Simulation test method compatible with multiple pumped storage power stations and computer device

Abstract

The invention discloses a simulation test method compatible with multiple pumped storage power stations and a computer device. The method comprises the following steps: constructing a multi-compatibility simulation model adaptive to different types of pumped storage power stations; the multi-compatibility simulation model comprises an unsupervised learning model and a semi-supervised learning model; training the unsupervised learning model, and taking the trained model as a first model; the unsupervised learning model comprises an initial generator and an initial discriminator; training the semi-supervised learning model, and taking the trained model as a second model; simulating operation behaviors under various assumed working conditions through the first model to obtain operation data of the target power station under various assumed working conditions; and inputting the operation data into the second model to obtain simulation operation information of the target power station under various assumed working conditions. According to the method, comprehensive, accurate and efficient simulation testing of the target power station can be realized, diversified requirements are met, and reliable decision support can be provided for regulation and control of the power station.

Description

Technical Field

[0001] The present disclosure relates to the technical field of simulation testing of pumped-storage power stations, and in particular to a simulation testing method and a computer device compatible with multiple pumped-storage power stations. Background Art

[0002] As a highly efficient way of storing energy, pumped-storage power stations work by using excess electricity to pump water to high-level reservoirs when electricity demand is low, and releasing the stored water to generate electricity when electricity demand peaks. This energy conversion process can not only balance the supply and demand of the power grid, but also improve the flexibility and stability of the entire power system.

[0003] However, the simulation test methods disclosed in the prior art are usually designed for specific types of power plant models, lacking sufficient flexibility and universality, and it is difficult to comprehensively evaluate the performance of the power plant under various operating conditions, which limits the accurate evaluation and optimization of the overall performance of the power plant. Summary of the invention

[0004] In view of this, the embodiments of the present disclosure provide a simulation test method and computer device that are compatible with multiple pumped-storage power stations, which can solve the problems existing in the prior art of poor accuracy, low efficiency, poor adaptability, poor flexibility, and inability to meet the diverse needs of pumped-storage power stations in the simulation test of the overall performance of pumped-storage power stations.

[0005] In a first aspect, the embodiment of the present disclosure provides a simulation test method compatible with multiple pumped storage power stations, which adopts the following technical solution:

[0006] Constructing a multi-compatibility simulation model; the multi-compatibility simulation model includes an unsupervised learning model and a semi-supervised learning model, and the multi-compatibility simulation model is suitable for different types of pumped storage power stations;

[0007] Training the unsupervised learning model, and using the trained unsupervised learning model as the first model; the unsupervised learning model includes an initial generator and an initial discriminator;

[0008] The semi-supervised learning model is trained, and the trained semi-supervised learning model is used as YY+232025P

[0009] For the second model;

[0010] Simulating the operating behaviors under various assumed working conditions by using the first model to obtain the operating data of the target power station under various assumed working conditions;

[0011] The operating data of the target power station under various assumed working conditions are input into the second model to obtain simulated operating information of the target power station under various assumed working conditions.

[0012] Optionally, the training of the unsupervised learning model and using the trained unsupervised learning model as the first model includes:

[0013] Preprocessing the collected historical operation data of the target power station to obtain a first training set;

[0014] The initial generator is trained based on the first training set, and the trained initial generator is used as the first generator; the virtual data generated by the first generator matches the operating characteristics of the target power station;

[0015] Training the initial discriminator based on the virtual data generated by the first generator and the real data from the corresponding operating conditions in the target power station, and using the initial discriminator after iterative training that meets the preset indicators as the target discriminator;

[0016] Input the virtual data generated by the first generator into the target discriminator, analyze the first generator according to the output information of the target discriminator and the preset loss function, and obtain the gradient information of the loss relative to the parameters of the first generator;

[0017] Update the network parameters of the first generator based on the gradient information;

[0018] Iteratively train the updated first generator, and use the first generator that meets the training conditions as the target generator;

[0019] The first model includes the target generator and the target discriminator.

[0020] Optionally, the preset indicators include one or more of an accuracy indicator, a precision indicator, a recall rate indicator, an F1 score indicator, a ROC curve indicator, and an AUC curve indicator.

[0021] Optionally, the training of the semi-supervised learning model and using the trained semi-supervised learning model as the second model includes:

[0022] Acquire key state information and state transition trend analysis information in the operation of the target power station; the key state information includes one or more of normal operation information, maintenance state information, and abnormal state information; the state transition trend analysis information includes transfer path information of transfer trends between different operating states, transferred state occurrence condition information, and transferred state occurrence probability information;

[0024] Constructing a second training set based on the key state information and the state transition trend analysis information;

[0025] The second training set is preprocessed, the semi-supervised learning model is trained according to the preprocessed second training set and a preset iterative algorithm, and the trained semi-supervised learning model is used as the second model.

[0026] In a second aspect, the embodiment of the present disclosure provides a simulation test method compatible with multiple pumped storage power stations, which adopts the following technical solution:

[0027] Build multi-compatibility simulation models;

[0028] Using a genetic algorithm to adjust the multi-compatibility simulation model to obtain an optimized multi-compatibility simulation model;

[0029] The optimized multi-compatibility simulation model includes an unsupervised learning model and a semi-supervised learning model, and the multi-compatibility simulation model is suitable for different types of pumped storage power stations;

[0030] Training the unsupervised learning model, and using the trained unsupervised learning model as the first model; the unsupervised learning model includes an initial generator and an initial discriminator;

[0031] Training the semi-supervised learning model, and using the trained semi-supervised learning model as the second model;

[0032] Simulating the operating behaviors under various assumed working conditions by using the first model to obtain the operating data of the target power station under various assumed working conditions;

[0033] The operating data of the target power station under various assumed working conditions are input into the second model to obtain simulated operating information of the target power station under various assumed working conditions.

[0034] Optionally, the adopting a genetic algorithm to adjust the multi-compatibility simulation model to obtain an optimized multi-compatibility simulation model includes:

[0035] Determine the decision variables that need to be adjusted in the multi-compatibility simulation model; the decision variables include one or more of power plant operation parameters, scheduling strategies, target equipment status, resource allocation strategies, task allocation strategies, target parameter optimization strategies, and environmental variables;

[0036] Construct a random key sequence;

[0037] Determine the mapping of random keys to the actual decision variable value space and fitness function;

[0038] Randomly generate a preset number of initial individuals to form an initial population;

[0039] The initial individuals are composed of random key sequences, and each of the initial individuals represents a potential YY+232025P

[0040] Solution:

[0041] Performing genetic operation processing on the initial individual, determining whether the processed initial individual meets the preset conditions, and if not, repeating the genetic operation processing until the processed initial individual meets the preset conditions, thereby obtaining the optimized multi-compatibility simulation model;

[0042] The genetic operation includes a selection operation, a crossover operation, and a mutation operation; the preset conditions include reaching a maximum number of iterations or convergence of fitness values.

[0043] Optionally, the constructing of the random key sequence includes: assigning a random key to each decision variable; each random key includes a value range of a corresponding single decision variable;

[0044] The unsupervised learning model is a generative adversarial network model; the semi-supervised learning model is a hidden Markov model.

[0045] In a third aspect, the embodiment of the present disclosure further provides a simulation test system compatible with multiple pumped storage power stations, including:

[0046] A construction module, used to construct a multi-compatibility simulation model; the multi-compatibility simulation model includes an unsupervised learning model and a semi-supervised learning model, and the multi-compatibility simulation model is suitable for different types of pumped storage power stations;

[0047] A first training module, used to train the unsupervised learning model, and use the trained unsupervised learning model as the first model; the unsupervised learning model includes an initial generator and an initial discriminator;

[0048] A second training module is used to train the semi-supervised learning model and use the trained semi-supervised learning model as a second model;

[0049] A simulation module, used to simulate the operation behaviors under various assumed working conditions through the first model, and obtain the operation data of the target power station under various assumed working conditions;

[0050] The acquisition module is used to input the operating data of the target power station under various assumed working conditions into the second model to obtain simulated operating information of the target power station under various assumed working conditions.

[0051] In a fourth aspect, the embodiment of the present disclosure further provides a simulation test system compatible with multiple pumped storage power stations, including:

[0052] A building block for building a multi-compatible simulation model;

[0053] an optimization module, configured to adjust the multi-compatibility simulation model by using a genetic algorithm to obtain an optimized multi-compatibility simulation model; the optimized multi-compatibility simulation model includes an unsupervised learning model and a semi-supervised learning model, and the multi-compatibility simulation model is suitable for different types of pumped-storage power stations;

[0054] A first training module, used to train the unsupervised learning model, and use the trained unsupervised learning model as the first model; the unsupervised learning model includes an initial generator and an initial discriminator;

[0055] A second training module is used to train the semi-supervised learning model and use the trained semi-supervised learning model as a second model;

[0056] A simulation module, used to simulate the operation behaviors under various assumed working conditions through the first model, and obtain the operation data of the target power station under various assumed working conditions;

[0057] The acquisition module is used to input the operating data of the target power station under various assumed working conditions into the second model to obtain simulated operating information of the target power station under various assumed working conditions.

[0058] In a fifth aspect, the embodiments of the present disclosure further provide a computer device, which adopts the following technical solution:

[0059] The computer device comprises:

[0060] at least one processor; and,

[0061] a memory communicatively connected to the at least one processor; wherein,

[0062] The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute any of the above-mentioned simulation test methods for compatible multiple pumped-storage power stations.

[0063] In a sixth aspect, an embodiment of the present disclosure further provides a computer-readable storage medium storing computer instructions for enabling a computer to execute any of the above-described simulation test methods for compatible multiple pumped-storage power stations.

[0064] In a seventh aspect, an embodiment of the present disclosure further provides a computer program product, comprising a computer program / instruction, which implements the steps of any of the above methods when executed by a processor.

[0065] The simulation test method compatible with multiple pumped storage power stations disclosed in this application can more comprehensively test and evaluate different types of pumped storage power stations, improve the adaptability and versatility of the simulation model, have good flexibility, and good data processing effect, and can effectively improve the accuracy of the test results, the efficiency of the test process, and the calculation speed and accuracy of the simulation test, thereby providing YY+232025P for the design, operation and maintenance of the power station.

[0066] Provide more reliable decision support.

[0067] Specifically, by constructing a multi-compatibility simulation model and combining unsupervised learning models and semi-supervised learning models, it is possible to adapt to pumped-storage power stations of different types and sizes and realize simulation testing and analysis of different power stations; by training unsupervised learning models and semi-supervised learning models, it is possible to better utilize data resources, improve model training efficiency and performance, and improve the model's understanding and prediction capabilities of power station behavior; by simulating the power station operation behavior under various assumed working conditions using the first model, the operation data of the power station under different conditions can be obtained, which is helpful for analyzing and evaluating the performance of the power station under various simulated working conditions; by inputting the simulated power station operation data into the second model, the simulated operation information of the power station under various assumed working conditions can be obtained, including prediction results, anomaly detection, etc., to provide a reference for decision-making; the simulated operation information obtained by this method can provide an important reference for power station operation management and decision-making, help optimize the operation strategy and scheduling plan of the power station, and improve the operation efficiency, operation performance and stability of the power station; using the simulation test method, testing and analysis can be carried out in a virtual environment, reducing the risks and costs that may be brought about in the actual test process, and improving the test efficiency and reliability.

[0068] The method disclosed in this application is more flexible and efficient, and is crucial to improving the performance evaluation and optimization of pumped-storage power stations. This can not only promote the efficient operation of the power stations, but also make important contributions to the stability and sustainable development of the power grid.

[0069] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the following preferred embodiments are specifically cited and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0071] Figure 1 A schematic flow chart of a simulation test method compatible with multiple pumped-storage power stations provided in an embodiment of the present disclosure.

[0072] Figure 2 A schematic flowchart of a method for obtaining a first model provided in an embodiment of the present disclosure.

[0073] Figure 3 A flowchart of a method for obtaining a second model provided in an embodiment of the present disclosure.

[0074] Figure 4 A schematic flow chart of another embodiment of a simulation test method compatible with multiple pumped-storage power stations provided in the present application.

[0075] Figure 5 A flowchart of a method for adjusting multiple compatibility simulation models provided in an embodiment of the present disclosure.

[0076] Figure 6 A principle block diagram of a simulation test system level compatible with multiple pumped-storage power stations provided in an embodiment of the present disclosure.

[0077] Figure 7 A schematic diagram of the structure of a computer device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0078] The embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.

[0079] It should be clear that the following embodiments of the present disclosure are described by specific specific examples, and those skilled in the art can easily understand other advantages and effects of the present disclosure from the contents disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all of the embodiments. The present disclosure can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present disclosure. It should be noted that the following embodiments and features in the embodiments can be combined with each other in the absence of conflict. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present disclosure.

[0080] It should be noted that various aspects of the embodiments within the scope of the appended claims are described below. It should be apparent that the aspects described herein may be embodied in a wide variety of forms, and any specific structure and / or function described herein is merely illustrative. Based on the present disclosure, it should be understood by those skilled in the art that an aspect described herein may be implemented independently of any other aspect, and two or more of these aspects may be combined in various ways. For example, any number of aspects described herein may be used to implement the device and / or practice the method. In addition, other structures and / or functionalities other than one or more of the aspects described herein may be used to implement this device and / or practice this method.

[0081] It should also be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present disclosure. The drawings only show components related to the present disclosure rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0082] Additionally, in the following description, specific details are provided to facilitate a thorough understanding of the examples. However, it will be understood by those skilled in the art that the aspects described may be practiced without these specific details.

[0083] Reference Figure 1 In a first aspect, the present application discloses a simulation test method compatible with multiple pumped storage power stations, comprising:

[0084] S100, constructing a multi-compatibility simulation model; the multi-compatibility simulation model includes an unsupervised learning model and a semi-supervised learning model, and the multi-compatibility simulation model is suitable for different types of pumped-storage power stations.

[0085] In this step, the constructed multi-compatibility simulation model is adaptable to different types of pumped-storage power stations. The parameters can be adjusted dynamically and optimized according to the specific characteristics and operating conditions of different power stations. It has high flexibility and strong versatility. At the same time, a more comprehensive simulation test can be carried out for a single pumped-storage power station.

[0086] Specifically, data from different types of pumped-storage power stations can be collected to build simulation models suitable for various power stations.

[0087] In this embodiment, the multi-compatibility simulation model also includes a data input layer, a decision layer and a feedback adjustment layer, wherein the unsupervised learning model and the semi-supervised learning model are both connected to the data input layer signal, the decision layer is connected to the unsupervised learning model and the semi-supervised learning model signal, and the feedback adjustment layer is connected to the decision layer, the unsupervised learning model, and the semi-supervised learning model signal.

[0088] Among them, the data input layer is used to receive data from different types of pumped-storage power stations, including power station operation data, environmental data, equipment status and other data; the unsupervised learning model uses unsupervised learning algorithms to perform tasks such as data clustering, dimensionality reduction, and anomaly detection to help discover potential patterns and associations in the data, and provide support for subsequent decision-making and optimization; the semi-supervised learning model is used to use a small amount of labeled data and a large amount of unlabeled data for training, and learn the distribution and characteristics of the data through a semi-supervised learning algorithm to improve the generalization ability and prediction accuracy of the model; the decision-making layer is used to make decisions based on the output results of the model after the corresponding learning model training is completed, optimize the power station's operating parameters, scheduling strategies, resource allocation strategies, etc., to achieve optimization and improvement of the pumped-storage power station; the feedback adjustment layer is configured with a feedback mechanism, which is used to continuously adjust the parameters and performance of the model through real-time monitoring and data updates to ensure the continuous optimization and adaptability of the model.

[0089] Furthermore, the multi-compatibility simulation model also includes a feature extraction layer connected to the data input layer signal. The feature extraction layer is used to extract key feature vectors from the input data received by the data input layer through a feature extraction algorithm for subsequent learning and prediction, or to pre-process the input data received by the data input layer according to a preset processing method to obtain data that meets the requirements for use in subsequent learning model training.

[0090] S200, training an unsupervised learning model, and using the trained unsupervised learning model as the first model; the unsupervised learning model includes an initial generator and an initial discriminator.

[0091] The initial generator and initial discriminator are two neural networks, which compete and cooperate to generate data. The initial generator is responsible for creating fake data that looks like real data, while the task of the initial discriminator is to distinguish between real data and fake data generated by the initial generator; this architecture makes the model very effective in generating complex, high-dimensional and continuous data, especially for simulating the behavior of power plants under unknown or unobserved conditions.

[0092] The trained initial generator and initial discriminator can help understand and generate unlabeled data and improve data utilization.

[0093] Specifically, unsupervised learning algorithms, such as generative adversarial networks (GANs), can be used to train pumped storage power station data.

[0094] S300, training the semi-supervised learning model, and using the trained semi-supervised learning model as the second model.

[0095] Training a semi-supervised learning model can better utilize a small amount of labeled data and a large amount of unlabeled data and improve the generalization ability of the model.

[0096] S400, simulating operating behaviors under various assumed working conditions through a first model to obtain operating data of a target power station under various assumed working conditions.

[0097] Obtaining the operating data of the power plant under various assumed conditions can help test the performance and stability of the power plant in different scenarios and provide data support for subsequent analysis.

[0098] Specifically, the trained unsupervised learning model is used to generate and simulate data under various hypothetical working conditions.

[0099] S500: Inputting the operation data of the target power station under various assumed working conditions into the second model to obtain simulated operation information of the target power station under various assumed working conditions.

[0100] In this step, the output data of the first model is used as the input data of the second model, that is, the first model and the second model are integrated, which can improve the accuracy of the simulation model, thereby helping to optimize the operating efficiency and reliability of the power plant and achieve a more comprehensive and accurate power plant operation simulation.

[0101] The semi-supervised learning model is used to process the data to obtain simulation operation information with more generalization ability, which helps to evaluate the performance of the power plant more comprehensively.

[0102] The simulation test method compatible with multiple pumped-storage power stations disclosed in the present application can more comprehensively test and evaluate different types of pumped-storage power stations, improve the adaptability and versatility of the simulation model, have good flexibility, and good data processing effect, and can effectively improve the accuracy of the test results, the efficiency of the test process, and the calculation speed and precision of the simulation test, thereby providing more reliable decision support for the design, operation and maintenance of the power station; specifically, by constructing a multi-compatible simulation model, combined with an unsupervised learning model and a semi-supervised learning model, it can adapt to pumped-storage power stations of different types and sizes, and realize simulation testing and analysis of different power stations; by training unsupervised learning models and semi-supervised learning models, it can better utilize data resources, improve model training efficiency and performance, and improve the model's understanding and Predictive capability: By using the first model to simulate the operation behavior of the power plant under various assumed working conditions, the operation data of the power plant under different conditions can be obtained, which is helpful for analyzing and evaluating the performance of the power plant under various simulated working conditions; by inputting the simulated power plant operation data into the second model, the simulated operation information of the power plant under various assumed working conditions can be obtained, including prediction results, anomaly detection, etc., to provide a reference for decision-making; the simulated operation information obtained by this method can provide an important reference for power plant operation management and decision-making, help optimize the operation strategy and scheduling plan of the power plant, and improve the operation efficiency, operation performance and stability of the power plant; the simulation test method can be used to perform testing and analysis in a virtual environment, reduce the risks and costs that may be brought about in the actual testing process, and improve the test efficiency and reliability.

[0103] In summary, the simulation test method compatible with multiple pumped-storage power stations disclosed in this embodiment combines unsupervised learning and semi-supervised learning models to simulate the operating behavior of power stations under various assumed working conditions, obtain simulated operation information, provide important support for power station operation management and decision-making, help improve the stability, efficiency and performance of the power station, and reduce testing risks and costs.

[0104] Among them, the unsupervised learning model can be preferably a generative adversarial network (GAN) model; the semi-supervised learning model can be preferably a hidden Markov model (HMM). In power plant simulation, HMM can be used to simulate the random changes in the internal state of the power plant and its impact on the operation of the power plant.

[0105] The unsupervised learning algorithm includes one or more of a generative adversarial network algorithm, a K-means algorithm, an autoencoder algorithm, and a principal component analysis algorithm.

[0106] The semi-supervised learning algorithm includes one or more of a self-training algorithm, a collaborative training algorithm, a kernel smoothing algorithm, and a generative semi-supervised learning algorithm.

[0107] Furthermore, the output data of the first model can be used as input data of the second model to obtain the operating information of the simulated power plant under various assumed operating conditions. Then, based on the operating information, the accuracy and applicability of the data generated by the first model can be evaluated, and the parameters of the first model and the second model can be optimized to improve the simulation effect of the overall model (i.e., the multi-compatibility simulation model).

[0108] Reference Figure 2 , the method for obtaining the first model specifically includes the following steps:

[0109] S210, preprocessing the collected historical operation data of the target power station to obtain a first training set.

[0110] Among them, the historical operation data includes the historical operation data of the target power station at different times, different environmental conditions and different operation modes, including but not limited to water level data, power output, energy consumption, environmental conditions (such as temperature, rainfall), etc.

[0111] Specifically, preprocessing includes data cleaning, exclusion of outliers and erroneous records to ensure data quality.

[0112] In this step, through preprocessing, clean and accurate data can be provided for subsequent training.

[0113] S220, training the initial generator based on the first training set, and using the trained initial generator as the first generator; the virtual data generated by the first generator matches the operating characteristics of the target power station.

[0114] It should be noted that the initial generator can capture and simulate the complexity and multi-dimensional characteristics of power plant data.

[0115] By learning the distribution characteristics of historical data through the generator model, the generated virtual data can be closer to the real data and the generation ability of the model can be improved.

[0116] S230, training the initial discriminator based on the virtual data generated by the first generator and the real data from the corresponding operating conditions in the target power station, and using the initial discriminator after iterative training that meets the preset indicators as the target discriminator.

[0117] By training the initial discriminator, the ability to identify virtual data and real data can be improved, which helps to improve the accuracy and stability of the model. That is, the target discriminator obtained through training can determine whether the input data is real power station data or data generated by the generator.

[0118] Among them, the preset indicators include one or more of accuracy indicator, precision indicator, recall rate indicator, F1 score indicator, ROC curve indicator, and AUC curve indicator.

[0119] S240, input the virtual data generated by the first generator into the target discriminator, analyze the first generator according to the output information of the target discriminator and the preset loss function, and obtain the gradient information of the loss relative to the parameters of the first generator.

[0120] Through the game process between the generator and the discriminator, the realism of the data generated by the generator can be improved, and the parameters of the generator can be adjusted according to the loss information to make the generated data more in line with the actual situation, that is, the trained first generator can generate virtual data that is as close to the real running data as possible.

[0121] S250, updating the network parameters of the first generator based on the gradient information.

[0122] The network parameters of the generator are updated according to the gradient information, so that the generator can better generate virtual data that meets the operating characteristics of the target power station and continuously optimize the performance of the generator.

[0123] S260, iteratively train the updated first generator, and use the first generator that meets the training conditions as the target generator; the first model includes a target generator and a target discriminator.

[0124] By continuously iterating and training the generator, the generation capability of the generator can be gradually improved so that the generated virtual data better matches the operating characteristics of the target power plant.

[0125] The method for obtaining the first model disclosed in this embodiment gradually optimizes the generation capability of the generator through a training game between the generator and the discriminator, so that the generated virtual data is closer to the real data, thereby improving the simulation capability of the model; by iteratively training the target generator and the target discriminator, a more accurate and stable first model can be obtained, and the generalization ability of the model in various scenarios can be improved, that is, it can simulate the operation mode of the power station under various unknown or unobserved working conditions, and provide strong data support for further simulation tests, that is, improve the reliability and accuracy of the multi-compatibility simulation test of the trained multi-compatibility simulation model.

[0126] Reference Figure 3 , the method for obtaining the second model specifically includes the following steps:

[0127] S310, obtaining key status information and status transition trend analysis information of the target power station during operation;

[0128] The key status information includes one or more of normal operation information, maintenance status information, and abnormal status information. The status transition trend analysis information includes the transition path information of the transition trend between different operating states, the condition information of the transferred state, and the probability information of the transferred state.

[0129] For example, the state transition trend analysis information includes the possibility of transition from the normal operation state to the maintenance state.

[0130] By collecting the key status information and status transition trend analysis information of the target power station, we can better understand the key status and status transition laws of the power station operation, providing an important basis for model construction.

[0131] S320: construct a second training set based on the key state information and the state transition trend analysis information.

[0132] It should be noted that the second training set is partially labeled data, and the second training set contains sufficient status information; for example, it may include representative data containing key status labels, such as operation logs and maintenance records.

[0133] By constructing a training set containing key state information and state transition trend information, it is helpful to improve the accuracy and robustness of the model in predicting and analyzing the power plant operation state transition process.

[0134] S330, preprocessing the second training set, training the semi-supervised learning model according to the preprocessed second training set and a preset iterative algorithm, and using the trained semi-supervised learning model as the second model.

[0135] By preprocessing the second training set and training the semi-supervised learning model, we can better utilize labeled and unlabeled data and improve the generalization ability and prediction accuracy of the model.

[0136] Among them, preprocessing includes data cleaning, formatting and other processing.

[0137] Specifically, an iterative algorithm (such as the Baum-Welch algorithm) can be used for model training to adjust model parameters until convergence to obtain the best state transition and emission probability estimates; the trained model (i.e., the second model) is used for simulation testing to predict the operation of the power station under different states.

[0138] The second model can be used to predict the changes and duration of the power plant status in each test scenario. Based on the prediction results, the possible operating problems or efficiency bottlenecks that may occur in the power plant under different conditions can be analyzed, providing a basis for optimizing the operation of the power plant.

[0139] The method for obtaining the second model disclosed in this embodiment can better understand the operating characteristics and state transition laws of the power station through the acquisition and training of the second model, which is helpful to simulate and predict the operating status of the power station under different working conditions, and improve the comprehensiveness and accuracy of the simulation model; combining the information of the first model and the second model, the actual operation of the power station and key status information can be comprehensively considered, providing a more comprehensive and in-depth analysis and evaluation for multi-compatibility simulation testing, which is helpful to optimize the power station operation strategy and decision-making.

[0140] Reference Figure 4 Another aspect of the present application discloses a simulation test method compatible with multiple pumped storage power stations, comprising:

[0141] S10, constructing a multi-compatibility simulation model.

[0142] S20, using a genetic algorithm to adjust the multi-compatibility simulation model to obtain an optimized multi-compatibility simulation model;

[0143] The optimized multi-compatibility simulation model includes an unsupervised learning model and a semi-supervised learning model, and the multi-compatibility simulation model is suitable for different types of pumped-storage power stations.

[0144] Specifically, the genetic algorithm is preferably a random bond genetic algorithm (RKGA).

[0145] S30, training the unsupervised learning model, and using the trained unsupervised learning model as the first model; the unsupervised learning model includes an initial generator and an initial discriminator.

[0146] It should be noted that the specific scheme of this step is consistent with the corresponding scheme in the test method disclosed in the first aspect of the present application, that is, consistent with the specific scheme of S200 therein, so it will not be repeated here one by one.

[0147] S40, training the semi-supervised learning model, and using the trained semi-supervised learning model as the second model.

[0148] It should be noted that the specific scheme of this step is consistent with the corresponding scheme in the test method disclosed in the first aspect of the present application, that is, consistent with the specific scheme of S300 therein, so it will not be repeated here one by one.

[0149] S50, simulating the operation behaviors under various assumed working conditions by using the first model to obtain the operation data of the target power station under various assumed working conditions.

[0150] It should be noted that the specific scheme of this step is consistent with the corresponding scheme in the test method disclosed in the first aspect of the present application, that is, consistent with the specific scheme of S400 therein, so it will not be repeated here one by one.

[0151] S60, inputting the operating data of the target power station under various assumed working conditions into the second model to obtain simulated operating information of the target power station under various assumed working conditions.

[0152] It should be noted that the specific scheme of this step is consistent with the corresponding scheme in the test method disclosed in the first aspect of the present application, that is, consistent with the specific scheme of S500 therein, so it will not be repeated here one by one.

[0153] On the other hand, the present application discloses a simulation test method compatible with multiple pumped storage power stations. By combining unsupervised learning and semi-supervised learning models, the overall solution can cover a variety of different types of pumped storage power stations and can perform comprehensive simulation tests under various assumed working conditions. This can more accurately simulate and predict the operating behavior of the target power station under various conditions, helping users to more comprehensively understand the performance and stability of the power station.

[0154] Specifically, by using the genetic algorithm to optimize the model, the overall solution has a certain degree of flexibility and adaptability, can adapt to different types of pumped storage power stations, and provide simulation test models suitable for various situations; through the training of unsupervised learning and semi-supervised learning models, the model can effectively utilize limited labeled data and a large amount of unlabeled data, improve generalization ability and prediction accuracy, which can make the model more reliable and effective in practical applications; by generating virtual data and analyzing the real data of the target power station, the overall solution can provide users with more experimental data and information to help users better understand the operating characteristics of the power station, and provide users with in-depth insights and decision support through analysis and comparison of data; the overall solution can provide users with rich simulation operation information, including target power station operation data and simulation results under various assumed conditions, which is of great significance for making optimization decisions and improving power station operation strategies. Users can evaluate and compare based on this information, so as to better optimize power station operation, improve efficiency and reduce costs; by simulating the power station operation behavior under various assumed conditions and conducting comprehensive analysis, the overall solution can conduct systematic evaluation and testing, help users fully understand the response capability and stability of the power station under various conditions, discover potential problems and put forward improvement suggestions, which is helpful to improve the overall performance of the power station.

[0155] The overall solution combines multiple models and algorithms, and through a reasonable training and optimization process, it can help users save cost and time in the design and testing process; through simulation testing, users can verify multiple situations in a virtual environment, reduce the need for actual experiments, and reduce the additional costs caused by trial and error; the combination of the overall solution utilizes different learning algorithms, laying the foundation for future application expansion and development; through in-depth research and improvement of the model, the accuracy and efficiency of simulation testing can be further improved, and expanded to applications and scenarios in more fields.

[0156] In summary, the simulation test method compatible with multiple pumped-storage power stations disclosed in this embodiment is highly comprehensive, comprehensive and flexible, can provide users with more accurate and reliable simulation test results, and provide strong decision-making support for power station operation and optimization, while saving costs and time, and has broad application prospects and potential.

[0157] Reference Figure 5 The method for adjusting the multi-compatibility simulation model specifically includes the following steps:

[0158] S21, determining the decision variables that need to be adjusted in the multi-compatibility simulation model; the decision variables include one or more of power plant operation parameters, scheduling strategies, target equipment status, resource allocation strategies, task allocation strategies, target parameter optimization strategies, and environmental variables.

[0159] Among them, the target device status includes the switch status, working status, operation mode, etc. of the equipment. The adjustment of these status may affect the performance and efficiency of the overall system; the resource allocation strategy includes human resource allocation / utilization strategy, material resource allocation / utilization strategy, etc., and the system operation efficiency can be improved by optimizing the resource allocation plan; the task allocation strategy includes the allocation and scheduling of tasks or work, including the order of task allocation, which resources are allocated, etc. Reasonable task allocation can improve production efficiency; the target parameter optimization strategy includes maximizing energy output, minimizing maintenance costs, etc. These parameters can be used as criteria for evaluating optimization plans; environmental variables include the impact of external environmental changes on system operation, such as weather, market demand, etc.

[0160] S22, construct a random key sequence.

[0161] Specifically, constructing a random key sequence includes: assigning a random key to each decision variable; each random key includes a value range of a corresponding single decision variable.

[0162] Among them, these random keys can be a string of random numbers or symbols.

[0163] Mapping decision variables to random key sequences, where each random key includes a value range, can introduce a certain degree of randomness and diversity, help avoid falling into local optimal solutions, and improve the diversity of the search process; by mapping each decision variable to a random key, the processing of the value range of decision variables in the entire optimization process can be simplified, making calculations and operations more convenient and quick; converting decision variables into random keys can protect the privacy and confidentiality of decision variables to a certain extent and reduce the risk of information leakage; through the construction of random key sequences, the value range of decision variables can be effectively organized and managed, which is conducive to the optimization algorithm to search for the global optimal solution more efficiently, thereby improving the search efficiency; the method of converting decision variables into random key sequences can better adapt to different optimization models and algorithms, and is easy to expand and apply to more optimization problems.

[0164] Constructing a random key sequence to assign a random key to each decision variable, including the corresponding value range, can effectively improve the search efficiency of the optimization algorithm, introduce randomness and diversity, simplify the processing process, and protect privacy information at the same time, which has certain optimization and practical value.

[0165] S23, determine to map the random key to the actual decision variable value space and fitness function.

[0166] Specifically, for the determination of the decision variable value space, first determine the length and encoding method of the random key, such as binary encoding, integer encoding, floating point encoding, etc.; then map the random key to the actual decision variable value space according to the encoding method; map it to the corresponding numerical range according to the encoding method and length of the random key to ensure that a reasonable decision variable value is obtained. For example, for binary encoding, the binary number can be converted to a decimal number and then mapped to the actual range; according to the range and precision of the decision variable set for the specific problem, the mapped value is standardized to ensure that it falls within a reasonable value range.

[0167] The determination of the fitness function specifically includes designing a fitness function or objective function according to the specific characteristics and goals of the problem to evaluate the quality of individuals; according to the decoded decision variable value, it is substituted into the designed fitness function for calculation to obtain the fitness value of each individual. Among them, the fitness function should be consistent with the optimization goal of the problem, such as maximizing or minimizing a certain indicator, so that individuals can be selected, crossed and mutated according to the fitness value during the evolution process.

[0168] S24, randomly generating a preset number of initial individuals to form an initial population;

[0169] The initial individuals consist of a random key sequence, and each initial individual represents a potential solution.

[0170] Among them, the initial population is a set of possible solutions.

[0171] Specifically, determine the encoding method of each decision variable, such as binary encoding, integer encoding, floating point encoding, etc.; determine the length of each random key to ensure that it can cover the value range of the decision variable; for each initial individual, generate a random key sequence of corresponding length as a genotype, representing a potential solution; randomly generate a certain number of initial individuals to form an initial population. The size of the initial population can be determined according to the specific problem.

[0172] A specific implementation example (taking binary coding as an example) is used to illustrate the problem. The problem is described as follows: Suppose a decision variable x is to be optimized, ranging from [0, 10], and each individual is represented by a binary string of length 8.

[0173] First, randomly generate the initial individual: for each individual, randomly generate a binary string of length 8 as a random key sequence. For example, an initial individual can be 10101011.

[0174] Then construct the initial population: generate a certain number of initial individuals, for example, generate 10 initial individuals, to constitute the initial population.

[0175] YY+232025P

[0176] Then the initial population examples include: 1. Individual 1: 10101011; 2. Individual 2: 11000110...4. Individual 10: 00111100.

[0177] In practical applications, the number of randomly generated initial individuals can be adjusted according to the complexity of the problem and the requirements of the population size. The random generation of initial individuals should ensure that the possible solution space is covered so that potential solutions can be better searched and optimized during the iterative process of the evolutionary algorithm.

[0178] Through the above specific implementation scheme, when the decision variables are represented by random key sequences, a preset number of initial individuals can be randomly generated to construct an initial population, providing a starting point for subsequent iterative optimization of the evolutionary algorithm and exploring possible solutions in the search space.

[0179] S25, performing genetic operation processing on the initial individual, determining whether the processed initial individual meets the preset conditions, and if not, repeating the genetic operation processing until the processed initial individual meets the preset conditions, thereby obtaining an optimized multi-compatibility simulation model;

[0180] Genetic operations include selection operations, crossover operations, and mutation operations; preset conditions include reaching the maximum number of iterations or fitness value convergence.

[0181] Among them, the selection operation may include selecting excellent individuals according to the fitness function; the crossover operation may include crossing the selected individuals to generate new individuals; the mutation operation may include randomly changing the random keys of certain individuals to introduce new characteristics.

[0182] Specifically, selection is performed based on the fitness value of the individuals, usually using selection operators such as roulette selection and tournament selection.

[0183] For the crossover operation, the parent individuals are selected and crossover operations are performed to generate offspring individuals. Common crossover methods include single-point crossover, multi-point crossover, etc. For the mutation operation, mutation operations are performed on the offspring individuals to increase the diversity of the population, usually by randomly changing certain gene values ​​of the individuals.

[0184] Judging whether the processed individuals meet the preset conditions includes setting the preset conditions, checking the convergence conditions, and judging whether the conditions are met. Among them, setting the preset conditions includes: defining reaching the maximum number of iterations or fitness value convergence as the preset conditions; checking the convergence conditions includes: judging whether the fitness value of the current population converges, or judging whether the number of iterations reaches the maximum value; judging whether the conditions are met includes: if the population meets the preset conditions, then stop the iteration, otherwise continue the genetic operation processing.

[0185] In the optimization process of the genetic algorithm, the initial individuals are processed by genetic operations and the fitness of the individuals is evaluated in each iteration. If an individual meets the preset conditions, that is, meets the optimization requirements, it can be regarded as an optimized solution (potential optimization model).

[0186] Specifically, in general, in genetic algorithms, we will judge whether an individual has achieved the expected optimization level based on the set optimization goals and termination conditions. When an individual meets the optimization goals and meets the set termination conditions, the individual can be regarded as the optimized solution, that is, the final model after adjustment and optimization for the multi-compatibility simulation model.

[0187] Therefore, "obtaining an optimized multi-compatibility simulation model" means that after the iterative optimization process of the genetic algorithm, an optimal solution that meets the optimization requirements is found. This solution represents the adjusted and optimized multi-compatibility simulation model. This process is to evaluate the fitness of individuals and continuously optimize the individuals in the population until the solution that best meets the optimization goal is found.

[0188] The following is a detailed description with reference to a specific implementation example. The problem is: Assume that a genetic algorithm is used to optimize a multi-compatibility simulation model, the maximum number of iterations is set to 100, and the fitness value converges to 0.001.

[0189] The genetic operation process includes: randomly selecting parent individuals, and performing crossover and mutation operations to obtain offspring individuals.

[0190] The judgment condition processing includes: checking whether the fitness value of the population converges or whether the number of iterations exceeds the maximum number. If the preset conditions are met, the iteration is stopped; otherwise, the genetic operation processing continues.

[0191] If the population does not meet the preset conditions, the genetic operation process is repeated until the preset conditions are met.

[0192] In the process of genetic manipulation, it is necessary to design appropriate selection, crossover and mutation operations according to the characteristics of the problem and the goal to ensure the optimization effect and the diversity of the population. The setting of the preset conditions should take into account the characteristics of the problem itself and the requirements of optimization to ensure that the algorithm can obtain satisfactory optimization results within a reasonable time.

[0193] Through the above specific implementation scheme, during the genetic operation processing, it is possible to dynamically determine whether the processed individuals meet the preset conditions, ensure that the optimized multi-compatibility simulation model meets the optimization requirements, and continuously find the optimal solution through an iterative process.

[0194] The method disclosed in this embodiment determines multiple decision variables that need to be adjusted in the multi-compatibility simulation model, including power plant operating parameters, scheduling strategies, target equipment status, etc. The overall solution can comprehensively consider various influencing factors, so that the optimized model is more comprehensive and diverse; constructs a random key sequence and randomly generates initial individuals, combined with the selection, crossover, and mutation processes of genetic operations, which can introduce a certain degree of randomness and diversity to avoid falling into local optimal solutions, thereby making it more likely to find better solutions; by performing genetic operations on the initial individuals and setting preset conditions, such as reaching the maximum number of iterations or fitness value convergence, the overall solution can quickly and effectively search for the optimal solution and improve the efficiency of the optimization process; through genetic operations such as selection, crossover, and mutation, the overall solution can continuously optimize in the search space, and finally obtain the optimized multi-compatibility simulation model, which can improve the performance and effect of the model; the overall solution can automatically find the optimal solution in continuous iterations through automated genetic operation processing, has a certain degree of intelligence and adaptability, and helps to provide users with more effective optimization solutions.

[0195] In summary, the overall solution can provide a comprehensive, efficient and automated optimization process for the adjustment of multi-compatible simulation models by combining multiple optimization methods and strategies and introducing randomness and genetic operations, thereby achieving better simulation results and performance improvement.

[0196] In order to better illustrate the process of optimizing the operation strategy of a pumped storage power station, a specific example is provided below.

[0197] To better illustrate this process, we assume that the goal is to optimize the operation strategy of a pumped storage power plant, including the operation time of the pumps and the schedule of electricity generation.

[0198] Define decision variables; decision variables include the start and stop time of the pump, power generation time, etc., which are used to describe the operation strategy of the pumped storage power station.

[0199] Assign random keys; specifically, assign a random key between 0 and 1 to each decision variable to represent different decision options.

[0200] Design a fitness function; specifically, design a fitness function to evaluate the overall energy efficiency of the power plant, taking into account multiple factors including power generation and energy consumption to evaluate the pros and cons of different decision-making schemes.

[0201] Randomly generate initial solutions; specifically, randomly generate a certain number of solutions, each solution contains a random key sequence, to form an initial population for starting the iterative optimization process; for example, randomly generate 100 solutions containing random keys.

[0202] Ensure population diversity; specifically, ensure that the solutions in the initial population are widely distributed in the decision space to increase the diversity of the search and help explore the solution space more comprehensively; for example, ensure that the 100 solutions are widely distributed in the decision space.

[0203] Perform a selection operation; specifically, evaluate the pros and cons of each solution according to the fitness function, and select a portion of solutions with higher fitness as parents for breeding the next generation. For example, according to the fitness function, select the top 20% of solutions as parents.

[0204] Perform a crossover operation; specifically, randomly select a parent solution for a crossover operation to generate a new offspring solution to retain excellent genes and introduce new changes, that is, randomly pair the selected parents, perform a crossover operation, and generate a new offspring.

[0205] Perform mutation operations; specifically, randomly select a part of individuals in the offspring solutions and perform mutation operations on their random keys to increase the diversity of the population and avoid falling into local optimal solutions, that is, randomly select some solutions in the offspring and fine-tune their random keys.

[0206] Perform iterative optimization; specifically, through multiple iterations, gradually improve the average fitness of the solution population to find a better solution; record and compare the best solution after each iteration, and iteratively gradually find the global optimal solution; set termination conditions, such as terminating the iteration when a certain number of iterations is reached or the fitness improvement is not obvious after several consecutive iterations, so as to end the algorithm and obtain the optimized power plant operation strategy, for example, terminating after 1000 iterations or when there is no significant fitness improvement after 50 consecutive iterations.

[0207] Through the above process, the random key genetic algorithm (RKGA) can effectively optimize the operation strategy of the pumped storage power station, so that the power station can find an efficient and stable operation plan while comprehensively considering energy efficiency and economy, and achieve efficient energy utilization and optimization of operating costs.

[0208] The simulation test method compatible with multiple pumped-storage power stations disclosed in the present application can be applied to large-scale pumped-storage power stations, that is, the technical solution can be used to simulate and test various working conditions of large-scale power stations to optimize the operation strategy and energy efficiency of the power stations; it can be applied to the testing of medium-sized pumped-storage power stations, that is, the simulation model parameters can be adjusted according to the specific needs of medium-sized power stations to improve the flexibility and economy of power station operation; it can be applied to the adaptability testing of small-scale pumped-storage power stations, that is, simulation tests can be customized for small-scale power stations to ensure stable operation and optimal energy utilization under different environmental conditions.

[0209] Reference Figure 6 The embodiment of the present disclosure also provides a simulation test system 10 compatible with multiple pumped storage power stations, including:

[0210] A construction module 11 is used to construct a multi-compatibility simulation model; the multi-compatibility simulation model includes an unsupervised learning model and a semi-supervised learning model, and the multi-compatibility simulation model is suitable for different types of pumped storage power stations;

[0211] A first training module 12 is used to train the unsupervised learning model and use the trained unsupervised learning model as the first model; the unsupervised learning model includes an initial generator and an initial discriminator;

[0212] A second training module 13 is used to train the semi-supervised learning model and use the trained semi-supervised learning model as the second model;

[0213] The simulation module 14 is used to simulate the operation behaviors under various assumed working conditions through the first model to obtain the operation data of the target power station under various assumed working conditions;

[0214] The acquisition module 15 is used to input the operating data of the target power station under various assumed working conditions into the second model to obtain the simulated operating information of the target power station under various assumed working conditions.

[0215] In a fourth aspect, the embodiment of the present disclosure further provides a simulation test system compatible with multiple pumped storage power stations, including:

[0216] A building block for building a multi-compatible simulation model;

[0217] An optimization module, used for adjusting the multi-compatibility simulation model by using a genetic algorithm to obtain an optimized multi-compatibility simulation model; the optimized multi-compatibility simulation model includes an unsupervised learning model and a semi-supervised learning model, and the multi-compatibility simulation model is suitable for different types of pumped storage power stations;

[0218] A first training module is used to train an unsupervised learning model, and use the trained unsupervised learning model as the first model; the unsupervised learning model includes an initial generator and an initial discriminator;

[0219] The second training module is used to train the semi-supervised learning model and use the trained semi-supervised learning model as the second model;

[0220] A simulation module, used to simulate the operation behaviors under various assumed working conditions through the first model, and obtain the operation data of the target power station under various assumed working conditions;

[0221] The acquisition module is used to input the operating data of the target power station under various assumed working conditions into the second model to obtain the simulated operating information of the target power station under various assumed working conditions.

[0222] The computer device according to the embodiment of the present disclosure includes a memory and a processor. The memory is used to store non-temporary computer-readable instructions. Specifically, the memory may include one or more computer program products, and the computer program product may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may, for example, include a random access memory (RAM) and / or a cache memory (cache), etc. The non-volatile memory may, for example, include a read-only memory (ROM), a hard disk, a flash memory, etc.

[0223] The processor may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the computer device to perform desired functions. In one embodiment of the present disclosure, the processor is used to run the computer-readable instructions stored in the memory, so that the computer device performs all or part of the steps of the simulation test method for compatible multi-pumped storage power stations of the aforementioned embodiments of the present disclosure.

[0224] Those skilled in the art should be able to understand that in order to solve the technical problem of how to obtain a good user experience, the present embodiment may also include well-known structures such as a communication bus and an interface, and these well-known structures should also be included in the protection scope of the present disclosure.

[0225] like Figure 7 A schematic diagram of the structure of a computer device provided in an embodiment of the present disclosure is shown, which is a schematic diagram of the structure of a computer device suitable for implementing the embodiment of the present disclosure. Figure 7 The computer device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0226] like Figure 7 As shown, the computer device may include a processor (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) or a program loaded from a storage device into a random access memory (RAM). In the RAM, various programs and data required for the operation of the computer device are also stored. The processor, ROM, and RAM are connected to each other via a bus. An input / output (I / O) interface is also connected to the bus.

[0227] Typically, the following devices can be connected to the I / O interface: input devices such as sensors or visual information acquisition devices; output devices such as display screens; storage devices such as tapes, hard disks, etc.; and communication devices. The communication device can allow the computer device to communicate with other devices (such as edge computing devices) wirelessly or by wire to exchange data. Figure 7 A computer device having various devices is shown, but it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed instead.

[0228] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device, or installed from a ROM. When the computer program is executed by a processor, all or part of the steps of the simulation test method for a compatible multi-pumped storage power station of an embodiment of the present disclosure are executed.

[0229] For detailed description of this embodiment, please refer to the corresponding description in the above embodiments.

[0230] No more details.

[0231] According to the computer-readable storage medium of the embodiment of the present disclosure, non-transitory computer-readable instructions are stored thereon. When the non-transitory computer-readable instructions are executed by a processor, all or part of the steps of the simulation test method for compatible multi-pumped storage power stations of the aforementioned embodiments of the present disclosure are executed.

[0232] The above-mentioned computer-readable storage media include, but are not limited to: optical storage media (e.g., CD-ROM and DVD), magneto-optical storage media (e.g., MO), magnetic storage media (e.g., magnetic tape or mobile hard disk), media with built-in rewritable non-volatile memory (e.g., memory card) and media with built-in ROM (e.g., ROM box).

[0233] For detailed description of this embodiment, reference may be made to the corresponding descriptions in the aforementioned embodiments, which will not be repeated here.

[0234] The basic principles of the present disclosure are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present disclosure are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present disclosure. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, and are not limitations. The above details do not limit the present disclosure to the necessity of adopting the above specific details to be implemented.

[0235] In the present disclosure, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. The block diagrams of the devices, devices, equipment, and systems involved in the present disclosure are only illustrative examples and are not intended to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagram. As will be appreciated by those skilled in the art, these devices, devices, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including", "comprising", "having", etc. are open words, referring to "including but not limited to", and can be used interchangeably with them. The words "or" and "and" used here refer to the words "and / or" and can be used interchangeably with them, unless the context clearly indicates otherwise. The words "such as" used here refer to the phrase "such as but not limited to", and can be used interchangeably with them.

[0236] Additionally, as used herein, "or" used in a list of items beginning with "at least one" indicates a disjunctive list, so that, for example, a list of "at least one of A, B, or C" means A or B or C, or AB or AC or BC, or ABC (i.e., A and B and C).

[0237] The term "exemplary" does not mean that the described example is preferred or better than other examples.

[0238] It should also be noted that in the system and method of the present disclosure, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present disclosure.

[0239] Various changes, substitutions, and modifications of the techniques described herein may be made without departing from the teachings defined by the appended claims. Furthermore, the scope of the claims of the present disclosure is not limited to the specific aspects of the processes, machines, manufactures, compositions of events, means, methods, and actions described above. Currently existing or later to be developed processes, machines, manufactures, compositions of events, means, methods, or actions that perform substantially the same functions or achieve substantially the same results as the corresponding aspects described herein may be utilized. Thus, the appended claims include such processes, machines, manufactures, compositions of events, means, methods, or actions within their scope.

[0240] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present disclosure. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present disclosure. Therefore, the present disclosure is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

[0241] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present disclosure to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. A simulation test method compatible with multiple pumped storage power stations, characterized in that: include: Build multi-compatibility simulation models; The multi-compatibility simulation model includes an unsupervised learning model and a semi-supervised learning model, and the multi-compatibility simulation model is suitable for different types of pumped storage power stations; Training the unsupervised learning model, and using the trained unsupervised learning model as the first model; the unsupervised learning model includes an initial generator and an initial discriminator; Training the semi-supervised learning model, and using the trained semi-supervised learning model as the second model; Simulating the operating behaviors under various assumed working conditions by using the first model to obtain the operating data of the target power station under various assumed working conditions; The operating data of the target power station under various assumed working conditions are input into the second model to obtain simulated operating information of the target power station under various assumed working conditions.

2. The simulation test method compatible with multiple pumped storage power stations according to claim 1 is characterized in that: The step of training the unsupervised learning model and using the trained unsupervised learning model as the first model includes: Preprocessing the collected historical operation data of the target power station to obtain a first training set; The initial generator is trained based on the first training set, and the trained initial generator is used as the first generator; the virtual data generated by the first generator matches the operating characteristics of the target power station; Training the initial discriminator based on the virtual data generated by the first generator and the real data from the corresponding operating conditions in the target power station, and using the initial discriminator after iterative training that meets the preset indicators as the target discriminator; Input the virtual data generated by the first generator into the target discriminator, analyze the first generator according to the output information of the target discriminator and the preset loss function, and obtain the gradient information of the loss relative to the parameters of the first generator; Update the network parameters of the first generator based on the gradient information; Iteratively train the updated first generator, and use the first generator that meets the training conditions as the target generator; The first model includes the target generator and the target discriminator.

3. The simulation test method compatible with multiple pumped storage power stations according to claim 2 is characterized in that: The preset indicators include one or more of an accuracy indicator, a precision indicator, a recall rate indicator, an F1 score indicator, a ROC curve indicator, and an AUC curve indicator.

4. The simulation test method compatible with multiple pumped storage power stations according to claim 2 is characterized in that: The step of training the semi-supervised learning model and using the trained semi-supervised learning model as the second model includes: Acquire key state information and state transition trend analysis information in the operation of the target power station; the key state information includes one or more of normal operation information, maintenance state information, and abnormal state information; the state transition trend analysis information includes transfer path information of transfer trends between different operating states, transferred state occurrence condition information, and transferred state occurrence probability information; Constructing a second training set based on the key state information and the state transition trend analysis information; The second training set is preprocessed, the semi-supervised learning model is trained according to the preprocessed second training set and a preset iterative algorithm, and the trained semi-supervised learning model is used as the second model.

5. A simulation test method compatible with multiple pumped storage power stations, characterized in that: include: Build multi-compatibility simulation models; Using a genetic algorithm to adjust the multi-compatibility simulation model to obtain an optimized multi-compatibility simulation model; The optimized multi-compatibility simulation model includes an unsupervised learning model and a semi-supervised learning model, and the multi-compatibility simulation model is suitable for different types of pumped storage power stations; Training the unsupervised learning model, and using the trained unsupervised learning model as the first model; the unsupervised learning model includes an initial generator and an initial discriminator; Training the semi-supervised learning model, and using the trained semi-supervised learning model as the second model; Simulating the operating behaviors under various assumed working conditions by using the first model to obtain the operating data of the target power station under various assumed working conditions; The operating data of the target power station under various assumed working conditions are input into the second model to obtain simulated operating information of the target power station under various assumed working conditions.

6. The simulation test method compatible with multiple pumped storage power stations according to claim 5, characterized in that: The adopting a genetic algorithm to adjust the multi-compatibility simulation model to obtain the optimized multi-compatibility simulation model comprises: Determine the decision variables that need to be adjusted in the multi-compatibility simulation model; the decision variables include one or more of power plant operation parameters, scheduling strategies, target equipment status, resource allocation strategies, task allocation strategies, target parameter optimization strategies, and environmental variables; YY+232025P Construct a random key sequence; Determine the mapping of random keys to the actual decision variable value space and fitness function; Randomly generate a preset number of initial individuals to form an initial population; The initial individuals are composed of a random key sequence, and each of the initial individuals represents a potential solution; Performing genetic operation processing on the initial individual, determining whether the processed initial individual meets the preset conditions, and if not, repeating the genetic operation processing until the processed initial individual meets the preset conditions, thereby obtaining the optimized multi-compatibility simulation model; The genetic operation includes a selection operation, a crossover operation, and a mutation operation; the preset conditions include reaching a maximum number of iterations or convergence of fitness values.

7. The simulation test method compatible with multiple pumped storage power stations according to claim 6 is characterized in that: The constructing of the random key sequence comprises: assigning a random key to each decision variable; each random key comprises a value range of a corresponding single decision variable; The unsupervised learning model is a generative adversarial network model; the semi-supervised learning model is a hidden Markov model.

8. A computer device, characterized in that: The computer device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the simulation test method for compatible multiple pumped-storage power stations described in any one of claims 1-7.

9. A computer-readable storage medium, characterized in that: The computer readable storage medium YY+232025P The device stores computer instructions, which are used to enable a computer to execute the simulation test method for compatible multiple pumped-storage power stations as described in any one of claims 1-7.

10. A computer program product comprising computer instructions, characterized in that: When the computer instructions are executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images