Intelligent energy management control method and system for energy storage system

Abstract

The invention discloses an energy storage system intelligent energy management control method and system, and relates to the technical field of energy storage system intelligent energy management control, and the method comprises the steps: obtaining an energy storage unit operation state parameter and an environment disturbance factor matrix through a multi-source data collection device, and carrying out the data preprocessing; and constructing a long-short-term memory neural network model, dynamically adjusting a prediction time window of the model and optimizing a target function weight coefficient based on the load predicted by the model, and determining a charging and discharging control instruction. And updating and optimizing the target function weight coefficient and the charging and discharging strategy library through a reinforcement learning algorithm to realize self-adaptive optimization adjustment of the charging and discharging strategy. According to the method, high-precision energy management and intelligent optimization control of the energy storage system in a complex environment are realized. The load prediction accuracy and the energy utilization rate of the system are improved, the aging rate of the battery is reduced, the service life of the battery is prolonged, the adjusting capacity of the energy storage system is improved, and the overall stability of the energy storage system in dynamic change is enhanced.

Description

Technical Field

[0001] This invention relates to the field of intelligent energy management and control technology for energy storage systems, specifically to an intelligent energy management and control method and system for energy storage systems. Background Technology

[0002] In recent years, new energy storage technologies such as lithium-ion batteries and supercapacitors have matured, leading to a continuous increase in the installed capacity of energy storage systems and placing higher demands on their intelligent energy management. Existing energy storage systems typically employ fixed strategies or simple linear models for charge and discharge control, making them ill-suited to complex and ever-changing environmental disturbances and load fluctuations. Meanwhile, with the rapid development of big data analytics and artificial intelligence technologies, improving the energy utilization efficiency and operational lifespan of energy storage systems through intelligent algorithms has become a hot research topic in the industry.

[0003] Existing intelligent energy management methods for energy storage systems generally suffer from shortcomings. They lack multi-source data fusion and efficient preprocessing capabilities, leading to inaccurate perception of the energy storage unit's operating status and environmental factors, thus affecting the effectiveness of subsequent control strategies. Load forecasting models mostly employ traditional regression or shallow machine learning algorithms, failing to effectively capture complex temporal characteristics, resulting in low prediction accuracy and an inability to dynamically adjust control strategies to cope with rapidly changing load demands. In the optimization control process, existing technologies are mostly based on static weight allocation and fixed charge / discharge strategies, lacking real-time adaptive adjustment capabilities. This makes it impossible to achieve optimal energy scheduling under multiple constraints, easily leading to problems such as low energy utilization and accelerated battery aging in energy storage systems. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by this invention is that existing intelligent energy management and control methods for energy storage systems suffer from inaccurate data perception, low load prediction accuracy, single control strategies, and a lack of adaptability, as well as the problem of how to achieve efficient energy utilization and battery life optimization in complex environments.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: an intelligent energy management and control method for an energy storage system, comprising acquiring the operating status parameters of the energy storage unit and the environmental disturbance factor matrix through a multi-source data acquisition device, and performing data preprocessing.

[0007] A long short-term memory neural network model is constructed. Based on the prediction time window of the load dynamic adjustment model predicted by the model and the weight coefficients of the optimized objective function, the charging and discharging control commands are determined.

[0008] By updating and optimizing the objective function weight coefficients and the charging / discharging strategy library using reinforcement learning algorithms, adaptive optimization and adjustment of the charging / discharging strategy can be achieved.

[0009] A policy gradient-based reinforcement learning algorithm is employed to construct a reward function that incorporates energy utilization rate, battery aging rate, and execution deviation of charge / discharge control commands. The weight coefficients of the objective function are dynamically adjusted. Based on the current energy storage unit operating state parameters and historical execution results, the reinforcement learning algorithm evaluates the effectiveness of policies in the charge / discharge policy library. When a policy's evaluation is superior to the current policy, it automatically replaces the corresponding policy in the library, enabling dynamic updates to the policy library. The policy library uses a hash index to manage optimal charge / discharge policies under different environmental disturbance scenarios. A policy invocation priority sorting mechanism enables rapid invocation of the optimal charge / discharge policy, adapting to changes in environmental disturbance factors and load demand.

[0010] As a preferred embodiment of the intelligent energy management and control method for the energy storage system described in this invention, the step of acquiring the operating status parameters of the energy storage unit and the environmental disturbance factor matrix through a multi-source data acquisition device includes: configuring a sensor device to collect the operating status parameters of the energy storage unit, including battery voltage, current, energy storage unit temperature and state of charge parameters; and collecting ambient temperature, humidity, grid frequency deviation and load change rate in real time through an environmental sensor.

[0011] As a preferred embodiment of the intelligent energy management and control method for the energy storage system described in this invention, the data preprocessing includes: removing outliers and filtering noise from the operating status parameters of the energy storage unit using a sliding time window mechanism; and synchronously organizing the ambient temperature, humidity, grid frequency deviation, and load change rate according to timestamps to form an environmental disturbance factor matrix.

[0012] As a preferred embodiment of the intelligent energy management and control method for the energy storage system described in this invention, the prediction time window and the weight coefficients of the optimized objective function of the load dynamic adjustment model based on model prediction include: constructing a long short-term memory neural network model; using historical load data, preprocessed energy storage unit operating state parameters, and an environmental disturbance factor matrix to predict the load change trend within a set time period; dynamically adjusting the prediction time window and control step size of the model based on the prediction results and the current state of charge; and plotting the predicted load curve. When the predicted load change rate exceeds a preset threshold, the prediction time window of the model predictive control is shortened and the control step size is increased. The weight coefficients of the optimized objective function are dynamically adjusted based on historical charge and discharge data, battery health status, and energy utilization rate using a gradient descent algorithm.

[0013] As a preferred embodiment of the intelligent energy management and control method for the energy storage system described in this invention, the determination of the charge / discharge control command includes: constructing an optimization objective function for charge / discharge control based on model prediction results, integrating the operating state parameters of the energy storage unit, the environmental disturbance factor matrix, and the predicted load curve; and solving for the optimal charge / discharge control command using a fast sequential quadratic programming algorithm. The optimization objective function considers the limitations of the energy storage unit's state of charge range, the safe threshold of the charge / discharge current, and the safe operating range of the battery temperature during the solution process. The determined charge / discharge control command is then sent to the energy storage unit for execution in real time via the control interface. The actual execution effect is dynamically monitored during the charge / discharge process. When the actual state of charge deviates from the predicted value by more than a set deviation range, the recalculation process of the charge / discharge control command is automatically triggered.

[0014] As a preferred embodiment of the intelligent energy management and control method for the energy storage system described in this invention, the step of updating and optimizing the objective function weight coefficients and the charging and discharging strategy library through a reinforcement learning algorithm includes: A policy gradient-based reinforcement learning algorithm is employed, defining a composite reward function that includes energy utilization rate, battery aging rate, and charge / discharge control command execution deviation. The weight coefficients of the objective function are dynamically and adaptively adjusted. The composite reward function calculates a comprehensive reward value based on the energy conversion efficiency, battery aging state changes, and charge / discharge control command deviation magnitude during the actual charge / discharge process of the energy storage unit. The reinforcement learning algorithm optimizes the current weight coefficient configuration based on this reward value. The algorithm uses real-time feedback of the energy storage unit's operating state parameters and historical charge / discharge strategy execution performance data, combined with a policy gradient update method, to calculate the optimal adjustment direction and step size for each weight coefficient. During the calculation, strategies that improve energy utilization rate beyond a preset threshold, reduce battery aging rate to the target range, and minimize charge / discharge control command deviation are prioritized and assigned higher weight adjustment ratios. The charge / discharge strategy library uses a hash index to classify and store the optimal charge / discharge strategies under different environmental disturbance scenarios. A multi-dimensional strategy call priority ranking mechanism is established in the strategy library. The matching degree is calculated by comparing the current environmental disturbance factor matrix with the feature vector of the energy storage unit's operating state, and the charge / discharge strategy with the highest matching degree is selected. When the reinforcement learning algorithm evaluates and finds that the current charging and discharging strategy performs better than the historical strategy, it automatically replaces the corresponding strategy record in the strategy library.

[0015] As a preferred embodiment of the intelligent energy management and control method for the energy storage system described in this invention, the adaptive optimization and adjustment of the charging and discharging strategy includes: setting a charging and discharging strategy priority calculation model; performing feature matching on the current energy storage unit operating state and environmental disturbance factors; and determining the optimal charging and discharging strategy based on the historical strategy execution effects. When the highest priority charging and discharging strategy is consistent with the current strategy, the existing charging and discharging strategy is maintained; when the highest priority charging and discharging strategy differs from the current strategy and the rate of change of the energy storage unit operating state reaches a switching threshold, the charging and discharging strategy switching process is automatically triggered.

[0016] Another objective of this invention is to provide an intelligent energy management and control system for an energy storage system. This system can use an optimization and adjustment module to update and optimize the objective function weight coefficients and the charging and discharging strategy library through reinforcement learning algorithms, thereby achieving adaptive optimization and adjustment of the charging and discharging strategy. This enables the energy storage system to achieve efficient energy scheduling and flexible strategy adjustment under complex environments and variable load demands, ultimately achieving the beneficial effects of reducing energy consumption, improving system stability, and enhancing adaptive decision-making capabilities.

[0017] As a preferred embodiment of the intelligent energy management and control system for the energy storage system described in this invention, it includes a data acquisition and preprocessing module, a calculation module, and an optimization and adjustment module.

[0018] The data acquisition and preprocessing module is used to acquire the operating status parameters of the energy storage unit and the environmental disturbance factor matrix through a multi-source data acquisition device, and to perform data preprocessing.

[0019] The computing module is used to construct a long short-term memory neural network model, and to determine the charging and discharging control command based on the model's predicted load dynamic adjustment model's prediction time window and the weight coefficients of the optimized objective function.

[0020] The optimization and adjustment module is used to update the weight coefficients of the objective function and the charging and discharging strategy library through reinforcement learning algorithms, so as to realize the adaptive optimization and adjustment of the charging and discharging strategy.

[0021] A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement a smart energy management and control method for an energy storage system.

[0022] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of an intelligent energy management and control method for an energy storage system.

[0023] The beneficial effects of this invention are as follows: The intelligent energy management and control method for energy storage systems provided by this invention improves the accuracy and stability of energy management through data acquisition and preprocessing, providing solid data support for energy storage scheduling under complex operating conditions. LSTM-based predictive and optimization control enhances the system's adaptability to load changes, effectively reduces battery aging rates, improves energy utilization and overall operating efficiency of the energy storage system, ultimately achieving the beneficial effects of improved economy and extended battery life. Reinforcement learning-based policy optimization enables efficient energy scheduling and flexible policy adjustment of the energy storage system under complex environments and variable load demands, ultimately achieving the beneficial effects of reduced energy consumption, improved system stability, and enhanced adaptive decision-making capabilities. Attached Figure Description

[0024] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 The first embodiment of the present invention provides an overall flowchart of an intelligent energy management and control method for an energy storage system.

[0026] Figure 2 The system flowchart is provided for a smart energy management and control method for an energy storage system according to the second embodiment of the present invention. Detailed Implementation

[0027] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0028] Example 1, referring to Figure 1 As an embodiment of the present invention, an intelligent energy management and control method for an energy storage system is provided, comprising: S1: Obtain the operating status parameters of the energy storage unit and the environmental disturbance factor matrix through a multi-source data acquisition device, and perform data preprocessing.

[0029] The device's sensors collect operating status parameters of the energy storage unit, including battery voltage, current, energy storage unit temperature, and state of charge parameters. Environmental sensors collect ambient temperature, humidity, grid frequency deviation, and load change rate in real time.

[0030] Furthermore, voltage, current, and temperature sensors are configured to collect the battery's voltage, current, and temperature values ​​in real time. State-of-charge (SOC) parameters are collected in real time using a SOC estimation device based on the open-circuit voltage method and coulomb measurement. Grid frequency deviation is recorded in real time by a grid frequency monitoring device, and the load change rate is obtained by collecting load power change data and calculating the change amplitude per unit time. The calculation method is: the absolute value of the current load power minus the previous load power value, divided by the previous load power value.

[0031] The operating status parameters of the energy storage unit are used to remove outliers and filter noise through a sliding time window mechanism. The ambient temperature, humidity, grid frequency deviation and load change rate are synchronously organized according to timestamps to form an environmental disturbance factor matrix.

[0032] A fixed sliding time window of 10 minutes is set. The mean and standard deviation are calculated based on historical data collected within the time window. Data points deviating from the current window mean by more than three times the standard deviation are considered outliers and removed. Noise filtering employs a Kalman filter algorithm to smooth continuously sampled data, eliminating instantaneous fluctuations caused by sensor noise and improving data stability and reliability. Preprocessed environmental temperature, humidity, grid frequency deviation, and load change rate data are synchronized and organized according to a unified timestamp, forming a complete environmental disturbance factor matrix. The environmental disturbance factor matrix is ​​arranged in chronological order of data collection. Each time point includes temperature, humidity, grid frequency deviation, and load change rate values ​​to ensure temporal consistency and completeness, providing accurate environmental input variables for subsequent model prediction and control strategy optimization.

[0033] It should be noted that S1 improves the accuracy of acquiring energy storage unit operating data and environmental factors through multi-source data acquisition and precise preprocessing, ensuring the reliability of input data for subsequent model predictions. By using a sliding time window and Kalman filtering to reduce the impact of abnormal fluctuations and improve data stability, it provides high-quality, time-consistent basic data support for charging and discharging strategy optimization and energy management control, thereby improving the overall decision-making accuracy and response capability of the system.

[0034] S2: Construct a long short-term memory neural network model, and determine the charging and discharging control command based on the prediction time window of the load dynamic adjustment model and the weight coefficients of the optimized objective function.

[0035] A long short-term memory neural network model is constructed to predict the load change trend within a set time period using historical load data, preprocessed energy storage unit operating status parameters, and environmental disturbance factor matrix. The prediction time window and control step size of the model are dynamically adjusted based on the prediction results and the current state of charge, and the predicted load curve is plotted.

[0036] Furthermore, in this invention, a long short-term memory neural network model is constructed to predict load change trends over a future period. Specifically, historical load data, preprocessed energy storage unit operating status parameters, and an environmental disturbance factor matrix are used as input variables for the model. The historical load data includes load power values ​​at multiple consecutive moments; the energy storage unit operating status parameters include battery voltage, current, temperature, and state of charge; and the environmental disturbance factor matrix includes ambient temperature, humidity, grid frequency deviation, and load change rate. Based on the input variables, the model calculates the predicted load power value for each future prediction moment, thereby plotting a load change trend curve. This provides fundamental data support for subsequent adjustments to the prediction time window and optimization of control strategies.

[0037] When the predicted load change rate exceeds a preset threshold, the prediction time window of the model predictive control is shortened and the control step size is increased. The weight coefficients of the optimized objective function are dynamically adjusted based on historical charge-discharge data, battery health status, and energy utilization rate using a gradient descent algorithm.

[0038] Furthermore, based on the predicted load change trend, the current predicted load change rate is calculated according to the changes in the predicted load power values ​​between two adjacent prediction times. Specifically, the predicted load power value at the current time is subtracted from the predicted load power value at the previous time, the absolute value is taken, and then divided by the load power value at the previous time. The result is the load change rate. When the calculated load change rate is greater than or equal to a set threshold (15% in this invention), it indicates that the load change trend has fluctuated drastically. In this case, the model's prediction time window is shortened to its minimum (10 minutes is recommended), and the model's control step size is increased to enhance its responsiveness to rapid load changes. Conversely, when the predicted load change rate is lower than the set threshold, it indicates that the load change is relatively stable. The default prediction time window length (30 minutes in this invention) and control step size are maintained to reduce computational complexity and maintain prediction stability.

[0039] Based on the model prediction results, and considering the operating state parameters of the energy storage unit, the environmental disturbance factor matrix, and the predicted load curve, an optimization objective function for charge and discharge control is constructed. The optimal charge and discharge control command is then solved using a fast sequential quadratic programming algorithm. The optimization objective function considers constraints such as the energy storage unit's state of charge range, the safe threshold for charge and discharge current, and the safe operating range for battery temperature during the solution process.

[0040] Furthermore, the objective function can be expressed as: , in, This indicates the optimization of the objective function value. This indicates the number of prediction steps. , , Indicates the first The weighting coefficients at each time point correspond to the state of charge deviation, charge / discharge power, and aging factor. The target state of charge is indicated; in this invention, it is set to 50%. Indicates time State of charge of the energy storage unit. Indicates time Energy storage unit charging and discharging power. Indicates time Battery aging factor is estimated based on actual operating current and temperature.

[0041] The optimal charge / discharge control command, obtained using a fast sequential quadratic programming algorithm, is expressed as follows: , in, This indicates the charging and discharging power value of the energy storage unit. Indicates time Charge and discharge power commands for the energy storage unit. Indicates the time of the energy storage unit battery Temperature value. Indicates time State of charge of the energy storage unit. The limit of the state of charge is indicated, which is set to 20% in this invention. This indicates the upper limit of the state of charge; in this invention, it is set to 80%. This represents the minimum charging and discharging power, which is set to 80% of the system's maximum charging and discharging capacity in this invention. This indicates the maximum charging and discharging power, which is set to 100% of the system's maximum charging and discharging capacity in this invention. This indicates the minimum safe temperature for the battery; in this invention, it is set to 0°C. This indicates the maximum safe temperature for the battery, which is set to 45°C in this invention.

[0042] Once a definite charge / discharge control command is obtained, it is sent to the energy storage unit for execution in real time through the control interface. The actual execution effect is dynamically monitored during the charge / discharge process. When the actual state of charge deviates from the predicted value by more than the set deviation range, the charge / discharge control command recalculation process is automatically triggered.

[0043] Furthermore, the deviation range is set to ±5% in this invention. When the actual state of charge deviates from the predicted value by more than ±5%, the recalculation process of the charge and discharge control command is triggered to correct the control strategy in a timely manner and ensure that the energy storage unit operates smoothly according to the predicted target.

[0044] It should be noted that S2, by introducing a long short-term memory neural network model, achieves accurate prediction of the load change trend of the energy storage unit. It also dynamically adjusts the model's prediction time window and control step size based on the prediction results, enhancing the system's rapid response capability to load fluctuations. Furthermore, by introducing an optimization objective function and a fast sequential quadratic programming algorithm, it achieves the calculation of optimal charge and discharge control commands under multiple constraints. Based on real-time monitoring of the state-of-charge deviation, it automatically adjusts the control strategy, effectively improving the energy utilization rate and control accuracy of the energy storage system. Compared to the shortcomings of existing technologies that rely on fixed control strategies and static parameter configurations, this step achieves adaptive optimization and dynamic adjustment of charge and discharge control, improving the intelligent decision-making capability and energy efficiency management level of the energy storage system in complex environments, and effectively extending battery life.

[0045] S3: The objective function weight coefficients and the charging / discharging strategy library are updated and optimized through reinforcement learning algorithms to achieve adaptive optimization and adjustment of the charging / discharging strategy.

[0046] A policy gradient-based reinforcement learning algorithm is adopted, and a composite reward function is defined that includes energy utilization rate index, battery aging rate index and charge / discharge control command execution deviation. The weight coefficients of the optimization objective function are dynamically and adaptively adjusted.

[0047] Furthermore, the composite reward function is expressed as: , in, This represents the combined reward value of the composite reward function. This represents the energy utilization bonus value, calculated as the ratio of the actual discharge energy to the theoretical maximum usable energy in this cycle. This represents the battery aging rate bonus value, calculated as the reciprocal of the rate of decrease in battery life during this cycle. This represents the reward value for deviation in the execution of charge / discharge control commands, calculated as the absolute value of the deviation between the actual charge / discharge power and the predicted charge / discharge power. , and The weighting coefficients for energy utilization rate, battery aging rate, and deviation reward are set to 0.5, 0.3, and 0.2, respectively, in this invention.

[0048] The composite reward function calculates a comprehensive reward value based on the energy conversion efficiency, battery aging state changes, and deviations in charge / discharge control commands during the actual charging and discharging process of the energy storage unit. The reinforcement learning algorithm optimizes the current weight coefficient configuration using this reward value. Furthermore, the reinforcement learning algorithm calculates the optimal adjustment direction and step size for each weight coefficient using real-time feedback of the energy storage unit's operating state parameters and historical charging / discharging strategy execution performance data, combined with the strategy gradient update method.

[0049] Furthermore, the weight coefficient update is expressed as: , in, Indicates the updated number The weight coefficients of the objective function are optimized. Indicates the number before the update Each weight coefficient. The learning rate is set to 0.01 in this invention to control the magnitude of weight updates. This represents the gradient change of the composite reward function with respect to the weight coefficients, indicating the degree of contribution of the current weight to the reward function.

[0050] During the calculation process, strategies that improve energy utilization beyond a preset threshold, reduce battery aging rate to the target range, and minimize charge / discharge control command deviations are given higher weight adjustments. The charge / discharge strategy library uses a hash index to categorize and store the optimal charge / discharge strategies under different environmental disturbance scenarios. A multi-dimensional strategy call priority ranking mechanism is established in the strategy library. The matching degree is calculated by comparing the current environmental disturbance factor matrix with the feature vector of the energy storage unit's operating state, and the charge / discharge strategy with the highest matching degree is selected. When the reinforcement learning algorithm evaluation finds that the current charge / discharge strategy performs better than historical strategies, the corresponding strategy record in the strategy library is automatically replaced.

[0051] A priority calculation model for charging and discharging strategies is established. Feature matching is performed on the current operating state of the energy storage unit and environmental disturbance factors, and the optimal charging and discharging strategy is determined based on the historical strategy execution results. When the highest priority charging and discharging strategy is consistent with the current strategy, the existing charging and discharging strategy is maintained. When the highest priority charging and discharging strategy is different from the current strategy and the rate of change of the energy storage unit's operating state reaches the switching threshold, the charging and discharging strategy switching process is automatically triggered.

[0052] Furthermore, the strategy priority calculation is expressed as: , in, Indicates the first in the strategy library The priority value of each strategy, with lower values ​​indicating higher priority. This represents the total number of environmental disturbance factors and energy storage unit state characteristics. Indicates the first The weighting factors of each feature are used to adjust the impact of different features on priority, and the optimal weights are set based on the historical strategy results. Indicates the current time. The actual value of each feature. Indicates the first in the strategy library The strategy corresponds to the first The historical best value of each feature.

[0053] The switching threshold is set to 5% in this invention. When the rate of change of the operating status of the energy storage unit is greater than 5%, the strategy switching is triggered.

[0054] It should be noted that S3, by introducing a policy gradient-based reinforcement learning algorithm, achieves dynamic adaptive updates of the objective function weight coefficients and the charge / discharge strategy library, thereby enhancing the intelligent decision-making capability of the energy storage system in complex environments. By designing a composite reward function that includes energy utilization rate, battery aging rate, and charge / discharge command execution deviation, the reinforcement learning algorithm can automatically optimize weight allocation and strategy selection based on real-time operating status and historical strategy effects, improving energy utilization rate, reducing battery aging rate, and minimizing control deviation. Compared to the shortcomings of existing technologies with fixed strategies and static weight configurations, this step achieves adaptive switching of charge / discharge strategies and rapid matching of optimal strategies, significantly enhancing the regulation flexibility and operational stability of the energy storage system under different loads and environmental changes, effectively extending battery life and improving the overall system energy efficiency.

[0055] Example 2, refer to Figure 2 As an embodiment of the present invention, an intelligent energy management and control system for an energy storage system is provided, including a data acquisition and preprocessing module 100, a calculation module 200, and an optimization and adjustment module 300.

[0056] S4: Data acquisition and preprocessing module 100 is used to acquire the operating status parameters of the energy storage unit and the environmental disturbance factor matrix through the multi-source data acquisition device, and to perform data preprocessing.

[0057] The data acquisition and preprocessing module 100 includes a data acquisition submodule 101 and a data preprocessing submodule 102.

[0058] Furthermore, the data acquisition submodule 101 is used to acquire the voltage, current, temperature, and state of charge parameters of the energy storage unit in real time by configuring voltage sensors, current sensors, temperature sensors, and a state of charge estimation device. It also acquires external disturbance factors such as ambient temperature, humidity, grid frequency deviation, and load change rate through environmental sensors. The data preprocessing submodule 102 is used to perform outlier removal and noise filtering on the acquired data using a sliding time window mechanism, employs a Kalman filter algorithm to improve data stability, and synchronously organizes the preprocessed data according to timestamps to form an environmental disturbance factor matrix, providing high-quality input data for subsequent load forecasting and control optimization.

[0059] It should be noted that the data acquisition submodule 101 is the starting point for the data acquisition and preprocessing module 100, and directly affects the accuracy of subsequent model predictions. The data preprocessing submodule 102 ensures the integrity and validity of the input data through data cleaning and feature extraction, providing a solid data foundation for model training and optimization decisions.

[0060] S5: The calculation module 200 is used to construct a long short-term memory neural network model, and determine the charging and discharging control commands based on the prediction time window of the load dynamic adjustment model and the weight coefficients of the optimized objective function.

[0061] The calculation module 200 includes a model prediction submodule 201 and a control instruction calculation submodule 202.

[0062] Furthermore, the model prediction submodule 201 is used to construct a long short-term memory neural network model. By inputting historical load data, preprocessed energy storage unit operating state parameters, and environmental disturbance factor matrix, it predicts the load change trend within a set time period in the future, dynamically adjusts the model's prediction time window and control step size, and plots the predicted load curve. The control command calculation submodule 202 is used to solve for the optimal charging and discharging control command that satisfies the constraints of state of charge, safe temperature, and charging and discharging current based on the prediction results and the optimization objective function, using a fast sequential quadratic programming algorithm. This command is then sent to the energy storage unit for execution in real time through the control interface.

[0063] It should be noted that the model prediction submodule 201 provides high-precision load change trend prediction results, which is the foundation for realizing dynamic optimization control. The control command calculation submodule 202 ensures that the charging and discharging commands meet the balance between safety and economy, thereby improving the energy utilization efficiency and battery life of the energy storage system.

[0064] S6: The optimization and adjustment module 300 is used to update and optimize the objective function weight coefficients and the charging and discharging strategy library through reinforcement learning algorithms, so as to realize the adaptive optimization and adjustment of the charging and discharging strategy.

[0065] The optimization and adjustment module 300 includes a weight coefficient optimization submodule 301 and a strategy library management submodule 302.

[0066] Furthermore, the weight coefficient optimization submodule 301 employs a policy gradient-based reinforcement learning algorithm, defining a composite reward function that includes energy utilization rate, battery aging rate, and charge / discharge control command deviation. It dynamically adjusts the weight coefficients of the objective function through real-time feedback and historical policy effects, improving the adaptability and optimization effect of the charge / discharge strategy. The policy library management submodule 302 uses a hash index and priority calculation model to classify, store, and quickly retrieve the optimal charge / discharge strategies under different environmental disturbance scenarios. When the rate of change of the energy storage unit's operating state exceeds a set threshold, it automatically triggers a policy switch, ensuring that the charge / discharge strategy remains in its optimal state under different environments.

[0067] It should be noted that the weight coefficient optimization submodule 301 is the core of the optimization and adjustment module 300, achieving adaptive optimization and adjustment of the charging and discharging strategy through reinforcement learning. The strategy library management submodule 302 improves the system's intelligent decision-making ability and energy scheduling efficiency in response to complex environmental changes through an efficient strategy storage and retrieval mechanism.

[0068] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part 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 several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0069] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing 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 (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0070] More specific examples (a non-exhaustive list) of computer-readable media include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which programs can be printed, because programs can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0071] 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 embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc. It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A smart energy management and control method for an energy storage system, characterized in that, include: The energy storage unit's operating status parameters and environmental disturbance factor matrix are acquired through a multi-source data acquisition device, and the data is preprocessed. A long short-term memory neural network model is constructed. Based on the prediction time window of the load dynamic adjustment model predicted by the model and the weight coefficients of the optimized objective function, the charging and discharging control commands are determined. By updating and optimizing the objective function weight coefficients and the charging and discharging strategy library through reinforcement learning algorithms, adaptive optimization and adjustment of the charging and discharging strategy can be achieved. The objective function weight coefficients and charge / discharge strategy library are updated and optimized through reinforcement learning algorithms. The algorithm is based on policy gradient and constructs a reward function that includes energy utilization rate, battery aging rate and charge / discharge control command execution deviation. The objective function weight coefficients are then dynamically adjusted. The reinforcement learning algorithm evaluates the effectiveness of the strategies in the charging and discharging strategy library based on the current operating status parameters of the energy storage unit and the historical execution results. When a strategy is evaluated as superior to the current strategy, it automatically replaces the corresponding strategy in the charging and discharging strategy library, thereby realizing the dynamic updating of the charging and discharging strategy library. The charging and discharging strategy library uses a hash index to manage the optimal charging and discharging strategies under different environmental disturbance scenarios. It achieves fast invocation of the optimal charging and discharging strategy through a strategy invocation priority sorting mechanism, adapting to changes in environmental disturbance factors and load demand.

2. The intelligent energy management and control method for an energy storage system as described in claim 1, characterized in that: The acquisition of energy storage unit operating status parameters and environmental disturbance factor matrix through multi-source data acquisition devices includes, The device's sensors collect operating status parameters of the energy storage unit, including battery voltage, current, energy storage unit temperature, and state of charge parameters. Environmental sensors collect ambient temperature, humidity, grid frequency deviation, and load change rate in real time.

3. The intelligent energy management and control method for an energy storage system as described in claim 1 or 2, characterized in that: The data preprocessing includes, The operating status parameters of the energy storage unit are used to remove outliers and filter noise through a sliding time window mechanism. The ambient temperature, humidity, grid frequency deviation and load change rate are synchronously organized according to timestamps to form an environmental disturbance factor matrix.

4. The intelligent energy management and control method for an energy storage system as described in claim 3, characterized in that: The prediction time window and optimization objective function weight coefficients of the model-based load dynamic adjustment model include, A long short-term memory neural network model is constructed. Historical load data, preprocessed energy storage unit operating status parameters, and environmental disturbance factor matrix are used to predict the load change trend within a set time period. The prediction time window and control step size of the model are dynamically adjusted according to the prediction results and the current state of charge, and the predicted load curve is plotted. When the predicted load change rate exceeds the preset threshold, the prediction time window of the model predictive control is shortened and the control step size is increased. The weight coefficients of the objective function are optimized based on historical charge and discharge data, battery health status, and energy utilization rate, and are dynamically adjusted through the gradient descent algorithm.

5. The intelligent energy management and control method for an energy storage system as described in any one of claims 1, 2, or 4, characterized in that: The determined charge / discharge control command includes, Based on the model prediction results, the operating status parameters of the energy storage unit, the environmental disturbance factor matrix, and the predicted load curve are integrated to construct the charging and discharging control optimization objective function, and the optimal charging and discharging control command is solved by the fast sequence quadratic programming algorithm. The optimization objective function takes into account the limitations of the energy storage unit's state of charge range, the safe threshold of charge and discharge current, and the safe operating range of battery temperature during the solution process. Once a definite charge / discharge control command is obtained, it is sent to the energy storage unit for execution in real time through the control interface. The actual execution effect is dynamically monitored during the charge / discharge process. When the actual state of charge deviates from the predicted value by more than the set deviation range, the charge / discharge control command recalculation process is automatically triggered.

6. The intelligent energy management and control method for an energy storage system as described in claim 5, characterized in that: The method of updating and optimizing the objective function weight coefficients and the charging / discharging strategy library through reinforcement learning algorithms includes, A policy gradient-based reinforcement learning algorithm is adopted, and a composite reward function is defined that includes energy utilization rate index, battery aging rate index and charge / discharge control command execution deviation. The weight coefficients of the optimization objective function are dynamically and adaptively adjusted. The composite reward function calculates a comprehensive reward value based on the energy conversion efficiency, battery aging status changes, and deviation of charge and discharge control commands during the actual charging and discharging process of the energy storage unit. The reinforcement learning algorithm optimizes the current weight coefficient configuration through the reward value. The reinforcement learning algorithm calculates the optimal adjustment direction and step size of each weight coefficient by combining real-time feedback of the energy storage unit's operating status parameters and historical charging and discharging strategy execution effect data with the strategy gradient update method. During the calculation process, strategies that improve energy utilization rate beyond the preset threshold, reduce battery aging rate to the target range, and minimize charging and discharging control command deviation are given priority in being assigned higher weight adjustment ratios. The charging and discharging strategy library uses a hash index to classify and store the optimal charging and discharging strategies under different environmental disturbance scenarios. A multi-dimensional strategy call priority sorting mechanism is established in the strategy library. The matching degree is calculated by the current environmental disturbance factor matrix and the feature vector of the energy storage unit's operating state, and the charging and discharging strategy with the highest matching degree is selected. When the reinforcement learning algorithm evaluates and finds that the current charging and discharging strategy performs better than the historical strategy, it automatically replaces the corresponding strategy record in the strategy library.

7. The intelligent energy management and control method for an energy storage system as described in any one of claims 1, 2, 4, or 6, characterized in that: The adaptive optimization and adjustment of the charging and discharging strategy includes... A priority calculation model for charging and discharging strategies is set up, and feature matching is performed on the current operating status of the energy storage unit and environmental disturbance factors. The optimal charging and discharging strategy is determined based on the historical strategy execution results. When the highest priority charging / discharging strategy is consistent with the current strategy, the existing charging / discharging strategy will continue to be executed. When the highest priority charging / discharging strategy is different from the current strategy and the change rate of the energy storage unit's operating state reaches the switching threshold, the charging / discharging strategy switching process will be automatically triggered.

8. An intelligent energy management and control system for an energy storage system, characterized in that: It includes a data acquisition and preprocessing module (100), a calculation module (200), and an optimization and adjustment module (300); The data acquisition and preprocessing module (100) is used to acquire the energy storage unit's operating status parameters and environmental disturbance factor matrix through a multi-source data acquisition device, and to perform data preprocessing. The calculation module (200) is used to construct a long short-term memory neural network model, and determine the charging and discharging control command based on the prediction time window of the load dynamic adjustment model and the weight coefficient of the optimization objective function. The optimization and adjustment module (300) is used to update the weight coefficients of the objective function and the charging and discharging strategy library through reinforcement learning algorithm, so as to realize the adaptive optimization and adjustment of the charging and discharging strategy.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the intelligent energy management and control method for the energy storage system as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the intelligent energy management and control method for the energy storage system as described in any one of claims 1 to 7.

Images