Energy storage cluster optimization scheduling method based on big data analysis

Abstract

The present invention discloses an energy storage cluster optimization scheduling method based on big data analysis, which relates to the technical field of energy storage cluster scheduling, including deploying intelligent monitoring equipment to collect operating data for preprocessing, transmitting the preprocessed operating data to a central database; extracting the operating data in the central database to construct a scheduling optimization objective function, and performing a global search on the scheduling optimization objective function through the Gray Wolf Optimization Algorithm to generate an optimal energy storage cluster scheduling strategy. The present invention deploys intelligent monitoring equipment to collect operating data for preprocessing, transmitting the preprocessed operating data to a central database; extracting the operating data in the central database to construct a scheduling optimization objective function, and performing a global search on the scheduling optimization objective function through the Gray Wolf Optimization Algorithm to generate an optimal energy storage cluster scheduling strategy; increasing the operating efficiency and intelligence level of the energy storage system, and improving the accuracy and real-time performance of the scheduling plan.

Description

Technical Field

[0001] The present invention relates to the technical field of energy storage cluster scheduling, and in particular to an energy storage cluster optimization scheduling method based on big data analysis. Background Art

[0002] With the rapid development of renewable energy and the transformation of the global energy structure, energy storage systems have become an indispensable part of the power system. They play an important role in energy regulation, balancing supply and demand, and improving the stability of the power system. With the rapid development of big data technology, energy storage cluster optimization scheduling methods based on big data have emerged. The core of these methods is to optimize the operating efficiency of the energy storage system through real-time monitoring and analysis of diverse data in the energy storage system. With the popularization of big data analysis technology, the optimization scheduling of energy storage clusters has gradually introduced data-driven intelligent optimization methods to utilize rich operating data to accurately evaluate the status of the energy storage system and predict future loads. Existing scheduling methods usually have shortcomings in data processing efficiency, real-time performance, and adaptability to multi-objective optimization, and are in urgent need of further innovation.

[0003] Existing technologies still have some significant shortcomings in big data-driven energy storage cluster scheduling. Many scheduling algorithms only use basic cost optimization methods and ignore important factors such as the health status and efficiency of energy storage units, which may lead to increased equipment aging and higher operating costs. Summary of the Invention

[0004] In view of the problems existing in the above-mentioned existing energy storage cluster optimization scheduling method based on big data analysis, the present invention is proposed.

[0005] Therefore, the problem to be solved by the present invention is that the multi-scheduling algorithm only uses basic cost optimization means, ignoring important factors such as the health status and efficiency of the energy storage unit, which may lead to increased equipment aging and increased operating costs.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: a method for optimizing and scheduling energy storage clusters based on big data analysis, comprising: deploying intelligent monitoring equipment to collect operating data for preprocessing, and transmitting the preprocessed operating data to a central database; extracting the operating data in the central database to construct a scheduling optimization objective function, and performing a global search on the scheduling optimization objective function through the gray wolf optimization algorithm to generate an optimal energy storage cluster scheduling strategy; the central database controls the execution of the optimal energy storage cluster scheduling strategy, and collects user feedback data to adjust the scheduling optimization objective function; and constructs a visual interface to display the execution strategy and operating data, and to store, collect, and analyze the generated operating data.

[0007] As a preferred solution of the energy storage cluster optimization scheduling method based on big data analysis described in the present invention, the deployment of intelligent monitoring equipment to collect operating data for pre-processing includes:

[0008] Deploy intelligent monitoring equipment on each energy storage unit to collect the operating data of the energy storage unit;

[0009] Collect data on the maximum amount of energy storage, electricity price cost, and cumulative usage time of smart monitoring devices from the usage logs of smart monitoring devices;

[0010] The operating data includes output power, input power and temperature data;

[0011] The preprocessing includes filling missing data using linear interpolation, screening and deleting outliers in the operation data using an anomaly detection algorithm, and normalizing the collected operation, maximum energy storage, electricity price cost and cumulative usage time data of the intelligent monitoring equipment using a standardized formula.

[0012] As a preferred solution of the energy storage cluster optimization scheduling method based on big data analysis described in the present invention, wherein: transmitting the pre-processed operation data to the central database refers to packaging the pre-processed operation, maximum energy storage, electricity price cost, and cumulative usage time data of the intelligent monitoring device into JSON format according to the MQTT topic structure, serializing the JSON format using UTF-8 encoding, and sending the packaged JSON data to the central database through the MQTT publish-subscribe mechanism;

[0013] The receiving end of the central database subscribes to the MQTT topic and parses the JSON data.

[0014] As a preferred solution of the energy storage cluster optimization scheduling method based on big data analysis described in the present invention, wherein: extracting the operating data in the central database to construct a scheduling optimization objective function, performing a global search on the scheduling optimization objective function using the gray wolf optimization algorithm, and generating the optimal energy storage cluster scheduling strategy includes:

[0015] Calculating the efficiency of energy storage units using charge and discharge efficiency ;

[0016] Calculate the health status of energy storage units using exponential functions ;

[0017] Calculate the electricity cost of the i-th energy storage unit using the power price method ;

[0018] Calculate the maintenance cost of the i-th energy storage unit using linear accumulation ;

[0019] The entropy weight method is used to calculate the weight coefficient of health status and the weight coefficient of electricity cost. The formula is:

[0020] , ,

[0021] , ,

[0022] in is the entropy value of the health state, is the entropy value of electricity cost, n is the total number of energy storage units, o is the counting variable, k is the normalization constant, is the weight coefficient of health status, is the weight coefficient of maintenance cost;

[0023] Based on the efficiency of the energy storage unit , health status , electricity costs and maintenance costs ,Construct the scheduling optimization objective function,, the formula is:

[0024] ,

[0025] Where J is the scheduling optimization objective function;

[0026] Set the constraints of the scheduling optimization objective function;

[0027] Substitute the position of each gray wolf into the scheduling optimization objective function to calculate the result after substituting it into the objective function, record it as the fitness value, sort the fitness values in ascending order, select the gray wolves with the top three fitness values, mark the gray wolf with the lowest fitness value as Alpha wolf, the gray wolf with the second-ranked fitness value as Beta wolf, and the gray wolf with the third-ranked fitness value as Delta wolf;

[0028] Generate an adversarial solution using the adversarial learning formula based on the current Alpha wolf position , the formula is:

[0029] ,

[0030] in is the current Alpha wolf position, is the resistance coefficient, is the position of the i-th non-Alpha wolf at time t;

[0031] Use Logistic Chaos Map to generate the initial position of the gray wolf, which includes the efficiency, health status, power cost, and maintenance cost of the energy storage unit;

[0032] Generate adversarial solutions Substitute the fitness value of the adversarial solution into the scheduling optimization objective function to calculate the fitness value of the adversarial solution, and compare the fitness value of the adversarial solution with the current Alpha wolf position. If the fitness value of the adversarial solution is greater than the fitness value of the adversarial solution of the current Alpha wolf position, then update the position of the gray wolf. If the fitness value of the adversarial solution is less than or equal to the fitness value of the adversarial solution of the current Alpha wolf position, then keep the current Alpha wolf position unchanged.

[0033] Use Euclidean distance to calculate the distance between the gray wolf's initial position and the Alpha wolf, Beta wolf, and Delta wolf;

[0034] Update the position of the i-th gray wolf using the position update formula based on the distance between Alpha wolf, Beta wolf and Delta wolf ;

[0035] The position of the i-th gray wolf will be updated Comparing with the constraints of the scheduling optimization objective function, the gray wolf positions that exceed the constraints of the scheduling optimization objective function are deleted using the cutting method;

[0036] Use the heuristic setting method to set the maximum number of iterations. When the position update reaches the maximum number of iterations, stop the iteration and output the Alpha wolf.

[0037] The output of Alpha Wolf is set as the optimal energy storage cluster scheduling strategy.

[0038] As a preferred solution of the energy storage cluster optimization scheduling method based on big data analysis described in the present invention, wherein: the central database controls the execution of the optimal energy storage cluster scheduling strategy including:

[0039] The central database reads the generated optimal energy storage cluster scheduling strategy, converts it into a control instruction using the Modbus protocol, and sends it to the intelligent monitoring device. After receiving the control instruction, the intelligent monitoring device performs the corresponding operation.

[0040] As a preferred solution of the energy storage cluster optimization scheduling method based on big data analysis described in the present invention, wherein: collecting user feedback data to adjust the scheduling optimization objective function refers to collecting user feedback data and preprocessing it;

[0041] Use statistical analysis methods to analyze the pre-processed feedback data, update the health status weight coefficient and maintenance cost weight coefficient of the scheduling optimization objective function based on the feedback analysis results, and use historical operation data for simulation testing;

[0042] The central database readjusts the control instructions based on the adjusted scheduling optimization objective function.

[0043] As a preferred solution of the energy storage cluster optimization scheduling method based on big data analysis described in the present invention, wherein: the construction of a visual interface to display the execution strategy and operating data refers to using the Dash framework of the Python visualization tool to build a visual interface, and using a data visualization tool to display the operating data and the optimal energy storage cluster scheduling strategy;

[0044] Users who have passed real-name verification are allowed to view the information.

[0045] As a preferred solution of the energy storage cluster optimization scheduling method based on big data analysis described in the present invention, the storage of the operating data collected and analyzed refers to storing the collected operating data and the optimal energy storage cluster scheduling strategy generated by the analysis in a central database, implementing data security protection measures, backing up the collected operating data and the optimal energy storage cluster scheduling strategy generated by the analysis in the cloud, and regularly performing integrity checks on the collected operating data, the optimal energy storage cluster scheduling strategy generated by the analysis, and the backup data.

[0046] A computer device comprises: a memory and a processor; the memory stores a computer program, and the processor implements the steps of the above-mentioned energy storage cluster optimization scheduling method based on big data analysis when executing the computer program.

[0047] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-mentioned energy storage cluster optimization scheduling method based on big data analysis.

[0048] The beneficial effects of the present invention are as follows: the present invention collects operating data by deploying intelligent monitoring equipment for preprocessing, and transmits the preprocessed operating data to a central database; extracts the operating data in the central database to construct a scheduling optimization objective function, and performs a global search on the scheduling optimization objective function through the gray wolf optimization algorithm to generate an optimal energy storage cluster scheduling strategy; increases the operating efficiency and intelligence level of the energy storage system, and improves the accuracy and real-time performance of the scheduling plan. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions of 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 paying any creative work.

[0050] Figure 1 Flowchart of the energy storage cluster optimization scheduling method based on big data analysis.

[0051] Figure 2This is a structural diagram of the energy storage cluster optimization scheduling method based on big data analysis. DETAILED DESCRIPTION

[0052] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0053] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0054] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it constitute a separate or selective embodiment that is mutually exclusive with other embodiments.

[0055] Example 1, with reference to Figure 1 and Figure 2 , which is the first embodiment of the present invention, provides an energy storage cluster optimization scheduling method based on big data analysis. The energy storage cluster optimization scheduling method based on big data analysis includes:

[0056] S1. Deploy intelligent monitoring equipment to collect operation data for pre-processing and transmit the pre-processed operation data to the central database;

[0057] Specifically, intelligent monitoring equipment is deployed on each energy storage unit to collect the operating data of the energy storage unit;

[0058] Collect data on the maximum amount of energy storage, electricity price cost, and cumulative usage time of smart monitoring devices from the usage logs of smart monitoring devices;

[0059] The operating data includes output power, input power and temperature data;

[0060] The preprocessing includes filling missing data using linear interpolation, screening and deleting outliers in the operation data using an anomaly detection algorithm, and normalizing the collected operation, maximum energy storage, electricity price cost and cumulative usage time data of the intelligent monitoring equipment using a standardized formula.

[0061] The deployment of intelligent monitoring equipment can realize real-time monitoring and data collection of the operating status of energy storage units. Such intelligent equipment can automatically obtain key operating data including output power, input power, temperature, etc. of energy storage units, providing basic support for subsequent status evaluation and system optimization. Through intelligent monitoring, it can effectively avoid the delay and error problems of manual monitoring, improve the response speed and accuracy of the system, and obtain the operating data of energy storage units in real time. It can not only be used to evaluate the current health status of the equipment, but also provide data support in scheduling decisions and optimize the overall energy efficiency of the system. Through comprehensive analysis of these log data, predictive analysis based on historical information can be carried out, the long-term operation management of the system can be optimized, and the charging and discharging process of energy storage units can be scheduled according to different electricity price periods, thereby saving costs. The use of linear interpolation can effectively fill these missing data and maintain the integrity and continuity of the data. The complete data set is very important for subsequent analysis and decision-making. The linear interpolation method enables data-driven analysis and optimization even when the data is incomplete, reducing the impact of missing data on the accuracy of the results. The anomaly detection algorithm can automatically identify and clean up these unconventional data points, ensuring the accuracy and consistency of the data set and improving the reliability of the analysis model. Through automated anomaly detection and cleaning, manual intervention can be reduced, data quality can be improved, and the stability of the system can be improved, reducing misjudgments or decision-making errors caused by data anomalies. Data of different dimensions may not be directly comparable. Through standardization, these data of different units and dimensions are unified to the same level, so that unified analysis and calculation can be performed between data. Standardized data provides more balanced input for subsequent machine learning and optimization models, making the optimization process smoother, avoiding certain variables dominating the results due to excessively large or small values, and helping to achieve more reasonable scheduling decisions and system control.

[0062] Furthermore, transmitting the pre-processed operation data to the central database means packaging the pre-processed operation, maximum energy storage capacity, electricity price cost, and accumulated usage time data of the intelligent monitoring device into JSON format according to the MQTT topic structure, serializing the JSON format using UTF-8 encoding, and sending the packaged JSON data to the central database through the MQTT publish-subscribe mechanism;

[0063] The receiving end of the central database subscribes to the MQTT topic and parses the JSON data.

[0064] Through the unified data format and transmission protocol, not only the efficiency of data transmission is ensured, but also the compatibility of data between different devices and systems is guaranteed. This process greatly reduces the risk of data loss and delay, improves the real-time nature of data, and provides accurate real-time data support for subsequent scheduling optimization algorithms, thereby improving the overall efficiency and reliability of the energy storage scheduling system. By selecting a specific topic to publish to the MQTT message queue when uploading data from each energy storage unit, the system's central database can only subscribe to topic data related to scheduling management, reducing the amount of redundant data transmission. The application of this mechanism in the energy storage cluster scheduling system can greatly reduce the system's communication costs, while improving the data receiving efficiency of the scheduling system, facilitating the real-time management and precise scheduling of large-scale energy storage units. The central database can directly parse data in JSON format and quickly Extracting key parameters from the data and avoiding complex data conversion processes significantly improves the system's data reading and parsing speed. The versatility and easy extensibility of the JSON format enable the energy storage system to easily adapt to future expansion needs and support the access of multiple types of devices and multiple data types. By uniformly adopting UTF-8 encoding, the standardization of data transmission is ensured, avoiding data decoding failures or garbled characters due to encoding differences. The standardized encoding strategy effectively guarantees the integrity and reliability of data in the energy storage cluster scheduling system, further improving the robustness of the system and the accuracy of the data. The central database provides a variety of data redundancy and fault-tolerant mechanisms to ensure the integrity and security of data during transmission. This centralized data management method enables the scheduling system to dynamically adjust the scheduling strategy based on accurate real-time data, thereby improving the operation and maintenance efficiency of the energy storage cluster.

[0065] S2. Extract the operating data from the central database to construct a scheduling optimization objective function. Use the Grey Wolf Optimization Algorithm to perform a global search on the scheduling optimization objective function to generate the optimal energy storage cluster scheduling strategy.

[0066] Specifically, the operation data in the central database is extracted to construct the scheduling optimization objective function. The scheduling optimization objective function is searched globally by the gray wolf optimization algorithm to generate the optimal energy storage cluster scheduling strategy, including:

[0067] Calculating the efficiency of energy storage units using charge and discharge efficiency , the formula is:

[0068] ,

[0069] in is the output power of the i-th energy storage unit, is the input power of the i-th energy storage unit;

[0070] Calculate the health status of energy storage units using exponential functions , the formula is:

[0071] ,

[0072] in is the maximum capacity of the i-th energy storage unit, is the temperature of the i-th energy storage unit, is the temperature sensitivity coefficient, which indicates the effect of temperature on the amount of electricity;

[0073] Calculate the electricity cost of the i-th energy storage unit using the power price method , the formula is:

[0074] ,

[0075] in is the electricity price cost of the energy storage unit, is the charging and discharging power of the i-th energy storage unit;

[0076] Calculate the maintenance cost of the i-th energy storage unit using linear accumulation , the formula is:

[0077] ,

[0078] in is the maintenance cost coefficient per unit time, The cumulative usage time of the smart monitoring device;

[0079] The entropy weight method is used to calculate the weight coefficient of health status and the weight coefficient of electricity cost. The formula is:

[0080] , ,

[0081] , ,

[0082] in is the entropy value of the health state, is the entropy value of electricity cost, n is the total number of energy storage units, o is a counting variable used to traverse all indicators, and k is a normalization constant , is the weight coefficient of health status, is the weight coefficient of maintenance cost;

[0083] The entropy weight method measures the discreteness of indicators through information entropy, and assigns different weights according to the health status and electricity cost distribution of each energy storage unit. The calculation of entropy value directly reflects the amount of information of the indicator, and can objectively quantify the relative importance of health status and electricity cost. The entropy weight method can measure the information distribution of these two indicators respectively through entropy value calculation to ensure the rationality of weight distribution. Compared with other methods, such as simple weight distribution or linear programming, it cannot handle the relative importance of multiple indicators at the same time, nor can it measure the distribution characteristics of these indicators as effectively as the entropy weight method. The entropy weight method can calculate the information entropy and quantify the information distribution of the health status and electricity cost according to the health status and electricity cost. The actual distribution of the system dynamically adjusts the weight to ensure that in actual operation, the weight of each energy storage unit can be adjusted according to the changes in real-time data, which improves the accuracy and flexibility of the model. In traditional energy storage scheduling systems, only a single indicator may be considered (such as the lowest cost or the highest efficiency). In contrast, the improvement scheme we proposed can process multiple independent indicators (health status and electricity cost) at the same time through the entropy weight method, and dynamically adjust the weight of each indicator through the entropy value. This improvement ensures that when optimizing multiple indicators at the same time, the weight distribution of each indicator can be reasonably balanced, so that the health status and electricity cost can be optimized synergistically. In energy storage optimization scheduling, the health status is the most important factor influencing the optimization of the energy storage system. The health and electricity cost data are usually unevenly distributed. It is possible that some energy storage units are in poor health but have low electricity costs, or some units are in good health but have high electricity costs. This complex distribution pattern requires the entropy weight method to dynamically analyze the contribution of each indicator. Other methods cannot adaptively calculate weights based on this uneven distribution. The entropy weight method relies on the distribution characteristics of the data itself to determine the weights, and can dynamically adapt to changes in the system environment and data distribution. This subjective weighting method will produce deviations in energy storage systems with complex data and uneven distribution. Therefore, only the entropy weight method can accurately capture data distribution information through entropy values and give more dynamic The objectivity and dynamism of weight calculation and its adaptability to complex system data distribution make this method more flexible and accurate than traditional technical solutions with fixed weights or subjective weight settings. Information entropy is traditionally used in the field of data communication, while this method innovatively introduces it into the scheduling optimization of the health status and electricity cost of energy storage clusters, realizing the function of automatic adjustment of weights by entropy values, which is innovative in this field. Through these improvements, the health status and electricity cost of the energy storage system can be more accurately and real-timely weighted in the optimization scheduling process, thereby improving the overall operating efficiency of the system.

[0084] Based on the efficiency of the energy storage unit , health status , electricity costs and maintenance costs , construct the scheduling optimization objective function to minimize the operating cost of the energy storage cluster, maximize energy efficiency and equipment health. The formula is:

[0085] ,

[0086] Where J is the scheduling optimization objective function;

[0087] The optimization objective function can consider multiple key factors (electricity cost, efficiency, health status and maintenance cost of energy storage units) at the same time. In the scheduling of energy storage clusters, it is necessary not only to minimize the operating cost, but also to maximize the energy efficiency of the energy storage unit, while considering the long-term health maintenance of the equipment. This comprehensive consideration makes the system operation more efficient and reliable in the long term. By using weight coefficients to represent the importance of health status and maintenance cost respectively, this method can flexibly adjust the impact of various factors on the final optimization goal. This weight adjustment method is not only highly adaptable, but also avoids the imbalance problem caused by other methods (such as only considering a single objective function). Traditional optimization methods often only consider the minimization of operating costs, but ignore the long-term health and maintenance of equipment. This method adds health By taking into account the weights of equipment health and maintenance costs, long-term goals (such as maximizing equipment life) are integrated into the optimization process, avoiding long-term inefficiency caused by short-term interests. Existing scheduling optimization technologies often only focus on a single goal (such as minimizing electricity costs) and ignore other important factors in the operation of energy storage systems, such as the health status of energy storage units and maintenance costs. The dynamic adjustment mechanism can better cope with the state changes in the operation of energy storage systems and ensure flexible scheduling in different scenarios. By adding considerations of equipment health and maintenance costs, the efficient and healthy operation of the energy storage cluster in the long term is guaranteed. In large-scale energy storage systems, maintaining a healthy state is the key to extending equipment life. This method innovatively combines equipment health with cost management to improve the overall efficiency of the system.

[0088] Setting the constraints of the scheduling optimization objective function includes:

[0089] ,

[0090] ,

[0091] ,

[0092] in and are the minimum and maximum powers of the i-th energy storage unit, and are the minimum and maximum power of the i-th energy storage unit, and are the minimum and maximum temperatures of the i-th energy storage unit respectively;

[0093] Substitute the position of each wolf into the scheduling optimization objective function and calculate the result after substituting it into the objective function. This is recorded as the fitness value. The fitness values are sorted in ascending order. The top three wolves are selected. The wolf with the lowest fitness value is marked as Alpha wolf, the wolf with the second-highest fitness value is marked as Beta wolf, and the wolf with the third-highest fitness value is marked as Delta wolf. The fitness value is to measure the performance of each wolf's current position in the optimization objective function. The goal is to minimize the total cost, energy efficiency loss, and maintenance cost. The smaller the fitness value, the better the performance of the solution in the optimization problem.

[0094] Generate an adversarial solution using the adversarial learning formula based on the current Alpha wolf position , the formula is:

[0095] ,

[0096] in is the current Alpha wolf position, is the antagonistic coefficient, which controls the position offset of the antagonistic solution. is the position of the i-th non-Alpha wolf at time t;

[0097] This adversarial solution generation method based on the Alpha wolf position can ensure that the generation of adversarial samples revolves around the optimal solution, thereby continuously approaching the target solution. By dynamically adjusting the position offset, the diversity and effectiveness of the adversarial samples are ensured, avoiding falling into the local optimum. The control parameter of the adversarial offset can determine the offset amplitude of the adversarial solution generation, ensuring that the adversarial solution generation process will not deviate too much or be too concentrated, thereby maintaining effective exploration of adversarial solutions in multi-dimensional space. Compared with other adversarial learning methods (such as fixed-step update methods), this method has stronger adaptability. Compared with directly generating random adversarial solutions or generating adversarial samples by gradient, this method can generate different adversarial solutions while retaining the main solution through adversarial coefficient adjustment, avoiding excessive randomness or excessive concentration of adversarial samples in other methods. To solve the problem of hitting a specific point, adversarial samples are generated by directly perturbing the input samples. This method simulates the behavior of the wolf pack intelligent algorithm, combines the position information of the Alpha wolf with the positions of other non-Alpha wolves, and dynamically generates adversarial samples. This dynamic position adjustment method can better simulate the optimization process in multi-dimensional space and improve the diversity and adaptability of adversarial samples. The introduction of the adversarial coefficient enables the model to adaptively adjust the amplitude and range of sample generation, thereby avoiding the invalid adversarial sample problem caused by excessive perturbation, and also avoids the invalid sample problem caused by insufficient perturbation. The position update mechanism of the Alpha wolf is introduced, which can generate more effective adversarial solutions by continuously approaching the optimal solution, rather than simply perturbing the input data, thereby improving the effect of adversarial learning.

[0098] Use logistic chaos mapping to generate the initial position of each gray wolf. The gray wolf position includes the efficiency, health status, power cost, and maintenance cost of the energy storage unit. Each gray wolf corresponds to an adjustment strategy.

[0099] Generate adversarial solutions Substitute the fitness value of the adversarial solution into the scheduling optimization objective function to calculate the fitness value of the adversarial solution, and compare the fitness value of the adversarial solution with the current Alpha wolf position. If the fitness value of the adversarial solution is greater than the fitness value of the adversarial solution of the current Alpha wolf position, then update the position of the gray wolf. If the fitness value of the adversarial solution is less than or equal to the fitness value of the adversarial solution of the current Alpha wolf position, then keep the current Alpha wolf position unchanged.

[0100] Use Euclidean distance to calculate the distance between the gray wolf's initial position and the Alpha wolf, Beta wolf, and Delta wolf;

[0101] Update the position of the i-th gray wolf using the position update formula based on the distance between Alpha wolf, Beta wolf and Delta wolf , the formula is:

[0102] ,

[0103] in 、 and are the distances to the Alpha wolf, Beta wolf, and Delta wolf, 、 and These are the positions of Alpha wolf, Beta wolf and Delta wolf respectively. 、 and are the random coefficients of the scaling step length;

[0104] The position of the i-th gray wolf will be updated Comparing with the constraints of the scheduling optimization objective function, the gray wolf positions that exceed the constraints of the scheduling optimization objective function are deleted using the cutting method;

[0105] Use the heuristic setting method to set the maximum number of iterations. When the position update reaches the maximum number of iterations, stop the iteration and output the Alpha wolf.

[0106] The output of Alpha Wolf is set as the optimal energy storage cluster scheduling strategy.

[0107] The initial gray wolf positions generated by the Logistic chaos map have complex and changeable dynamics. It not only avoids the insufficient search efficiency caused by the low randomness of the initial solution, but also provides nonlinear chaos characteristics, making the initial layout of the gray wolf more dispersed. This feature can effectively improve the initial population quality of the gray wolf algorithm and make the subsequent optimization process more refined. Chaotic mapping has natural adaptability when processing high-dimensional data and is suitable for processing energy storage unit scheduling problems in multi-dimensional parameter space, ensuring that the scheduling scheme has high stability under different environmental conditions. The detailed calculation of key parameters such as charging and discharging efficiency, health status, electricity cost and maintenance cost makes the scheduling strategy not only applicable to the current energy storage state, but also predicts and adapts to the storage state. The system can proactively plan the maintenance and update cycles of energy storage units through dynamic tracking of health status, effectively avoiding system losses caused by sudden failures. The data support of refined modeling enables the dispatching strategy to be quickly adapted when facing system expansion or equipment updates, thereby greatly improving the adaptability and reliability of the energy storage system. Through multi-objective optimization of efficiency, health status, electricity cost and maintenance cost, the dispatching strategy can be flexibly adjusted in a variety of application scenarios. The flexibility of the multi-objective optimization objective function enables the energy storage system to adapt to market changes in real time and improve profit potential. With the introduction of more factors, the objective function can be further expanded to adapt to the growing functional requirements of the energy storage cluster. Through Alpha Wolf and Bet The algorithm has a three-layer structure of a wolf and Delta wolf. The algorithm can perform multi-angle search adjustments in each iterative step, so that the algorithm can quickly approach the global optimum in a complex solution space. The high fault tolerance also enables the gray wolf optimization algorithm to maintain the stability of the scheduling strategy under variable operating conditions. Even when individual energy storage units fail, the system can still maintain the optimization path, further improving the robustness of the energy storage cluster system. The entropy weight method automatically calculates the weights of different indicators, objectively and reasonably reflects the relative importance of health status and electricity cost in the scheduling strategy, reducing the interference of human factors. The dynamic adjustment mechanism of the entropy weight method enables the energy storage system to automatically update the weight according to real-time operating data, maintaining the stability of the system in various usage scenarios. The application of the entropy weight method not only improves the accuracy of the scheduling plan, but also enhances the adaptability of the system, especially in highly volatile power grids and constantly changing energy storage states. The entropy weight method can ensure the long-term effectiveness of the strategy. The adversarial solution generated by the adversarial learning strategy has an efficient exploration effect on the scheduling optimization objective function. By increasing the breadth and depth of the search through adversarial learning, the system can more comprehensively analyze the fitness of the optimization solution and further improve the global optimal possibility of the scheduling strategy. This mechanism is particularly effective in complex multi-variable scheduling, reducing the risk of falling into local optimality. The adversarial solution can also maintain optimization stability between different scenarios. Even in the case of sudden changes in the system operating load, the system can still effectively generate the optimal scheduling plan and improve the overall flexibility of the system.Using Euclidean distance calculation to update the gray wolf position can achieve a reasonable iterative update step in the solution space, ensuring that the algorithm has higher efficiency and stability when exploring the optimal solution. Through the distance calculation of Alpha wolf, Beta wolf and Delta wolf, the system can ensure that the gray wolf position will not converge too quickly in multiple updates, thereby retaining a sufficient range of solution space exploration. This process effectively avoids the risk of the algorithm falling into a local optimal solution in the early stage, so that the system can still achieve relatively accurate global optimal scheduling even when facing a high-complexity energy storage cluster. Through the heuristic setting method, the scheduling system not only improves the accuracy under the premise of high computing efficiency, but also effectively controls the energy consumption of the system, and has high economic The optimization algorithm can flexibly adjust the scheduling strategy based on changes in system load, energy demand, or market electricity prices. This adaptability not only ensures the system's operating efficiency under high-load conditions, but also enhances the application potential of energy storage clusters in different scenarios such as new energy power stations and backup power sources, making the scheduling system more widely applicable. This optimization design combines the economic efficiency and long-term stability of the energy storage cluster, allowing the scheduling strategy to focus not only on current economic benefits but also on the long-term benefits of the energy storage system. Compared with traditional single-indicator optimization scheduling methods, the multi-objective scheduling strategy of this invention can achieve a longer service life with lower maintenance costs, significantly improving the lifecycle management of the energy storage cluster system.

[0108] S3, the central database controls and executes the optimal energy storage cluster scheduling strategy, collects user feedback data and adjusts the scheduling optimization objective function;

[0109] Specifically, the central database controls and executes the optimal energy storage cluster scheduling strategy, including:

[0110] The central database reads the generated optimal energy storage cluster scheduling strategy, converts it into a control instruction using the Modbus protocol, and sends it to the intelligent monitoring device. After receiving the control instruction, the intelligent monitoring device performs the corresponding operation.

[0111] As the information processing and management center of the energy storage system, the central database has a high degree of centralized control advantage by reading the optimal energy storage cluster scheduling strategy and uniformly managing its execution process. This can ensure the consistency and coherence of the scheduling strategy, effectively reducing the decision-making inconsistency and execution deviation problems caused by the decentralized management of multiple units. The centralized management mode enables the system to maintain the uniformity of scheduling execution when facing multiple energy storage units, ensuring the overall coordination and consistency of the energy storage system. The Modbus protocol has the advantages of simple data transmission, clear structure and low latency. It is particularly suitable for industrial control scenarios such as energy storage systems that require high reliability and low latency. Transmitting instructions through the Modbus protocol not only ensures the accuracy of scheduling instructions It can also conveniently add more energy storage units when the system is expanded without affecting the data transmission speed, making the scheduling system highly scalable and adaptable. This real-time response not only ensures the scheduling accuracy of the energy storage units, but also enables the system to quickly adjust the charging and discharging status to cope with electricity price fluctuations or changes in energy storage demand. This fast and real-time response mechanism is of great significance for grid load balancing and energy storage cost control, ensuring that the system can respond flexibly in a complex and changing market environment, and improving the overall economic benefits of the energy storage system. The high efficiency of centralized scheduling enables the system to charge collectively when electricity prices are low and discharge collectively when electricity prices are peak, thereby achieving significant cost optimization effects and further improving the overall economy of the energy storage system.

[0112] Furthermore, collecting user feedback data to adjust the scheduling optimization objective function refers to collecting user feedback data and preprocessing it;

[0113] Use statistical analysis methods to analyze the pre-processed feedback data, update the health status weight coefficient and maintenance cost weight coefficient of the scheduling optimization objective function based on the feedback analysis results, and use historical operation data for simulation testing;

[0114] The central database readjusts the control instructions based on the adjusted scheduling optimization objective function.

[0115] By collecting user feedback, the system can continuously optimize the scheduling strategy to make it more practical and adaptable. The preprocessing step can effectively remove redundant or abnormal data, provide a more reliable data basis for subsequent statistical analysis, ensure the accuracy and availability of feedback data, and thus improve the optimization effect of the entire system. The statistical analysis method comprehensively analyzes the feedback data, which can not only reveal the core issues that users are concerned about during the operation of the system, but also identify the shortcomings of the system in specific scenarios. Through statistical analysis, the system can dynamically adjust the health status weight and maintenance cost weight in the scheduling optimization objective function to further improve the accuracy and applicability of the scheduling strategy. This flexible data analysis process makes the system more insightful and helps to identify and improve the weak links in the scheduling strategy. When users feedback that the energy storage unit is operating relatively stably, the weight of the health status can be appropriately reduced to improve Improve the flexibility of the scheduling strategy. Such a dynamic weight adjustment mechanism not only improves the accuracy and response speed of the scheduling strategy, but also improves the long-term stability of the energy storage cluster system. It not only reduces the potential risks caused by strategy adjustments, but also can perform multiple rounds of adjustments without affecting the actual system operation, ensuring the high adaptability and accuracy of the strategy. The simulation test effectively reduces the time and cost of direct debugging and improves the overall operation and maintenance efficiency of the system. The user feedback data reflects the actual needs and pain points of users. When updating the scheduling strategy, the system makes adjustments based on this feedback information, making the optimized strategy more in line with user expectations and reducing user operating costs. This user-oriented feedback mechanism is innovative in the energy storage cluster system, providing valuable data support for the user experience of future intelligent scheduling systems, making the system more in line with market demand and having broad commercial application prospects.

[0116] S4. Build a visual interface to display execution strategies and operation data, and store the operation data generated by collection and analysis;

[0117] Specifically, building a visual interface to display execution strategies and operating data involves using the Python Dash framework to build a visual interface and using data visualization tools to display operating data and the optimal energy storage cluster scheduling strategy.

[0118] Users who have passed real-name verification are allowed to view the information.

[0119] By using the Dash framework, the present invention can achieve dynamic updating and presentation of real-time data, so that users can quickly understand the operating status and scheduling strategy of the energy storage cluster. The Dash framework supports multiple interactive functions. Users can customize data filtering conditions in the interface and conduct in-depth data analysis and exploration. Through visualization, users can intuitively and in real time observe the operating status of the energy storage cluster and adjust the scheduling strategy. This function not only improves the operability of the energy storage system, but also makes users more reliable when making scheduling decisions, and improves the overall intelligence of the system. The visual display of the scheduling strategy can help users better understand the system's optimization goals and strategy execution effects, so that they can adjust the scheduling strategy in time when needed. This transparent display method not only improves the user's trust in the system, but also enhances the user's control over the system's operation. The function of displaying operating data and scheduling strategies allows users to accurately manage and optimize the energy storage cluster according to real-time conditions. It not only helps users understand the current operating status, but also provides important data support for future system adjustments, helping users make more efficient management decisions and introduce real-name registration. The verification mechanism protects sensitive data in the energy storage system and prevents unauthorized access, thereby improving system security and data privacy. The real-name verification mechanism also enhances the system's traceability, recording each user's operational behavior for subsequent tracking and review. Users can customize different time ranges, the status of different energy storage units, and related performance indicators, making data display more personalized and flexible. The interactive capabilities provided by the Dash framework allow users to not only view data but also make real-time adjustments to the system based on actual needs, improving the system's adaptability. Users can quickly make adjustment decisions based on actual operating conditions and the execution results of the scheduling strategy, thereby ensuring that the system always operates in the optimal state.

[0120] Furthermore, storing the collected and analyzed operating data means storing the collected operating data and the optimal energy storage cluster scheduling strategy generated by the analysis in a central database, implementing data security protection measures, backing up the collected operating data and the optimal energy storage cluster scheduling strategy generated by the analysis in the cloud, and regularly performing integrity checks on the collected operating data, the optimal energy storage cluster scheduling strategy generated by the analysis, and the backup data.

[0121] By storing the collected operating data and the dispatching strategies derived from the analysis in a central database, users can quickly access and analyze the data when needed. The central database provides stable and reliable data support for subsequent automated decision-making, dispatching optimization, and operation management, avoiding the data redundancy and query complexity problems brought about by decentralized storage. Through cloud backup, operating data and dispatching strategies can be securely stored on remote servers. Cloud backup provides redundant protection for data storage, ensuring that data can be restored from the cloud even in the event of local device failure, damage, or data loss, greatly improving the system's fault tolerance and reliability. Regular integrity testing can ensure that the data stored in the central database and cloud backup is complete and has not been tampered with. Integrity testing checks the accuracy of the data by generating hash values or checksums. Once the data integrity is found to be damaged, the system can immediately initiate a repair mechanism to prevent erroneous data from being used in decision-making or optimization processes. This is especially important for long-term energy storage systems because system decisions rely on accurate historical data. Any data anomalies or damage may have a negative impact on the system's dispatch results.

[0122] In Example 2, if the functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage media include various media capable of storing program code, such as USB flash drives, mobile hard drives, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0123] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0124] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, and then editing, interpreting, or processing in another suitable manner as necessary, and then storing it in a computer memory.

[0125] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having logic gate circuits for implementing logic functions on data signals, an application-specific integrated circuit having suitable combinational logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

Claims

1. An energy storage cluster optimization scheduling method based on big data analysis, characterized by: include, Deploy intelligent monitoring equipment to collect operating data for pre-processing and transmit the pre-processed operating data to the central database; Extracting operating data from the central database to construct a scheduling optimization objective function, the Grey Wolf Optimization Algorithm is used to perform a global search of the scheduling optimization objective function to generate the optimal energy storage cluster scheduling strategy. The central database controls and executes the optimal energy storage cluster scheduling strategy, collects user feedback data and adjusts the scheduling optimization objective function; Build a visual interface to display execution strategies and operation data, and store, collect, and analyze the operation data generated; The deployment of intelligent monitoring equipment to collect operation data for pre-processing includes: Deploy intelligent monitoring equipment on each energy storage unit to collect the operating data of the energy storage unit; Collect data on the maximum amount of energy storage, electricity price cost, and cumulative usage time of smart monitoring devices from the usage logs of smart monitoring devices; The operating data includes output power, input power and temperature data; The preprocessing includes filling missing data using linear interpolation, screening and deleting outliers in the operating data using an anomaly detection algorithm, and normalizing the collected operating data, maximum energy storage capacity, electricity price cost, and accumulated usage time of intelligent monitoring equipment using a standardized formula; Transmitting the pre-processed operation data to the central database means packaging the pre-processed operation, maximum energy storage capacity, electricity price cost, and accumulated usage time data of the intelligent monitoring device into JSON format according to the MQTT topic structure, serializing the JSON format using UTF-8 encoding, and sending the packaged JSON data to the central database through the MQTT publish-subscribe mechanism; The receiving end of the central database subscribes to the MQTT topic and parses the JSON data; The operation data in the central database is extracted to construct a scheduling optimization objective function, and the scheduling optimization objective function is globally searched by the gray wolf optimization algorithm to generate the optimal energy storage cluster scheduling strategy, including: Calculating the efficiency of energy storage units using charge and discharge efficiency ; Calculate the health status of energy storage units using exponential functions ; Calculate the electricity cost of the i-th energy storage unit using the power price method ; Calculate the maintenance cost of the i-th energy storage unit using linear accumulation ; The entropy weight method is used to calculate the weight coefficient of health status and the weight coefficient of electricity cost. The formula is: ， ， ， ， in is the entropy value of the health state, is the entropy value of electricity cost, n is the total number of energy storage units, o is the counting variable, k is the normalization constant, is the weight coefficient of health status, is the weight coefficient of maintenance cost; Based on the efficiency of the energy storage unit , health status , electricity costs and maintenance costs , construct the scheduling optimization objective function, the formula is: ， Where J is the scheduling optimization objective function; Set the constraints of the scheduling optimization objective function; Substitute the position of each gray wolf into the scheduling optimization objective function to calculate the result after substituting it into the objective function, record it as the fitness value, sort the fitness values in ascending order, select the gray wolves with the top three fitness values, mark the gray wolf with the lowest fitness value as Alpha wolf, the gray wolf with the second-ranked fitness value as Beta wolf, and the gray wolf with the third-ranked fitness value as Delta wolf; Generate an adversarial solution using the adversarial learning formula based on the current Alpha wolf position , the formula is: ， in is the current Alpha wolf position, is the resistance coefficient, is the position of the i-th non-Alpha wolf at time t; Use Logistic Chaos Map to generate the initial position of the gray wolf, which includes the efficiency, health status, power cost, and maintenance cost of the energy storage unit; Generate adversarial solutions Substitute the fitness value of the adversarial solution into the scheduling optimization objective function to calculate the fitness value of the adversarial solution, and compare the fitness value of the adversarial solution with the current Alpha wolf position. If the fitness value of the adversarial solution is greater than the fitness value of the adversarial solution of the current Alpha wolf position, then update the position of the gray wolf. If the fitness value of the adversarial solution is less than or equal to the fitness value of the adversarial solution of the current Alpha wolf position, then keep the current Alpha wolf position unchanged. Use Euclidean distance to calculate the distance between the gray wolf's initial position and the Alpha wolf, Beta wolf, and Delta wolf; Update the position of the i-th gray wolf using the position update formula based on the distance between Alpha wolf, Beta wolf and Delta wolf ; The position of the i-th gray wolf will be updated Comparing with the constraints of the scheduling optimization objective function, the gray wolf positions that exceed the constraints of the scheduling optimization objective function are deleted using the cutting method; Use the heuristic setting method to set the maximum number of iterations. When the position update reaches the maximum number of iterations, stop the iteration and output the Alpha wolf. The output of Alpha Wolf is set as the optimal energy storage cluster scheduling strategy.

2. The energy storage cluster optimization scheduling method based on big data analysis according to claim 1, characterized in that: The central database controls and executes the optimal energy storage cluster scheduling strategy, including: The central database reads the generated optimal energy storage cluster scheduling strategy, converts it into a control instruction using the Modbus protocol, and sends it to the intelligent monitoring device. After receiving the control instruction, the intelligent monitoring device performs the corresponding operation.

3. The energy storage cluster optimization scheduling method based on big data analysis according to claim 2 is characterized in that: Collecting user feedback data and adjusting the scheduling optimization objective function refers to collecting user feedback data and performing preprocessing; Use statistical analysis methods to analyze the pre-processed feedback data, update the health status weight coefficient and maintenance cost weight coefficient of the scheduling optimization objective function based on the feedback analysis results, and use historical operation data for simulation testing; The central database readjusts the control instructions based on the adjusted scheduling optimization objective function.

4. The energy storage cluster optimization scheduling method based on big data analysis according to claim 3 is characterized in that: Constructing a visual interface to display the execution strategy and operation data refers to using the Dash framework of the Python visualization tool to construct a visual interface, and using a data visualization tool to display the operation data and the optimal energy storage cluster scheduling strategy; Users who have passed real-name verification are allowed to view the information.

5. The energy storage cluster optimization scheduling method based on big data analysis according to claim 4 is characterized in that: The storage of the operating data collected and analyzed refers to storing the collected operating data and the optimal energy storage cluster scheduling strategy generated by the analysis in a central database, implementing data security protection measures, backing up the collected operating data and the optimal energy storage cluster scheduling strategy generated by the analysis in the cloud, and regularly performing integrity checks on the collected operating data, the optimal energy storage cluster scheduling strategy generated by the analysis, and the backup data.

6. A computer device comprising: memory and processor; The memory stores a computer program, characterized in that when the processor executes the computer program, the steps of the energy storage cluster optimization scheduling method based on big data analysis described in any one of claims 1 to 5 are implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the energy storage cluster optimization scheduling method based on big data analysis according to any one of claims 1 to 5 are implemented.

Images