Operation strategy optimization method for energy storage EMS system
By building a load adaptive predictor and optimization algorithm, combined with the operation characteristics of the energy storage system, the operation strategy of the energy storage EMS system is determined and optimized, and the problems of low efficiency and insufficient stability caused by inaccurate operation strategies in the existing technology are solved, and the problem of inaccurate operation strategies is achieved. Efficient and stable energy management.
Patent Information
- Application Number
- CN202410941638.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-15
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2044-07-15
AI Technical Summary
The existing energy storage EMS system has insufficient operating strategies for energy conversion and regulation, resulting in low energy utilization efficiency, slow system response speed and insufficient stability.
By obtaining the power load prediction requirements, building a load adaptive predictor for power load prediction, and determining the operation strategy based on the operation characteristics analysis of the energy storage system, building a strategy parameter solution space, configuring an optimization algorithm, and using the optimization algorithm to find optimization in the solution space. Dynamic feedback adjustment is performed to improve system efficiency and stability.
It realizes accurate prediction of power load, improves the efficiency of energy conversion and regulation processes, enhances the stability and response speed of the system, and meets the needs of modern energy management.
Smart Images

Figure CN118472981B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of energy management technology, specifically to the field of energy storage system regulation technology, and in particular to an operation strategy optimization method for an energy storage EMS system. Background Art
[0002] With the growing global energy demand and the rapid development of renewable energy technology, energy storage systems, as a key technology for energy conversion and regulation, are receiving increasing attention. Energy storage systems can effectively solve the volatility and intermittency problems of renewable energy, improve energy utilization efficiency, and ensure the stable operation of the power grid. However, existing energy storage EMS systems often have problems with inaccurate and inefficient operation strategies during energy conversion and regulation. This leads to technical problems such as low energy utilization efficiency, slow system response speed, and insufficient stability, which cannot meet the needs of modern energy management. Summary of the invention
[0003] This application provides an operating strategy optimization method for an energy storage EMS system, aiming to solve the technical problems in the prior art of the energy storage EMS system in the process of energy conversion and regulation, such as low energy utilization efficiency, slow system response speed and insufficient stability due to the inaccurate and optimized operating strategy.
[0004] In view of the above problems, the present application provides an operation strategy optimization method for an energy storage EMS system.
[0005] The present application provides an operation strategy optimization method for an energy storage EMS system, the method comprising: obtaining a power load forecasting demand, constructing a load adaptive predictor according to the power load forecasting demand, and performing load forecasting for a preset time period on a target power system based on the load adaptive predictor to obtain power load forecasting information; performing an operation characteristic analysis on a target energy storage EMS system to obtain operation characteristic element information of the energy storage system, performing an operation strategy analysis based on the power load forecasting information and the operation characteristic element information of the energy storage system, and determining an operation strategy of the target energy storage system; constructing an energy storage system operation strategy parameter solution space according to the target energy storage system operation strategy, and performing configuration analysis on strategy optimization algorithm parameters based on the energy storage system operation characteristic element information to obtain an optimization algorithm for the target energy storage system; performing parameter optimization analysis in the energy storage system operation strategy parameter solution space based on the target energy storage system optimization algorithm to determine an operation strategy control parameter set, and performing dynamic feedback adjustment on the target energy storage EMS system based on the operation strategy control parameter set.
[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages:
[0007] The above-mentioned operation strategy optimization method for the energy storage EMS system analyzes and determines the power load forecasting requirements of the power system. According to these requirements, a load forecaster capable of adaptive changes is constructed to accurately predict the power load in a certain period of time in the future. By obtaining these forecast information, the changes in power demand can be better understood. Subsequently, the operation characteristics of the target energy storage EMS system are deeply analyzed to master its key elements. Combined with the power load forecasting information obtained previously, the operation strategy is analyzed to find the operation strategy that best suits the current power demand and the state of the energy storage system. After that, based on the selected operation strategy, a parameter solution space is constructed to find the optimal control parameters. At the same time, the optimization algorithm is parameterized according to the characteristics of the energy storage system to ensure the effectiveness of the algorithm. Then, the optimization algorithm is used to find the best set of control parameters in the parameter solution space, and these parameters are applied to the energy storage EMS system to ensure the efficient and stable operation of the system through dynamic feedback adjustment. The whole process aims to improve the energy utilization efficiency of the energy storage system, speed up the response speed, and enhance the stability of the system.
[0008] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS
[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0010] Figure 1 A schematic diagram of a flow chart of an operation strategy optimization method for an energy storage EMS system in one embodiment;
[0011] Figure 2 A schematic diagram of constructing a load adaptive predictor for an operation strategy optimization method for an energy storage EMS system in one embodiment. DETAILED DESCRIPTION
[0012] The embodiment of the present application provides an operation strategy optimization method for an energy storage EMS system to solve the technical problems in the prior art of low energy utilization efficiency, slow system response speed and insufficient stability due to the inaccurate and optimized operation strategy of the energy storage EMS system during the energy conversion and regulation process.
[0013] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.
[0014] It should be noted that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or inherent to these processes, methods, products or devices.
[0015] Embodiment 1
[0016] like Figure 1 , Figure 2 As shown, the present application provides an operation strategy optimization method for an energy storage EMS system, the method comprising:
[0017] S1: Obtain power load forecasting requirements, construct a load adaptive forecaster according to the power load forecasting requirements, perform load forecasting for a preset time period on a target power system based on the load adaptive forecaster, and obtain power load forecasting information;
[0018] With the transformation of the global energy structure and the widespread use of renewable energy, energy management is facing unprecedented challenges and opportunities. As an important part of energy management, the optimization of the operation strategy of the energy storage EMS system is of great significance to improving energy utilization efficiency and ensuring the stability of power supply.
[0019] In an embodiment of the present application, in order to manage power resources more effectively, the system terminal first understands the predicted demand for power load. These demands are obtained from power grid operators, energy management companies or relevant government departments. Subsequently, the system terminal performs target analysis on the power load forecast demand, and processes the historical power load data set according to the analysis results to obtain a load forecast data set. After that, the load forecast data set is classified by feature identification, and the recurrent neural network structure is trained according to the classification results to construct a load adaptive predictor. This predictor can learn and adapt to the changing law of power load, so as to more accurately predict future power load. After obtaining the load adaptive predictor, the system terminal uses this load adaptive predictor to predict the load of the target power system for a preset time period. The predicted time period is set according to actual needs. For example, if it is for real-time power dispatching and balancing, the preset time period can be set to 1 hour, etc., so that the short-term change trend of power demand can be more accurately grasped. Finally, the system terminal uses the preset time period to predict the load of the target power system and obtains power load forecast information. This forecast information contains the expected situation of power demand in a certain period of time in the future. In summary, according to the power load forecasting requirements, an adaptive forecaster is constructed to predict the power load of the target power system in the future, so as to obtain power load forecasting information. This information is of great significance for the planning, scheduling and operation of the power system, and helps to achieve the optimal configuration and efficient utilization of power resources.
[0020] Further, the present application provides the method of constructing a load adaptive predictor in S1, and further comprising:
[0021] S11: performing target analysis on the power load forecasting demand to obtain forecasting timeliness and forecasting accuracy, and periodically preprocessing the historical power load data set according to the forecasting timeliness and forecasting accuracy to obtain a load forecasting data set;
[0022] S12: Obtain load forecast factor information, where the load forecast factor information includes peak factors, seasonal factors, weather factors, and special day factors;
[0023] Preferably, for the power load forecasting requirements, the system terminal first clarifies the forecasting time range, such as one hour or one day in short-term forecasting, one week or one month in long-term forecasting, etc. Then set the expected accuracy of the forecast results, such as the control requirements of the error rate, etc., to ensure that the forecast results can meet the actual needs. Subsequently, identify and process the abnormal points in the historical power load data set, such as abnormal load values caused by equipment failure or data recording errors. Then check whether there are missing parts. If there are missing data, the system terminal calculates the mean of the adjacent 4 data and uses the calculation result as the filling value to supplement the missing data. After that, according to the forecast timeliness, filter out the historical load data in the corresponding time period. For example, if the forecast timeliness is 12 hours, extract the historical load data at the same time point every day. Then extract other factor data related to the forecast target, such as weather data, holiday information, etc., to prepare for subsequent analysis. Then, according to the forecast accuracy requirements, average the data. For example, if the forecast accuracy requirement is hourly load, multiple data points in the same hour are integrated into an average value. A structured load forecasting data set is formed to facilitate subsequent model training and forecasting analysis. Finally, after the above periodic preprocessing, a cleaned, extracted and integrated load forecasting dataset will be obtained. This dataset will be used as the input of the subsequent forecasting model for training and testing the model to achieve accurate prediction of future power load.
[0024] During the forecasting process, the system terminal considers multiple factors that affect the load, which are extracted when screening historical load data, including peak factors, seasonal factors, weather factors, and special day factors. Among them, peak factors focus on the surge in electricity demand caused by production, commercial and residential activities, especially during a specific time period. Seasonal factors are related to electricity demand in different seasons, such as cooling and heating demand in summer and winter. Weather factors include temperature, humidity, wind speed, and rainfall, which have a direct impact on power load. Special day factors include holidays or power supply days for specific industrial equipment, which may cause fluctuations in power load.
[0025] In summary, by clarifying the prediction timeliness and accuracy, periodically preprocessing historical data, and combining multiple factors such as peaks, seasons, weather, and special days, it is possible to more accurately predict power load and provide strong support for the planning and operation of power systems.
[0026] S13: performing feature identification classification on the load prediction data set based on the load prediction factor information to obtain a multi-load factor feature data set;
[0027] S14: Using a recurrent neural network structure to perform training fusion optimization on the multi-load factor feature data set to construct the load adaptive predictor.
[0028] Preferably, the system terminal performs feature identification and classification on the load prediction data set according to the extracted load prediction factor information. These factors include peak load, seasonal changes, weather conditions and special days, which are all important factors affecting the power load. Specifically, the system terminal analyzes each factor to understand the way and degree of influence of these factors on the power load, so as to accurately perform feature identification in the future. Subsequently, the peak load time period of each day, such as morning, noon or evening, is identified and marked, and the load value during the peak period is extracted as the peak load feature. Then, the load data is classified according to the four seasons of spring, summer, autumn and winter, and the average load and load change trend of each season are extracted. After that, the weather data of the time period corresponding to the load data is obtained, including temperature, humidity, wind speed, rainfall, etc., and the weather data is matched with the load data to extract the load characteristics under different weather conditions. Special dates such as holidays and industrial equipment power supply days are then identified and marked. The impact of these special dates on the load is analyzed and relevant features are extracted. Then, the extracted features are normalized to eliminate the dimensional differences between different features. The identified and standardized features are then classified and integrated according to different factors to form a multi-load factor feature data set.
[0029] After obtaining the multi-load factor feature data set, the system terminal uses a recurrent neural network structure to further process the multi-load factor feature data set. A recurrent neural network is a neural network structure suitable for processing sequence data, which can capture the time dependency and long-term dependency in the data. By training the multi-load factor feature data set, the neural network can learn the interaction and influence between different factors, thereby obtaining a set of load forecasting branch models. Subsequently, the decision fusion optimization is performed on each branch model in the load forecasting branch model set to construct an adaptive load forecaster. This load adaptive forecaster can automatically adjust the forecasting strategy based on historical data and current load factor information to more accurately predict future power load conditions. Its construction process is a process of continuous optimization and improvement. By continuously training and adjusting network parameters, the performance and accuracy of the predictor can be gradually improved.
[0030] In summary, by classifying the load prediction factors and using the recurrent neural network structure for training and fusion optimization, a predictor that can adaptively predict the power load is constructed. This predictor can fully consider various factors that affect the load and automatically adjust the prediction strategy to provide more accurate and reliable power load forecast information.
[0031] Further, the present application provides that the load adaptive predictor is constructed in S14, and the method further includes:
[0032] Using a recurrent neural network structure to train the multiple load factor feature data sets respectively to obtain a load forecasting branch model set;
[0033] Assigning weight factors to the load forecast factor information to obtain load factor weight factor information, and verifying and obtaining branch model prediction accuracy information of the load forecast branch model set;
[0034] Optionally, the system terminal uses a recurrent neural network structure to train a multi-load factor feature data set, and an independent load forecasting branch model can be constructed for each load factor. Specifically, the system terminal traces back the feature identification classification process to ensure that the data set has been classified and identified according to the load factor. Subsequently, a suitable recurrent neural network structure, such as a long short-term memory network (LSTM) or a gated recurrent unit (GRU), is selected according to the characteristics of the data set. For example, such changes in the morning rush hour or the evening rush hour often have a short time dependency, that is, the load at the current moment is closely related to the load at the previous moment or the previous moments. In this case, since the peak factor mainly focuses on changes in the short term, the gated recurrent unit (GRU) is selected as the recurrent neural network structure because the GRU has a simpler structure and a faster training speed. Afterwards, the number of nodes and activation functions of the input layer, hidden layer, and output layer of the selected multiple recurrent neural network structures are determined. And the loss function and optimizer of the model are set so that parameters can be adjusted during the training process. Then, based on each load factor and the corresponding recurrent neural network structure, an independent recurrent neural network model is constructed, and the feature data of the corresponding load factor is used as input to train the model. During the training process, the loss value and accuracy of the model are monitored to ensure that the model performance is gradually improved. After the model training is completed, the system terminal adjusts the model parameters, such as learning rate, batch size, etc., according to the performance during the training process, or tries different network structures or increases the network depth to further improve the prediction performance of the model. Finally, the adjusted branch models are integrated into a set, namely the load forecasting branch model set, and each model in this set corresponds to a load factor.
[0035] After obtaining the load forecast branch model set, in order to more accurately reflect the impact of different load factors on the power load, the system terminal assigns weight factors to each load forecast factor based on historical experience. These weight factors reflect the relative importance of each factor to the power load. Subsequently, the existing weight factors are reviewed to ensure their rationality and consistency. If it is found that some weight factors may be biased or unreasonable, the system terminal fine-tunes them according to professional knowledge and actual conditions to obtain the final weight allocation result, that is, the load factor weight factor information. After that, the system terminal divides a part of the load forecast data set as a validation set, which is not used when training the model. Then, each branch model in the load forecast branch model set is used to predict the validation set to obtain the prediction result. Then, the prediction result of each branch model is compared with the actual load value in the validation set, the average absolute error of each branch model is calculated, and the calculation result is used as the prediction accuracy. The smaller the average absolute error, the smaller the prediction error of the model and the higher the prediction accuracy. This step helps the system terminal understand the advantages and disadvantages of each branch model and provides a basis for subsequent model optimization and selection.
[0036] In summary, the recurrent neural network structure is used to train the multi-load factor feature data set to obtain a set of load forecasting branch models. At the same time, through the weight factor allocation and prediction accuracy verification, we further understand the importance of different load factors in the prediction and the performance of the branch model. This information will help build a more accurate and reliable power load forecasting model.
[0037] Perform parameter weighted fusion on the load factor weight factor information and the branch model prediction accuracy information to determine model fusion coefficient information;
[0038] Based on the model fusion coefficient information, each branch model in the load prediction branch model set is subjected to decision fusion optimization to generate the load adaptive predictor.
[0039] Optionally, performing parameter weighted fusion of the load factor weight factor information and the branch model prediction accuracy information to determine the model fusion coefficient information is a key step in building a load adaptive predictor. Specifically, the system terminal multiplies the weight factor of each load factor with the prediction accuracy information of the corresponding branch model to form a comprehensive consideration standard, namely, the model fusion coefficient information. In this way, in the subsequent decision fusion optimization, the prediction results of different branch models can be weighted according to these fused coefficients to obtain a more accurate and reliable load prediction result. Subsequently, the system terminal fuses the branch models in parallel, so that multiple branch models process the input data in parallel to generate a load adaptive predictor. When multiple branch models output results, the load adaptive predictor combines the model fusion coefficient information to perform weighted summation on their output results to obtain the final prediction result. During the fusion process, the load adaptive predictor can adaptively adjust the weights of each branch model according to the feedback of real-time data and historical data to optimize the prediction performance. Moreover, this predictor can also automatically select a suitable model or model combination for prediction according to different situations and conditions, thereby improving the accuracy and stability of the prediction.
[0040] In summary, this process is to optimize and improve the accuracy and reliability of load forecasting by combining the importance of load factors and the predictive ability of the model. In this way, a more intelligent and efficient load adaptive forecaster can be built to provide more powerful support for the operation and management of the power system.
[0041] S2: Analyze the operation characteristics of the target energy storage EMS system, obtain the operation characteristic element information of the energy storage system, perform operation strategy analysis based on the power load forecast information and the operation characteristic element information of the energy storage system, and determine the operation strategy of the target energy storage system;
[0042] In one embodiment, the analysis of the operating characteristics of the target energy storage EMS system is a key step in determining its operating strategy. First, by analyzing the operating characteristics of the energy storage system and parsing the operating characteristic element information, the system terminal can obtain a series of element information, which is of great significance for understanding the performance and behavior of the target energy storage EMS system. Subsequently, combined with the power load forecast information, the system terminal conducts an in-depth analysis of the operating strategy of the target energy storage EMS system. The power load forecast information can show the changing trend of future power demand, while the operating characteristic element information of the target energy storage EMS system provides the system terminal with the capabilities and limitations of the target energy storage EMS system itself. Based on these two aspects of information, a preliminary strategy can be determined, namely the target energy storage system operating strategy. This strategy will be used for subsequent parameter optimization analysis to better guide how the target energy storage EMS system can better respond to power demand, optimize energy utilization, and improve the efficiency and stability of the target energy storage EMS system in future operations.
[0043] Furthermore, the present application provides the method for determining the energy storage system operation strategy in S2, and further includes:
[0044] S21: parsing the energy storage system operation characteristic element information to obtain system operation configuration parameters and system operation function requirements;
[0045] S22: constructing an energy storage system operation strategy knowledge base, matching and dividing the energy storage system operation strategy knowledge base based on the system operation configuration parameters and the system operation function requirements, and determining a matching system operation strategy set;
[0046] Preferably, by parsing the operation characteristic element information of the energy storage system, the system terminal obtains the system operation configuration parameters and system operation function requirements of the energy storage system. Among them, the system operation configuration parameters include the system hardware configuration, capacity, power and renewable energy access operation configuration parameters, which reflect the basic capabilities and technical characteristics of the energy storage system. The system operation function requirements refer to the role played by the energy storage system in the power grid, such as peak shaving and valley filling, backup power supply, etc. After having the system operation configuration parameters and system operation function requirements of the energy storage system, the system terminal uses these parameters and requirements to match and divide the information in the pre-built energy storage system operation strategy knowledge base. Through this process, multiple operation strategies that match the current energy storage system can be determined, and these strategies are sorted to generate a system operation strategy set. Among them, the energy storage system operation strategy knowledge base is obtained by the system terminal in advance by searching research reports, technical documents, expert opinions, historical operation data, etc., and sorting and classifying the retrieved data. Then use the knowledge base management tool to design the framework and structure of the knowledge base, and fill the classified data into the knowledge base according to the framework. In summary, by analyzing the configuration parameters and functional requirements of the energy storage system and matching them with the strategy information in the knowledge base, we can find the operating strategy set that best suits the current energy storage system, providing an important basis for optimizing the operation of the energy storage system.
[0047] S23: performing trend similarity analysis on the power load forecast information and the matching system operation strategy set to obtain a strategy similarity set;
[0048] S24: fine-tuning the matching system operation strategy set based on the strategy similarity set to determine the target energy storage system operation strategy.
[0049] Preferably, the system terminal performs trend similarity analysis on the power load forecast information and the matching system operation strategy set in order to evaluate the degree of consistency between each operation strategy and the future power load change trend. Specifically, the system terminal extracts key features from the power load forecast information, such as load peak, peak-valley period, growth trend, etc. Then, for each strategy in the matching system operation strategy set, the corresponding operation features are also extracted, such as charging / discharging period, power adjustment range, etc. Subsequently, a blank distance matrix is created to store the distance between each point in the power load forecast information and each point in the matching system operation strategy set. Then, the power load forecast information and the matching system operation strategy set are traversed, and the distance between each corresponding point between them is calculated using the absolute difference, and the blank distance matrix is filled. After that, the dynamic time warping (DTW) algorithm is used to find the best alignment between the power load forecast information and the matching system operation strategy set. DTW constructs a cumulative distance matrix, gradually compares the points in the sequence, and finds the path that minimizes the cumulative distance. This process allows the sequence to be stretched or compressed in time to better match each other's shapes. Then, the DTW algorithm outputs a cumulative distance value through the above process of finding the path with the minimum cumulative distance, which represents the minimum distance between the prediction sequence and the strategy response sequence. The smaller the cumulative distance, the more similar the two sequences are. In order to obtain the similarity score, the system terminal performs a reciprocal operation on the cumulative distance value and converts the distance into a similarity score. The higher the similarity score, the more the strategy matches the power load forecast information. Finally, the above calculation process is repeated for each operating strategy to obtain the similarity score between each strategy and the power load forecast information, thereby obtaining a strategy similarity set. This set reflects the degree of match between different strategies and the predicted load.
[0050] After obtaining the strategy similarity set, the system terminal optimizes and fine-tunes the matching system operation strategy set based on the strategy similarity set. This means that the system terminal selects the operation strategy that best matches the forecast trend based on the similarity between the strategy and the power load trend, and makes necessary fine-tuning to adapt to the specific system conditions and needs. This process is designed to ensure that the selected strategy can not only effectively cope with future power load changes, but also achieve the best results in actual operation. Through this series of analysis and optimization steps, the system terminal can determine the final operation strategy of the target energy storage system. This strategy is not only highly matched with the power load forecast information, but also fine-tuned to adapt to the specific operating conditions of the system, thereby ensuring that the energy storage system can perform at its best in future operations.
[0051] Further, the present application provides that the method for determining the target energy storage system operation strategy in S24 further includes:
[0052] Based on the strategy similarity set, the matching system operation strategy set is optimized in descending order to determine the basic energy storage system operation strategy;
[0053] Determine a basic matching similarity according to the basic energy storage system operation strategy, and perform deviation analysis based on the basic matching similarity to generate a strategy deviation characteristic parameter;
[0054] Optionally, the system terminal selects the matching system operation strategy set in descending order according to the strategy similarity set calculated previously. This process is to sort from high to low according to the similarity, and select the strategy with the highest similarity as the first choice. The strategy selected in this way is the operation strategy that best matches the predicted power load trend, and the selected operation strategy is defined as the basic energy storage system operation strategy. Subsequently, the system terminal uses this highest similarity score as the basic matching similarity, which represents the best matching degree between the basic strategy and the power load forecast information. After that, the system terminal simulates the operation of this basic energy storage system operation strategy, collects data of the energy storage system in operation, and compares it with the actual power load data. Then, the same method as the aforementioned method for obtaining the strategy similarity set is used to calculate the similarity score in operation and compare it with the basic matching similarity. Then, the difference between the actual similarity and the basic matching similarity is analyzed to find out the key factors that may cause deviations, such as energy storage system performance, operating environment changes, etc. Based on the results of the analysis, the deviation between the strategy and the actual situation is quantified to generate strategy deviation characteristic parameters. These parameters include the size of the deviation, the direction of the deviation, the duration of the deviation, etc. For example, the difference in similarity scores, whether the similarity score is too high or too low, etc., can quantify the difference between the basic strategy and the actual power load situation, and help the system terminal to have a deeper understanding of the performance of the strategy and possible problems. In summary, this process is to first find the operation strategy that best matches the power load forecast trend as the basic strategy, and then further understand the performance of the strategy by analyzing the similarity and deviation characteristic parameters of the basic strategy, providing a basis for subsequent fine-tuning or optimization.
[0055] Performing fine-tuning correlation analysis on the strategy deviation characteristic parameters to obtain the strategy fine-tuning pace;
[0056] The basic energy storage system operation strategy is fine-tuned and optimized based on the strategy fine-tuning pace to determine the target energy storage system operation strategy.
[0057] Optionally, the system terminal performs fine-tuning correlation analysis on the policy deviation characteristic parameters, mainly to find out the operation policy parameters that are closely related to these deviation parameters, and determine the range or threshold value of these parameters that need to be adjusted. This process provides specific adjustment directions and steps for policy optimization. The system terminal uses the correlation analysis method to calculate the correlation coefficient between each operation policy parameter and the policy deviation characteristic parameter. The size and sign of the correlation coefficient can reflect the degree and direction of association between the parameters. Subsequently, according to the calculated correlation coefficient, the correlation relationship between the operation policy parameter and the policy deviation characteristic parameter is determined. A correlation coefficient close to 1 indicates a positive strong correlation, that is, as the value of the operation policy parameter increases, the value of the policy deviation characteristic parameter will also increase. In order to reduce the policy deviation, the value of the operation policy parameter should be reduced. A correlation coefficient close to -1 indicates a reverse strong correlation, which is opposite to the positive strong correlation. A correlation coefficient close to 0 indicates a weak correlation or no correlation. Afterwards, according to the results of the correlation analysis, the operation policy parameters that are significantly associated with the policy deviation characteristic parameters are screened out. These parameters will be the focus of subsequent fine-tuning optimization. Then, according to the adjustment direction and expected effect of the key parameters, a specific strategy fine-tuning pace is formulated. The fine-tuning pace includes the adjustment range, adjustment amplitude and adjustment sequence of each key parameter to ensure the orderly fine-tuning process. For example, if it is found that the power supply level is significantly correlated with certain deviation parameters, then it is necessary to consider increasing the power supply level to optimize the strategy.
[0058] After obtaining the strategy fine-tuning pace, the system terminal fine-tunes and optimizes the basic energy storage system operation strategy based on the obtained strategy fine-tuning pace. This means that the system terminal should make appropriate adjustments to the relevant operation strategy parameters based on the original strategy and the suggestions of the fine-tuning pace. Through such fine-tuning optimization, a target energy storage system operation strategy that better meets the actual power load demand can be obtained.
[0059] In summary, this process is to analyze the deviation characteristic parameters, find out the strategy parameters that need to be adjusted and their adjustment range, and then fine-tune the basic strategy to finally determine a more accurate and efficient target energy storage system operation strategy.
[0060] S3: constructing an energy storage system operation strategy parameter solution space according to the target energy storage system operation strategy, and performing configuration analysis on the strategy optimization algorithm parameters based on the energy storage system operation characteristic element information to obtain the target energy storage system optimization algorithm;
[0061] In one embodiment, the system terminal constructs a strategy parameter solution space according to the operation strategy of the target energy storage system. This solution space is a set of all possible strategy parameter combinations, which reflects the range in which the strategy parameters can vary during the adjustment process. By constructing this solution space, various possible strategy parameter combinations can be explored more comprehensively to find the optimal operation strategy. Subsequently, the system terminal performs operation optimization and improvement on this solution space. Afterwards, the improved solution space is evaluated for optimization performance, and the target energy storage system optimization algorithm is preferentially configured based on the evaluation results. Through the above steps, the system terminal can obtain an optimization algorithm for the target energy storage system. This algorithm will use the constructed strategy parameter solution space to perform parameter optimization analysis, thereby optimizing the operation strategy of the energy storage system.
[0062] Furthermore, the present application provides an optimization algorithm for obtaining the target energy storage system in S3, and the method further includes:
[0063] S31: Determine a list of candidate optimization algorithms and a set of energy storage system operation scenario instances according to the energy storage system operation characteristic element information;
[0064] S32: simulating the target energy storage EMS system based on the energy storage system operation scenario instance set to obtain an energy storage system scenario instance operation status set;
[0065] Preferably, according to the operation characteristic element information of the energy storage system, the system terminal first determines a list of candidate optimization algorithms, which are selected according to the specific characteristics and requirements of the energy storage system, including various machine learning algorithms, optimization algorithms, etc. At the same time, the system terminal also constructs a set of energy storage system operation scenario instances, which contains various representative test cases, which not only include the operation scenarios under normal conditions, but also cover the abnormal conditions or extreme conditions that may occur. Subsequently, the system terminal uses this set of operation scenario instances to simulate the target energy storage EMS system. Simulation is a method of simulating the operation of the actual target energy storage EMS system on a computer, which can help the system terminal predict and analyze the performance of the target energy storage EMS system in different scenarios. Through simulation, a set of energy storage system scenario instance operation status is obtained, which contains the operation data, performance indicators and other information of the energy storage system in various scenarios. Through this process, not only can the operation status of the target energy storage EMS system under various conditions be fully understood, but also an important basis can be provided for the subsequent selection and configuration of the optimization algorithm. Because different optimization algorithms may perform differently in different scenarios, it is necessary to select a suitable optimization algorithm according to the actual operation scenario and configure it accordingly. At the same time, the results of these simulations can also be used as a basis for verifying and optimizing the effectiveness of the algorithm, helping system terminals to continuously improve and optimize the operating strategies of the energy storage system.
[0066] S33: respectively using each optimization algorithm in the candidate optimization algorithm list to perform operation optimization and improvement on the energy storage system scenario instance operation status set to obtain a candidate optimization algorithm improvement test information set;
[0067] S34: performing optimization performance evaluation on the candidate optimization algorithm improvement test information set to obtain a candidate optimization algorithm improvement performance set, and preferentially configuring the target energy storage system optimization algorithm based on the candidate optimization algorithm improvement performance set.
[0068] Preferably, the system terminal uses each optimization algorithm in the candidate optimization algorithm list to optimize the operation of the previously obtained energy storage system scenario instance operation status set. This means that the system terminal will use each optimization algorithm to improve the operation efficiency or performance of the energy storage system in different scenarios, and record their respective performance. This process generates a candidate optimization algorithm improvement test information set, which contains the optimization effect data of various algorithms in different scenarios. Subsequently, the system terminal conducts an in-depth optimization performance evaluation on this candidate optimization algorithm improvement test information set. The evaluation dimensions include optimization effect, convergence speed and stability. The optimization effect refers to the actual effectiveness of the algorithm in improving the operation efficiency of the energy storage system and reducing deviations; the convergence speed is a measure of the time required for the algorithm to find the optimal solution; and the stability focuses on whether the algorithm performs stably and reliably under various conditions. Based on this optimization performance evaluation result, the system terminal obtains a candidate optimization algorithm improvement performance set. This set not only shows the specific scores of each optimization algorithm on various performance indicators, but also provides an intuitive comparison and selection basis. Finally, based on the information in this set, the optimization algorithm for configuring the target energy storage system is selected. This algorithm performs well in multiple performance indicators and can effectively improve the operating performance and stability of the energy storage system. Through this process, the system terminal successfully found the optimization algorithm that best suits the target energy storage system and configured the best parameter settings for it. This will help the energy storage system achieve better performance in actual operation and provide a strong guarantee for its long-term stable and efficient operation.
[0069] Based on the target energy storage system optimization algorithm, parameter optimization analysis is performed in the energy storage system operation strategy parameter solution space, an operation strategy control parameter set is determined, and dynamic feedback adjustment is performed on the target energy storage EMS system based on the operation strategy control parameter set.
[0070] In one embodiment, the system terminal further performs parameter optimization analysis in the energy storage system operation strategy parameter solution space based on the determined target energy storage system optimization algorithm. Specifically, the system terminal applies the target energy storage system optimization algorithm to the energy storage system operation strategy parameter solution space and starts searching for the optimal strategy parameter combination. The optimization algorithm will continuously try different parameter combinations in the solution space according to its internal logic and configuration parameters, and evaluate the performance of each combination. For each attempted parameter combination, the optimization algorithm will use a simulation model to perform performance evaluation. The evaluation indicators also include optimization effect, convergence speed, stability and other performance. According to the evaluation results, the algorithm will adjust its search strategy and gradually approach the optimal solution. The optimization algorithm will continuously iterate this process, and gradually find a solution with better performance by continuously trying and adjusting parameter combinations. When the optimization algorithm reaches the preset maximum number of iterations, it will stop searching and output the optimal strategy parameter combination currently found. This combination is the operation strategy control parameter set determined by parameter optimization analysis in the solution space. This set contains the key parameter values that enable the energy storage system to achieve the best operating performance. Subsequently, the system terminal dynamically feedback-adjusts the target energy storage EMS system based on this operation strategy control parameter set. By comparing the difference between the feedback parameters and the preset system operation effects, variation comparison is performed to further dynamically adjust the energy storage system. In summary, this process is to search for the optimal parameters in the solution space through the optimization algorithm, obtain the control parameter set, and then adjust the energy storage EMS system in real time to ensure that it is always in the best operating state. This not only improves the performance of the energy storage system, but also enhances its ability to cope with different operating scenarios.
[0071] Further, the present application provides that in S4, the target energy storage EMS system is dynamically feedback-regulated based on the operation strategy control parameter set, and the method further includes:
[0072] S41: Performing operation regulation and control on the target energy storage EMS system based on the operation strategy control parameter set to obtain energy storage system operation control feedback parameters;
[0073] S42: If the energy storage system operation control feedback parameters do not achieve the preset system operation effect, optimizing and analyzing the energy storage system operation control feedback parameters to obtain the control parameter variation direction;
[0074] Preferably, after determining the set of operation strategy control parameters, the system terminal applies these parameters to the target energy storage EMS system to achieve operation regulation and control of the energy storage system. In this process, the system terminal will make corresponding operation adjustments according to these control parameters and collect feedback parameters for the operation control of the energy storage system. These feedback parameters can reflect the performance and effect of the energy storage system in actual operation. Subsequently, the system terminal conducts a detailed analysis of these data, compares the difference between the actual operation effect and the preset effect, and identifies which parameters have a large deviation from the preset effect. After that, based on the results of the data analysis, the key issues affecting the operation effect of the energy storage system are determined. Based on the analysis of the key issues, the control parameters that need to be adjusted and their variation directions are determined, that is, whether the parameters should be increased or decreased, and how large the adjustment should be. For example, if the energy storage efficiency is low, it may be necessary to increase the charging current of the energy storage device or adjust the charging strategy; if the temperature is too high, it may be necessary to optimize the heat dissipation system or reduce the charging power. When determining the direction of parameter variation, the system terminal also comprehensively considers the overall performance and stability of the energy storage system to avoid the adverse effects of the adjustment of a single parameter on other parameters. In summary, this process is to regulate and control the energy storage system based on the operating strategy control parameter set, and optimize the feedback parameters according to the comparison results between the feedback parameters and the preset effects to obtain the variation direction of the control parameters, thereby guiding the subsequent further adjustment and optimization of the operating strategy.
[0075] A parameter variation function is set according to the variation direction of the control parameter, and a variation optimization comparison is performed on the operation strategy control parameter set based on the parameter variation function to obtain an operation strategy optimization control parameter set for dynamic adjustment of the energy storage system.
[0076] Preferably, based on the variation direction of the control parameters, the system terminal first sets a parameter variation function. The function of this function is to generate a series of new parameter values according to the known variation direction. These new parameter values are an optimization attempt for the original operation strategy control parameter set, with the aim of finding a parameter combination that can better improve the operation effect of the energy storage system. Subsequently, the system terminal uses this parameter variation function to perform variation optimization comparison on the operation strategy control parameter set. This means that the system terminal will generate multiple different parameter combinations and test the effects of these parameter combinations in the simulation model. By comparing the operation effects of the energy storage system under different parameter combinations, it is evaluated which parameter combinations are better, that is, they can bring higher energy storage efficiency, better stability or other preset operation effects. After such a comparison process, an operation strategy optimization control parameter set will eventually be determined. This parameter set is obtained based on the results of the variation optimization comparison, and it contains parameter combinations that can enable the energy storage system to achieve the best operation effect. Afterwards, the energy storage system is dynamically adjusted using this operation strategy optimization control parameter set. This means that in actual operation, the energy storage system will be adjusted in real time according to these optimized parameters to cope with various operation scenarios, thereby ensuring that the system always remains in the best operating state.
[0077] Furthermore, the present application provides the parameter variation function, and the method further comprises:
[0078] The parameter variation function is specifically:
[0079] ;
[0080] in, is the position after parameter variation, is the position of the original control parameter, is a random number between 0 and 1. is the mean of the parametric normal distribution.
[0081] Optionally, the parameter mutation function is a method for generating new parameter values, which mutates the original control parameters to find a better parameter combination. The parameter mutation function is as follows:
[0082] ;
[0083] in, It indicates the new position after the parameter is mutated, which is the optimized and adjusted parameter required by the system terminal. is the position of the original control parameter, that is, the initial parameter to be adjusted. It is a randomly generated number between 0 and 1. This random number is used to introduce randomness into the parameter variation process, helping the system terminal to explore in different directions to find possible better solutions. is the mean of the normal distribution of the parameter. It represents the central tendency or average level of the parameter in the process of variation. By adjusting this value, the range and degree of parameter variation can be controlled. is a tangent function, which is based on The tangent function is periodic and can produce large changes within its period, which allows parameter variation to be carried out in a larger range, which helps to find better parameter combinations. The parameter variation function first determines the original control parameter. Subsequently, a random number between 0 and 1 is automatically generated. This random number will be used to introduce the randomness of the variation so that each variation can produce different results. After that, an intermediate value is calculated using the mean of the normal distribution of the parameter, the mean is adjusted to a range centered on 0.5, and the corresponding tangent value is calculated. Then, the generated random number is multiplied by the calculated tangent value to obtain a random variation. The size and direction of this variation are random and are affected by the random number, mean u and tangent function. Finally, this variation is added to 1 and multiplied by the original control parameter to obtain the new parameter value after variation. This new parameter value will be used as the optimized control parameter for dynamic adjustment of the energy storage system. By repeatedly executing this calculation process, multiple mutated parameter values can be generated and their effects can be tested in the simulation model to find the optimal control parameter combination, thereby optimizing the operation effect of the energy storage system.
[0084] In summary, the embodiments of the present application have at least the following technical effects:
[0085] The embodiment of the present application constructs a load adaptive predictor according to the power load forecasting demand, performs load forecasting on the target power system, and obtains power load forecasting information. Subsequently, the target energy storage EMS system is analyzed for operating characteristics, and the target energy storage system operating strategy is determined in combination with the power load forecasting information. After that, the energy storage system operating strategy parameter solution space is constructed, and the optimization algorithm parameters are configured to obtain the target energy storage system optimization algorithm. Then, the optimization algorithm is used to find the best in the parameter solution space, determine the operating strategy control parameter set, and dynamically feedback adjust the target energy storage EMS system. When constructing the load adaptive predictor, the prediction is realized by training and fusing multiple load factor feature data sets, and the recurrent neural network structure is used to achieve prediction. When determining the energy storage system operating strategy, the system operating configuration parameters and functional requirements are matched and divided with the knowledge base, and the strategy is optimized and fine-tuned based on the power load forecasting information. When obtaining the target energy storage system optimization algorithm, the candidate optimization algorithm is determined according to the operating characteristic element information, and the optimization configuration is optimized through simulation and performance evaluation. During dynamic feedback adjustment, if the feedback parameter does not achieve the preset effect, the parameter variation function is used to perform variation optimization comparison, and the optimization control parameter set is obtained for dynamic adjustment. These technical effects jointly solve the technical problems in the existing technology that the energy storage EMS system has low energy utilization efficiency, slow system response speed and insufficient stability due to the inaccurate and optimized operation strategy during the energy conversion and regulation process. It realizes the accurate prediction of power load, improves the efficiency of the energy conversion and regulation process, and enhances the technical effect of system stability.
[0086] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. The processes depicted in the accompanying drawings do not necessarily require the specific order and continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0087] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.
[0088] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.
Claims
1. An operation strategy optimization method for an energy storage EMS system, characterized in that: The method comprises: S1: Obtain power load forecasting requirements, construct a load adaptive forecaster according to the power load forecasting requirements, perform load forecasting for a preset time period on a target power system based on the load adaptive forecaster, and obtain power load forecasting information; S2: Analyze the operation characteristics of the target energy storage EMS system, obtain the operation characteristic element information of the energy storage system, perform operation strategy analysis based on the power load forecast information and the operation characteristic element information of the energy storage system, and determine the operation strategy of the target energy storage system; S3: constructing an energy storage system operation strategy parameter solution space according to the target energy storage system operation strategy, and performing configuration analysis on the strategy optimization algorithm parameters based on the energy storage system operation characteristic element information to obtain the target energy storage system optimization algorithm; S4: performing parameter optimization analysis in the energy storage system operation strategy parameter solution space based on the target energy storage system optimization algorithm, determining an operation strategy control parameter set, and dynamically feedback-adjusting the target energy storage EMS system based on the operation strategy control parameter set; The step S4 includes dynamically feedback-adjusting the target energy storage EMS system based on the operation strategy control parameter set, including: S41: Performing operation regulation and control on the target energy storage EMS system based on the operation strategy control parameter set to obtain energy storage system operation control feedback parameters; S42: If the energy storage system operation control feedback parameter does not achieve the preset system operation effect, optimizing and analyzing the energy storage system operation control feedback parameter to obtain the control parameter variation direction; S43: Setting a parameter variation function according to the control parameter variation direction, performing variation optimization comparison on the operation strategy control parameter set based on the parameter variation function, and obtaining an operation strategy optimization control parameter set to dynamically adjust the energy storage system.
2. The method according to claim 1, characterized in that The load adaptive predictor is constructed in S1, including: S11: performing target analysis on the power load forecasting demand to obtain forecasting timeliness and forecasting accuracy, and periodically preprocessing the historical power load data set according to the forecasting timeliness and forecasting accuracy to obtain a load forecasting data set; S12: Obtain load forecast factor information, where the load forecast factor information includes peak factors, seasonal factors, weather factors, and special day factors; S13: performing feature identification classification on the load prediction data set based on the load prediction factor information to obtain a multi-load factor feature data set; S14: Using a recurrent neural network structure to perform training fusion optimization on the multi-load factor feature data set to construct the load adaptive predictor.
3. The method according to claim 2, characterized in that The step of constructing the load adaptive predictor in S14 includes: Using a recurrent neural network structure to train the multiple load factor feature data sets respectively to obtain a load forecasting branch model set; Assigning weight factors to the load forecast factor information to obtain load factor weight factor information, and verifying and obtaining branch model prediction accuracy information of the load forecast branch model set; Perform parameter weighted fusion on the load factor weight factor information and the branch model prediction accuracy information to determine model fusion coefficient information; Based on the model fusion coefficient information, each branch model in the load prediction branch model set is subjected to decision fusion optimization to generate the load adaptive predictor.
4. The method according to claim 1, characterized in that Determining the energy storage system operation strategy in S2 includes: S21: parsing the energy storage system operation characteristic element information to obtain system operation configuration parameters and system operation function requirements; S22: constructing an energy storage system operation strategy knowledge base, matching and dividing the energy storage system operation strategy knowledge base based on the system operation configuration parameters and the system operation function requirements, and determining a matching system operation strategy set; S23: performing trend similarity analysis on the power load forecast information and the matching system operation strategy set to obtain a strategy similarity set; S24: fine-tuning the matching system operation strategy set based on the strategy similarity set to determine the target energy storage system operation strategy.
5. The method according to claim 4, characterized in that Determining the target energy storage system operation strategy in S24 includes: Based on the strategy similarity set, the matching system operation strategy set is optimized in descending order to determine the basic energy storage system operation strategy; Determine a basic matching similarity according to the basic energy storage system operation strategy, and perform deviation analysis based on the basic matching similarity to generate a strategy deviation characteristic parameter; Performing fine-tuning correlation analysis on the strategy deviation characteristic parameters to obtain the strategy fine-tuning pace; The basic energy storage system operation strategy is fine-tuned and optimized based on the strategy fine-tuning pace to determine the target energy storage system operation strategy.
6. The method according to claim 1, characterized in that The target energy storage system optimization algorithm obtained in S3 includes: S31: Determine a list of candidate optimization algorithms and a set of energy storage system operation scenario instances according to the energy storage system operation characteristic element information S32: simulating the target energy storage EMS system based on the energy storage system operation scenario instance set to obtain an energy storage system scenario instance operation status set; S33: respectively using each optimization algorithm in the candidate optimization algorithm list to perform operation optimization and improvement on the energy storage system scenario instance operation status set to obtain a candidate optimization algorithm improvement test information set; S34: performing optimization performance evaluation on the candidate optimization algorithm improvement test information set to obtain a candidate optimization algorithm improvement performance set, and preferentially configuring the target energy storage system optimization algorithm based on the candidate optimization algorithm improvement performance set.
7. The method according to claim 1, characterized in that The parameter variation function is specifically: ; in, is the position after parameter variation, is the position of the original control parameter, is a random number between 0 and 1. is the mean of the parametric normal distribution.
