A new energy design optimization method and system

Abstract

The present invention discloses a new energy design optimization method and system, the method includes: according to the initial design parameters of the new energy system, initializing the parameter initial value to generate a multivariate parameter candidate set; iteratively optimizing according to the multivariate parameter candidate set to obtain an optimized parameter group; classifying the optimized parameter group, respectively training the corresponding deep learning prediction sub-models of the sub-performance of the new energy system; according to each type of optimization parameter, predicting each sub-performance of the new energy system based on each deep learning prediction sub-model; if there is a prediction result of a sub-performance that does not meet the preset standard, the optimization parameters of the sub-performance that does not meet the preset standard are used as a new multivariate parameter candidate set, and returning to execute the step of iteratively optimizing according to the multivariate parameter candidate set to obtain the optimized parameter group, until the prediction results of the sub-performance all meet the preset standard. By using the embodiments of the present invention, the scientificity and efficiency of parameter optimization can be improved, and the accuracy and reliability of performance prediction can be enhanced.

Description

Technical Field

[0001] The present invention belongs to the field of new energy technology, and in particular to a new energy design optimization method and system. Background Art

[0002] As the world pays more and more attention to renewable energy, the development of new energy technology is advancing rapidly. The application of renewable energy such as wind energy, solar energy, geothermal energy, etc. not only plays an important role in mitigating global warming and reducing greenhouse gas emissions, but also becomes a key research area in achieving sustainable use of energy and ensuring energy security. The design and optimization of new energy systems are directly related to their economy and efficiency, so it is particularly important to develop an effective optimization method. Summary of the invention

[0003] The purpose of the present invention is to provide a new energy design optimization method and system to address the deficiencies in the prior art. By comprehensively considering multiple parameter candidate sets, iterative optimization and deep learning technology, it can not only improve the scientificity and efficiency of parameter optimization, but also enhance the accuracy and reliability of performance prediction.

[0004] An embodiment of the present application provides a new energy design optimization method, the method comprising:

[0005] According to the initial design parameters of the new energy system, the initial values ​​of the parameters are initialized to generate a multivariate parameter candidate set;

[0006] According to the current multivariate parameter candidate set, an optimized parameter group is obtained by iterative optimization through an evolutionary process;

[0007] Fuzzy clustering technology is used to classify the optimization parameter groups, and deep learning prediction sub-models are trained to predict the sub-performance of the new energy system corresponding to each type of optimization parameter;

[0008] According to the classified optimization parameters of each category, the sub-performance of the new energy system is predicted based on the trained deep learning prediction sub-models;

[0009] If there is a sub-performance whose prediction result fails to meet the preset standard, the optimization parameters corresponding to the sub-performance that fails to meet the preset standard are used as a new multivariate parameter candidate set, and the process returns to execute the iterative optimization based on the current multivariate parameter candidate set through the evolution process to obtain the step of optimizing the parameter group until the prediction results of all sub-performances meet the preset standard.

[0010] Optionally, the initializing of the initial values ​​of the parameters according to the initial design parameters of the new energy system to generate a multivariate parameter candidate set includes:

[0011] Normalize the initial design parameters and convert them to the [0,1] interval to meet the input requirements of chaotic mapping;

[0012] Using the preset chaotic mapping function, the normalized initial design parameters are used as input to generate the corresponding chaotic sequence. The chaotic sequence is obtained by iterating the chaotic mapping function multiple times, and the new values ​​generated in each iteration will be used as the input for the next iteration.

[0013] The generated chaotic sequence is denormalized to map back to the original parameter range, and each chaotic sequence value is mapped one by one to the specific parameters of the new energy system as a multivariate parameter candidate set.

[0014] Optionally, the iterative optimization is performed through an evolutionary process according to the current multivariate parameter candidate set to obtain an optimized parameter group, including:

[0015] Constructing an initial population according to the multivariate parameter candidate set, wherein each individual in the population represents a specific parameter combination;

[0016] Calculating the fitness value for each individual using a preset fitness function, wherein the fitness function uses a weighted method to normalize and combine the performance indicators of each new energy system;

[0017] A biologically inspired selection mechanism is used to select individuals whose fitness values ​​are higher than the preset fitness threshold, and new individuals are generated through crossover and mutation operations, and the selected individuals and new individuals are used as the new generation population;

[0018] For the new generation population, returning to the step of calculating the fitness value for each individual using a preset fitness function to perform iterative optimization until the iteration reaches a preset maximum number of generations or the fitness change is less than a preset change threshold;

[0019] Get the individual with the highest fitness in the final generation as the optimization parameter group.

[0020] Optionally, the method of classifying the optimization parameter groups by using fuzzy clustering technology includes:

[0021] Obtaining an objective function applied to fuzzy clustering and determining the number of clusters, wherein the objective function represents the degree of membership of each optimization parameter group and each cluster center;

[0022] By calculating the membership degree of each optimization parameter with each cluster, the optimization parameter is assigned to the cluster with the smallest membership degree with the optimization parameter.

[0023] Another embodiment of the present application provides a new energy design optimization system, the system comprising:

[0024] A processing module, used for initializing the initial values ​​of the parameters according to the initial design parameters of the new energy system, and generating a multivariate parameter candidate set;

[0025] The optimization module is used to perform iterative optimization through an evolutionary process according to the current multivariate parameter candidate set to obtain an optimized parameter group;

[0026] A training module, for classifying the optimization parameter groups using fuzzy clustering technology, and respectively training a deep learning prediction sub-model for predicting the sub-performance of the new energy system corresponding to each type of optimization parameter;

[0027] The prediction module is used to predict the sub-performance of the new energy system based on the optimized parameters of each category after classification and the trained deep learning prediction sub-models;

[0028] The iterative module is used to, if the prediction result of a sub-performance does not meet the preset standard, use the optimization parameters corresponding to the sub-performance that does not meet the preset standard as a new multivariate parameter candidate set, return to execute the iterative optimization according to the current multivariate parameter candidate set through the evolutionary process to obtain the step of optimizing the parameter group until the prediction results of all sub-performances meet the preset standard.

[0029] Yet another embodiment of the present application provides a storage medium, wherein the storage medium stores a computer program, wherein the computer program is configured to execute any of the above methods when running.

[0030] Yet another embodiment of the present application provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute any of the methods described above.

[0031] Compared with the prior art, the present invention provides a new energy design optimization method, which initializes the initial values ​​of the parameters according to the initial design parameters of the new energy system to generate a multivariate parameter candidate set; iteratively optimizes through an evolutionary process according to the current multivariate parameter candidate set to obtain an optimization parameter group; classifies the optimization parameter group using fuzzy clustering technology, and trains deep learning prediction sub-models for predicting the sub-performance of the new energy system corresponding to each type of optimization parameter; predicts each sub-performance of the new energy system based on the trained deep learning prediction sub-models according to each type of optimization parameter after classification; if there is a sub-performance whose prediction result does not meet the preset standard, the optimization parameter corresponding to the sub-performance that does not meet the preset standard is used as a new multivariate parameter candidate set, and returns to execute iterative optimization through an evolutionary process according to the current multivariate parameter candidate set, thereby being able to improve the scientificity and efficiency of parameter optimization and enhance the accuracy and reliability of performance prediction by comprehensively considering the multivariate parameter candidate set, iterative optimization and deep learning technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 A hardware structure block diagram of a computer terminal for a new energy design optimization method provided by an embodiment of the present invention;

[0033] Figure 2 A schematic diagram of a new energy design optimization method provided by an embodiment of the present invention;

[0034] Figure 3 A schematic diagram of the structure of a new energy design optimization system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0035] The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, but should not be construed as limiting the present invention.

[0036] The embodiment of the present invention first provides a new energy design optimization method, which can be applied to electronic devices, such as computer terminals, specifically ordinary computers, etc.

[0037] The following describes it in detail by taking running on a computer terminal as an example. Figure 1 The hardware structure block diagram of a computer terminal of a new energy design optimization method provided by an embodiment of the present invention. Figure 1 As shown, the computer terminal may include one or more ( Figure 1 Only one is shown in the figure) a processor 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data. Optionally, the computer terminal may also include a transmission device 106 for communication functions and an input and output device 108. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above-mentioned computer terminal. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0038] The memory 104 can be used to store software programs and modules of application software, such as program instructions / modules corresponding to the new energy design optimization method in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, implementing the above method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the computer terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0039] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of a computer terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (Radio Frequency, RF) module, which is used to communicate with the Internet wirelessly.

[0040] See also Figure 2 , an embodiment of the present invention provides a new energy design optimization method, which may include the following steps:

[0041] S201, initializing the initial values ​​of the parameters according to the initial design parameters of the new energy system to generate a multivariate parameter candidate set;

[0042] In the design process of new energy systems, the initial design parameters are usually set by designers based on experience and basic theory. These initial parameter values ​​have a direct impact on the performance of the system. However, these initial parameters may not be optimal and may require further optimization and adjustment. By initializing these parameters, multiple candidate parameter sets can be generated to provide basic data for subsequent optimization.

[0043] The generated multivariate parameter candidate set lays the foundation for the subsequent optimization steps. It makes the design process more flexible, and can explore a wider design space based on the initial settings to find possible optimization directions. At the same time, this process can also provide diversity for the performance prediction of new energy systems, which helps to improve the reliability and efficiency of the final design.

[0044] Specifically, the initial design parameters can be normalized and converted to the [0,1] interval to meet the input requirements of the chaotic mapping;

[0045] Normalization is to convert parameters of different dimensions to a unified scale so that they can be treated equally in subsequent operations. By scaling the parameter values ​​to the [0,1] interval, the differences in actual values ​​can be eliminated, ensuring the stability and effectiveness of the chaotic mapping process. Normalization helps to improve the stability of the subsequent chaotic mapping algorithm, avoid calculation errors caused by differences in parameter values, and ensure that the generated chaotic sequence can accurately reflect the characteristics of the initial parameters.

[0046] Using the preset chaotic mapping function, the normalized initial design parameters are used as input to generate the corresponding chaotic sequence. The chaotic sequence is obtained by iterating the chaotic mapping function multiple times, and the new values ​​generated in each iteration will be used as the input for the next iteration.

[0047] Chaotic mapping function is a mathematical function used to generate nonlinear and random sequences, including Logistic mapping, etc. By iterating the normalized initial design parameters multiple times, a series of chaotic sequences with good distribution characteristics can be obtained. These sequences contain diversity and potential effective parameter combinations. The generated chaotic sequence can effectively break the limitations of the parameter space, and then explore more potential optimization space, providing rich data support for subsequent parameter optimization and prediction.

[0048] The generated chaotic sequence is denormalized to map back to the original parameter range, and each chaotic sequence value is mapped one by one to the specific parameters of the new energy system as a multivariate parameter candidate set.

[0049] The denormalization process is to convert the generated chaotic sequence from the interval [0,1] back to the actual range of the initial design parameters. By mapping the chaotic sequence values ​​back to the original parameter range, candidate parameters that meet the actual conditions can be obtained, ensuring the practical applicability of the optimization results. This process makes the candidate parameters generated by the optimization not only theoretically extensive, but also feasible in practical applications, ensuring that every parameter generated during the design process can be applied to the actual design of the new energy system.

[0050] S202, performing iterative optimization through an evolutionary process according to the current multivariate parameter candidate set to obtain an optimized parameter group;

[0051] The evolutionary process refers to the iterative optimization of a multivariate parameter candidate set by simulating mechanisms such as natural selection, crossover, and mutation using the idea of ​​biological evolution. This process usually starts from an initial population, where each individual represents a specific parameter combination. Through multiple generations of selection, crossover, and mutation, individuals with higher fitness are gradually selected, ultimately forming a high-quality optimization parameter set.

[0052] Specifically, an initial population may be constructed according to the multivariate parameter candidate set, wherein each individual in the population represents a specific parameter combination;

[0053] In evolutionary algorithms (such as genetic algorithms, etc.), the construction of the initial population is the basis of the optimization process. Each individual is a specific parameter configuration generated by a random combination of a multivariate parameter candidate set. These individuals together constitute the search space and can be selected and reproduced in subsequent optimization. The diversity of the initial population directly affects the efficiency and effect of the subsequent evolutionary process. By constructing an initial population with rich diversity, the breadth of the search can be increased, the possibility of finding the global optimal solution can be increased, and the quality of the optimization results can be improved.

[0054] The diversity of the initial population directly affects the efficiency and effect of the subsequent evolution process. By constructing an initial population with rich diversity, the breadth of the search can be increased, the possibility of finding the global optimal solution can be increased, and the quality of the optimization results can be improved.

[0055] Calculating the fitness value for each individual using a preset fitness function, wherein the fitness function uses a weighted method to normalize and combine the performance indicators of each new energy system;

[0056] The fitness function is the key in the evolutionary algorithm, which is used to evaluate the quality of each individual. By normalizing multiple performance indicators and combining them with weighting factors, the overall contribution of each parameter combination to the new energy system can be comprehensively evaluated. The fitness value can effectively guide the selection process and give priority to those individuals who contribute more to the improvement of system performance. Therefore, the appropriate fitness function design can improve the search efficiency and help to quickly find a better parameter configuration.

[0057] A biologically inspired selection mechanism is used to select individuals whose fitness values ​​are higher than the preset fitness threshold, and new individuals are generated through crossover and mutation operations, and the selected individuals and new individuals are used as the new generation population;

[0058] In biological evolution, survival of the fittest is the core of the selection mechanism. By selecting individuals with fitness higher than the threshold, they have a greater chance of being selected into the next round of evolution, and generating new individuals through crossover and mutation operations, thereby introducing new genetic features and increasing the diversity of the population. This selection mechanism combined with crossover and mutation operations can effectively avoid premature phenomena, enhance the exploration ability of the population, ensure the continuity and effectiveness of the optimization process, and enable the final parameter group to cover a wider parameter space.

[0059] You can first set a fitness threshold and select individuals with fitness values ​​above the threshold. Then, randomly select two individuals for crossover to form new individuals, for example, take a partial parameter combination of two individuals. Then, apply a mutation operation with a certain probability to the new individuals to introduce randomness and ensure the diversity of the new generation of population.

[0060] For the new generation population, returning to the step of calculating the fitness value for each individual using a preset fitness function to perform iterative optimization until the iteration reaches a preset maximum number of generations or the fitness change is less than a preset change threshold;

[0061] This process is the iterative mechanism of the evolutionary algorithm. In each generation, the performance of the new individual is evaluated by calculating the fitness value, and then deciding whether to continue evolving until the stopping condition is reached. Through continuous iterative optimization, the evolutionary algorithm can gradually improve the accuracy and efficiency of finding the optimal parameter combination. After reaching the stopping condition, it can effectively ensure that the optimized parameter group obtained is the result of sufficient exploration and verification, and has practical application value.

[0062] After each round of iteration, you can check whether the current generation has reached the preset maximum generation, or monitor whether the fitness change is less than the preset change threshold, that is, convergence. If any condition is met, the iteration is stopped; otherwise, fitness calculation and selection operations are continued to form new individuals and populations. Finally, the individual with the highest fitness is selected as the final optimization parameter group.

[0063] Get the individual with the highest fitness in the final generation as the optimization parameter group.

[0064] The main purpose of this optimization process is to find the best parameter combination to improve the performance of the new energy system. Through multiple iterations and interactions, the parameter space can be effectively explored to find some excellent configurations that are difficult to find using traditional methods. This not only improves design efficiency, but also ensures that the final parameter group has better performance in actual applications.

[0065] S203, classifying the optimization parameter groups using fuzzy clustering technology, and respectively training deep learning prediction sub-models for predicting the sub-performance of the new energy system corresponding to each type of optimization parameter;

[0066] Fuzzy clustering technology is an unsupervised learning algorithm that can classify parameter groups into different categories based on their similarities. Each category contains parameter combinations that have similar characteristics or behaviors to some extent. By training a deep learning model on the parameters of each category, the impact of these parameter combinations on the sub-performance of new energy systems can be captured more accurately.

[0067] By classifying the optimization parameter groups, the accuracy of subsequent performance prediction can be effectively improved. By training independent deep learning prediction sub-models for each type of parameter, more detailed performance analysis can be performed using the features of different categories, thereby improving the prediction accuracy of the overall performance of the new energy system. This method can also help identify the potential impact of different parameter combinations in practical applications, thereby providing a more scientific basis for the optimal design of the power system.

[0068] Specifically, by using fuzzy clustering technology to classify the optimization parameter groups, an objective function applied to fuzzy clustering can be obtained, and the number of clusters can be determined, wherein the objective function represents the degree of membership of each optimization parameter group and each cluster center;

[0069] The objective function of fuzzy clustering usually determines the degree of membership of each data point in each cluster by calculating the distance between each data point and each cluster center. The objective function can be to minimize the sum of distances within the cluster, and determine which cluster each parameter combination is assigned to based on the value of the membership. By defining a clear objective function, the clustering process can be made more effective and the scientificity and rationality of the clustering results can be ensured. Determining a reasonable number of clusters is crucial for subsequent analysis, and can help the model find a moderate balance in performance and avoid overfitting or underfitting.

[0070] You can use common fuzzy clustering algorithms such as FCM (Fuzzy C-means), set a cluster number K, and then calculate the distance between each optimized parameter group and the K cluster centers. According to the calculated membership, the cluster center is repeatedly adjusted until the convergence condition is reached, and finally the cluster classification of each parameter group is determined.

[0071] By calculating the membership degree of each optimization parameter with each cluster, the optimization parameter is assigned to the cluster with the smallest membership degree with the optimization parameter.

[0072] The core of this step is to calculate the membership based on the distance (or similarity) between each optimized parameter group and the cluster center. Each parameter group will be assigned to the cluster with the highest membership, thus forming the final classification result. By performing accurate membership calculations, it can be ensured that each parameter combination is assigned to the most appropriate cluster, which helps with subsequent deep learning model training. Ensuring that the parameter combinations of each category are similar in characteristics can better improve the learning efficiency and accuracy of the model, thereby improving the performance prediction of new energy systems.

[0073] S204, predicting various sub-performances of the new energy system based on each of the classified optimization parameters and the trained deep learning prediction sub-models;

[0074] After the optimization parameters are classified, each category of optimization parameters corresponds to a specially trained deep learning prediction sub-model. These sub-models are based on the characteristics of each type of optimization parameters, and are learned and trained through the relationship between historical data and performance indicators, so as to convert the input parameters into corresponding new energy system sub-performance predictions. In this way, targeted predictions for different types of parameter configurations can be achieved, improving the accuracy and reliability of performance evaluation.

[0075] By performing deep learning prediction based on the classified optimization parameters, the fine-grained features of the model can be fully utilized to accurately predict the corresponding sub-performance for the different features of each parameter combination. This method can effectively improve the overall design efficiency of new energy systems, while reducing R&D costs and time costs, and provide decision makers with more reliable performance evaluations. In addition, the prediction results can help designers adjust the system configuration in a timely manner to meet the preset performance standards.

[0076] Specifically, each type of optimization parameter and the corresponding system sub-performance can be regarded as a node in a graph. The features of each node can include the numerical value of the parameter, historical performance data and their relationship with each other (such as correlation, weight, etc.), and the related nodes are connected by edges to form a graph structure. The node features are aggregated through the graph convolutional network (GCN). In the training phase, the feature vector of the node is calculated by constructing a graph for each category of parameters, so that the influence of adjacent nodes can be propagated, the accuracy of the representation is improved, and a global understanding is formed.

[0077] Based on GNN, a multi-task learning framework is built to construct multiple output layers for different sub-performance indicators. The parameters of the graph convolution layer are shared during training. At the same time, a dedicated loss function and optimization strategy are designed for the characteristics of each sub-performance to ensure that the model can find a balance between different sub-performances. By training the sub-performance of each type of parameter combination, an appropriate optimization algorithm (such as Adam) is used to minimize the loss function of each sub-performance prediction. At the same time, in the verification stage, the model is evaluated using cross-validation and hyperparameters are adjusted to prevent overfitting. After the model training is completed, a new combination of optimized parameters is input, and the graph neural network is used for inference to output the prediction results of each sub-performance.

[0078] This multi-task learning method based on graph neural network can make full use of the relationship between parameters in the optimization design process to achieve efficient and accurate sub-performance prediction. It has good scalability and is suitable for complex new energy system design scenarios.

[0079] S205, if there is a sub-performance whose prediction result does not meet the preset standard, the optimization parameters corresponding to the sub-performance that does not meet the preset standard are used as a new multi-parameter candidate set, and the process returns to execute the iterative optimization step according to the current multi-parameter candidate set through the evolution process to obtain the optimization parameter group, until the prediction results of all sub-performances meet the preset standard.

[0080] In the process of new energy design optimization, the accuracy of predicted sub-performance is directly related to the overall effectiveness of the system. When the predicted sub-performance results do not meet the preset standards, the program needs to re-evaluate the failed parameter combinations. Using the optimization parameters corresponding to the sub-performance outputs that do not meet the standards as new candidate sets means focusing on those parameters that have potential but have not yet achieved the optimization effect. This iterative process can gradually approach the optimal solution by continuously adjusting parameters and exploring the optimization space.

[0081] Through this feedback mechanism, the optimization process can adjust itself to ensure that the design results can meet the actual needs. This not only improves the performance and efficiency of the system, but also effectively reduces the waste of resources. In addition, the iterative optimization process enables designers to deeply understand the impact of each set of parameters on system performance, ultimately helping them better master and apply these parameters and optimize the overall design strategy.

[0082] Specifically, all sub-performance criteria can be regarded as a partially ordered set, in which each combination of optimization parameters is an element and the sub-performance target is an attribute. A priority model is created by comparing the relative advantages and disadvantages between each set of parameter combinations and the preset standards. By traversing all prediction results, those parameter combinations that do not meet the preset standards are screened out to build a new multivariate parameter candidate set. Only the optimization parameters that do not meet the standards are marked as "to be optimized" and returned to execute iterative optimization based on the current multivariate parameter candidate set through the evolution process. After each iteration, the prediction results of each optimization parameter combination are re-evaluated and compared with the standards in the partially ordered set. If all sub-performances meet the preset standards, the algorithm terminates; otherwise, the above steps are repeated to form a self-optimization closed loop to ensure that the design solution is improved through continuous iteration.

[0083] It can be seen that according to the initial design parameters of the new energy system, the initial values ​​of the parameters are initialized to generate a multivariate parameter candidate set; according to the current multivariate parameter candidate set, iterative optimization is performed through the evolutionary process to obtain an optimized parameter group; the optimized parameter group is classified using fuzzy clustering technology, and the deep learning prediction sub-models used to predict the sub-performance of the new energy system corresponding to each type of optimization parameter are trained separately; according to each type of classified optimization parameter, based on the trained deep learning prediction sub-models, the various sub-performances of the new energy system are predicted; if there is a sub-performance whose prediction result does not meet the preset standard, the optimization parameter corresponding to the sub-performance that does not meet the preset standard is used as a new multivariate parameter candidate set, and it is returned to execute according to the current multivariate parameter candidate set, through the evolutionary process, iterative optimization is performed, so that by comprehensively considering the multivariate parameter candidate set, iterative optimization and deep learning technology, not only the scientificity and efficiency of parameter optimization are improved, but also the accuracy and reliability of performance prediction are enhanced.

[0084] Another embodiment of the present invention provides a new energy design optimization system, see Figure 3 , the system may include:

[0085] The processing module 301 is used to initialize the initial values ​​of the parameters according to the initial design parameters of the new energy system to generate a multivariate parameter candidate set;

[0086] The optimization module 302 is used to perform iterative optimization through an evolutionary process according to the current multivariate parameter candidate set to obtain an optimized parameter group;

[0087] A training module 303 is used to classify the optimization parameter groups using fuzzy clustering technology, and to train deep learning prediction sub-models for predicting the sub-performance of the new energy system corresponding to each type of optimization parameter;

[0088] The prediction module 304 is used to predict various sub-performances of the new energy system based on the classified optimization parameters of each category and the trained deep learning prediction sub-models;

[0089] Iterative module 305 is used to, if there is a sub-performance whose prediction result does not meet the preset standard, use the optimization parameters corresponding to the sub-performance that does not meet the preset standard as a new multi-parameter candidate set, and return to execute the step of iteratively optimizing the current multi-parameter candidate set through the evolutionary process to obtain the optimized parameter group, until the prediction results of all sub-performances meet the preset standard.

[0090] It can be seen that according to the initial design parameters of the new energy system, the initial values ​​of the parameters are initialized to generate a multivariate parameter candidate set; according to the current multivariate parameter candidate set, iterative optimization is performed through the evolutionary process to obtain an optimized parameter group; the optimized parameter group is classified using fuzzy clustering technology, and the deep learning prediction sub-models used to predict the sub-performance of the new energy system corresponding to each type of optimization parameter are trained separately; according to each type of classified optimization parameter, based on the trained deep learning prediction sub-models, the various sub-performances of the new energy system are predicted; if there is a sub-performance whose prediction result does not meet the preset standard, the optimization parameter corresponding to the sub-performance that does not meet the preset standard is used as a new multivariate parameter candidate set, and it is returned to execute according to the current multivariate parameter candidate set, through the evolutionary process, iterative optimization is performed, so that by comprehensively considering the multivariate parameter candidate set, iterative optimization and deep learning technology, not only the scientificity and efficiency of parameter optimization are improved, but also the accuracy and reliability of performance prediction are enhanced.

[0091] An embodiment of the present invention further provides a storage medium, in which a computer program is stored, wherein the computer program is configured to execute the steps of any of the above method embodiments when running.

[0092] Specifically, in this embodiment, the above storage medium may be configured to store a computer program for performing the following steps:

[0093] S201, initializing the initial values ​​of the parameters according to the initial design parameters of the new energy system to generate a multivariate parameter candidate set;

[0094] S202, performing iterative optimization through an evolutionary process according to the current multivariate parameter candidate set to obtain an optimized parameter group;

[0095] S203, classifying the optimization parameter groups using fuzzy clustering technology, and respectively training deep learning prediction sub-models for predicting the sub-performance of the new energy system corresponding to each type of optimization parameter;

[0096] S204, predicting various sub-performances of the new energy system based on each of the classified optimization parameters and the trained deep learning prediction sub-models;

[0097] S205, if there is a sub-performance whose prediction result does not meet the preset standard, the optimization parameters corresponding to the sub-performance that does not meet the preset standard are used as a new multi-parameter candidate set, and the process returns to execute the iterative optimization step according to the current multi-parameter candidate set through the evolution process to obtain the optimization parameter group, until the prediction results of all sub-performances meet the preset standard.

[0098] It can be seen that according to the initial design parameters of the new energy system, the initial values ​​of the parameters are initialized to generate a multivariate parameter candidate set; according to the current multivariate parameter candidate set, iterative optimization is performed through the evolutionary process to obtain an optimized parameter group; the optimized parameter group is classified using fuzzy clustering technology, and the deep learning prediction sub-models used to predict the sub-performance of the new energy system corresponding to each type of optimization parameter are trained separately; according to each type of classified optimization parameter, based on the trained deep learning prediction sub-models, the various sub-performances of the new energy system are predicted; if there is a sub-performance whose prediction result does not meet the preset standard, the optimization parameter corresponding to the sub-performance that does not meet the preset standard is used as a new multivariate parameter candidate set, and it is returned to execute according to the current multivariate parameter candidate set, through the evolutionary process, iterative optimization is performed, so that by comprehensively considering the multivariate parameter candidate set, iterative optimization and deep learning technology, not only the scientificity and efficiency of parameter optimization are improved, but also the accuracy and reliability of performance prediction are enhanced.

[0099] An embodiment of the present invention further provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0100] Specifically, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0101] Specifically, in this embodiment, the processor may be configured to perform the following steps through a computer program:

[0102] S201, initializing the initial values ​​of the parameters according to the initial design parameters of the new energy system to generate a multivariate parameter candidate set;

[0103] S202, performing iterative optimization through an evolutionary process according to the current multivariate parameter candidate set to obtain an optimized parameter group;

[0104] S203, classifying the optimization parameter groups using fuzzy clustering technology, and respectively training deep learning prediction sub-models for predicting the sub-performance of the new energy system corresponding to each type of optimization parameter;

[0105] S204, predicting various sub-performances of the new energy system based on each of the classified optimization parameters and the trained deep learning prediction sub-models;

[0106] S205, if there is a sub-performance whose prediction result does not meet the preset standard, the optimization parameters corresponding to the sub-performance that does not meet the preset standard are used as a new multi-parameter candidate set, and the process returns to execute the iterative optimization step according to the current multi-parameter candidate set through the evolution process to obtain the optimization parameter group, until the prediction results of all sub-performances meet the preset standard.

[0107] It can be seen that according to the initial design parameters of the new energy system, the initial values ​​of the parameters are initialized to generate a multivariate parameter candidate set; according to the current multivariate parameter candidate set, iterative optimization is performed through the evolutionary process to obtain an optimized parameter group; the optimized parameter group is classified using fuzzy clustering technology, and the deep learning prediction sub-models used to predict the sub-performance of the new energy system corresponding to each type of optimization parameter are trained separately; according to each type of classified optimization parameter, based on the trained deep learning prediction sub-models, the various sub-performances of the new energy system are predicted; if there is a sub-performance whose prediction result does not meet the preset standard, the optimization parameter corresponding to the sub-performance that does not meet the preset standard is used as a new multivariate parameter candidate set, and it is returned to execute according to the current multivariate parameter candidate set, through the evolutionary process, iterative optimization is performed, so that by comprehensively considering the multivariate parameter candidate set, iterative optimization and deep learning technology, not only the scientificity and efficiency of parameter optimization are improved, but also the accuracy and reliability of performance prediction are enhanced.

[0108] The above describes in detail the structure, features and effects of the present invention based on the embodiments shown in the drawings. The above is only a preferred embodiment of the present invention, but the present invention is not limited to the scope of implementation shown in the drawings. Any changes made according to the concept of the present invention, or modifications to equivalent embodiments with equivalent changes, which still do not exceed the spirit covered by the description and drawings, should be within the protection scope of the present invention.

Claims

1. A new energy design optimization method, characterized in that: The method comprises: Initializing the initial values ​​of the parameters according to the initial design parameters of the new energy system to generate a multivariate parameter candidate set; Initializing the initial values ​​of the parameters according to the initial design parameters of the new energy system to generate a multivariate parameter candidate set includes: Normalize the initial design parameters and convert them to the [0,1] interval to meet the input requirements of chaotic mapping; Using the preset chaotic mapping function, the normalized initial design parameters are used as input to generate the corresponding chaotic sequence. The chaotic sequence is obtained by iterating the chaotic mapping function multiple times, and the new values ​​generated in each iteration will be used as the input for the next iteration. The generated chaotic sequence is denormalized to map back to the original parameter range, and each chaotic sequence value is mapped one by one to the specific parameters of the new energy system as a multivariate parameter candidate set; According to the current multivariate parameter candidate set, an optimized parameter group is obtained by iterative optimization through an evolutionary process; wherein the evolutionary process starts from an initial population, each individual represents a parameter combination, and through multiple generations of selection, crossover and mutation, individuals with higher fitness are gradually screened out to finally form an optimized parameter group, wherein the initial population is constructed according to the multivariate parameter candidate set; Fuzzy clustering technology is used to classify the optimization parameter groups, and deep learning prediction sub-models are trained to predict the sub-performance of the new energy system corresponding to each type of optimization parameter; According to the classified optimization parameters of each category, the sub-performance of the new energy system is predicted based on the trained deep learning prediction sub-models; If there is a sub-performance whose prediction result fails to meet the preset standard, the optimization parameters corresponding to the sub-performance that fails to meet the preset standard are used as a new multivariate parameter candidate set, and the process returns to execute the iterative optimization based on the current multivariate parameter candidate set through the evolution process to obtain the step of optimizing the parameter group until the prediction results of all sub-performances meet the preset standard.

2. The method according to claim 1, characterized in that The iterative optimization is performed according to the current multivariate parameter candidate set through the evolution process to obtain the optimized parameter group, including: Constructing an initial population according to the multivariate parameter candidate set, wherein each individual in the population represents a parameter combination; Calculating the fitness value for each individual using a preset fitness function, wherein the fitness function uses a weighted method to normalize and combine the performance indicators of each new energy system; A biologically inspired selection mechanism is used to select individuals whose fitness values ​​are higher than the preset fitness threshold, and new individuals are generated through crossover and mutation operations, and the selected individuals and new individuals are used as the new generation population; For the new generation population, returning to the step of calculating the fitness value for each individual using a preset fitness function to perform iterative optimization until the iteration reaches a preset maximum number of generations or the fitness change is less than a preset change threshold; Get the individual with the highest fitness in the final generation as the optimization parameter group.

3. The method according to claim 2, characterized in that The method of using fuzzy clustering technology to classify the optimization parameter groups includes: Obtaining an objective function applied to fuzzy clustering and determining the number of clusters, wherein the objective function represents the degree of membership of each optimization parameter group and each cluster center; By calculating the membership degree of each optimization parameter with each cluster, the optimization parameter is assigned to the cluster with the smallest membership degree with the optimization parameter.

4. A new energy design optimization system, characterized in that: The system comprises: The processing module is used to initialize the initial values ​​of the parameters according to the initial design parameters of the new energy system to generate a multivariate parameter candidate set; the processing module is specifically used to: Normalize the initial design parameters and convert them to the [0,1] interval to meet the input requirements of chaotic mapping; Using the preset chaotic mapping function, the normalized initial design parameters are used as input to generate the corresponding chaotic sequence. The chaotic sequence is obtained by iterating the chaotic mapping function multiple times, and the new values ​​generated in each iteration will be used as the input for the next iteration. The generated chaotic sequence is denormalized to map back to the original parameter range, and each chaotic sequence value is mapped one by one to the specific parameters of the new energy system as a multivariate parameter candidate set; An optimization module is used to perform iterative optimization through an evolutionary process according to the current multivariate parameter candidate set to obtain an optimized parameter group; wherein the evolutionary process starts from an initial population, each individual represents a parameter combination, and through multiple generations of selection, crossover and mutation, individuals with higher fitness are gradually screened out to finally form an optimized parameter group, wherein the initial population is constructed according to the multivariate parameter candidate set; A training module, for classifying the optimization parameter groups using fuzzy clustering technology, and respectively training a deep learning prediction sub-model for predicting the sub-performance of the new energy system corresponding to each type of optimization parameter; The prediction module is used to predict the sub-performance of the new energy system based on the optimized parameters of each category after classification and the trained deep learning prediction sub-models; The iterative module is used to, if the prediction result of a sub-performance does not meet the preset standard, use the optimization parameters corresponding to the sub-performance that does not meet the preset standard as a new multivariate parameter candidate set, return to execute the iterative optimization according to the current multivariate parameter candidate set through the evolutionary process to obtain the step of optimizing the parameter group until the prediction results of all sub-performances meet the preset standard.

5. The system according to claim 4, characterized in that The optimization module is specifically used for: Constructing an initial population according to the multivariate parameter candidate set, wherein each individual in the population represents a parameter combination; Calculating the fitness value for each individual using a preset fitness function, wherein the fitness function uses a weighted method to normalize and combine the performance indicators of each new energy system; A biologically inspired selection mechanism is used to select individuals whose fitness values ​​are higher than the preset fitness threshold, and new individuals are generated through crossover and mutation operations, and the selected individuals and new individuals are used as the new generation population; For the new generation population, returning to the step of calculating the fitness value for each individual using a preset fitness function to perform iterative optimization until the iteration reaches a preset maximum number of generations or the fitness change is less than a preset change threshold; Get the individual with the highest fitness in the final generation as the optimization parameter group.

6. The system according to claim 5, characterized in that The training module is specifically used for: Obtaining an objective function applied to fuzzy clustering and determining the number of clusters, wherein the objective function represents the degree of membership of each optimization parameter group and each cluster center; By calculating the membership degree of each optimization parameter with each cluster, the optimization parameter is assigned to the cluster with the smallest membership degree with the optimization parameter.

7. A storage medium, characterized in that: The storage medium stores a computer program, wherein the computer program is configured to execute the method according to any one of claims 1 to 3 when executed.

8. An electronic device, comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to run the computer program to perform the method according to any one of claims 1 to 3.

Images