model. This represents the terminal voltage regulation of the i-th generator obtained from the voltage stability control model. Let represent the active power of the i-th generator in the actual optimal control decision parameters. Let represent the terminal voltage of the i-th generator in the actual optimal control decision parameters. This represents the reactive power compensation amount of the j-th reactive power compensation device obtained from the voltage stability control model. Let represent the reactive power output of the j-th reactive power compensation device in the actual optimal control decision parameters. This represents the minimum safety margin requirement for static voltage stability. The node voltage sample data is represented as The active power of nodes in the node power sample data is The node reactive power in the node power sample data is Load margin sample, This indicates that the ReLU operation only incurs a penalty when the margin is insufficient. This represents the total active power imbalance in the system. This represents the total reactive power imbalance of the system. This represents the active power demand of the k-th load. This represents the reactive power demand of the k-th load. This indicates the reactive power generated by the generator due to voltage regulation.

[0051] Preferably, the weight coefficients mentioned above can be determined through hyperparameter tuning, and the active power injection and reactive power injection at the nodes can be calculated based on the predicted optimal control decision parameters and power flow equations currently output by the model.

[0052] In this preferred embodiment, training sample data for model training is obtained based on several sets of historical power data, thereby enabling the training of the voltage stability control model.

[0053] In another preferred embodiment, the above-mentioned calculation based on all second load margin predictions, historical node voltages, and historical node powers yields several sets of training sample data for training the preset voltage stability control model, including: Obtain the single switching cost of the reactive power compensation equipment and the base power of the generator. Based on the preset minimum load margin threshold, preset generator active power output cost coefficient, single switching cost and benchmark power, a control strategy optimization model and corresponding constraints are constructed with the goal of minimizing operating costs. Specifically, the objective function of the above control strategy optimization model is: In the formula, F represents the total operating cost. Represents a set of generators. , and This represents the cost coefficient for different active power outputs of the i-th generator. This represents the active power of the i-th generator obtained after the power system executes the control decision parameters. This represents a collection of reactive power compensation devices. Indicates the reference power. This indicates the cost per cut. This represents the reactive power compensation amount obtained by the reactive power compensation device n after the power system executes the control decision parameters.

[0054] It should be noted that after solving the above control strategy optimization model, the optimal value under the current second load margin prediction will be obtained. and The generator terminal voltage under the current second load margin forecast can be determined based on... and The parameters are obtained by combining power flow calculations (this is existing technology and will not be elaborated here). Since these three parameters are the result quantities obtained after executing a set of control decision parameters, the corresponding optimal control decision parameters can be deduced based on this result quantity and the quantity before execution.

[0055] Specifically, the constraints are: In the formula, This represents the second load margin prediction value output by the preset load margin prediction model. This represents the minimum load margin threshold.

[0056] Based on all the predicted values ​​of the second load margin, historical node voltage and historical node power, under the above constraints, the control strategy optimization model is solved to obtain several optimal control decision parameters, as well as the predicted values ​​of the second load margin, historical node voltage and historical node power corresponding to each optimal control decision parameter. Each optimal control decision parameter is used as the actual optimal control decision parameter; the second load margin prediction value corresponding to each optimal control decision parameter is used as the corresponding load margin sample; the historical node voltage corresponding to each optimal control decision parameter is used as the corresponding node voltage sample data; and the historical node power corresponding to each optimal control decision parameter is used as the corresponding node power sample data, thereby obtaining several sets of training sample data.

[0057] Preferably, by using the PGNN end-to-end control framework, neural network modeling of power balance and voltage constraints is achieved, which enhances the generalization ability and physical consistency of the model and ensures that the control results meet the operating rules of the power system.

[0058] In this preferred embodiment, training sample data for training the preset voltage stability control model is obtained by combining all the second load margin predictions, historical node voltages and historical node powers, the control strategy optimization model constructed with the goal of minimizing operating costs, and the corresponding constraints.

[0059] In another preferred embodiment, the above-mentioned adjustment of the model parameters in the current voltage stability control model includes: Obtain the current model parameters of the current voltage stability control model, as well as the current value of the second loss function; Based on the current value of the second loss function, backpropagation optimization is performed on the current model parameters of the current voltage stability control model to obtain the adjusted model parameters of the voltage stability control model.

[0060] Specifically, the model parameters of the voltage stability control model are optimized using backpropagation, and the update process is expressed as follows: In the formula, This represents the model parameters of the voltage stability control model, including the encoder and decoder.

[0061] In this preferred embodiment, the model parameters of the voltage stability control model are optimized using a backpropagation optimization algorithm.

[0062] Step S104: Perform voltage stability control on the power system based on the above control decision parameters.

[0063] Specifically, after obtaining the current control decision parameters of the power system, the power output of the corresponding generator, the terminal voltage of the generator, and the reactive power compensation of the reactive power compensation equipment are adjusted according to the generator output adjustment, generator terminal voltage adjustment, and reactive power compensation, so as to achieve voltage stability control of the power system.

[0064] Based on the above method embodiments, the present invention provides corresponding apparatus embodiments.

[0065] like Figure 2 As shown, an embodiment of the present invention provides a power system static voltage stability prevention and control device, comprising: The module includes a data acquisition module, a load margin prediction module, a control decision parameter output module, and a voltage stability control module. The aforementioned data acquisition module is used to acquire the current power data of the power system; wherein, the aforementioned power data includes: generator active power output, generator terminal voltage, reactive power compensation, line load, node voltage and node power; The aforementioned load margin prediction module is used to input the active power output of the generator, the terminal voltage, the reactive power compensation, and the line load into a preset load margin prediction model constructed based on a convex neural network, so that the preset load margin prediction model can predict the first load margin prediction value of the power system. The aforementioned control decision parameter output module is used to input the aforementioned first load margin prediction value, node voltage and node power into the preset voltage stability control model, so that the encoder built into the preset voltage stability control model maps the aforementioned first load margin prediction value, node voltage and node power to obtain the mapping result, and inputs the mapping result into the decoder built into the preset voltage stability control model to obtain the current control decision parameters of the power system. The voltage stability control module described above is used to control the voltage stability of the power system based on the aforementioned control decision parameters.

[0066] It should be noted that the device embodiments described above are merely illustrative. The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationship between modules indicates that they have a communication connection, which can be implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without creative effort. The above schematic diagrams are merely examples of a power system static voltage stability prevention and control device and do not constitute a limitation on a power system static voltage stability prevention and control device. It may include more or fewer components than shown, or combine certain components, or use different components.

[0067] Based on the above method embodiments, the present invention provides corresponding terminal device embodiments.

[0068] Another embodiment of the present invention provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the power system static voltage stability prevention and control method described in any embodiment of the present invention.

[0069] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the device. The aforementioned terminal devices may be computing devices such as desktop computers, laptops, handheld computers, and cloud servers. These devices may include, but are not limited to, processors and memory. The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. This processor is the control center of the device, connecting various parts of the device via various interfaces and lines. The aforementioned memory can be used to store the aforementioned computer programs and / or modules. The aforementioned processor implements various functions of the aforementioned device by running or executing the computer programs and / or modules stored in the aforementioned memory, and by calling data stored in the memory. The aforementioned memory may mainly include a program storage area and a data storage area, wherein the program storage area may store the operating system, at least one application program required for a function, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0070] Based on the above method embodiments, the present invention provides corresponding storage medium embodiments.

[0071] Another embodiment of the present invention provides a storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the storage medium is located to execute the power system static voltage stability prevention and control method described in any embodiment of the present invention. In this embodiment, the storage medium is a computer-readable storage medium, and the computer program includes computer program code, which may be in the form of source code, object code, executable file, or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0072] The above are preferred embodiments of the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for preventing and controlling static voltage stability in a power system, characterized in that, include: Acquire current power data of the power system; wherein, the power data includes: generator active power output, generator terminal voltage, reactive power compensation, line load, node voltage and node power; The generator's active power output, terminal voltage, reactive power compensation, and line load are input into a preset load margin prediction model built on a convex neural network, so that the preset load margin prediction model can predict the first load margin prediction value of the power system. The first load margin prediction value, node voltage, and node power are input into the preset voltage stability control model, so that the encoder built into the preset voltage stability control model maps the first load margin prediction value, node voltage, and node power to obtain the mapping result. The mapping result is then input into the decoder built into the preset voltage stability control model to obtain the current control decision parameters of the power system. Voltage stability control of the power system is performed based on the control decision parameters.

2. The method for preventing and controlling static voltage stability in a power system according to claim 1, characterized in that, The training of the preset load margin prediction model includes: Acquire several sets of power sample data with a first real label; wherein, the power sample data includes: generator active power output sample data, generator terminal voltage sample data, reactive power compensation sample data, and line load sample data; the first real label is used to represent the actual load margin of each set of power sample data; Several sets of power sample data are input into the load margin prediction model to be trained for iterative training until the first loss function converges, and the trained preset load margin prediction model is obtained. In each iteration of training, the current load margin prediction sample value is obtained based on the current group of power sample data; the current first loss function is calculated based on the current load margin prediction sample value and the corresponding first true label; it is determined whether the current first loss function has converged. If it has converged, the current load margin prediction model is used as the preset load margin prediction model; otherwise, the model parameters of the current load margin prediction model are adjusted and training continues.

3. The method for preventing and controlling static voltage stability in a power system according to claim 2, characterized in that, The acquisition of several sets of power sample data with a first real label includes: Based on the preset value range of each power sample data, determine the limit value of each power sample data; For each power sample data, the width of the sub-interval used for interval division is calculated based on the limit value of the current power sample data; The corresponding preset value range is divided according to the width to obtain several sub-intervals; Uniform sampling is performed across all sub-intervals to obtain several sets of power sample data; For each set of power sample data, power flow calculation is performed to obtain the corresponding actual load margin, and the actual load margin is used as the first true label of the corresponding set of power sample data.

4. The method for preventing and controlling static voltage stability in a power system according to claim 3, characterized in that, The adjustment of the model parameters of the current load margin forecasting model includes: Obtain the current model parameters of the current load margin prediction model, as well as the current value of the first loss function; Based on the current first loss function value, the current model parameters of the current load margin prediction model are updated by projecting gradients to obtain the adjusted model parameters of the load margin prediction model.

5. The method for preventing and controlling static voltage stability in a power system according to claim 4, characterized in that, The training of the preset voltage stability control model includes: Acquire several sets of historical power data; wherein, the historical power data includes: historical generator active power output, historical generator terminal voltage, historical reactive power compensation, historical line load, historical node voltage, and historical node power; For each set of historical power data, the historical generator active power output, historical generator terminal voltage, historical reactive power compensation, and historical line load in the current set of historical power data are input into the preset load margin prediction model to obtain the second load margin prediction value corresponding to the current set of historical power data. Based on all the predicted values ​​of the second load margin, historical node voltages and historical node powers, several sets of training sample data are calculated for training the preset voltage stability control model; each set of training sample data includes: load margin samples, node voltage sample data, node power sample data and actual optimal control decision parameters. All training sample data are input into the voltage stability control model to be trained for iterative training until the second loss function converges, and the trained preset voltage stability control model is obtained. In each iteration, the current predicted optimal control decision parameters are obtained based on the current load margin sample, node voltage sample data, and node power sample data. The current second loss function is calculated based on the current predicted optimal control decision parameters and the corresponding actual optimal control decision parameters. It is then determined whether the current second loss function has converged. If it has converged, the current voltage stability control model is used as the pre-trained preset voltage stability control model. Otherwise, the model parameters in the current voltage stability control model are adjusted, and training continues.

6. The method for preventing and controlling static voltage stability in a power system according to claim 5, characterized in that, The process involves calculating several sets of training sample data for training the preset voltage stability control model based on all predicted second load margin values, historical node voltages, and historical node power. These data include: Obtain the single switching cost of the reactive power compensation equipment and the base power of the generator. Based on the preset minimum load margin threshold, preset generator active power output cost coefficient, single switching cost and benchmark power, a control strategy optimization model and corresponding constraints are constructed with the goal of minimizing operating costs. Based on all the predicted values ​​of the second load margin, historical node voltage and historical node power, under the constraints, the control strategy optimization model is solved to obtain several optimal control decision parameters, as well as the predicted value of the second load margin, historical node voltage and historical node power corresponding to each optimal control decision parameter. Each optimal control decision parameter is taken as the actual optimal control decision parameter; the second load margin prediction value corresponding to each optimal control decision parameter is taken as the corresponding load margin sample; the historical node voltage corresponding to each optimal control decision parameter is taken as the corresponding node voltage sample data; and the historical node power corresponding to each optimal control decision parameter is taken as the corresponding node power sample data, thereby obtaining several sets of training sample data.

7. The method for preventing and controlling static voltage stability in a power system according to claim 6, characterized in that, The adjustment of model parameters in the current voltage stability control model includes: Obtain the current model parameters of the current voltage stability control model, as well as the current value of the second loss function; Based on the current value of the second loss function, backpropagation optimization is performed on the current model parameters of the current voltage stability control model to obtain the adjusted model parameters of the voltage stability control model.

8. A static voltage stability prevention and control device for a power system, characterized in that, include: The module includes a data acquisition module, a load margin prediction module, a control decision parameter output module, and a voltage stability control module. The data acquisition module is used to acquire the current power data of the power system; wherein, the power data includes: generator active power output, generator terminal voltage, reactive power compensation, line load, node voltage and node power; The load margin prediction module is used to input the generator active power output, terminal voltage, reactive power compensation and line load into a preset load margin prediction model based on a convex neural network, so that the preset load margin prediction model can predict the first load margin prediction value of the power system. The control decision parameter output module is used to input the first load margin prediction value, node voltage and node power into the preset voltage stability control model, so that the encoder built into the preset voltage stability control model maps the first load margin prediction value, node voltage and node power to obtain the mapping result, and inputs the mapping result into the decoder built into the preset voltage stability control model to obtain the current control decision parameters of the power system. The voltage stability control module is used to control the voltage stability of the power system according to the control decision parameters.

9. A terminal device, characterized in that, The system includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements a power system static voltage stability prevention and control method as described in any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device where the storage medium is located to perform a power system static voltage stability prevention and control method as described in any one of claims 1 to 7.

Images