New energy power grid electromechanical-electromagnetic simulation track mapping method and system based on residual long short-term memory network

Abstract

The invention discloses a new energy power grid electromechanical-electromagnetic simulation track mapping method and system based on a residual long short-term memory network. Comprising the following steps: respectively carrying out digital simulation on an electromechanical transient process and an electromagnetic transient process of a high-proportion new energy power grid system to obtain electromechanical and electromagnetic transient time sequence track samples and carrying out preprocessing; a time sequence track mapping model based on the residual long short-term memory network is constructed, the model comprises an input layer, a plurality of LSTM layers, a full connection layer and a regression layer, and the residual network is introduced between the LSTM layers; and training the model by using the preprocessed electromechanical and electromagnetic transient time sequence track samples to realize mapping from an electromechanical transient simulation time sequence track to an electromagnetic transient simulation time sequence track. According to the method, the dynamic response process of the system under the electromagnetic transient scale can be predicted only based on the existing electromechanical transient simulation trajectory data, and meanwhile, the safety and stability of the system under the power electronic high-frequency dynamic condition are reflected.

Description

Technical Field

[0001] The present invention belongs to the field of energy technology, and relates to a new energy power grid electromechanical-electromagnetic simulation trajectory mapping method and system, and in particular to a new energy power grid electromechanical-electromagnetic simulation trajectory mapping method and system based on residual long short-term memory network. Background Art

[0002] As the proportion of new energy such as wind power and photovoltaic power in the power grid continues to increase, the power grid's electronic characteristics are highlighted, the system's operating characteristics are more complex, and the stability form has undergone profound changes. In order to cope with the new challenges faced by a high proportion of new energy power grids, accurate and efficient power grid simulation technology can effectively evaluate the stability of new energy grid connection and system operating characteristics, and is one of the key means to ensure the safe and stable operation of the power grid. However, the access of a large number of power electronic equipment has greatly increased the scale and complexity of the power grid, and the system's component composition and dynamic process have become increasingly complex, which has put forward high requirements on the simulation component modeling accuracy and computing power.

[0003] Transient simulation of new energy power grids can usually be divided into electromechanical transient simulation and electromagnetic transient simulation. Electromechanical transient simulation mainly reflects the operating conditions of the system near the power frequency, such as the change in the mechanical motion of the rotor caused by the change in the electromagnetic torque of the generator and the motor. The calculation model usually uses fundamental phasor description, which cannot simulate the high-frequency characteristics of power electronic equipment and its control loop. With the development of DC transmission and the continuous improvement of the penetration rate of new energy, electromechanical transients are difficult to adapt to the stability analysis needs of high-proportion new energy power grids. Relatively speaking, electromagnetic transients can more accurately reflect the dynamic characteristics of power electronic equipment, but the electromagnetic model considers smaller time scale dynamics, which greatly increases the calculation amount and calculation time of the simulation. For a certain scale of system, it is usually difficult to carry out electromagnetic transient simulation, and it is impossible to provide real-time basis for decision-making and scheduling of new energy power grids.

[0004] Considering the current situation that the dispatch of new energy power grid is mainly aimed at the dynamics of electromechanical time scale, the present invention provides a method and system for mapping electromechanical-electromagnetic simulation trajectories of new energy power grid based on residual long short-term memory network. The present invention extracts the correlation characteristics and mapping relationship of simulation time series trajectories at two time scales through data-driven learning of the difference in dynamic response of the system at different electromechanical and electromagnetic time scales, and constructs a time series trajectory mapping model from electromechanical transient to electromagnetic transient, realizing a reasonable prediction method for the transient process of the system at the electromagnetic transient time scale based only on the dynamic response characteristics of the system at the electromechanical time scale, thereby realizing the safety and stability evaluation of the system under the high-frequency dynamics of power electronics to a certain extent. Summary of the invention

[0005] To solve the above problems, the present invention proposes a new energy power grid electromechanical-electromagnetic simulation trajectory mapping method and system based on residual long short-term memory network, which solves the problem that the new energy power grid is difficult to apply the original electromechanical and electromagnetic simulation under the background of increasing new energy penetration rate.

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

[0007] A new energy power grid electromechanical-electromagnetic simulation trajectory mapping method based on residual long short-term memory network includes the following steps:

[0008] A new energy power grid electromechanical-electromagnetic simulation trajectory mapping method based on residual long short-term memory network includes the following steps:

[0009] Digitally simulate the electromechanical transient process and electromagnetic transient process of the high-proportion new energy power grid system, obtain electromechanical and electromagnetic transient time series trajectory samples, and pre-process the electromechanical and electromagnetic transient time series trajectory samples;

[0010] A time series trajectory mapping model based on a residual long short-term memory network (R-LSTM) is constructed, the model input is an electromechanical transient time series trajectory, and the output is an electromagnetic transient time series trajectory. The time series trajectory mapping model includes an input layer, several LSTM layers, a fully connected layer and a regression layer, and a residual network is introduced between the input of the first LSTM layer and the output of the last LSTM layer;

[0011] The preprocessed electromechanical and electromagnetic transient timing trajectory samples are used to train the timing trajectory mapping model to achieve mapping from the electromechanical transient simulation timing trajectory to the electromagnetic transient simulation timing trajectory.

[0012] Furthermore, the computing element model in the electromechanical transient simulation timing trajectory of the high-proportion new energy power grid system is described by the fundamental wave vector, and the computing element model in the electromagnetic transient simulation timing trajectory is described by the abc three-phase instantaneous value.

[0013] Furthermore, the electromechanical and electromagnetic transient time series trajectory samples are preprocessed, specifically: based on the fault time scale and frequency domain response characteristics, the electromechanical and electromagnetic transient time series trajectory samples are resampled and linearly interpolated to obtain electromechanical and electromagnetic discrete time series data with the same sequence length and corresponding to the key fault nodes.

[0014] Furthermore, several LSTM layers in the time series trajectory mapping model are connected in series to abstract the input data features layer by layer and extract the features required for the task.

[0015] Furthermore, the residual network uses identity mapping to directly transmit the input signal of the first LSTM layer to the last LSTM layer as part of the feature vectors finally output by several LSTM layers.

[0016] Furthermore, the feature vector includes the output signal of the LSTM layer and the input signal transmitted through the residual network, and its expression is:

[0017]

[0018] in, is the output of the residual LSTM network, G(X) is the feature vector output by several LSTM layers, is the input signal transmitted by the residual network, and W r denote the activation function and projection matrix respectively.

[0019] Furthermore, the training process of the time series trajectory mapping model is specifically as follows:

[0020] Supervised learning is used to predict the electromagnetic transient timing trajectory data based on the input electromechanical transient timing trajectory data. The mapping error between the predicted electromagnetic transient timing trajectory data and the actual electromagnetic transient timing trajectory data is calculated using the loss function. The mapping error is back-propagated based on the gradient descent algorithm of machine learning to update the network parameters.

[0021] Furthermore, the loss function used in the regression layer of the time series trajectory mapping model is the HuberLoss loss function. When the prediction deviation is greater than a threshold parameter, a square error is used, and when the prediction deviation is less than the threshold parameter, a linear error is used.

[0022] A new energy power grid electromechanical-electromagnetic simulation trajectory mapping system based on residual long short-term memory network, comprising:

[0023] Data acquisition module: used to digitally simulate the electromechanical transient process and electromagnetic transient process of the high-proportion new energy power grid system, obtain electromechanical and electromagnetic transient time series trajectory samples, and pre-process the electromechanical and electromagnetic transient time series trajectory samples;

[0024] Model construction module: used to construct a time series trajectory mapping model based on residual long short-term memory network. The model input is the electromechanical transient time series trajectory, and the output is the electromagnetic transient time series trajectory. The time series trajectory mapping model includes an input layer, several LSTM layers, a fully connected layer and a regression layer, and a residual network introduced between the input of the first LSTM layer and the output of the last LSTM layer;

[0025] Model training module: used to train the timing trajectory mapping model using the preprocessed electromechanical and electromagnetic transient timing trajectory samples to achieve mapping from the electromechanical transient simulation timing trajectory to the electromagnetic transient simulation timing trajectory.

[0026] A computer device comprising:

[0027] one or more processors;

[0028] A memory for storing one or more programs;

[0029] When the one or more programs are executed by the one or more processors, the one or more processors implement the above-mentioned electromechanical-electromagnetic simulation trajectory mapping method for new energy power grid based on residual long short-term memory network.

[0030] The beneficial effects of the present invention are:

[0031] The present invention provides a method for mapping electromechanical-electromagnetic simulation trajectories of new energy power grids based on residual long short-term memory networks. The method aims to identify the differences in the dynamic responses of the system at different electromechanical and electromagnetic time scales through a data-driven method, and to construct a time trajectory mapping model from electromechanical transients to electromagnetic transients by extracting the associated features and mapping relationships of the simulation timing trajectories at different time scales, thereby achieving a reasonable prediction of the dynamic response process at the electromagnetic transient time scale based only on the electromechanical time scale response characteristics of the system. The present invention can provide a basis for the analysis of the physical mechanism of transient stability and the dominant instability characteristics of a high-proportion new energy power grid taking into account the dynamics of power electronics. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is a flow chart of time series trajectory mapping model training and prediction in an embodiment of the present invention.

[0033] Figure 2 LSTM unit structure diagram in an embodiment of the present invention.

[0034] Figure 3 This is a diagram of an R-LSTM model that introduces a residual network connection in an embodiment of the present invention.

[0035] Figure 4 This is a modeling diagram of a three-machine nine-node wind turbine grid-connected system in an embodiment of the present invention.

[0036] Figure 5 Graph showing model training error and validation error in an embodiment of the present invention.

[0037] Figure 6 It is a diagram of simulation waveform and mapping model prediction results in an embodiment of the present invention. DETAILED DESCRIPTION

[0038] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0039] A new energy power grid electromechanical-electromagnetic simulation trajectory mapping method based on residual long short-term memory network includes the following steps:

[0040] S1. Perform digital simulation on the electromechanical transient process and electromagnetic transient process of the high-proportion new energy power grid system, obtain the electromechanical and electromagnetic transient time sequence trajectory samples of the high-proportion new energy power grid system at different time scales, and pre-process the electromechanical and electromagnetic transient time sequence trajectory samples. The specific steps are: for the electromechanical transient modeling of the high-proportion new energy power grid system, its calculation element model is described by the fundamental wave phasor. Measure the voltage and current amplitude V of each new energy grid connection point at any time s rms (t), I rms (t). Among them, the magnitude of the controlled current source current of the double-fed wind turbine connected to the grid is The grid-connected power S(t) is calculated by the asynchronous motor equation and the grid-side impedance equation.

[0041] For electromagnetic transient modeling of high-proportion renewable energy grids, the calculation component model is described by the abc three-phase instantaneous value. The three-phase instantaneous value v of voltage and current at any time at each renewable energy grid connection point is measured. sabc and i abc and effective value I emt (t).

[0042] Assume Δt emt and Δt rms are the simulation steps of electromagnetic transient and electromechanical transient simulation respectively, the total simulation time is T, and the electromechanical transient timing trajectory is expressed as follows:

[0043]

[0044] The electromagnetic transient timing trajectory is expressed as follows:

[0045]

[0046] in, They represent the electromechanical transient timing trajectories of voltage and current at the renewable energy grid-connected nodes, They represent the electromagnetic transient time series trajectory of voltage and current of the new energy grid-connected node. (k×Δt rms )、(k×Δt emt ) represent the sampling time on the electromechanical and electromagnetic trajectories corresponding to the kth sampling point, respectively.

[0047] In order to solve the problem of inconsistent parameter settings such as electromechanical and electromagnetic simulation step size and fault time, the electromechanical-electromagnetic time series trajectory data of the new energy system is preprocessed to obtain discrete time series data X=[x(1),…,x(K)] and Y=[y(1),…,y(K)] with the same sequence length and corresponding to the key fault nodes, where X and Y are the discrete time series data of electromechanical and electromagnetic, respectively, and K is the sequence length.

[0048] S2. Construct a time series trajectory mapping model based on the residual long short-term memory network (R-LSTM), the model input is the electromechanical transient time series trajectory, and the output is the electromagnetic transient time series trajectory. The time series trajectory mapping model includes an input layer, several LSTM layers, a fully connected layer and a regression layer, and a residual network introduced between the input of the first LSTM layer and the output of the last LSTM layer. The specific steps are:

[0049] (1) Construct a time series mapping model based on a long short-term memory neural network (LSTM). Through n layers of LSTM in series, the input data features are abstracted layer by layer to extract the time series feature representation. The basic structure of the LSTM unit is as follows: Figure 2 As shown;

[0050] (2) A branch is introduced to connect the input of the first layer LSTM and the output of the nth layer LSTM. This connection signal branch is the residual connection, and its structure is as follows: Figure 3 , the residual network is composed of a series of residual blocks. The general representation of a residual block is:

[0051]

[0052] Where G(X) is the feature vector output by several LSTM layers, is the input signal transmitted by the residual network, is the activation function of the residual link. This branch allows the identity mapping of the input signal to be directly propagated to the final layer of the LSTM as a feature vector As a part of the network, deeper networks avoid overfitting and network degradation problems and are easier to optimize.

[0053] (3) R-LSTM network output feature vector That is, the dimensionality reduction representation of the simulation sequence of the original physical quantity in the electromechanical transient simulation. The feature vector retains the main information and modes in the original data, while removing redundancy and noise, making subsequent processing and analysis more efficient and accurate. The feature vector is passed through a layer of fully connected network to obtain the final prediction result.

[0054] Furthermore, the HuberLoss loss function is introduced into the time series mapping model to calculate the model regression error, including:

[0055] Instead of the traditional RMSE loss function, the HuberLoss loss function with a hyperparameter δ is introduced in the model regression layer:

[0056] L(y,f(x))=RMSE+L δ(y,f(x))

[0057]

[0058] Where y is the true value of the electromagnetic transient simulation trajectory, f(x) is the predicted value of the electromagnetic transient simulation trajectory output by the model, and δ determines the sensitivity of the Huber Loss function to outliers. When the prediction deviation is greater than δ, the square error is used to increase the sensitivity to the peak fitting error. When the prediction deviation is less than δ, the linear error is used to make the training more robust.

[0059] According to the characteristics of short duration and large variation of key extreme value data such as impulse current or voltage drop depth in power grid transient process, the fitting accuracy of model extreme points is improved by extracting peak features in time series trajectory and adaptively adjusting loss weights.

[0060] S3. Use the pre-processed electromechanical and electromagnetic transient timing trajectory samples to train the timing trajectory mapping model to achieve mapping from electromechanical transient simulation timing trajectory to electromagnetic transient simulation timing trajectory. The specific steps are:

[0061] Based on the electromechanical transient time sequence trajectory X = [x(1), ..., x(K)] and the electromagnetic transient time sequence trajectory Y = [y(1), ..., y(K)] of the above-mentioned high-proportion new energy power grid system, the time sequence trajectory mapping model is trained by supervised learning. The model predicts the electromagnetic transient time sequence trajectory Y of each electrical quantity based on the input electromechanical transient time sequence trajectory X, and the prediction result is expressed as Y p =[y p (1),…,y p (K)] indicates that the mapping error of the model is calculated according to the above loss function, and the network parameters are updated by back-propagating the mapping error based on the gradient descent algorithm of machine learning. The training and prediction flow chart of the time series trajectory mapping model is as follows: Figure 1 shown.

[0062] The present invention provides a method for mapping electromechanical-electromagnetic simulation trajectories of new energy power grids based on residual long short-term memory networks. The method aims to identify the differences in dynamic responses of new energy power grid systems at different electromechanical and electromagnetic time scales through a data-driven method, and to construct a time trajectory mapping model from electromechanical transients to electromagnetic transients by extracting the correlation characteristics and mapping relationships of simulation time series trajectories at different time scales, thereby achieving a reasonable prediction of the dynamic response process at the electromagnetic transient time scale based only on the electromechanical time scale response characteristics of the system. The present invention can provide a basis for the analysis of the physical mechanism of transient stability and the dominant instability characteristics of high-proportion new energy power grids taking into account the dynamics of power electronics.

[0063] A specific embodiment of the present invention is as follows:

[0064] The effectiveness of the proposed method is verified in the IEEE-3 9-bus doubly-fed wind turbine grid-connected system. Figure 4 As shown in the figure, the IEEE 9-Bus system model under electromechanical transient and electromagnetic transient is established respectively, and a double-fed wind turbine generator is added at Bus3 to replace part of the output of the synchronous generator at Bus3. Based on the load curve and the wind turbine output curve, a variety of working conditions are randomly generated, and the power flow calculation is performed for each working condition to determine the active output P of each synchronous generator. Gi and reactive power output Q Gi .

[0065] Electromechanical and electromagnetic transient simulations are performed on the low-throughput scenario of the double-fed wind turbine grid-connected system. A three-phase short-circuit fault is set at a random position on different lines. The fault is cleared after 0.2s. The voltage Vs, current Is, active output PDFIG and reactive output QDFIG of the wind turbine grid-connected point before and after the fault are observed. The simulation time is T = 2s; the electromagnetic and electromechanical simulation step sizes are set to Δt emt =25us and Δt rms =5ms, and the electromagnetic electromechanical simulation timing trajectory under different working conditions and faults is obtained. s =0.5ms is used as the time interval for sampling, and the electromechanical and electromagnetic transient time series samples (X, Y) with a sequence length of K=4000 are obtained, and stored as data samples for subsequent analysis. The model and training parameter settings are shown in Table 1. 9000 data samples containing various physical quantities are obtained by Simulink simulation. After randomly shuffling the data set, 30% of the samples are taken as the training set of the model, and 70% of the samples are taken as the test set of the model. The input sequence length of the model is the electromechanical transient simulation sequence length, and the output sequence length of the model is the electromagnetic transient simulation sequence length that the model needs to predict. The parameter setting of the LSTM network refers to the setting of the parameters of the classic LSTM model, and is finally determined by multiple experimental tests. Among them, the number of hidden units N in the LSTM network h Set to 200, which is the feature vector output by the R-LSTM network The dimension of the model is , the dimensionality reduction ratio is 20, the number of fully connected layer units is set to 50, the random inactivation rate is set to 20%, and the LSTM output activation function uses ReLU.

[0066] The method of the present invention is used to perform simulation calculations on the embodiment, and the results are as follows:

[0067] The prediction effects of R-LSTM neural network, commonly used LSTM neural network and convolutional neural network (CNN) were compared on a PC platform with an I5-3.4GHz CPU and 8GB RAM. The results are listed in Table 2.

[0068] Table 1 Comparison of calculation results of different models

[0069]

[0070] Comparing different algorithms, CNN has certain advantages in training time and computation time over LSTM. When the training time is similar, the prediction accuracy of residual connection LSTM is significantly improved. For different physical quantities, the residual connection LSTM neural network achieves the minimum RMSE error. Among them, the model has the highest prediction accuracy for the EMT dynamic trajectory of the voltage and current at the grid connection point. The training error and verification error of the model are shown in Figure 2. Figure 5 As shown, relatively speaking, the R-LSTM model training converges faster and performs better in the validation set, proving that the model has a certain ability to resist overfitting. The electromechanical / electromagnetic simulation waveforms and prediction results of the wind turbine grid-connected point current and voltage are shown in Figure 6. As can be seen from the figure, the timing trajectory mapping model of the present invention has good recognition ability and fitting accuracy for different physical quantities. Therefore, the introduction of the residual network in the model effectively improves the generalization ability of the model when predicting electromagnetic simulation sequences of other physical quantities; at the same time, the model can effectively extract the peak features of the sequence and adjust the loss weights, which improves the fitting accuracy of the mapping model in the peak area of ​​the EMT simulation trajectory, so that the model can better predict the maximum current and lowest voltage that can only be reflected in the dynamics of the electromagnetic model.

[0071] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0072] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 Process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0073] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0074] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0075] The above specific implementation modes are used to explain the present invention rather than to limit the present invention. Any modification and change made to the present invention within the spirit of the present invention and the protection scope of the claims shall fall within the protection scope of the present invention.

Claims

1. A new energy power grid electromechanical-electromagnetic simulation trajectory mapping method based on residual long short-term memory network, characterized in that: The following steps are involved: Digitally simulate the electromechanical transient process and electromagnetic transient process of the high-proportion new energy power grid system, obtain electromechanical and electromagnetic transient time series trajectory samples, and pre-process the electromechanical and electromagnetic transient time series trajectory samples; A time series trajectory mapping model based on a residual long short-term memory network is constructed, wherein the model input is an electromechanical transient time series trajectory, and the output is an electromagnetic transient time series trajectory. The time series trajectory mapping model includes an input layer, several LSTM layers, a fully connected layer, and a regression layer, and a residual network introduced between the input of the first LSTM layer and the output of the last LSTM layer; The preprocessed electromechanical and electromagnetic transient timing trajectory samples are used to train the timing trajectory mapping model to achieve mapping from the electromechanical transient simulation timing trajectory to the electromagnetic transient simulation timing trajectory.

2. According to a method for mapping the electromechanical-electromagnetic simulation trajectory of a new energy power grid based on a residual long short-term memory network according to claim 1, it is characterized in that: The computing element model in the electromechanical transient simulation timing trajectory of the high-proportion new energy power grid system is described by the fundamental wave vector, and the computing element model in the electromagnetic transient simulation timing trajectory is described by the abc three-phase instantaneous value.

3. According to a method for mapping the trajectory of electromechanical-electromagnetic simulation of a new energy power grid based on a residual long short-term memory network according to claim 1, it is characterized in that: The preprocessing of the electromechanical and electromagnetic transient time series trajectory samples is specifically: based on the fault time scale and frequency domain response characteristics, the electromechanical and electromagnetic transient time series trajectory samples are resampled and linearly interpolated to obtain electromechanical and electromagnetic discrete time series data with the same sequence length and corresponding to the key fault nodes.

4. According to a method for mapping the trajectory of electromechanical-electromagnetic simulation of a new energy power grid based on residual long short-term memory network in claim 1, it is characterized in that: In the time series trajectory mapping model, several LSTM layers are connected in series to abstract the input data features layer by layer and extract the features required for the task.

5. According to a method for mapping the trajectory of electromechanical-electromagnetic simulation of a new energy power grid based on a residual long short-term memory network according to claim 1, it is characterized in that: The residual network uses identity mapping to directly transmit the input signal of the first LSTM layer to the last LSTM layer as part of the feature vector finally output by several LSTM layers.

6. According to claim 5, a new energy power grid electromechanical-electromagnetic simulation trajectory mapping method based on residual long short-term memory network is characterized in that: The feature vector includes the output signal of the LSTM layer and the input signal transmitted through the residual network, and its expression is: in, is the output of the residual LSTM network, G(X) is the feature vector output by several LSTM layers, is the input signal transmitted by the residual network, and W r denote the activation function and projection matrix respectively.

7. According to claim 1, a new energy power grid electromechanical-electromagnetic simulation trajectory mapping method based on residual long short-term memory network is characterized in that: The training process of the time series trajectory mapping model is specifically as follows: Supervised learning is used to predict the electromagnetic transient timing trajectory data based on the input electromechanical transient timing trajectory data. The mapping error between the predicted electromagnetic transient timing trajectory data and the actual electromagnetic transient timing trajectory data is calculated using the loss function. The mapping error is back-propagated based on the gradient descent algorithm of machine learning to update the network parameters.

8. According to a method for mapping the electromechanical-electromagnetic simulation trajectory of a new energy power grid based on a residual long short-term memory network according to claim 1, it is characterized in that: The loss function used in the regression layer of the time series trajectory mapping model is the HuberLoss loss function. When the prediction deviation is greater than the threshold parameter, the square error is used, and when the prediction deviation is less than the threshold parameter, the linear error is used.

9. A new energy power grid electromechanical-electromagnetic simulation trajectory mapping system based on residual long short-term memory network, characterized in that: include: Data acquisition module: used to digitally simulate the electromechanical transient process and electromagnetic transient process of the high-proportion new energy power grid system, obtain electromechanical and electromagnetic transient time series trajectory samples, and pre-process the electromechanical and electromagnetic transient time series trajectory samples; Model construction module: used to construct a time series trajectory mapping model based on residual long short-term memory network. The model input is the electromechanical transient time series trajectory, and the output is the electromagnetic transient time series trajectory. The time series trajectory mapping model includes an input layer, several LSTM layers, a fully connected layer and a regression layer, and a residual network introduced between the input of the first LSTM layer and the output of the last LSTM layer; Model training module: used to train the timing trajectory mapping model using the preprocessed electromechanical and electromagnetic transient timing trajectory samples to achieve mapping from the electromechanical transient simulation timing trajectory to the electromagnetic transient simulation timing trajectory.

10. A computer device, characterized in that: The computer device comprises: one or more processors; A memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement a new energy power grid electromechanical-electromagnetic simulation trajectory mapping method based on a residual long short-term memory network as described in any one of claims 1-8.

Images