Dynamic optimization method of hydrogen energy power coupling system

Abstract

The invention discloses a dynamic optimization method for a hydrogen energy power coupling system, and belongs to the technical field of energy system scheduling, and the method comprises the steps: S1, carrying out the preprocessing of the historical operation data of the hydrogen energy power coupling system, and obtaining a training data set and a test data set; s2, identifying parameters of the hydrogen energy power coupling system based on a convolutional neural network and an extreme learning machine network, and constructing and training a hydrogen energy power coupling model according to an improved badger algorithm and the training data set; s3, whether the trained hydrogen energy power coupling model is qualified or not is verified based on the parameters of the hydrogen energy power coupling system and the test data set, if yes, the optimal hydrogen energy power coupling model is output, and if not, the S1 is executed; s4, inputting real-time operation data of the hydrogen energy power coupling system into the optimal hydrogen energy power coupling model to obtain optimal parameters; s5, the hydrogen energy power coupling system responds to the optimal parameter to execute a dynamic optimization action; and the stability of the hydrogen energy power coupling system can be obviously improved.

Description

Technical Field

[0001] The present invention relates to the technical field of energy system scheduling, and in particular to a dynamic optimization method for a hydrogen energy and electric power coupling system. Background Art

[0002] With the development of renewable energy power generation technology, more and more renewable energy power generation devices have replaced traditional thermal power generation devices in the power grid. Due to the volatility, randomness, and anti-peak characteristics of renewable energy power generation, the stability of the corresponding power grid has deteriorated and the quality of power transmission has declined. The existing technology usually uses batteries for energy storage regulation, but this is limited by environmental pollution, conversion efficiency and other unchangeable properties of the battery itself. Therefore, some scholars have proposed the use of hydrogen energy with higher conversion efficiency and more environmentally friendly to regulate the output characteristics of renewable energy power generation devices, such as the use of hydrogen fuel cells. However, the dynamic characteristics of hydrogen fuel cells also lead to dynamic changes and difficult control problems in the corresponding renewable energy power generation devices.

[0003] Chinese patent, publication number: CN117411037A, publication date: January 16, 2024, discloses a method for active power dispatch control of hybrid energy storage participating in power system considering dynamic adjustable capacity, including: based on the load operation characteristics of the energy storage power station, establishing a dynamic adjustable capacity analysis model of hybrid energy storage considering the load changes of the energy storage power station; reconstructing the dispatch control constraints of hybrid energy storage participating in power system including lithium-ion batteries and hydrogen fuel cells; establishing a model for hybrid energy storage participating in power system dispatch considering dispatch benefits, and analyzing the situation of hybrid energy storage participating in power grid dispatch under different conditions; issuing dispatch instructions to hybrid energy storage, and using the peak-valley arbitrage of hybrid energy storage to improve the dispatch economy of the power system; based on the instruction tracking error and the state of charge of the lithium battery, establishing an active power control model for hybrid energy storage responding to the dispatch instruction; using model predictive control to perform rolling optimization on the output of energy storage, and using the feedback correction link to correct the output error of energy storage; however, the invention does not consider the dynamic characteristics brought by hydrogen fuel cells to the power system, resulting in poor stability of the power system. Summary of the invention

[0004] The purpose of the present invention is to address the problem that the prior art does not take into account the dynamic characteristics brought by hydrogen fuel cells to the power coupling system, resulting in poor stability of the hydrogen energy power coupling system; a dynamic optimization method for a hydrogen energy power coupling system is proposed, by preprocessing the historical operation data of the hydrogen energy power coupling system to obtain a training data set and a test data set, identifying the corresponding hydrogen energy power coupling system parameters based on the historical operation data, constructing and training a hydrogen energy power coupling model based on the improved honey badger algorithm and the training data set, then verifying the performance of the hydrogen energy power coupling model based on the hydrogen energy power coupling system parameters and the test data set, finally inputting the real-time operation data of the hydrogen energy power coupling system into the hydrogen energy power coupling model with qualified performance verification to obtain optimal parameters, and the hydrogen energy power coupling system performs dynamic optimization actions in response to the optimal parameters, which significantly improves the stability of the hydrogen energy power coupling system.

[0005] In a first aspect, a technical solution provided in an embodiment of the present invention is a dynamic optimization method for a hydrogen energy power coupling system, comprising the following steps: S1. Obtain historical operation data of the hydrogen energy and power coupling system, and preprocess the historical operation data to obtain a training data set and a test data set; S2. Based on the convolutional neural network and extreme learning machine network, the hydrogen energy power coupling system parameters corresponding to the historical operation data are identified, and the hydrogen energy power coupling model is constructed and trained according to the honey badger algorithm improved based on the chaotic learning mechanism and the training data set; S3, based on the hydrogen energy power coupling system parameters and the test data set, verify whether the trained hydrogen energy power coupling model is qualified, if so, output the optimal hydrogen energy power coupling model, if not, execute S1; S4, collecting real-time operation data of the hydrogen energy and electricity coupling system, and inputting the real-time operation data into the optimal hydrogen energy and electricity coupling model to obtain optimal parameters; S5. The hydrogen energy and electricity coupling system performs dynamic optimization actions in response to the optimal parameters.

[0006] In this scheme, the historical operation data is preprocessed to obtain a training data set and a test data set, which can ensure the integrity and accuracy of the data. In addition, the parameters of the hydrogen energy power coupling system will change with the operation time during the operation process, that is, the parameters of the hydrogen energy power coupling system have a time series structure, so the hydrogen energy power coupling system parameters corresponding to the historical operation data can be identified based on the convolutional neural network and the extreme learning machine network, and the historical state of the hydrogen energy power coupling system represented by the hydrogen energy power coupling system parameters in the past closest to the current time scale is clarified. At the same time, according to the honey badger algorithm improved based on the chaotic learning mechanism and the training data set, a hydrogen energy power coupling model that can predict the state change of the hydrogen energy power coupling system is constructed and trained, and the hydrogen energy power coupling system parameters and the test data set are used to verify the performance of the hydrogen energy power coupling model, ensuring the effectiveness of the hydrogen energy power coupling model and its data processing capabilities. Finally, the collected real-time operation data is input into the hydrogen energy power coupling model to obtain the optimal parameters corresponding to the predicted state change of the hydrogen energy power coupling system. The hydrogen energy power coupling system performs dynamic optimization actions in response to the optimal parameters, which significantly improves the stability of the hydrogen energy power coupling system.

[0007] Preferably, in S1, the specific process of preprocessing the historical operation data to obtain the training data set and the test data set is: S111, cleaning the historical operation data to obtain a complete data set, and integrating the complete data set based on a collection source of the historical operation data to obtain an integrated data set; S112: Divide the integrated data set into a training data set and a test data set based on a preset training-test ratio.

[0008] In this solution, the historical operation data can be cleaned by using statistical methods and linear interpolation methods, wherein the statistical method can find outliers and missing values ​​in the data. After deleting the outliers, only the missing values ​​are left in the data. The linear interpolation method can fill in the missing values ​​in the data to ensure the continuity and completeness of the data. However, the data at this time is too complex, and statistics can be performed again based on the data collection source, such as sensors, measuring equipment, etc., and different labels can be established for data with the same name but different meanings, or with the same name but different attributes to distinguish them. The same label can be established for data with the same name and synonyms, or with the same name and attributes as a whole, and duplicate data can be deleted to obtain a representative and neat integrated data set. At the same time, in order to meet the subsequent steps, 80% of the data in the integrated data set can be divided as a training data set, and 20% of the data in the integrated data set can be divided as a test data set.

[0009] Preferably, in S2, the specific process of identifying the parameters of the hydrogen energy power coupling system corresponding to the historical operation data based on the convolutional neural network and the extreme learning machine network is: S211, converting the historical operation data into a two-dimensional matrix data set according to the time variation scale of the historical operation data and the convolutional neural network input data requirements; S212. Input the two-dimensional matrix data set into a convolutional neural network to obtain an operating data feature set, and input the operating data feature set into an extreme learning machine network to obtain corresponding hydrogen energy and power coupling system parameters.

[0010] In this scheme, since the input data of the convolutional neural network is usually two-dimensional data or three-dimensional data, combined with the historical operation data corresponding to the hydrogen energy and electricity coupling system parameters and time scale, the historical operation data can be converted into a two-dimensional matrix data set according to the time change scale of the historical operation data and the corresponding hydrogen energy and electricity coupling system parameters, ensuring that the convolutional neural network can effectively process the historical operation data. However, the convolutional neural network can only extract the features contained in the data, so the extreme learning machine network can be used to further process the features extracted by the convolutional neural network to accurately analyze the hydrogen energy and electricity coupling system parameters corresponding to the historical operation data.

[0011] Preferably, in S2, the specific process of constructing and training the hydrogen energy power coupling model according to the honey badger algorithm improved based on the chaotic learning mechanism and the training data set is: S221, improving the population initialization formula of the honey badger algorithm based on the adversarial learning mechanism to obtain the individual generation formula, and improving the random value update formula of the honey badger algorithm based on the chaos mechanism to obtain the random value chaos update formula; Synchronously, an objective function is established based on the equipment characteristics of the electrical equipment in the hydrogen energy and power coupling system corresponding to the training data set; S222, a hydrogen energy and power coupling model is constructed based on the individual generation formula, the random value chaos update formula and the traditional honey badger algorithm; S223, inputting the training data set into the constructed hydrogen energy and electricity coupling model for training; The hydrogen energy and electricity coupling model at least includes an odor intensity formula, a position update formula, a random value chaos update formula, an individual generation formula and an objective function.

[0012] In this scheme, in order to avoid falling into the local optimal situation, the population initialization formula of the honey badger algorithm is improved based on the adversarial learning mechanism to obtain the individual generation formula, providing more exploratory results, but this also leads to the corresponding solution space becoming larger and the time to explore the optimal solution is longer. Therefore, the random value update formula of the honey badger algorithm is improved based on the chaos mechanism to obtain the random value chaos update formula, which can improve the speed of exploring the optimal solution in the solution space. Secondly, the objective function is established based on the equipment characteristics of the electrical equipment in the hydrogen energy and power coupling system corresponding to the training data set, so that the hydrogen energy and power coupling model is closely linked with the corresponding hydrogen energy and power coupling system, and the hydrogen energy and power coupling model effectively improves the prediction ability of the operating status of the hydrogen energy and power coupling system.

[0013] Preferably, in S221, the individual generation formula is specifically: z j =lb j +r1×(ub j -lb j ); In the formula, z j is the candidate solution corresponding to the j-th honey badger individual, is the upper limit of the candidate solution, z -j is the lower bound of the candidate solution, is the opposite solution of the candidate solution, lb j is the lower bound of the candidate solution search space, ub j is the upper bound of the candidate solution search space, r1 is the first random value, and j represents the jth honey badger individual.

[0014] Preferably, in S221, the random value chaotic update formula is specifically: r1 new =r1+δ i ×r1; δ i+1 =4δ i ×(1-δ i ); In the formula, r1 new is the update value of the first random value, r1 is the first random value, is the updated value of the second random value, r2 is the second random value, is the updated value of the third random value, r3 is the third random value, δ i is the chaos value of the i-th generation, δ i+1 is the chaos value of the i+1th generation.

[0015] Preferably, in S223, the specific process of inputting the training data set into the constructed hydrogen energy and electricity coupling model for training is: A1. Establish a honey badger population based on the training data set and the individual generation formula; A2, determining the prey odor intensity based on the honey badger individuals in the honey badger population and the odor intensity formula; A3, updating the position of individual honey badgers in the honey badger population according to the prey odor intensity and position update formula; Synchronously, judging whether the training is completed based on the number of position updates of the honey badger individual and the preset maximum number of iterations, if the number of position updates is greater than or equal to the maximum number of iterations, the training is determined to be completed; if the number of position updates is less than the maximum number of iterations, the training is determined to be incomplete and A2 is executed; A4. When it is determined that the training is completed, the fitness values ​​of the honey badgers in the honey badger population that have completed the training are evaluated based on the objective function of the hydrogen energy and electricity coupling model, and the honey badger individual with the highest fitness value is marked as the optimal individual.

[0016] In this scheme, whether the training is completed is judged based on the number of position updates of the honey badger individual and the preset maximum number of iterations, which can ensure that the hydrogen energy and electricity coupling model converges within a reasonable time, effectively avoid the waste of computing resources, and prevent the hydrogen energy and electricity coupling model from falling into an infinite loop. By calculating the fitness value of the honey badger individual, the honey badger individual can be quantified and the optimal individual can be identified.

[0017] Preferably, in S3, the specific process of verifying whether the trained hydrogen energy power coupling model is qualified based on the hydrogen energy power coupling system parameters and the test data set is: S311, inputting the test data set into the trained hydrogen energy and electricity coupling model to obtain test parameters, and establishing a test honey badger individual according to the test parameters and the population initialization formula of the hydrogen energy and electricity coupling model; Simultaneously, a honey badger individual is established and verified according to the hydrogen energy and electricity coupling system parameters and the population initialization formula of the hydrogen energy and electricity coupling model; S312, evaluating the fitness of individual honey badgers based on the objective function of the hydrogen energy and electricity coupling model to obtain test fitness, and evaluating and verifying the fitness of individual honey badgers based on the objective function of the hydrogen energy and electricity coupling model to obtain verified fitness; S313, based on the test fitness and the verification fitness, a comparison is made. If the test fitness is greater than the verification fitness, the trained hydrogen energy power coupling model is determined to be qualified; if the test fitness is less than or equal to the verification fitness, the trained hydrogen energy power coupling model is determined to be unqualified; Simultaneously, the hydrogen energy and electricity coupling model that is judged to be qualified is marked as the optimal hydrogen energy and electricity coupling model.

[0018] In this solution, the objective function of the hydrogen energy and power coupling model evaluates the fitness values ​​of the test data set and the hydrogen energy and power coupling system parameters, and before the evaluation, the test data set needs to be input into the hydrogen energy and power coupling model to calculate the test parameters, and the hydrogen energy and power coupling system parameters are directly extracted from the historical operating data. Therefore, the essence of evaluating the fitness value is to calculate the predicted operating state of the hydrogen energy and power coupling system corresponding to the test data set and the historical operating state of the hydrogen energy and power coupling system corresponding to the hydrogen energy and power coupling system parameters. By comparing the predicted operating state with the historical operating state, it is effectively guaranteed that the performance of the hydrogen energy and power coupling model meets the actual working conditions.

[0019] Preferably, in S3, if the hydrogen energy and electricity coupling model verification fails and before executing S1, a cycle stop judgment needs to be performed, and the specific process of the cycle stop judgment is: A comparison is made based on the current verification times of the hydrogen energy and electricity coupling model and the preset cycle threshold. If the current verification times are less than the cycle threshold, the cycle is determined to be incomplete and the cycle continues; if the current verification times are greater than or equal to the cycle threshold, the cycle is determined to be complete and the cycle is stopped.

[0020] In this scheme, the cycle stop judgment can prevent the hydrogen energy and electricity coupling model from falling into a meaningless infinite training process, effectively avoiding the waste of computing resources.

[0021] Preferably, in S4, the specific process of inputting the real-time operation data into the optimal hydrogen energy and electricity coupling model to obtain the optimal parameters is: S411, cleaning the real-time operation data to obtain a real-time complete data set, and integrating the real-time complete data set based on a collection source of the real-time data to obtain a real-time integrated data set; S412, inputting the real-time integrated data set into the optimal hydrogen energy and electricity coupling model to obtain optimal parameters.

[0022] The present invention has at least the following substantial effects: The present application preprocesses the historical operation data to obtain a training data set and a test data set, which can ensure the integrity and accuracy of the data. In addition, the parameters of the hydrogen energy power coupling system will change with the operation time during the operation process, that is, the parameters of the hydrogen energy power coupling system have a time series structure, so the hydrogen energy power coupling system parameters corresponding to the historical operation data can be identified based on the convolutional neural network and the extreme learning machine network, and the historical state of the hydrogen energy power coupling system represented by the hydrogen energy power coupling system parameters in the past closest to the current time scale is clarified. At the same time, according to the honey badger algorithm improved based on the chaotic learning mechanism and the training data set, a hydrogen energy power coupling model that can predict the state change of the hydrogen energy power coupling system is constructed and trained, and the hydrogen energy power coupling system parameters and the test data set are used to verify the performance of the hydrogen energy power coupling model, ensuring the effectiveness of the hydrogen energy power coupling model and its data processing capabilities. Finally, the collected real-time operation data is input into the hydrogen energy power coupling model to obtain the optimal parameters corresponding to the predicted state change of the hydrogen energy power coupling system. The hydrogen energy power coupling system performs dynamic optimization actions in response to the optimal parameters, which significantly improves the stability of the hydrogen energy power coupling system. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Other features, objects and advantages of the present invention will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings. The drawings are only for the purpose of illustrating preferred embodiments and are not to be considered as limiting the present invention. Also, the same reference symbols are used throughout the drawings to represent the same parts.

[0024] Figure 1 It is a schematic flow chart of a dynamic optimization method for a hydrogen energy and electric power coupling system of the present invention. DETAILED DESCRIPTION

[0025] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific implementation method described herein is only an optimal embodiment of the present invention, which is only used to explain the present invention and does not limit the scope of protection of the present invention. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0026] Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flow charts. Although the flow charts describe the operations (or steps) as sequential processes, many of the operations (or steps) therein can be implemented in parallel, concurrently, or simultaneously. In addition, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but can also have additional steps not included in the drawings; the process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0027] Embodiment 1: like Figure 1 As shown, this embodiment provides a dynamic optimization method for a hydrogen energy power coupling system, comprising the following steps: S1. Obtain historical operation data of the hydrogen energy and power coupling system, and preprocess the historical operation data to obtain a training data set and a test data set.

[0028] In one embodiment, the specific process of preprocessing the historical operation data to obtain the training data set and the test data set is as follows: S111, cleaning the historical operation data to obtain a complete data set, and integrating the complete data set based on a collection source of the historical operation data to obtain an integrated data set; S112: Divide the integrated data set into a training data set and a test data set based on a preset training-test ratio.

[0029] In this embodiment, a statistical method can be used to count the historical operation data, so as to find missing values ​​and abnormal values ​​in the historical operation data, and delete the abnormal values, so that the historical operation data is in a state where only missing values ​​exist. Then, a linear interpolation method can be used to fill the missing values ​​based on the linear relationship in the historical operation data to obtain a complete and accurate complete data set. Secondly, the historical operation data may be collected by multiple different sensors, measuring devices, etc. The historical operation data can be counted based on the collection source of the historical operation data, and data with the same name but different meanings, or the same name but different attributes can be found and different labels can be established, and data with synonyms but different names, or the same name but different attributes can be found to establish the same label, and duplicate data can be found and deleted, thereby obtaining a representative and uncomplicated integrated data set. Finally, the integrated data set is divided into a training data set and a test data set according to a training-test ratio of 80% for training and 20% for testing.

[0030] S2. Based on the convolutional neural network and extreme learning machine network, the hydrogen energy power coupling system parameters corresponding to the historical operation data are identified, and the hydrogen energy power coupling model is constructed and trained according to the honey badger algorithm improved based on the chaotic learning mechanism and the training data set.

[0031] In one embodiment, the specific process of identifying the parameters of the hydrogen energy power coupling system corresponding to the historical operation data based on the convolutional neural network and the extreme learning machine network is as follows: S211, converting the historical operation data into a two-dimensional matrix data set according to the time variation scale of the historical operation data and the convolutional neural network input data requirements; S212. Input the two-dimensional matrix data set into a convolutional neural network to obtain an operating data feature set, and input the operating data feature set into an extreme learning machine network to obtain corresponding hydrogen energy and power coupling system parameters.

[0032] In this embodiment, the convolutional neural network CNN includes three types of perception: local receptive field, time / space subsampling and shared connection weights. These perceptions can be divided into two layers, namely, a convolutional layer and a subsampling layer, wherein the convolutional layer is directly related to the local receptive field and the shared connection weight, and the subsampling layer is directly related to the time / space subsampling; the extreme learning machine network ELM is a single hidden layer network that can improve efficiency; specifically, the historical operation data is sorted according to the time change scale of the historical operation data, and a two-dimensional matrix is ​​established with the time node as the row header and the corresponding hydrogen energy power coupling system parameter as the column header, and because the time change scale of the historical operation data is continuous, the two-dimensional matrix is ​​also continuous, and they have the same time change scale, and the width, height and other related parameters of the two-dimensional matrix need to be based on the convolutional neural network CNN. The input requirements are set, and then the obtained two-dimensional matrix data set can be divided into a training matrix set and a test matrix set according to a preset parameter ratio, and the parameter ratio is specifically 8:2. At the same time, the convolutional neural network CNN is trained based on the training matrix set to ensure the effectiveness of the convolutional neural network CNN in this embodiment, and the test matrix set is input into the trained convolutional neural network CNN to extract the features of the historical operation data to obtain an operation data feature set. Similarly, in order to ensure the effectiveness of the extreme learning machine network ELM in this embodiment, the operation data feature set can be divided into a training feature set and a test feature set according to a preset parameter ratio, and the extreme learning machine network ELM is trained based on the training feature set, and the test feature set is input into the trained extreme learning machine network ELM to accurately identify the parameters that meet the specific timing structure of the hydrogen energy power coupling system.

[0033] In one embodiment, the specific process of constructing and training the hydrogen energy power coupling model according to the honey badger algorithm improved based on the chaotic learning mechanism and the training data set is: S221, improving the population initialization formula of the honey badger algorithm based on the adversarial learning mechanism to obtain the individual generation formula, and improving the random value update formula of the honey badger algorithm based on the chaos mechanism to obtain the random value chaos update formula; Synchronously, an objective function is established based on the equipment characteristics of the electrical equipment in the hydrogen energy and power coupling system corresponding to the training data set; S222, a hydrogen energy and power coupling model is constructed based on the individual generation formula, the random value chaos update formula and the traditional honey badger algorithm; S223, inputting the training data set into the constructed hydrogen energy and electricity coupling model for training; The hydrogen energy and electricity coupling model at least includes an odor intensity formula, a position update formula, a random value chaos update formula, an individual generation formula and an objective function.

[0034] In this embodiment, the honey badger algorithm is a global optimization algorithm, including two stages: exploration and development. The exploration stage includes population initialization, description of prey odor intensity, local optimal escape and density factor update, and the development stage includes excavation and honey collection. Specifically, the initialization formula corresponding to the population initialization is: z j =lb j +r1×(ub j -lb j ); In the formula, z j is the candidate solution corresponding to the jth honey badger individual, lb j is the lower bound of the candidate solution search space, ub j is the upper bound of the candidate solution search space, r1 is the first random value, and j represents the jth honey badger individual; The odor intensity formula corresponding to the description of the prey odor intensity is: L=(z j -z j+1 ) 2 ; S j =Lz j ; In the formula, I j is the odor intensity of the jth prey, r2 is the second random value, L is the location of the prey, s j is the distance between the jth honey badger and its prey, z j is the position of the jth honey badger, z j+1 is the position of the j+1th honey badger; The escape formula corresponding to the local optimal escape is: In the formula, F is the exploration direction change flag, r6 is the sixth random value, if combined with else means that when the sixth random value r6 is less than or equal to 0.5, the exploration direction change flag F is 1, and when the sixth random value r6 is greater than 0.5, the exploration direction change flag F is -1; The density factor formula corresponding to the density factor update is: Where γ is the density factor, A is equal to or greater than 1, the default value of A is 2, t is the number of iterations, t max is the maximum number of iterations, exp is the density factor control function; The mining formula corresponding to the mining is: z new1 =z prey +F×α×I×z prey +F×r3×γ×s j ×|cos(2πr4)×[1-cos(2πr5)]|; In the formula, z new1 is the best prey position after excavation, z prey is the location of the best prey, I is the odor intensity, α is the ability of the honey badger to capture prey, the default value of α is 6, r3 is the third random value, r4 is the fourth random value, and r5 is the fifth random value; The honey collection formula corresponding to the honey collection is: z new =z prey +F×r7×γ×s i ; In the formula, z prey is the position of the prey, z new is the new position of the honey badger, and r7 is the seventh random value.

[0035] However, the honey badger algorithm may converge prematurely, resulting in a local optimal problem. In this embodiment, the adversarial learning mechanism OBL based on adversarial learning inspired by Eastern philosophical theory is first used to improve the population initialization formula to obtain the individual generation formula, which can randomly generate the opposite value of the candidate solution to provide a higher chance of finding the optimal solution. In one embodiment, the individual generation formula is specifically: z j =lb j +r1×(ub j -lb j ); In the formula, z j is the candidate solution corresponding to the j-th honey badger individual, is the upper limit of the candidate solution, z -j is the lower bound of the candidate solution, is the opposite solution of the candidate solution, lb j is the lower bound of the candidate solution search space, ub j is the upper bound of the candidate solution search space, r1 is the first random value, and j represents the jth honey badger individual; Next, chaos theory is used to generate pseudo-random values ​​to improve the speed of the honey badger algorithm, that is, the random value update formula is improved based on the chaos mechanism to obtain a random value chaos update formula, specifically involving the first random value r1, the second random value r2 and the third random value r3; in one embodiment, the random value chaos update formula is specifically: r1 new =r1+δ i ×r1; δ i+1 =4δ i ×(1-δ i ); In the formula, r1 new is the update value of the first random value, r1 is the first random value, is the updated value of the second random value, r2 is the second random value, is the updated value of the third random value, r3 is the third random value, δ i is the chaos value of the i-th generation, δ i+1 is the chaotic value of the i+1th generation; Finally, a corresponding objective function can also be constructed based on the equipment characteristics of the electrical equipment in the hydrogen energy and power coupling system, such as current, voltage, power, etc., which have been marked on the nameplate of the electrical equipment.

[0036] In one embodiment, the specific process of inputting the training data set into the constructed hydrogen energy and electricity coupling model for training is: A1. Establish a honey badger population based on the training data set and the individual generation formula; A2, determining the prey odor intensity based on the honey badger individuals in the honey badger population and the odor intensity formula; A3, updating the position of individual honey badgers in the honey badger population according to the prey odor intensity and position update formula; Synchronously, judging whether the training is completed based on the number of position updates of the honey badger individual and the preset maximum number of iterations, if the number of position updates is greater than or equal to the maximum number of iterations, the training is determined to be completed; if the number of position updates is less than the maximum number of iterations, the training is determined to be incomplete and A2 is executed; A4. When it is determined that the training is completed, the fitness values ​​of the honey badgers in the honey badger population that have completed the training are evaluated based on the objective function of the hydrogen energy and electricity coupling model, and the honey badger individual with the highest fitness value is marked as the optimal individual.

[0037] In this embodiment, the maximum number of iterations is set to 800, which can be modified according to actual working conditions.

[0038] S3. Verify whether the trained hydrogen energy and power coupling model is qualified based on the hydrogen energy and power coupling system parameters and the test data set. If so, output the optimal hydrogen energy and power coupling model. If not, execute S1.

[0039] In one embodiment, the specific process of verifying whether the trained hydrogen energy power coupling model is qualified based on the hydrogen energy power coupling system parameters and the test data set is as follows: S311, inputting the test data set into the trained hydrogen energy and electricity coupling model to obtain test parameters, and establishing a test honey badger individual according to the test parameters and the population initialization formula of the hydrogen energy and electricity coupling model; Simultaneously, a honey badger individual is established and verified according to the hydrogen energy and electricity coupling system parameters and the population initialization formula of the hydrogen energy and electricity coupling model; S312, evaluating the fitness of individual honey badgers based on the objective function of the hydrogen energy and electricity coupling model to obtain test fitness, and evaluating and verifying the fitness of individual honey badgers based on the objective function of the hydrogen energy and electricity coupling model to obtain verified fitness; S313, based on the test fitness and the verification fitness, a comparison is made. If the test fitness is greater than the verification fitness, the trained hydrogen energy power coupling model is determined to be qualified; if the test fitness is less than or equal to the verification fitness, the trained hydrogen energy power coupling model is determined to be unqualified; Simultaneously, the hydrogen energy and electricity coupling model that is judged to be qualified is marked as the optimal hydrogen energy and electricity coupling model.

[0040] In one embodiment, if the hydrogen energy power coupling model verification fails and before executing S1, a cycle stop judgment needs to be performed. The specific process of the cycle stop judgment is: A comparison is made based on the current verification times of the hydrogen energy and electricity coupling model and the preset cycle threshold. If the current verification times are less than the cycle threshold, the cycle is determined to be incomplete and the cycle continues; if the current verification times are greater than or equal to the cycle threshold, the cycle is determined to be complete and the cycle is stopped.

[0041] In this embodiment, the cycle threshold is set to 30, which can be modified according to actual working conditions.

[0042] S4, collecting real-time operation data of the hydrogen energy and electricity coupling system, and inputting the real-time operation data into the optimal hydrogen energy and electricity coupling model to obtain optimal parameters; S411, cleaning the real-time operation data to obtain a real-time complete data set, and integrating the real-time complete data set based on a collection source of the real-time data to obtain a real-time integrated data set; S412, inputting the real-time integrated data set into the optimal hydrogen energy and electricity coupling model to obtain optimal parameters.

[0043] S5. The hydrogen energy and electricity coupling system performs dynamic optimization actions in response to the optimal parameters.

[0044] In this embodiment, the hydrogen energy and power coupling system adjusts the control parameters of the corresponding equipment according to the optimal parameters, optimizes the real-time operating state of the hydrogen energy and power coupling system, and then controls the real-time power of the corresponding power grid, effectively improving the stability of the power grid.

[0045] This embodiment has at least the following substantial effects: In this embodiment, the historical operation data is preprocessed to obtain a training data set and a test data set, which can ensure the integrity and accuracy of the data. In addition, the parameters of the hydrogen energy power coupling system will change with the operation time during the operation process, that is, the parameters of the hydrogen energy power coupling system have a time series structure. Therefore, the hydrogen energy power coupling system parameters corresponding to the historical operation data can be identified based on the convolutional neural network and the extreme learning machine network, and the historical state of the hydrogen energy power coupling system represented by the hydrogen energy power coupling system parameters closest to the current time scale in the past can be clarified. At the same time, according to the honey badger algorithm improved based on the chaotic learning mechanism and the training data set, the hydrogen energy power coupling system can be predicted. A hydrogen energy power coupling model based on the state changes of the energy power coupling system is developed, and the performance of the hydrogen energy power coupling model is verified by using the hydrogen energy power coupling system parameters and test data sets to ensure the effectiveness of the hydrogen energy power coupling model and its data processing capabilities. Finally, the collected real-time operation data is input into the hydrogen energy power coupling model to obtain the optimal parameters corresponding to the predicted state changes of the hydrogen energy power coupling system. The hydrogen energy power coupling system adjusts the control parameters of the corresponding equipment according to the optimal parameters to optimize the real-time operation state of the hydrogen energy power coupling system, thereby controlling the real-time power of the corresponding power grid and significantly improving the stability of the hydrogen energy power coupling system.

[0046] The above specific embodiments are preferred embodiments of the present invention, and are not intended to limit the specific implementation scope of the present invention. The scope of the present invention includes but is not limited to the present specific embodiments. All equivalent changes made in accordance with the shape, structure, and method of the present invention are within the protection scope of the present invention.

Claims

1. A dynamic optimization method for a hydrogen energy and power coupling system, characterized in that: The method comprises the following steps: S1. Obtain historical operation data of the hydrogen energy and power coupling system, and preprocess the historical operation data to obtain a training data set and a test data set; S2. Based on the convolutional neural network and extreme learning machine network, the hydrogen energy power coupling system parameters corresponding to the historical operation data are identified, and the hydrogen energy power coupling model is constructed and trained according to the honey badger algorithm improved based on the chaotic learning mechanism and the training data set; S3, based on the hydrogen energy power coupling system parameters and the test data set, verify whether the trained hydrogen energy power coupling model is qualified, if so, output the optimal hydrogen energy power coupling model, if not, execute S1; S4, collecting real-time operation data of the hydrogen energy and electricity coupling system, and inputting the real-time operation data into the optimal hydrogen energy and electricity coupling model to obtain optimal parameters; S5. The hydrogen energy and electricity coupling system performs dynamic optimization actions in response to the optimal parameters.

2. The dynamic optimization method of a hydrogen energy and power coupling system according to claim 1 is characterized in that: In S1, the specific process of preprocessing the historical operation data to obtain the training data set and the test data set is: S111, cleaning the historical operation data to obtain a complete data set, and integrating the complete data set based on a collection source of the historical operation data to obtain an integrated data set; S112: Divide the integrated data set into a training data set and a test data set based on a preset training-test ratio.

3. The dynamic optimization method of a hydrogen energy and power coupling system according to claim 1 is characterized in that: In S2, the specific process of identifying the parameters of the hydrogen energy power coupling system corresponding to the historical operation data based on the convolutional neural network and the extreme learning machine network is as follows: S211, converting the historical operation data into a two-dimensional matrix data set according to the time variation scale of the historical operation data and the convolutional neural network input data requirements; S212. Input the two-dimensional matrix data set into a convolutional neural network to obtain an operating data feature set, and input the operating data feature set into an extreme learning machine network to obtain corresponding hydrogen energy and power coupling system parameters.

4. The dynamic optimization method of a hydrogen energy and power coupling system according to claim 1 is characterized in that: In S2, the specific process of constructing and training the hydrogen energy power coupling model according to the honey badger algorithm improved based on the chaotic learning mechanism and the training data set is: S221, improving the population initialization formula of the honey badger algorithm based on the adversarial learning mechanism to obtain the individual generation formula, and improving the random value update formula of the honey badger algorithm based on the chaos mechanism to obtain the random value chaos update formula; synchronously, establishing an objective function based on equipment characteristics of electrical equipment in the hydrogen energy-electricity coupling system corresponding to the training data set; S222. Constructing a hydrogen energy and power coupling model based on the individual generation formula, random value chaos update formula and traditional honey badger algorithm; S223, inputting the training data set into the constructed hydrogen energy and electricity coupling model for training; The hydrogen energy and electricity coupling model at least includes an odor intensity formula, a position update formula, a random value chaos update formula, an individual generation formula and an objective function.

5. The dynamic optimization method of a hydrogen energy and power coupling system according to claim 4 is characterized in that: In S221, the individual generation formula is specifically: z j =lb j +r1×(ub j -lb j ); In the formula, z j is the candidate solution corresponding to the j-th honey badger individual, is the upper limit of the candidate solution, z -j is the lower bound of the candidate solution, is the opposite solution of the candidate solution, lb j is the lower bound of the candidate solution search space, ub j is the upper bound of the candidate solution search space, r1 is the first random value, and j represents the jth honey badger individual.

6. The dynamic optimization method of a hydrogen energy and power coupling system according to claim 4 is characterized in that: In S221, the random value chaotic update formula is specifically: r1 new =r1+δ i ×r1; d i+1 =4d i ×(1-d i ); In the formula, r1 new is the update value of the first random value, r1 is the first random value, is the updated value of the second random value, r2 is the second random value, is the updated value of the third random value, r3 is the third random value, δ i is the chaos value of the i-th generation, δ i+1 is the chaos value of the i+1th generation.

7. The dynamic optimization method of a hydrogen energy and power coupling system according to claim 4 is characterized in that: In S223, the specific process of inputting the training data set into the constructed hydrogen energy and electricity coupling model for training is as follows: A1. Establish a honey badger population based on the training data set and the individual generation formula; A2, determining the prey odor intensity based on the honey badger individuals in the honey badger population and the odor intensity formula; A3, updating the position of individual honey badgers in the honey badger population according to the prey odor intensity and position update formula; Synchronously, judging whether the training is completed based on the number of position updates of the honey badger individual and the preset maximum number of iterations, if the number of position updates is greater than or equal to the maximum number of iterations, the training is determined to be completed; if the number of position updates is less than the maximum number of iterations, the training is determined to be incomplete and A2 is executed; A4. When it is determined that the training is completed, the fitness values ​​of the honey badgers in the honey badger population that have completed the training are evaluated based on the objective function of the hydrogen energy and electricity coupling model, and the honey badger individual with the highest fitness value is marked as the optimal individual.

8. The dynamic optimization method of a hydrogen energy and power coupling system according to claim 1 is characterized in that: In S3, the specific process of verifying whether the trained hydrogen energy power coupling model is qualified based on the hydrogen energy power coupling system parameters and the test data set is: S311, inputting the test data set into the trained hydrogen energy and electricity coupling model to obtain test parameters, and establishing a test honey badger individual according to the test parameters and the population initialization formula of the hydrogen energy and electricity coupling model; Simultaneously, a honey badger individual is established and verified according to the hydrogen energy and electricity coupling system parameters and the population initialization formula of the hydrogen energy and electricity coupling model; S312, evaluating the fitness of individual honey badgers based on the objective function of the hydrogen energy and electricity coupling model to obtain test fitness, and evaluating and verifying the fitness of individual honey badgers based on the objective function of the hydrogen energy and electricity coupling model to obtain verified fitness; S313, based on the test fitness and the verification fitness, a comparison is made. If the test fitness is greater than the verification fitness, the trained hydrogen energy power coupling model is determined to be qualified; if the test fitness is less than or equal to the verification fitness, the trained hydrogen energy power coupling model is determined to be unqualified; Simultaneously, the hydrogen energy and electricity coupling model that is judged to be qualified is marked as the optimal hydrogen energy and electricity coupling model.

9. The dynamic optimization method of a hydrogen energy and power coupling system according to claim 1, characterized in that: In S3, if the hydrogen energy and electricity coupling model verification fails and before executing S1, a cycle stop judgment needs to be performed. The specific process of the cycle stop judgment is: A comparison is made based on the current verification times of the hydrogen energy and electricity coupling model and the preset cycle threshold. If the current verification times are less than the cycle threshold, the cycle is determined to be incomplete and the cycle continues; if the current verification times are greater than or equal to the cycle threshold, the cycle is determined to be complete and the cycle is stopped.

10. The dynamic optimization method of a hydrogen energy and power coupling system according to claim 1, characterized in that: In S4, the specific process of inputting the real-time operation data into the optimal hydrogen energy and electricity coupling model to obtain the optimal parameters is: S411, cleaning the real-time operation data to obtain a real-time complete data set, and integrating the real-time complete data set based on a collection source of the real-time data to obtain a real-time integrated data set; S412, inputting the real-time integrated data set into the optimal hydrogen energy and electricity coupling model to obtain optimal parameters.

Images