Method and system for transient stability assessment of new energy system based on machine learning and data driving

Abstract

This invention discloses a method and system for transient stability assessment of new energy systems based on machine learning and data-driven approaches. It includes: performing time-domain simulations of electromechanical and electromagnetic transient models of the new energy system to obtain a sample set; constructing a system stability classification model based on a long short-term memory network; training the system stability classification model using a data-driven method; and achieving the assessment of electromagnetic transient stability through monitoring curves from electromechanical transient simulations. This invention establishes a mapping relationship between electromechanical and electromagnetic time scales through data-driven approaches, enabling the prediction of system stability at the electromagnetic transient scale based solely on the system's operation at the electromechanical transient scale. This solves the dual problems of low accuracy in electromechanical transient modeling of high-proportion new energy systems, which fails to accurately reflect real instability scenarios, and slow and inefficient electromagnetic transient simulation, thus achieving transient stability assessment of new energy systems.

Description

Technical Field

[0001] This invention belongs to the field of energy technology and relates to a method and system for evaluating the electromechanical-electromagnetic transient stability of power systems, and more particularly to a method and system for evaluating the transient stability of new energy systems based on machine learning and data-driven approaches. Background Technology

[0002] As the proportion of new energy sources such as wind power and photovoltaics in the power system continues to increase, the connection of a large number of power electronic devices has greatly increased the scale and complexity of the new energy power system. The component composition and dynamic process of the system are becoming increasingly complex, which puts forward higher requirements for the accuracy of component modeling and computing power in simulation.

[0003] For power systems with a high proportion of renewable energy (hereinafter referred to as renewable energy systems), traditional electromechanical transient simulation mainly reflects the system's operating status near the power frequency, and cannot simulate the high-frequency characteristics of power electronic equipment and its control loop, making it difficult to meet the needs of system transient stability analysis. Although electromagnetic transient simulation can more accurately reflect the dynamic characteristics of power electronic equipment, the nonlinearity and higher-order terms of the electromagnetic model greatly increase the computational load and time of the simulation, making it difficult to apply to the real-time simulation of large-scale power systems. To solve the dual problems of low accuracy in electromechanical transient modeling of renewable energy systems, which cannot accurately reflect real instability scenarios, and slow speed and low efficiency of electromagnetic transient simulation, this invention discloses a transient stability assessment method and system for renewable energy systems based on machine learning and data-driven approaches. Summary of the Invention

[0004] To address the aforementioned issues, this invention proposes a transient stability assessment method and system for new energy systems based on machine learning and data-driven approaches, thereby resolving the transient stability assessment problem for high-proportion new energy systems.

[0005] The technical solution adopted in this invention is as follows:

[0006] A transient stability assessment method for new energy systems based on machine learning and data-driven approaches includes the following steps:

[0007] Time-domain simulations were performed on electromechanical and electromagnetic transient models of power systems with a high proportion of new energy sources. Typical operating conditions and fault scenarios were designed, and a reliable sample set for electromechanical and electromagnetic transient time-domain simulations was constructed.

[0008] A system stability classification model based on a long short-term memory network (LSTM) is constructed. The model input is the electromechanical transient simulation monitoring curve, and the model output is the electromagnetic transient simulation stability classification probability value.

[0009] The reliable sample set of electromechanical and electromagnetic transient time-domain simulations is preprocessed, and the system stability classification model is trained using the preprocessed sample set based on a data-driven method, so as to realize the evaluation of electromagnetic transient stability by monitoring the electromechanical transient simulation curves.

[0010] Furthermore, the electromagnetic transient model adopts a three-phase model, wherein the computational element model is described by three-phase instantaneous values ​​(abc), and the model includes detailed modeling of the phase-locked loop control structure and the power electronic converter; the electromechanical transient model adopts a phasor model, wherein the computational element model is described by the fundamental phasor, and the model ignores the fast transient processes of converter switching dynamics and phase-locked loop dynamics.

[0011] Furthermore, the design of typical operating conditions and fault scenarios, and the construction of a reliable time-domain simulation sample set, specifically includes: considering the system load rate and the output of new energy sources, designing a typical operating condition set for the new energy system; considering the fault type and fault location, designing a typical fault set for the new energy system; performing electromechanical transient and electromagnetic transient simulations on the new energy system respectively, traversing all feasible combinations of the typical operating condition set and the typical fault set, and obtaining a sample set of electromechanical transient simulation monitoring curves and a sample set of electromagnetic transient simulation stability results.

[0012] Furthermore, the system stability classification model includes an LSTM network consisting of multiple LSTM units connected in a chain structure.

[0013] Furthermore, the preprocessing of the reliable sample set for electromechanical and electromagnetic transient time-domain simulations specifically includes:

[0014] The electromechanical transient simulation curve samples are segmented, retaining data for a period of time before and after the fault, and then normalized. The electromagnetic transient simulation stability result samples are converted into Boolean type, where the system stability result corresponds to 0 and the system instability result corresponds to 1.

[0015] Furthermore, the step of training the system stability classification model using the preprocessed sample set based on a data-driven method specifically involves:

[0016] During the forward propagation process, the system stability features in the input electromechanical transient simulation monitoring curve are extracted through the LSTM network, and the stability features are mapped to a one-dimensional classification probability value through the linear transformation of the fully connected layer. The fully connected layer uses the sigmoid activation function to make the classification probability value between 0 and 1.

[0017] During backpropagation, the Adam optimizer with weight decay is used to calculate the classification error based on the binary cross-entropy (BCELoss) loss function. The backpropagation algorithm based on gradient descent is used to backpropagate the classification error to calculate the parameter gradient and update the model parameters.

[0018] Furthermore, the Adam optimizer with weight decay specifically includes: updating model parameters using an adaptive Adam optimizer, adding a weight decay term when calculating the gradient, adjusting the learning rate of each parameter by calculating the first moment estimate and second moment estimate of the gradient, calculating the update amount of each parameter based on the learning rate, and applying the update amount to update the parameters.

[0019] Furthermore, the binary classification cross-entropy loss function is:

[0020] L = -ylog(y p )-(1-y)log(1-y p )

[0021] Where L is the classification error of the model, and y is the actual electromagnetic transient simulation stability result. p This represents the electromagnetic transient simulation stability classification probability value output by the model.

[0022] Furthermore, the output of the system stability classification model is a classification probability value between 0 and 1. The classification probability value is compared with a preset threshold. When the classification probability value is lower than the threshold, the evaluation result is that the system is stable; when it is higher than the threshold, the evaluation result is that the system is unstable.

[0023] A transient stability assessment system for new energy systems based on machine learning and data-driven methods, comprising:

[0024] Data acquisition module: used to perform time-domain simulations of electromechanical transient models and electromagnetic transient models of power systems with a high proportion of new energy sources, design typical operating conditions and fault scenarios, and construct a reliable sample set for electromechanical and electromagnetic transient time-domain simulations;

[0025] Model building module: used to build a system stability classification model based on long short-term memory network. The model input is the electromechanical transient simulation monitoring curve, and the model output is the electromagnetic transient simulation stability classification probability value.

[0026] Model training module: used to preprocess the reliable sample set of electromechanical and electromagnetic transient time-domain simulation, and use the preprocessed sample set to train the system stability classification model based on the data-driven method, so as to realize the evaluation of electromagnetic transient stability by monitoring the electromechanical transient simulation curve.

[0027] The beneficial effects of this invention are:

[0028] This invention constructs a system stability classification model based on machine learning. By establishing a data-driven mapping relationship between electromechanical and electromagnetic time scales, it enables the prediction of system stability at the electromagnetic transient scale based solely on the system's operation at the electromechanical transient scale. This addresses, to some extent, the dual challenges of low accuracy in electromechanical transient modeling of high-proportion renewable energy systems, which fails to accurately reflect real instability scenarios, and slow and inefficient electromagnetic transient simulation. This allows for transient stability assessment of renewable energy systems. This invention can provide a foundation for analyzing the transient stability physical mechanisms and dominant instability characteristics of high-proportion renewable energy systems. Attached Figure Description

[0029] Figure 1 This is a flowchart illustrating the training and prediction process of the system stability classification model in this embodiment of the invention.

[0030] Figure 2 This is a structural diagram of an LSTM unit in an embodiment of the present invention.

[0031] Figure 3 This is a flowchart of the preprocessing process in an embodiment of the present invention.

[0032] Figure 4 This is a topology diagram of a new energy system in an embodiment of the present invention.

[0033] Figure 5 This is an error curve diagram of the model training process in an embodiment of the present invention. Detailed Implementation

[0034] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0035] A transient stability assessment method for new energy systems based on machine learning and data-driven approaches includes the following steps:

[0036] (1) Time-domain simulations were performed on the electromechanical transient model and electromagnetic transient model of the power system with a high proportion of new energy, respectively. Typical operating conditions and fault scenarios were designed, and a reliable sample set of electromechanical and electromagnetic transient time-domain simulations was constructed.

[0037] The electromagnetic transient model adopts a three-phase model, in which the computational element model is described by the instantaneous values ​​of the three phases abc. The model includes detailed modeling of control structures such as phase-locked loops and power electronic converters. The electromechanical transient model adopts a phasor model, in which the computational element model is described by the fundamental phasor. The model ignores fast transient processes such as converter switching dynamics and phase-locked loop dynamics.

[0038] The design typical operating conditions and fault scenarios are used to construct a reliable sample set for time-domain simulation, specifically including:

[0039] 1) Considering factors such as system load rate and new energy output, design a typical operating condition set for the new energy system;

[0040] 2) Considering factors such as fault type and fault location, design a typical fault set for the new energy system;

[0041] 3) Perform electromechanical and electromagnetic transient simulations on the new energy system respectively, traverse all feasible combinations of typical operating condition sets and typical fault sets, and obtain sample sets of electromechanical transient simulation monitoring curves and electromagnetic transient simulation stability results.

[0042] The specific steps for obtaining electromechanical transient simulation monitoring curves include:

[0043] Based on the electromechanical transient model of the new energy system, electromechanical transient simulation is run; system operating data such as power angle, voltage, and frequency at each moment are recorded as monitoring curves. The electromechanical transient simulation monitoring curves include, but are not limited to: the generator's maximum power angle difference, the system's minimum / maximum bus voltage, and the system's minimum / maximum bus frequency.

[0044] The specific steps for obtaining the stable results of electromagnetic transient simulation include:

[0045] Based on the electromagnetic transient model of the new energy system, electromagnetic transient simulation is run; it is determined whether the new energy system can maintain stable operation after being subjected to fault disturbance. The stability criteria used include power angle, voltage and frequency criteria. If any instability criterion is triggered, the system is determined to be unstable, otherwise the system is stable.

[0046] Specifically, the power angle instability criterion refers to the generator's maximum power angle difference exceeding a certain angle; the voltage instability criterion refers to the system voltage failing to recover to the allowable range after a fault; and the frequency instability criterion refers to the system frequency failing to recover to the allowable range after a fault. In addition, system stability criteria may also include criteria for new energy source disconnection or repeated low-voltage ride-through, commutation failure, and DC blocking.

[0047] (2) Construct a system stability classification model based on Long Short-Term Memory (LSTM) network. The model input is the electromechanical transient simulation monitoring curve X = [x(1), ..., x(T)], and the model output is the electromagnetic transient simulation stability classification probability value, denoted by Y. p ∈(0,1) represents.

[0048] The system stability classification model includes an LSTM network composed of multiple LSTM units connected in a chain structure, as shown in the LSTM unit structure diagram. Figure 2 As shown, by introducing forget gate, input gate and output gate, the problem of long-term dependence and short-term memory can be effectively handled. It is suitable for processing time series data. The weight parameters and biases of the LSTM network are determined through model training.

[0049] (3) The reliable sample set of electromechanical and electromagnetic transient time-domain simulations is preprocessed. Using the preprocessed sample set, a data-driven method is used to train the system stability classification model, enabling the evaluation of electromagnetic transient stability through electromechanical transient simulation monitoring curves. The model training and prediction flowchart is shown below. Figure 1 As shown.

[0050] like Figure 3 As shown, the preprocessing of the reliable sample set for electromechanical and electromagnetic transient time-domain simulations specifically includes:

[0051] 1) The electromechanical transient simulation curve sample is segmented, retaining only the data for a period of time before and after the fault; for multiple monitoring curves in the same sample, one or more of them are selected as input;

[0052] 2) Convert the electromagnetic transient simulation stability results samples into Boolean type, where the system stability result corresponds to 0 (False) and the system instability result corresponds to 1 (True);

[0053] 3) Normalization: The samples are normalized using standardization or maximum-minimum normalization methods; electromagnetic transient samples do not require normalization since they have already been converted to Boolean type.

[0054] 4) Sample set partitioning: Randomly shuffle the samples and divide them into training set, test set and validation set according to the proportions;

[0055] 5) Data type conversion: Convert data types to CPU tensors or GPU tensors depending on the machine learning environment (using CPU or GPU).

[0056] The process of training the system stability classification model using a data-driven method based on the preprocessed sample set specifically involves:

[0057] During the forward propagation, the system stability characteristics in the input electromechanical transient simulation monitoring curves are extracted using an LSTM network. The stability features are mapped to one-dimensional classification probability values ​​through a linear transformation of a fully connected layer. This fully connected layer employs a sigmoid activation function, ensuring that the classification probability value Y... p Between 0 and 1;

[0058] During backpropagation, the Adam optimizer with weight decay is introduced. The classification error is calculated using the binary cross-entropy loss function BCELoss. The backpropagation algorithm based on gradient descent is used to backpropagate the classification error to calculate the parameter gradient and update the model parameters.

[0059] The Adam (Adaptive Momentum) optimizer with introduced weight decay specifically includes: updating model parameters using an adaptive Adam optimizer; adding a weight decay term when calculating the gradient; adjusting the learning rate of each parameter by calculating the first and second moment estimates of the gradient; calculating the update amount of each parameter based on the learning rate; and applying the update amount to update the model parameters. The adaptive learning rate makes the Adam optimizer more flexible in handling different parameter updates; adding the weight decay term to the gradient of the Adam optimizer enhances the model's generalization ability and solves the overfitting problem.

[0060] The binary cross-entropy loss function is used to evaluate the difference between the model output and the actual electromagnetic simulation results. The calculation formula is as follows:

[0061] L = -ylog(y p )-(1-y)log(1-y p )

[0062] Where L is the classification error of the model, and y is the actual electromagnetic transient simulation stability result. p This represents the electromagnetic transient simulation stability classification probability value output by the model.

[0063] After the model is trained, the trained system stability classification model is used to predict the stability of the current system under electromagnetic transient scales, including:

[0064] 1) Input the electromechanical transient simulation monitoring curve to be evaluated into the trained system stability classification model, extract features through an n-layer LSTM network, map the features into a one-dimensional result through a fully connected layer, and ensure that the output is between 0 and 1 through a sigmoid activation function;

[0065] 2) Compare the model output with the preset threshold to determine the stability of the current system under the electromagnetic transient scale: if it is below the threshold, the evaluation result is that the system is stable; if it is above the threshold, the evaluation result is that the system is unstable.

[0066] This invention discloses a method and system for transient stability assessment of new energy systems based on machine learning and data-driven approaches. The invention constructs a system stability classification model based on machine learning and establishes a mapping relationship between electromechanical and electromagnetic time scales through data-driven approaches. This enables the prediction of system stability at the electromagnetic transient scale based solely on the system's operation at the electromechanical transient scale. This addresses, to some extent, the dual challenges of low accuracy in electromechanical transient modeling of high-proportion new energy systems, which fails to accurately reflect real instability scenarios, and slow and inefficient electromagnetic transient simulation. This invention provides a foundation for analyzing the transient stability physical mechanisms and dominant instability characteristics of high-proportion new energy systems.

[0067] A specific embodiment of the present invention is as follows:

[0068] In such Figure 4 The effectiveness of the proposed method is verified in a high-proportion AC / DC hybrid system of renewable energy sources. The system includes six wind farms and six photovoltaic power plants, with a total renewable energy output of 1800MW. Multiple operating conditions are randomly generated based on load curves and renewable energy unit output curves to form a typical operating condition set for the system. Power flow calculations are used to determine the output of each synchronous generator and renewable energy equipment. Three-phase short-circuit faults are set at different lines to form a typical fault set for the system.

[0069] Electromechanical and electromagnetic transient simulations were performed on the system under various operating conditions and fault scenarios. A three-phase short-circuit fault occurred at t=1s, and the circuit breaker tripped 0.1s later, disconnecting the faulty line. For electromechanical simulation, the maximum power angle difference, minimum bus voltage, and maximum bus frequency of the system at each moment were recorded, with a sampling step size of ΔT=10ms. For electromagnetic simulation, the system stability was determined using the criteria shown in Table 1. Triggering any instability criterion constituted system instability, while the system was considered stable otherwise. All feasible combinations of the typical operating condition set and typical fault set of the system were traversed to construct a reliable sample set for time-domain simulation.

[0070] Table 1 Criteria for System Instability

[0071]

[0072] The sample data is preprocessed by extracting the portion from t=0 to 3s from the electromechanical transient data; the electromagnetic transient simulation results are converted to Boolean type, with 0 (False) corresponding to system stability and 1 (True) corresponding to system instability; the data is processed using the maximum-minimum normalization method; the samples are randomly shuffled and divided into training, testing and validation sets in an 8:1:1 ratio; since the machine learning environment is GPU, the data type is converted to GPU tensor.

[0073] The model structure parameters and training parameters are shown in Table 2. During forward propagation, the LSTM network extracts system stability features from the input data and maps these features to a one-dimensional classification result through a linear transformation of the fully connected layer. A sigmoid activation function is used to ensure the output result is between 0 and 1. During backpropagation, a binary cross-entropy loss function is used to calculate the classification error. A gradient descent-based backpropagation algorithm is employed to backpropagate the classification error to calculate the parameter gradient and update the model parameters. The model is trained using a data-driven approach. The training set is used as input, and the validation set is used to evaluate the training performance after each training round. The Adam optimizer is used, and weight decay is introduced to improve the model's generalization ability. The weight decay coefficient is set to 1e-5.

[0074] Table 2 LSTM Model Structure Parameters and Training Parameters

[0075]

[0076]

[0077] The transient stability of the embodiments was evaluated using the method of the present invention, and the results are as follows:

[0078] Figure 5 Error curves for the training process of the system stability classification model are presented. The results show that both the training set error and the validation set error decrease with the number of iterations, eventually reaching a low level, verifying the effectiveness of the model training. The trained model was evaluated, achieving a training set accuracy of 99.18% and a test set accuracy of 96.67%, where accuracy refers to the percentage of correctly classified samples relative to the total number of samples. The model's accuracy under different system load rates is shown in Table 3. The results demonstrate that the method designed in this invention can accurately evaluate the transient stability of a system on an electromagnetic scale and possesses a certain degree of generalization ability.

[0079] Table 3. Model accuracy under different system load rates

[0080]

[0081] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied 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.

[0082] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0083] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0084] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0085] The above specific embodiments are used to explain and illustrate the present invention, but not to limit the present invention. Any modifications and changes made to the present invention within the spirit and scope of the claims shall fall within the protection scope of the present invention.

Claims

1. A transient stability assessment method for a new energy system based on machine learning and data-driven approaches, characterized in that, Includes the following steps: Time-domain simulations were performed on electromechanical and electromagnetic transient models of power systems with a high proportion of new energy sources. Typical operating conditions and fault scenarios were designed, and a reliable sample set for electromechanical and electromagnetic transient time-domain simulations was constructed. A system stability classification model based on a long short-term memory network is constructed. The model input is the electromechanical transient simulation monitoring curve, and the model output is the electromagnetic transient simulation stability classification probability value. The reliable sample set of electromechanical and electromagnetic transient time-domain simulation is preprocessed, and the system stability classification model is trained based on the data-driven method using the preprocessed sample set, so as to realize the evaluation of electromagnetic transient stability by monitoring the electromechanical transient simulation curve; The specific steps for obtaining electromechanical transient simulation monitoring curves include: running electromechanical transient simulation based on the electromechanical transient model of the new energy system; recording the power angle, voltage, and frequency system operating data at each moment as monitoring curves; the electromechanical transient simulation monitoring curves include, but are not limited to: the generator's maximum power angle difference, the system's minimum / maximum bus voltage, and the system's minimum / maximum bus frequency. The specific steps to obtain the stability results of electromagnetic transient simulation include: running electromagnetic transient simulation based on the electromagnetic transient model of the new energy system; determining whether the new energy system can maintain stable operation after being subjected to fault disturbance. The stability criteria used include power angle, voltage and frequency criteria. Triggering any one of the instability criteria is considered as system instability, otherwise the system is stable. Preprocessing the reliable sample set of electromechanical transient time-domain simulation includes: cutting the electromechanical transient simulation curve sample and retaining only the data for a period of time before and after the fault; for multiple monitoring curves in the same sample, selecting one or several of them as input.

2. The transient stability assessment method for a new energy system based on machine learning and data-driven approaches according to claim 1, characterized in that, The electromagnetic transient model adopts a three-phase model, wherein the computational element model adopts... abc The three-phase instantaneous value description includes detailed modeling of the phase-locked loop control structure and the power electronic converter; the electromechanical transient model adopts the phasor model, wherein the computational element model adopts the fundamental phasor description, and the model ignores the fast transient processes of converter switching dynamics and phase-locked loop dynamics.

3. The transient stability assessment method for a new energy system based on machine learning and data-driven approaches according to claim 1, characterized in that, The design includes typical operating conditions and fault scenarios, and the construction of a reliable time-domain simulation sample set. Specifically, this includes: considering the system load rate and the output of new energy sources, designing a typical operating condition set for the new energy system; considering the fault type and fault location, designing a typical fault set for the new energy system; performing electromechanical transient and electromagnetic transient simulations on the new energy system respectively, traversing all feasible combinations of the typical operating condition set and the typical fault set, and obtaining a sample set of electromechanical transient simulation monitoring curves and a sample set of electromagnetic transient simulation stability results.

4. The transient stability assessment method for a new energy system based on machine learning and data-driven approaches according to claim 1, characterized in that, The system stability classification model includes an LSTM network consisting of multiple LSTM units connected in a chain structure.

5. The transient stability assessment method for a new energy system based on machine learning and data-driven approaches according to claim 3, characterized in that, The reliable sample set for electromagnetic transient time-domain simulation is preprocessed as follows: The electromagnetic transient simulation stability result sample is converted into a Boolean type, where the system stability result corresponds to 0 and the system instability result corresponds to 1.

6. The transient stability assessment method for a new energy system based on machine learning and data-driven approaches according to claim 1, characterized in that, The process of training the system stability classification model using a data-driven method based on the preprocessed sample set specifically involves: During the forward propagation process, the system stability features in the input electromechanical transient simulation monitoring curve are extracted through the LSTM network, and the stability features are mapped to a one-dimensional classification probability value through the linear transformation of the fully connected layer. The fully connected layer uses the sigmoid activation function to make the classification probability value between 0 and 1. During backpropagation, the Adam optimizer with weight decay is used to calculate the classification error based on the binary cross-entropy loss function. The backpropagation algorithm based on gradient descent is used to backpropagate the classification error to calculate the parameter gradient and update the model parameters.

7. The transient stability assessment method for a new energy system based on machine learning and data-driven approaches according to claim 6, characterized in that, The Adam optimizer with introduced weight decay specifically includes: updating model parameters using an adaptive Adam optimizer, adding a weight decay term when calculating the gradient, adjusting the learning rate of each parameter by calculating the first moment estimate and second moment estimate of the gradient, calculating the update amount of each parameter based on the learning rate, and applying the update amount to update the model parameters.

8. The transient stability assessment method for a new energy system based on machine learning and data-driven methods according to claim 6, characterized in that, The binary classification cross-entropy loss function is: ， Where L is the classification error of the model, y The results are based on actual electromagnetic transient simulation stability. y p This represents the electromagnetic transient simulation stability classification probability value output by the model.

9. The transient stability assessment method for a new energy system based on machine learning and data-driven approaches according to claim 1, characterized in that, The output of the system stability classification model is a classification probability value between 0 and 1. The classification probability value is compared with a preset threshold. When the classification probability value is lower than the threshold, the evaluation result is that the system is stable, and when it is higher than the threshold, the evaluation result is that the system is unstable.

10. A transient stability assessment system for a new energy system based on machine learning and data-driven methods, used to implement the method as described in any one of claims 1-9, characterized in that, include: Data acquisition module: used to perform time-domain simulations of electromechanical transient models and electromagnetic transient models of power systems with a high proportion of new energy sources, design typical operating conditions and fault scenarios, and construct a reliable sample set for electromechanical and electromagnetic transient time-domain simulations; Model building module: used to build a system stability classification model based on long short-term memory network. The model input is the electromechanical transient simulation monitoring curve, and the model output is the electromagnetic transient simulation stability classification probability value. Model training module: used to preprocess the reliable sample set of electromechanical and electromagnetic transient time-domain simulation, and use the preprocessed sample set to train the system stability classification model based on the data-driven method, so as to realize the evaluation of electromagnetic transient stability by monitoring the electromechanical transient simulation curve.

Images