A method for evaluating power grid transient voltage stability based on impedance information learning

Abstract

The present invention discloses a method for evaluating transient voltage stability of a power grid based on impedance information learning, including: extracting the electrical quantity time series trajectory of key load nodes based on time domain simulation, and calibrating the transient voltage stability / instability state of the power grid; calculating the load impedance time series trajectory and the Thevenin equivalent impedance time series trajectory of the key load nodes; combining the electrical quantity time series trajectory, impedance information and the transient voltage stability / instability state of the power grid based on a transient voltage stability impedance criterion with physical significance; learning the electrical quantity and impedance information of a convolutional neural network according to the obtained transient sample set, and obtaining a transient voltage stability evaluation model based on impedance information learning; analyzing the impedance information of the key load nodes based on the transient voltage stability evaluation model, and outputting the transient voltage stability state evaluation result of the system. The present invention can improve the performance of transient voltage stability evaluation and realize accurate and efficient online evaluation of transient voltage stability.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid dynamic stability analysis and evaluation, and in particular to a power grid transient voltage stability evaluation method based on impedance information learning. Background Art

[0002] In recent years, with the continuous expansion of the scale of power grids and the continuous growth of power load levels. Against this background, the form and structure of the power grid have undergone major changes. On the other hand, with the increasing proportion of renewable energy and power electronic equipment, the power grid has gradually shown the "double high" characteristics of a high proportion of renewable energy and a high proportion of power electronic equipment. These trends have caused the operating characteristics and stability mechanisms of traditional power grids to change, such as reduced system inertia and increased source-load uncertainty. The above characteristics and mechanisms make the voltage stability of the system face severe challenges after the power grid suffers transient disturbances. If rapid, accurate, and adaptive evaluations are not taken and relevant control measures are not taken in a timely manner, the power grid is very prone to voltage instability or even voltage collapse accidents.

[0003] In recent years, many scholars have conducted relevant research on the problem of transient voltage stability assessment of power grids, and proposed transient voltage stability assessment methods based on energy functions, time domain simulation, machine learning and other methods. The existing transient voltage stability assessment methods based on machine learning directly use the electrical quantity data obtained by the synchronous phasor measurement device as input to learn the deep learning model. This method lacks the learning of relevant knowledge and information about the transient voltage instability mechanism, resulting in low accuracy of transient voltage stability assessment based on machine learning.

[0004] Therefore, the prior art needs to be improved. Summary of the invention

[0005] The technical problem to be solved by the present invention is that, in view of the defects of the prior art, the present invention provides a power grid transient voltage stability assessment method based on impedance information learning to solve the problem of low accuracy of transient voltage stability assessment in the existing transient voltage stability assessment method based on machine learning.

[0006] The technical solution adopted by the present invention to solve the technical problem is as follows:

[0007] In a first aspect, the present invention provides a method for evaluating transient voltage stability of a power grid based on impedance information learning, comprising:

[0008] Based on the actual operation mode and transient faults of the power grid, time domain simulation is performed to extract the time series trajectory of electrical quantities at key load nodes and calibrate the transient voltage stability / instability state of the power grid;

[0009] The load impedance timing trajectory and the Thevenin equivalent impedance timing trajectory of the key load node are calculated according to the electrical quantity timing trajectory;

[0010] Based on a transient voltage stability impedance criterion with physical significance, the electrical quantity time series trajectory, impedance information and the transient voltage stability / instability state of the power grid are combined to obtain a transient sample set;

[0011] Constructing a convolutional neural network, and performing electrical quantity and impedance information learning on the convolutional neural network according to the transient sample set to obtain a transient voltage stability assessment model based on impedance information learning;

[0012] The impedance information of the key load nodes is analyzed based on the transient voltage stability assessment model, and a system transient voltage stability state assessment result is output.

[0013] In one implementation, the time domain simulation is performed based on the actual operation mode and transient faults of the power grid, the electrical quantity time sequence trajectory of the key load nodes is extracted, and the transient voltage stability / instability state of the power grid is calibrated, including:

[0014] The power grid model is constructed using power grid simulation software, and the parameters of the power grid model are set based on the steady-state operation strategy and transient accidents of the actual power grid to obtain a power grid model set;

[0015] Call the power grid simulation software to perform time domain simulation on the power grid model set. During each simulation, record the time length after the fault occurs as t win The voltage of the key load nodes of the power grid Active power P, reactive power Q, and obtain the time series trajectory of the electrical quantity;

[0016] The electrical quantity time sequence trajectory is automatically calibrated according to the transient voltage stability engineering criterion adopted by the power grid to obtain the calibrated electrical quantity time sequence trajectory.

[0017] In one implementation, the step of calculating the load impedance timing trajectory and the Thevenin equivalent impedance timing trajectory of the key load node according to the electrical quantity timing trajectory includes:

[0018] Based on the time length t after the transient fault occurs at the critical load node win Voltage Active power P, reactive power Q, according to the formula Calculate the current Where j represents the imaginary part of the complex number, represents voltage as a vector;

[0019] The circuit equations at two adjacent measurement moments are combined to obtain dI·Z=dV-dE; where ||dV||≥||dE|| and ||dE||≈0, then dI·Z=dV; Z represents impedance, and E represents power supply potential;

[0020] Based on the simultaneous equations and the least squares method, the formula The Thevenin equivalent impedance of the key load node at time n is calculated:

[0021] in n represents the time corresponding to the impedance calculation; m represents the length of the time frame selected for the impedance calculation;

[0022] According to the Thevenin equivalent impedance of the key load node, the Thevenin equivalent impedance trajectory is obtained. And based on the formula The load impedance time series trajectory of the key load node is calculated; wherein, represents load impedance; Indicates load voltage; Indicates the load current.

[0023] In one implementation, the transient voltage stability impedance criterion based on physical significance includes:

[0024] Based on the transient voltage stability impedance criterion with physical significance, extracting impedance information related to transient voltage instability from the load impedance time series trajectory and the Thevenin equivalent impedance time series trajectory;

[0025] The impedance information includes: the slope dZ of the load impedance time series trajectory L , the slope dZ of the Thevenin equivalent impedance time series trajectory n and the area Z enclosed by the Thevenin equivalent impedance and the load impedance locus A .

[0026] In one implementation, the constructing of a convolutional neural network, performing electrical quantity and impedance information learning on the convolutional neural network according to the transient sample set, and obtaining a transient voltage stability assessment model based on impedance information learning, includes:

[0027] Converting the time series information in the transient sample set into a three-dimensional format to obtain input information matching the input layer of the convolutional neural network;

[0028] Extract relevant features from the transient sample set using two convolution blocks; wherein each of the convolution blocks includes: a convolution layer, a random dropout layer, a batch normalization layer, and an activation function layer;

[0029] The feature graphs derived from the two convolution blocks are linked to the classification output layer through a fully connected layer, and associated with the discrete output of the transient voltage stability / instability of the power grid, so as to construct the convolutional neural network;

[0030] The sample information of the transient sample set is input into the convolutional neural network for learning until the accuracy and loss function of the convolutional neural network reach the convergence standard, thereby obtaining the transient voltage stability assessment model based on impedance information learning.

[0031] In one implementation, the electrical quantity and impedance information learning of the convolutional neural network is performed according to the transient sample set to obtain a transient voltage stability assessment model based on impedance information learning, and further includes:

[0032] The performance of the transient voltage stability evaluation model is evaluated based on the transient sample set, and the total number of all correctly evaluated samples in the transient sample set is counted. r , calculating the assessment accuracy of the transient voltage stability assessment model;

[0033] The final transient voltage stability assessment model is determined based on the calculated assessment accuracy.

[0034] In one implementation, the analyzing the impedance information of the key load node based on the transient voltage stability assessment model and outputting the system transient voltage stability state assessment result includes:

[0035] The key load nodes of the power grid are monitored in real time. If a transient fault occurs during operation, the time length after the fault is extracted as t win Voltage Active power P and reactive power Q;

[0036] Based on the voltage The active power P and the reactive power Q are used to calculate the Thevenin equivalent impedance and load impedance of the corresponding node;

[0037] Extract impedance information and convert the voltage The active power P, the reactive power Q and the impedance information are input into the transient voltage stability assessment model, and the grid transient voltage stability state Label′ is output as a grid transient voltage stability assessment result.

[0038] In a second aspect, the present invention provides a power grid transient voltage stability assessment device based on impedance information learning, comprising:

[0039] The time domain simulation module is used to perform time domain simulation based on the actual operation mode and transient faults of the power grid, extract the time sequence trajectory of the electrical quantities of the key load nodes, and calibrate the transient voltage stability / instability state of the power grid;

[0040] An impedance timing trajectory module, used to calculate the load impedance timing trajectory and the Thevenin equivalent impedance timing trajectory of the key load node according to the electrical quantity timing trajectory;

[0041] An impedance information extraction module is used to combine the electrical quantity time series trajectory, impedance information and the transient voltage stability / instability state of the power grid based on a transient voltage stability impedance criterion with physical significance to obtain a transient sample set;

[0042] An impedance information learning module is used to construct a convolutional neural network, and perform electrical quantity and impedance information learning on the convolutional neural network according to the transient sample set to obtain a transient voltage stability assessment model based on impedance information learning;

[0043] The transient voltage stability assessment module is used to analyze the impedance information of the key load node based on the transient voltage stability assessment model and output the system transient voltage stability state assessment result.

[0044] In a third aspect, the present invention provides a terminal, comprising: a processor and a memory, wherein the memory stores a power grid transient voltage stability assessment program based on impedance information learning, and when the power grid transient voltage stability assessment program based on impedance information learning is executed by the processor, it is used to implement the operation of the power grid transient voltage stability assessment method based on impedance information learning as described in the first aspect.

[0045] In a fourth aspect, the present invention further provides a medium, which is a computer-readable storage medium, and which stores a power grid transient voltage stability assessment program based on impedance information learning. When the power grid transient voltage stability assessment program based on impedance information learning is executed by a processor, it is used to implement the operation of the power grid transient voltage stability assessment method based on impedance information learning as described in the first aspect.

[0046] The present invention adopts the above technical solution to achieve the following effects:

[0047] The present invention combines the mechanism-based impedance information with the measured electrical quantity information, and performs feature extraction and analysis on the power grid from the perspective of transient voltage stability from both the knowledge and data perspectives. Unlike the traditional data-driven method that only processes from the data perspective, the present invention extracts impedance information from the transient voltage stability criterion based on physical significance, fuses the impedance information and the measured electrical quantity data, and extracts key features of the impedance information and electrical quantity data through a convolutional neural network. The impedance information learning method of the present invention can not only improve the accuracy of transient voltage stability assessment, but can also be further applied to the mechanistic explanation of transient voltage stability in the future, providing reliable support for preventing transient voltage instability accidents and establishing efficient transient voltage stability control solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying creative work.

[0049] Figure 1 It is a flow chart of the method for evaluating transient voltage stability of a power grid based on impedance information learning in the present invention.

[0050] Figure 2 It is a flow chart of transient voltage stability assessment of a power grid in one implementation of the present invention.

[0051] Figure 3 It is a structural diagram of the convolutional neural network model in the present invention.

[0052] Figure 4 It is a functional principle diagram of a terminal in one implementation of the present invention.

[0053] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings in conjunction with the embodiments. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solution and advantages of the present invention clearer and more specific, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0055] Exemplary Methods

[0056] In recent years, many scholars have conducted relevant research on the problem of transient voltage stability assessment of power grids, and proposed transient voltage stability assessment methods based on energy functions, time domain simulation, machine learning and other methods. The existing transient voltage stability assessment methods based on machine learning directly use the electrical quantity data obtained by the synchronous phasor measurement device as input to learn the deep learning model. This method lacks the learning of relevant knowledge and information about the transient voltage instability mechanism, resulting in low accuracy of transient voltage stability assessment based on machine learning.

[0057] In view of the above technical problems, an embodiment of the present invention provides a method for evaluating transient voltage stability of a power grid based on impedance information learning. The method extracts the electrical quantity time series trajectory of key load nodes based on time domain simulation, and calibrates the transient voltage stability / instability state of the power grid; calculates the load impedance time series trajectory and the Thevenin equivalent impedance time series trajectory of the key load nodes; based on the transient voltage stability impedance criterion with physical significance, the electrical quantity time series trajectory, impedance information and the transient voltage stability / instability state of the power grid are combined; the convolutional neural network is used to learn the electrical quantity and impedance information according to the obtained transient sample set, and a transient voltage stability evaluation model based on impedance information learning is obtained; the impedance information of the key load nodes is analyzed based on the transient voltage stability evaluation model, and the transient voltage stability state evaluation result of the system is output. The embodiment of the present invention can improve the performance of transient voltage stability evaluation and realize accurate and efficient online evaluation of transient voltage stability.

[0058] like Figure 1 As shown, an embodiment of the present invention provides a method for evaluating transient voltage stability of a power grid based on impedance information learning, comprising the following steps:

[0059] Step S100, performing time domain simulation based on the actual operation mode and transient faults of the power grid, extracting the time sequence trajectory of the electrical quantities of the key load nodes, and calibrating the transient voltage stability / instability state of the power grid.

[0060] In this embodiment, a transient voltage stability impedance criterion with physical significance is introduced. Based on the criterion, impedance information related to the transient voltage stability of the load node is extracted, the system response timing information and the impedance information are combined and fused, and a convolutional neural network is used to learn the above electrical quantity input information. Thus, a neural network classification model based on impedance information learning is constructed to achieve fast and reliable online evaluation of the transient voltage stability state of the power grid.

[0061] like Figure 2 As shown, the method in this embodiment is mainly divided into the following aspects: 1) time domain simulation; 2) basic electrical quantity data acquisition; 3) data labeling; 4) impedance calculation and impedance feature extraction; 5) construction of a short-term voltage stability classification model based on impedance learning; 6) online monitoring, online evaluation of the short-term voltage stability of the power grid.

[0062] In this embodiment, firstly, a time domain simulation is performed based on the actual operation mode and transient faults of the power grid, and basic electrical quantity data are collected during the time domain simulation to obtain simulation data; then, the electrical quantity timing trajectory of the key load nodes is extracted from the time domain simulation data, and the electrical quantity timing trajectory of the key load nodes is data labeled, that is, the transient voltage stability / instability state of the power grid is calibrated to obtain the calibrated electrical quantity timing trajectory.

[0063] The actual operation mode of the power grid can be obtained from the power grid system, for example, power grid configuration parameters, operation mode and corresponding electrical parameters, etc.; the transient fault can be obtained from system logs or historical operation data.

[0064] Specifically, in an implementation of this embodiment, step S100 includes the following steps:

[0065] Step S101, using power grid simulation software to build a power grid model, and setting power grid model parameters based on the steady-state operation strategy and transient accidents of the actual power grid to obtain a power grid model set;

[0066] Step S102, calling the power grid simulation software to perform time domain simulation on the power grid model set, and in each simulation process, recording the time length after the fault occurs as t win The voltage of the key load nodes of the power grid Active power P, reactive power Q, and obtain the time series trajectory of the electrical quantity;

[0067] Step S103, automatically calibrating the electrical quantity time series trajectory according to the transient voltage stability engineering criterion adopted by the power grid to obtain a calibrated electrical quantity time series trajectory.

[0068] In this embodiment, a power grid model is constructed using power grid simulation software, and after the power grid model parameters are set, a total number of N power grid model sets are obtained; as an example, assuming that the power grid has a steady-state operation modes, b transient fault locations, and c transient fault durations, then the total number of power grid model sets is N = a*b*c.

[0069] After obtaining the power grid model set, the power grid simulation software is called based on MatLab software to perform time domain simulation on the power grid model set. In each simulation process, the time length after the fault occurs is recorded as t win (For example, t win =10s) of the voltage at the key load node of the power grid Active power P, reactive power Q. According to the recorded electrical information of the key load nodes, the samples of the collected key load nodes (i.e., the electrical quantity time series trajectory of the key load nodes) are automatically calibrated using the transient voltage stability engineering criterion adopted by the power grid; wherein, the transient voltage stability engineering criterion adopted by the power grid can be set according to corresponding empirical data, and the samples calibrated as transient voltage stability are characterized by Label=1, and the samples calibrated as transient voltage instability are characterized by Label=0; thus, the calibrated electrical quantity time series trajectory of the key load nodes is obtained.

[0070] like Figure 1As shown, an embodiment of the present invention provides a method for evaluating transient voltage stability of a power grid based on impedance information learning, comprising the following steps:

[0071] Step S200, calculating the load impedance timing trajectory and the Thevenin equivalent impedance timing trajectory of the key load node according to the electrical quantity timing trajectory.

[0072] In this embodiment, after data labeling, impedance calculation and impedance feature extraction are performed based on the calibrated electrical quantity time series trajectory, that is, the load impedance time series trajectory and the Thevenin equivalent impedance time series trajectory of the load node are calculated, and impedance information is extracted to perform machine learning based on these impedance features.

[0073] Specifically, in an implementation of this embodiment, step S200 includes the following steps:

[0074] Step S201, based on the time length t after the transient fault occurs at the critical load node win Voltage Active power P, reactive power Q, according to the formula Calculate the current Where j represents the imaginary part of the complex number, represents voltage as a vector;

[0075] Step S202, the circuit equations at two adjacent measurement moments are combined to obtain dI·Z=dV-dE; wherein, ||dV||≥||dE|| and ||dE||≈0, then dI·Z=dV; Z represents impedance, and E represents power supply potential;

[0076] Step S203, based on the simultaneous equations and using the least squares method, the formula is obtained: The Thevenin equivalent impedance of the key load node at time n is calculated:

[0077] in n represents the time corresponding to the impedance calculation; m represents the length of the time frame selected for the impedance calculation;

[0078] Step S204, obtaining the Thevenin equivalent impedance trajectory according to the Thevenin equivalent impedance of the key load node And based on the formula The load impedance time series trajectory of the key load node is calculated; wherein, represents load impedance; Indicates load voltage; Indicates the load current.

[0079] In this embodiment, the time series trajectory of the Thevenin equivalent impedance of important load nodes in the power grid is calculated based on the basic electrical quantities collected by simulation. Specifically refer to the above steps S201 to S204, and based on the formula The load impedance timing trajectory of the load node is calculated.

[0080] In this embodiment, by calculating the Thevenin equivalent impedance time series trajectory and load impedance time series trajectory of important load nodes, these characteristics can be judged based on the transient voltage stability impedance criterion with physical significance, thereby obtaining relevant knowledge information about the transient voltage instability mechanism, so that the convolutional neural network model can learn information with physical significance.

[0081] like Figure 1 As shown, an embodiment of the present invention provides a method for evaluating transient voltage stability of a power grid based on impedance information learning, comprising the following steps:

[0082] Step S300, based on a transient voltage stability impedance criterion with physical significance, the electrical quantity time series trajectory, impedance information and the transient voltage stability / instability state of the power grid are combined to obtain a transient sample set.

[0083] Specifically, in an implementation of this embodiment, step S300 includes the following steps:

[0084] Step S301, based on the transient voltage stability impedance criterion with physical significance, extracting impedance information related to transient voltage instability from the load impedance time series trajectory and the Thevenin equivalent impedance time series trajectory.

[0085] In this embodiment, based on the transient voltage stability impedance criterion with physical significance, the load impedance time series trajectory and the Thevenin equivalent impedance time series trajectory of the key load node calculated in the above step S200 are characterized to extract impedance information related to transient voltage instability.

[0086] The transient voltage stability impedance criterion with physical significance refers to the load impedance modulus value |Z L | process to determine whether it is transient voltage instability; the specific determination process is: if the load impedance modulus value |Z L |Continuously decreases and eventually becomes less than the system's Thevenin equivalent impedance modulus value |Z n |Then it can be considered that the transient voltage is unstable.

[0087] In this embodiment, the extracted impedance information related to transient voltage instability includes: the slope dZ of the load impedance time series trajectory L , the slope dZ of the Thevenin equivalent impedance time series trajectory n and the area Z enclosed by the Thevenin equivalent impedance and the load impedance locus AThis information is used as supplementary input for training the transient voltage stability assessment model so that the trained model can accurately extract this impedance information.

[0088] In this embodiment, the information obtained by time domain simulation is collected, including voltage The time series trajectory of active power P and reactive power Q and the grid stability state Label. The impedance trajectory is calculated by calculating the timing trajectory corresponding to the active power P and reactive power Q. and And further extract impedance information. The grid stability state Label and impedance information are combined to obtain a transient sample. The transient samples obtained by N time domain simulations are collected to obtain a transient sample set containing N samples.

[0089] like Figure 1 As shown, an embodiment of the present invention provides a method for evaluating transient voltage stability of a power grid based on impedance information learning, comprising the following steps:

[0090] Step S400, constructing a convolutional neural network, performing electrical quantity and impedance information learning on the convolutional neural network according to the transient sample set, and obtaining a transient voltage stability assessment model based on impedance information learning.

[0091] Specifically, in an implementation of this embodiment, step S400 includes the following steps:

[0092] Step S401, converting the time series information in the transient sample set into a three-dimensional format to obtain input information matching the input layer of the convolutional neural network;

[0093] Step S402, extracting relevant features from the transient sample set using two convolution blocks; wherein each of the convolution blocks includes: a convolution layer, a random dropout layer, a batch normalization layer, and an activation function layer;

[0094] Step S403, linking the feature graphs derived from the two convolution blocks to the classification output layer through a fully connected layer, and associating them to the discrete output of the transient voltage stability / instability of the power grid, so as to construct the convolutional neural network;

[0095] Step S404: inputting the sample information of the transient sample set into the convolutional neural network for learning until the accuracy and loss function of the convolutional neural network reach the convergence standard, thereby obtaining the transient voltage stability assessment model based on impedance information learning.

[0096] In this embodiment, a convolutional neural network algorithm is used to learn the time series information of the obtained transient sample set, and the sample information of the transient sample set is converted into a three-dimensional format to directly match the input layer of the neural network; the three dimensions are of great significance, namely "the number of buses", "the time series length of the input information" and "the number of input variables"; among them, "the number of buses" represents the key load nodes monitored in the power grid, "the time series length of the input information" represents the length of the learning transient voltage stability characteristic information, and "the number of input variables" represents the input electrical quantity The number of various physical parameters such as P, Q and impedance information.

[0097] After converting the format of the sample information of the transient sample set, two convolution blocks are used to extract relevant features from the power grid data. Figure 3 As shown, each convolution block includes a convolution layer, a random dropout layer, a batch normalization layer, and an activation function layer. Among them, the convolution layer is used to extract key features from the input data, and the random dropout layer prevents overfitting by randomly deactivating neurons during training. Then, the input value of the previous layer is normalized through the batch normalization layer to speed up the training speed and enhance the model performance, and the corresponding activation function is selected to enhance the expressive ability of the model. Finally, the feature map exported by the convolution block is linked to the classification output layer through the fully connected layer, and associated with the discrete output of the transient voltage stability / instability of the power grid, thereby completing the construction of the convolutional neural network model required by the present invention. The structure of the convolutional neural network model constructed in this embodiment is as follows Figure 3 shown.

[0098] The sample information of the transient sample set in step S300 is input into the convolutional neural network model for learning until the accuracy and loss function of the model reach the convergence standard, thereby obtaining a transient voltage stability classification evaluation model based on impedance learning.

[0099] As an example, the convolutional neural network algorithm in this embodiment is implemented using the Pytorch tool library in Python.

[0100] Specifically, in an implementation of this embodiment, step S400 further includes the following steps:

[0101] Step S405: Based on the transient sample set, the transient voltage stability assessment model is evaluated for performance, and a total number of correctly evaluated samples in the transient sample set is counted. r , calculating the assessment accuracy of the transient voltage stability assessment model;

[0102] Step S406: determining a final transient voltage stability assessment model according to the calculated assessment accuracy.

[0103] In this embodiment, the performance of the transient voltage stability classification evaluation model obtained in step S404 is evaluated based on the transient sample set generated in step S300, and the total number of correctly evaluated samples in the transient sample set is counted. r , and calculate the model's evaluation accuracy Acc = N r / N×100%, if the model meets the performance requirements (i.e. Acc≥96%), then subsequent online monitoring will be performed. If the model does not meet the performance requirements (i.e. Acc<96%), then the transient sample set will be re-learned until the model performance meets the requirements and then online monitoring will be performed.

[0104] like Figure 1 As shown, an embodiment of the present invention provides a method for evaluating transient voltage stability of a power grid based on impedance information learning, comprising the following steps:

[0105] Step S500: analyzing the impedance information of the key load nodes based on the transient voltage stability assessment model, and outputting a system transient voltage stability state assessment result.

[0106] In this embodiment, when a transient fault occurs in the online operation of the power grid, the electrical quantity time series trajectory of the key load nodes is obtained in real time, and the impedance time series trajectory is calculated to extract the impedance information, which is input into the transient voltage stability assessment model to perform an online assessment of the transient voltage stability state of the system.

[0107] Specifically, in an implementation of this embodiment, step S500 includes the following steps:

[0108] Step S501: Real-time monitoring of the key load nodes of the power grid is performed. If a transient fault occurs during operation, the time length after the fault is extracted as t win Voltage Active power P and reactive power Q;

[0109] Step S502, based on the voltage The active power P and the reactive power Q are used to calculate the Thevenin equivalent impedance and load impedance of the corresponding node;

[0110] Step S503, extracting impedance information, converting the voltage The active power P, the reactive power Q and the impedance information are input into the transient voltage stability assessment model, and the grid transient voltage stability state Label′ is output as a grid transient voltage stability assessment result.

[0111] In this embodiment, the key load nodes of the corresponding power grid are monitored in real time. If a transient fault occurs during operation, the time length after the fault is extracted as t win Voltage Active power P, reactive power Q, and based on voltage Active power P, reactive power Q calculate the Thevenin equivalent impedance and load impedance of key nodes, and then further extract impedance information, convert voltage The active power P, reactive power Q and impedance information are input into the transient voltage classification assessment model obtained in step S400, and the grid transient voltage stability state Label′ is output as the transient voltage stability assessment result of the grid.

[0112] This embodiment achieves the following technical effects through the above technical solution:

[0113] This embodiment combines the mechanism-based impedance information with the measured electrical quantity information, and performs feature extraction and analysis on the power grid from the perspective of transient voltage stability from both the knowledge and data perspectives. Unlike the traditional data-driven method that only processes from the data perspective, this embodiment extracts impedance information from the transient voltage stability criterion based on physical significance, fuses the impedance information and the measured electrical quantity data, and extracts key features of the impedance information and electrical quantity data through a convolutional neural network. The impedance information learning method of this embodiment can not only improve the accuracy of transient voltage stability assessment, but can also be further applied to the mechanistic explanation of transient voltage stability in the future, providing reliable support for preventing transient voltage instability accidents and establishing efficient transient voltage stability control solutions.

[0114] Exemplary Devices

[0115] Based on the above embodiments, the present invention further provides a power grid transient voltage stability assessment device based on impedance information learning, comprising:

[0116] The time domain simulation module is used to perform time domain simulation based on the actual operation mode and transient faults of the power grid, extract the time sequence trajectory of the electrical quantities of the key load nodes, and calibrate the transient voltage stability / instability state of the power grid;

[0117] An impedance timing trajectory module, used to calculate the load impedance timing trajectory and the Thevenin equivalent impedance timing trajectory of the key load node according to the electrical quantity timing trajectory;

[0118] An impedance information extraction module is used to combine the electrical quantity time series trajectory, impedance information and the transient voltage stability / instability state of the power grid based on a transient voltage stability impedance criterion with physical significance to obtain a transient sample set;

[0119] An impedance information learning module is used to construct a convolutional neural network, and perform electrical quantity and impedance information learning on the convolutional neural network according to the transient sample set to obtain a transient voltage stability assessment model based on impedance information learning;

[0120] The transient voltage stability assessment module is used to analyze the impedance information of the key load node based on the transient voltage stability assessment model and output the system transient voltage stability state assessment result.

[0121] This embodiment achieves the following technical effects through the above technical solution:

[0122] This embodiment combines the mechanism-based impedance information with the measured electrical quantity information, and performs feature extraction and analysis on the power grid from the perspective of transient voltage stability from both the knowledge and data perspectives. Unlike the traditional data-driven method that only processes from the data perspective, this embodiment extracts impedance information from the transient voltage stability criterion based on physical significance, fuses the impedance information and the measured electrical quantity data, and extracts key features of the impedance information and electrical quantity data through a convolutional neural network. The impedance information learning method of this embodiment can not only improve the accuracy of transient voltage stability assessment, but can also be further applied to the mechanistic explanation of transient voltage stability in the future, providing reliable support for preventing transient voltage instability accidents and establishing efficient transient voltage stability control solutions.

[0123] Based on the above embodiment, the present invention further provides a terminal, whose principle block diagram can be as follows: Figure 4 shown.

[0124] The terminal includes: a processor, a memory, an interface, a display screen and a communication module connected through a system bus; wherein the processor of the terminal is used to provide computing and control capabilities; the memory of the terminal includes a storage medium and an internal memory; the storage medium stores an operating system and a computer program; the internal memory provides an environment for the operation of the operating system and the computer program in the storage medium; the interface is used to connect to external devices; the display screen is used to display corresponding information; and the communication module is used to communicate with a cloud server or other devices.

[0125] When the computer program is executed by a processor, it is used to implement the operation of a power grid transient voltage stability assessment method based on impedance information learning.

[0126] It can be understood by those skilled in the art that Figure 4 The principle block diagram shown in the figure is only a block diagram of a partial structure related to the scheme of the present invention, and does not constitute a limitation on the terminal to which the scheme of the present invention is applied. The specific terminal may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0127] In one embodiment, a terminal is provided, which includes: a processor and a memory, the memory storing a power grid transient voltage stability assessment program based on impedance information learning, and the power grid transient voltage stability assessment program based on impedance information learning is used to implement the operation of the above-mentioned power grid transient voltage stability assessment method based on impedance information learning when executed by the processor.

[0128] In one embodiment, a storage medium is provided, wherein the storage medium stores a power grid transient voltage stability assessment program based on impedance information learning, and when the power grid transient voltage stability assessment program based on impedance information learning is executed by a processor, it is used to implement the operation of the above-mentioned power grid transient voltage stability assessment method based on impedance information learning.

[0129] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the computer program can be stored in a non-volatile storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention can include non-volatile and volatile memory.

[0130] In summary, the present invention provides a method for evaluating transient voltage stability of a power grid based on impedance information learning, the method comprising: extracting the electrical quantity time series trajectory of key load nodes based on time domain simulation, and calibrating the transient voltage stability / instability state of the power grid; calculating the load impedance time series trajectory and the Thevenin equivalent impedance time series trajectory of the key load nodes; combining the electrical quantity time series trajectory, impedance information, and the transient voltage stability / instability state of the power grid based on a transient voltage stability impedance criterion with physical significance; learning the electrical quantity and impedance information of a convolutional neural network based on the obtained transient sample set, and obtaining a transient voltage stability evaluation model based on impedance information learning; analyzing the impedance information of the key load nodes based on the transient voltage stability evaluation model, and outputting the transient voltage stability state evaluation result of the system. The present invention can improve the performance of transient voltage stability evaluation and realize accurate and efficient online evaluation of transient voltage stability.

[0131] It should be understood that the application of the present invention is not limited to the above examples. For ordinary technicians in this field, improvements or changes can be made based on the above description. All these improvements and changes should fall within the scope of protection of the claims attached to the present invention.

Claims

1. A method for evaluating transient voltage stability of a power grid based on impedance information learning, characterized in that: include: Based on the actual operation mode and transient faults of the power grid, time domain simulation is performed to extract the time series trajectory of electrical quantities at key load nodes and calibrate the transient voltage stability / instability state of the power grid; The load impedance timing trajectory and the Thevenin equivalent impedance timing trajectory of the key load node are calculated according to the electrical quantity timing trajectory; Based on a transient voltage stability impedance criterion with physical significance, the electrical quantity time series trajectory, impedance information and the transient voltage stability / instability state of the power grid are combined to obtain a transient sample set; Constructing a convolutional neural network, and performing electrical quantity and impedance information learning on the convolutional neural network according to the transient sample set to obtain a transient voltage stability assessment model based on impedance information learning; Analyzing the impedance information of the key load nodes based on the transient voltage stability assessment model, and outputting a system transient voltage stability state assessment result; The step of calculating the load impedance timing trajectory and the Thevenin equivalent impedance timing trajectory of the key load node according to the electrical quantity timing trajectory includes: Based on the time length after the transient fault of the critical load node occurs Voltage , Active Power , reactive power , according to the formula Calculate the current ;in, represents the imaginary part of a complex number, represents voltage as a vector; The circuit equations of two adjacent measurement moments are combined to obtain ;in, and , then ; represents impedance, Indicates the power supply potential; Based on the simultaneous equations and the least squares method, the formula , calculate the time The Thevenin equivalent impedance of the key load node: in , ; represents the time corresponding to the calculated impedance; Indicates the length of the time frame selected for impedance calculation; According to the Thevenin equivalent impedance of the key load node, the Thevenin equivalent impedance trajectory is obtained. , and based on the formula The load impedance time series trajectory of the key load node is calculated; wherein, represents load impedance; Indicates load voltage; Indicates load current; The transient voltage stability impedance criterion based on physical significance includes: Based on the transient voltage stability impedance criterion with physical significance, extracting impedance information related to transient voltage instability from the load impedance time series trajectory and the Thevenin equivalent impedance time series trajectory; The impedance information includes: the slope of the load impedance timing trajectory , the slope of the Thevenin equivalent impedance timing trajectory and the area enclosed by the Thevenin equivalent impedance and the load impedance locus ; The transient voltage stability impedance criterion with physical significance is: the load impedance modulus value in the transient process The process is judged to determine whether it is transient voltage instability.

2. The method for evaluating transient voltage stability of a power grid based on impedance information learning according to claim 1, characterized in that: The time domain simulation is performed based on the actual operation mode and transient faults of the power grid, the electrical quantity time sequence trajectory of the key load nodes is extracted, and the transient voltage stability / instability state of the power grid is calibrated, including: The power grid model is constructed using power grid simulation software, and the parameters of the power grid model are set based on the steady-state operation strategy and transient accidents of the actual power grid to obtain a power grid model set; Call the power grid simulation software to perform time domain simulation on the power grid model set. During each simulation, record the time length after the fault occurs. The voltage of the key load nodes of the power grid , Active Power , reactive power , obtain the electrical quantity time series trajectory; The electrical quantity time sequence trajectory is automatically calibrated according to the transient voltage stability engineering criterion adopted by the power grid to obtain the calibrated electrical quantity time sequence trajectory.

3. The method for evaluating transient voltage stability of a power grid based on impedance information learning according to claim 1, characterized in that: The convolutional neural network is constructed, and electrical quantity and impedance information learning is performed on the convolutional neural network according to the transient sample set to obtain a transient voltage stability assessment model based on impedance information learning, including: Converting the time series information in the transient sample set into a three-dimensional format to obtain input information matching the input layer of the convolutional neural network; Extract relevant features from the transient sample set using two convolution blocks; wherein each of the convolution blocks includes: a convolution layer, a random dropout layer, a batch normalization layer, and an activation function layer; The feature graphs derived from the two convolution blocks are linked to the classification output layer through a fully connected layer, and are associated with the discrete output of transient voltage stability / instability of the power grid, so as to construct the convolutional neural network; The sample information of the transient sample set is input into the convolutional neural network for learning until the accuracy and loss function of the convolutional neural network reach the convergence standard, thereby obtaining the transient voltage stability assessment model based on impedance information learning.

4. The method for evaluating transient voltage stability of a power grid based on impedance information learning according to claim 1, characterized in that: The step of performing electrical quantity and impedance information learning on the convolutional neural network according to the transient sample set to obtain a transient voltage stability assessment model based on impedance information learning also includes: The transient voltage stability evaluation model is evaluated based on the transient sample set, and the total number of correctly evaluated samples in the transient sample set is counted. , calculating the assessment accuracy of the transient voltage stability assessment model; The final transient voltage stability assessment model is determined based on the calculated assessment accuracy.

5. The method for evaluating transient voltage stability of a power grid based on impedance information learning according to claim 1, characterized in that: The step of analyzing the impedance information of the key load node based on the transient voltage stability assessment model and outputting a system transient voltage stability state assessment result includes: The key load nodes of the power grid are monitored in real time. If a transient fault occurs during operation, the length of time after the fault is extracted. Voltage , Active Power And reactive power ; Based on the voltage The active power And the reactive power , calculate the Thevenin equivalent impedance and load impedance of the corresponding node; Extract impedance information and convert the voltage The active power , the reactive power The impedance information is input into the transient voltage stability assessment model, and the transient voltage stability state of the power grid is obtained as output. , as the transient voltage stability assessment result of the power grid.

6. A power grid transient voltage stability assessment device based on impedance information learning, used to implement the power grid transient voltage stability assessment method based on impedance information learning as claimed in any one of claims 1 to 5, characterized in that: include: The time domain simulation module is used to perform time domain simulation based on the actual operation mode and transient faults of the power grid, extract the time sequence trajectory of the electrical quantities of the key load nodes, and calibrate the transient voltage stability / instability state of the power grid; An impedance timing trajectory module, used to calculate the load impedance timing trajectory and the Thevenin equivalent impedance timing trajectory of the key load node according to the electrical quantity timing trajectory; An impedance information extraction module is used to combine the electrical quantity time series trajectory, impedance information and the transient voltage stability / instability state of the power grid based on a transient voltage stability impedance criterion with physical significance to obtain a transient sample set; An impedance information learning module is used to construct a convolutional neural network, and perform electrical quantity and impedance information learning on the convolutional neural network according to the transient sample set to obtain a transient voltage stability assessment model based on impedance information learning; The transient voltage stability assessment module is used to analyze the impedance information of the key load node based on the transient voltage stability assessment model and output the system transient voltage stability state assessment result.

7. A terminal, characterized in that: include: A processor and a memory, wherein the memory stores a power grid transient voltage stability assessment program based on impedance information learning, and when the power grid transient voltage stability assessment program based on impedance information learning is executed by the processor, it is used to implement the operation of the power grid transient voltage stability assessment method based on impedance information learning as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a power grid transient voltage stability assessment program based on impedance information learning, and when the power grid transient voltage stability assessment program based on impedance information learning is executed by a processor, it is used to implement the operation of the power grid transient voltage stability assessment method based on impedance information learning as described in any one of claims 1-5.

Images