A method for collecting and compressing energy storage detailed data based on batch-stream fusion

Abstract

The present invention discloses a method for collecting and compressing energy storage detailed data based on batch-stream fusion, which relates to the technical field of data compression control and comprises the following steps: obtaining the operating state parameters of the energy storage system and performing state perception and hierarchical mapping on the energy storage system to obtain a state level; adaptively adjusting the state level based on a deep reinforcement learning method to obtain compression control parameters; adjusting the compression control strategy of the energy storage system based on the fuzzy reasoning method and performing feedback optimization according to the compression control parameters; the present invention realizes adaptive adjustment of the compression strategy by constructing an operating state potential function and combining deep reinforcement learning with fuzzy reasoning, thereby solving the problems of fixed compression strategy, delayed regulation response and insufficient control accuracy of existing energy storage systems under multi-state working conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of data compression control, and more specifically, to a method for collecting and compressing energy storage detailed data based on batch-stream fusion. Background Art

[0002] As the proportion of new energy access continues to increase, energy storage systems, as important regulating devices for power system peak and frequency regulation, voltage support, energy balance, and coordinated control of distributed energy, are becoming increasingly prominent in automated control systems. To ensure the dynamic controllability and system responsiveness of energy storage devices in multi-source heterogeneous operation scenarios, it is necessary to frequently collect and continuously monitor detailed energy storage data (such as voltage, current, temperature, state of charge (SOC), power changes, etc.) during the operation of battery clusters, battery management systems, converters, and auxiliary equipment. However, in actual operation, energy storage systems face the following major problems:

[0003] Energy storage systems typically sample detailed energy storage data at the second or even millisecond level. This sampling frequency needs to be further increased, especially under conditions of frequent grid fluctuations or rapid load changes. This results in a sharp increase in data traffic between the control system's edge nodes and the central processing unit, causing network communication congestion, high data redundancy, and low query efficiency.

[0004] To cope with the pressure brought by high-frequency data collection and continuous monitoring, some energy storage systems have introduced data compression mechanisms to reduce the dimension, abstract or simplify the original monitoring data before uploading.

[0005] For example, the invention patent announcement with announcement number: 202411948955.X discloses a running speed safety control method based on a reinforcement learning hybrid control strategy, which includes the following steps: S1. Constructing a high-speed train operation process model; S2. Constructing a reward function for a hybrid control strategy based on reinforcement learning and fractional-order PID; S3. Constructing a hybrid control strategy based on reinforcement learning and fractional-order PID, and generating the final control signal through superposition processing.

[0006] For example, the invention patent announcement with announcement number: 201810142764.2 discloses an AGC control method and device for a thermal power unit based on a fuzzy control strategy, which includes the following steps: obtaining a fuzzy quantity by performing fuzzy algorithm processing on the active power deviation and the change in active power deviation, and deriving a valid fuzzy set based on the optimized membership function and pre-established fuzzy rules, and obtaining the actual value by using anti-fuzzy calculation, and generating a corresponding adjustment signal.

[0007] The above disclosed technical solutions have at least the following technical problems:

[0008] Current compression control methods are typically static and do not dynamically sense and adjust according to the operating status and environmental changes of the energy storage system. Since the energy storage system has different requirements for data acquisition and compression under different operating conditions, static control strategies cannot meet the system's requirements for compression accuracy and storage efficiency in each state, resulting in unstable compression effects. In addition, existing methods mostly rely on fixed rules and lack intelligent optimization mechanisms based on adaptive adjustment of system status. In addition, the operating status and environmental factors of the energy storage system are relatively complex, making it difficult for traditional experience-based and rule-based methods to achieve efficient and accurate compression control. In response to the above problems, the present invention proposes a solution. Summary of the Invention

[0009] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a method for collecting and compressing energy storage detailed data based on batch-stream fusion. By constructing an operating state potential function and combining deep reinforcement learning with fuzzy reasoning, adaptive adjustment of the compression strategy is achieved, which solves the problems of fixed compression strategy, delayed control response and insufficient control accuracy of existing energy storage systems under multi-state working conditions.

[0010] To achieve the above object, the present invention provides the following technical solutions:

[0011] A method for collecting and compressing energy storage detailed data based on batch-stream fusion includes the following steps: obtaining operating status parameters of an energy storage system and performing state perception and hierarchical mapping on the energy storage system to obtain a state level; adaptively adjusting the state level based on a deep reinforcement learning method to obtain compression control parameters; and adjusting the compression control strategy of the energy storage system based on the fuzzy inference method according to the compression control parameters and performing feedback optimization.

[0012] In a preferred embodiment, the operating state parameters of the energy storage system are obtained and the energy storage system is state-aware and graded mapped to obtain the state level, specifically: the operating state parameters of the energy storage system at the current moment are obtained, and a state vector is constructed, and the state vectors of adjacent time points are differentially processed to obtain a state perception vector; a sliding time window sequence is constructed based on the state perception vector based on the current moment and a preset window length; for each type of state perception vector in the window sequence, its evolution trend in the time dimension is extracted and a state evolution rate vector is formed; a state potential function is constructed through the state evolution rate vector based on the principal component analysis method; and state perception and state graded mapping are performed according to the state potential function to obtain the state level.

[0013] In a preferred embodiment, for each type of state perception vector in the window sequence, its evolution trend in the time dimension is extracted respectively and a state evolution rate vector is formed, specifically: a state variable sequence is extracted according to the state perception vector in the window sequence, wherein the state variable sequence includes the state of charge, health state estimation value and battery cell temperature; a fitting model is obtained by performing time trend fitting operation on each variable in the state variable sequence based on the least squares method; and a fitting slope is extracted according to the fitting model as the evolution rate of the state variable sequence to obtain a state evolution rate vector.

[0014] In a preferred embodiment, state perception and state grading mapping are performed according to the state potential function to obtain the state level, specifically: a state potential function sequence within a preset time period is obtained, and sorted in ascending order to construct a set of candidate partitioning points, wherein each candidate partitioning point is the median of two adjacent state potential function values; for each candidate partitioning point, the state potential function sequence is divided into a left subset and a right subset, and the entropy values ​​of the left subset and the right subset are calculated respectively; the total partition entropy corresponding to each candidate partitioning point is calculated based on the number of samples through the entropy value, and the information gain of the current candidate partitioning point is obtained based on the original entropy of the entire sequence; the partitioning point with the largest information gain is selected from the candidate partitioning point set as the first-level grading threshold of the state potential function; on the basis of the first-level grading threshold division, the information gain acquisition and selection manipulation are repeated, and recursive partitioning is performed iteratively until the preset number of grading layers is reached to obtain a state potential function grading threshold set; based on the state potential function grading threshold set, the state potential function value is graded and mapped to obtain the state level.

[0015] In a preferred embodiment, the compression control parameters are obtained by adaptively adjusting the state level based on a deep reinforcement learning method, specifically by: constructing a Markov decision process model based on the state level;

[0016] The Markov decision process model is iteratively trained using a deep deterministic policy gradient algorithm, and interactive optimization of policy evaluation and policy improvement is achieved through a dual neural network architecture until the cumulative discounted reward value of the policy function converges to a preset threshold;

[0017] The trained strategy function is deployed as the compression control strategy mapping function to establish a dynamic mapping relationship between the state level and the compression parameter group. The mapping relationship is implemented through an improved radial basis function neural network, and parameter boundary constraints are set to ensure the feasibility of the compression operation.

[0018] In a preferred embodiment, the compression control strategy of the energy storage system is adjusted and feedback optimized based on the fuzzy reasoning method according to the compression control parameters. Specifically, the compression control parameters are fuzzified, and continuous input values ​​are mapped into fuzzy linguistic values ​​using a triangular membership function to form a fuzzy set; a fuzzy rule base is established based on the fuzzy set, and the fuzzy linguistic values ​​and their memberships are used as input items. Fuzzy reasoning is performed based on the rule base, and a fuzzy output variable is obtained using a Mamdani reasoning mechanism to obtain a lossless priority value; the lossless priority value is defuzzified to obtain a clear quantized value, and a judgment is made based on a preset lossless priority threshold; a compression control strategy adjustment signal is obtained based on the judgment result, and a corresponding lossless or lossy compression strategy is selected for the data collected by the energy storage system; feedback parameters after the compression round are collected to construct a compression effect scoring function to obtain a compression strategy performance score value, and the feedback parameters include an actual compression ratio, an actual reconstruction error, and an actual compression delay; if the score value is lower than a preset threshold for several consecutive cycles, the score value is fed back to the fuzzy reasoning system to trigger optimization of the fuzzy rule base.

[0019] In a preferred embodiment, the optimization of the fuzzy rule base is triggered, specifically as follows: based on a genetic algorithm, the fuzzy rules are encoded as genetic individuals in the form of rule vectors, and each fuzzy rule includes a matching relationship between the fuzzy linguistic value of the input parameter and the output strategy priority; the compression effect scoring function is used as the fitness function of the genetic algorithm, and the individuals are iteratively evolved; after iterating for a preset number of evolutionary rounds, the set of fuzzy rule individuals with the largest fitness value is selected as the updated fuzzy rule base, and the rules in the current inference system are replaced; the iterative evolution operation is repeated to form a closed-loop fuzzy rule optimization process driven by compression feedback to optimize the fuzzy rule base.

[0020] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0021] 1. By introducing a state-aware and hierarchical mapping mechanism, the state evolution rate vector and state potential function are constructed based on the energy storage system's operating parameters, enabling dynamic trend modeling and hierarchical classification of the operating state. Compared to traditional static compression strategies, this approach offers significant technical advantages such as strong feedforward capabilities, sensitive response, and high discrimination. It can identify system fluctuation trends and potential anomalies based on the rate of state change, accurately mapping the current state level. This enables the compression control strategy to possess hierarchical awareness and dynamic adaptability, effectively improving data compression efficiency and fidelity under multi-state conditions, and facilitating differentiated retention and intelligent control of critical data.

[0022] 2. By combining deep reinforcement learning with fuzzy reasoning mechanisms, adaptive adjustment of the state level and feedback closed-loop optimization of the compression control parameters of the energy storage system are achieved, with the following significant technical advantages: On the one hand, the reinforcement learning strategy dynamically maps the compression parameters according to the system operating status, achieving a strong correlation and flexible adaptability between the compression strategy and the operating status, effectively avoiding the problem of decoupling the compression strength from the system status in traditional methods; on the other hand, the fuzzy reasoning mechanism is combined with the genetic algorithm to achieve precise judgment and adaptive evolution of the compression strategy, with strong nonlinear processing capabilities and uncertainty resistance, and can maintain high robustness and efficiency of the compression strategy in a changing environment. The overall solution constructs a closed-loop optimization system driven by the operating status and based on the compression effect, which significantly improves the accuracy of compression strategy judgment, system resource utilization efficiency, and operational stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 A flow chart of a method for collecting and compressing energy storage detailed data based on batch-stream fusion provided in an embodiment of the present application. DETAILED DESCRIPTION

[0024] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0025] Example 1, Figure 1 A flow chart of a method for collecting and compressing energy storage detailed data based on batch-stream fusion provided in an embodiment of the present application includes the following steps:

[0026] S1, obtain the operating status parameters of the energy storage system and perform state perception and hierarchical mapping of the energy storage system to obtain the state level.

[0027] In this embodiment, the operating state of the energy storage system directly influences the dynamic characteristic distribution of the data. Unlike traditional static sampling compression strategies, this approach obtains energy storage system operating state parameters to provide the raw input for state trend modeling and the construction of perception indicators. Furthermore, by obtaining the state level, the energy storage system's current state range can be accurately calibrated, enabling the compression control strategy to implement level-aware adaptive control.

[0028] The acquisition of the energy storage system operating status parameters and the performance of state perception and hierarchical mapping of the energy storage system to obtain the state level are specifically as follows:

[0029] Obtaining the current operating status parameters of the energy storage system, including voltage, current, state of charge, estimated state of health, and battery cell temperature;

[0030] A state vector is constructed based on the operating state parameters of the energy storage system, and the state vectors at adjacent time points are differentially processed to obtain a state perception vector;

[0031] Construct a sliding time window sequence based on the current moment and the preset window length according to the state perception vector;

[0032] For each type of state perception vector in the window sequence, extract its evolution trend in the time dimension and form a state evolution rate vector;

[0033] The state potential function is constructed through the state evolution rate vector based on principal component analysis;

[0034] The state level is obtained by performing state perception and state classification mapping based on the state potential function, specifically:

[0035] Obtain the state potential function sequence within a preset time period, sort it in ascending order, and construct a set of candidate partition points, where each candidate partition point is the median of two adjacent state potential function values;

[0036] For each candidate partition point, the state potential function sequence is divided into a left subset and a right subset, and the entropy values ​​of the left subset and the right subset are calculated respectively;

[0037] The total partition entropy corresponding to each candidate partition point is calculated based on the number of samples through the entropy value, and the information gain of the current candidate partition point is obtained based on the original entropy of the entire sequence;

[0038] Select the partition point with the largest information gain from the candidate partition point set as the first-level classification threshold of the state potential function;

[0039] On the basis of the first-level threshold division, the information gain acquisition and selection manipulation are repeated, and the recursive division is iterated until the preset number of hierarchical layers is reached to obtain the state potential function threshold set, which includes the first-level threshold, the second-level threshold and the third-level threshold, that is, 、 、 ;

[0040] Based on the state potential function classification threshold set, the state potential function value is hierarchically mapped to obtain the state level, which includes level one, i.e. stable operation state; level two, i.e. slight fluctuation of operation state; level three, i.e. deviation from normal operation state; and level four, i.e. abnormal operation state warning;

[0041] like , which is level one;

[0042] like , which is level 2

[0043] like , which is level three

[0044] like , which is level four

[0045] in is the state potential function value.

[0046] The entropy values ​​of the left subset and the right subset are calculated as follows:

[0047]

[0048]

[0049] The total partition entropy is specifically calculated as follows:

[0050]

[0051] The specific calculation formula of the information gain is as follows:

[0052]

[0053] Where, is the entropy of the left subset, is the entropy of the right subset, is the category of the status level defined in the subset, is the first The probability of the class state level appearing, is the first The probability of the class state level appearing, The total partition entropy at the current partition point, is the number of samples in the left subset, is the number of samples in the right subset, is the total sample size, is the information gain, is the raw entropy of the entire sequence.

[0054] It should be noted that in the state classification or control strategy selection of energy storage systems, relying solely on the current state value is not forward-looking. The state perception vector obtained through differential processing can predict whether the system is tending to be stable or deviating from the normal range, and can explicitly characterize the evolution rate, fluctuation direction and degree of each parameter in a short time. Compared with using only static values, it helps to detect system anomalies earlier, enhances the state perception's ability to respond to short-term drastic changes and non-stationary characteristics, and provides time structure information for subsequent evolution modeling.

[0055] The evolution trend of each state perception vector in the window sequence in the time dimension is extracted and a state evolution rate vector is formed, which is specifically:

[0056] Extracting a state variable sequence based on the state perception vector in the window sequence, wherein the state variable sequence includes a state of charge, a health state estimation value, and a battery cell temperature;

[0057] Based on the least square method, the time trend fitting operation is performed on each variable in the state variable sequence to obtain the fitting model;

[0058] The fitting slope is extracted according to the fitting model as the evolution rate of the state variable sequence to obtain a state evolution rate vector.

[0059] The specific calculation formula of the fitting model is as follows:

[0060] ,

[0061] The state evolution rate vector is specifically:

[0062]

[0063] The specific calculation formula of the state potential function is as follows:

[0064]

[0065] Where, To fit the model, is the fitting slope, is the intercept of the fitted line at time zero, is the time position within the sliding time window, For the current moment, is the sliding window width, is the state evolution rate vector set, is the state of charge at time The evolution rate near Estimates of health status at time points The evolution rate near is the cell temperature at time point The evolution rate near is the state potential function value, 、 、 are weight coefficients respectively.

[0066] It should be noted that by extracting the state variable sequence—i.e., the state of charge, estimated state of health, and cell temperature—within a window, we can quantitatively characterize the short-term and medium- to long-term operational evolution paths of the energy storage system and construct a dynamic benchmark for operational evolution. The extracted state evolution rate vector serves as the primary input for the subsequent state potential function construction, determining the current data retention intensity and the model's key focus areas. Compared to static threshold rules, trend modeling offers greater predictive and feedforward capabilities, enabling more flexible and intelligent scheduling and anomaly prevention.

[0067] It should be noted that the state evolution rate vector is used to describe the changing trend of the core state variables of the energy storage system, providing input basis for subsequent state judgment and compression decision-making.

[0068] S2, based on the deep reinforcement learning method combined with the state level, adaptively adjusts the compression control parameters.

[0069] In this embodiment, based on the classification of the energy storage system's operating status, a state space is constructed. The state perception vector and the state evolution rate vector are used to jointly characterize the system's operating status. A reinforcement learning strategy network is used to output compression control parameters based on different state levels (stable, slightly fluctuating, deviation from normal, and abnormal warning). This makes the compression strategy highly sensitive and adaptable to changes in the operating status, solving the problem of decoupling the compression intensity from the operating status in traditional compression methods.

[0070] Based on the deep reinforcement learning method and the state level, the compression control parameters are adaptively adjusted to obtain the following parameters:

[0071] Construct a Markov decision process model based on state hierarchy;

[0072] The Markov decision process model is iteratively trained using a deep deterministic policy gradient algorithm, and interactive optimization of policy evaluation and policy improvement is achieved through a dual neural network architecture until the cumulative expected reward value of the policy function converges to a preset threshold;

[0073] The trained strategy function is deployed as the compression control strategy mapping function, and a dynamic mapping relationship between the state level and the compression parameter group is established to obtain the compression control parameters corresponding to the state level of the current energy storage system. The mapping relationship is implemented through an improved radial basis function neural network, and parameter boundary constraints are set to ensure the feasibility of the compression operation.

[0074] The Markov decision process model includes:

[0075] A state space generated by data state classification rules, where the state levels are dynamically defined based on data type, data priority, and real-time requirements;

[0076] An action space composed of compression control parameters, the compression control parameters including a compression rate adjustment coefficient, an error tolerance threshold, and a sampling interval adjustment factor, wherein the sampling interval adjustment factor controls the sampling density in both the time dimension and the space dimension;

[0077] A multi-dimensional composite reward objective function that integrates the compressed data recovery accuracy evaluation index, data storage consumption cost index, and error accumulation rate index in a linear weighted manner. The recovery accuracy evaluation index adopts a fusion calculation model of structural similarity (SSIM) and peak signal-to-noise ratio (PSNR).

[0078] The specific calculation formula of the strategy function is as follows:

[0079]

[0080] The specific calculation formula of the reward objective function is as follows:

[0081]

[0082] The specific calculation formula of the objective function of maximizing the cumulative expected reward value is as follows:

[0083]

[0084] Where, is the policy function, is the status level, For compression control action, is the reward objective function, For data recovery accuracy, For data storage consumption, is the error rate, 、 、 are the weight parameters of the reward function, To maximize the cumulative expected reward, Discount factor, is the sampling mean of the policy function during the state transition process, is the time limit for policy optimization and state trajectory evaluation, For the current moment.

[0085] It should be noted that the reward objective function is used to measure the impact of the compressed data after taking the corresponding action on the system availability, accuracy, and resource occupancy indicators. At each state level, the action output by the strategy mapping function is the precise value of the compression control parameter defined in the action space. The sampling mean of the strategy function during the state transition process is used to ensure that the compression is effective for all running trajectories and avoid overfitting a certain state. The time length boundary of the strategy optimization and state trajectory evaluation is used to control the evaluation time range of the compression strategy optimization and balance real-time performance with global performance. The compression rate is the ratio of the amount of compressed data to the amount of original data, which is used to adjust the aggressiveness of the compression algorithm. The error tolerance is the maximum deviation range allowed between the compressed data and the original data. The sampling interval adjustment factor is to dynamically adjust the sampling interval based on the original sampling period to control the frequency of data updates.

[0086] S3, adjusting the compression control strategy of the energy storage system based on the fuzzy reasoning method according to the compression control parameters and performing feedback optimization.

[0087] In this embodiment, the compression control parameters include a compression rate, an error tolerance, and a sampling interval adjustment factor.

[0088] According to the compression control parameters, the compression control strategy of the energy storage system is adjusted based on the fuzzy reasoning method and feedback optimization is performed. Specifically:

[0089] Fuzzy processing is performed on the compression control parameters, and a triangular membership function is used to map continuous input values ​​into fuzzy language values ​​to form fuzzy sets, wherein each fuzzy set is divided based on the fuzzy language value and a preset numerical interval, the fuzzy set of the compression rate includes low, medium, and high, the fuzzy set of the error tolerance includes very small, small, medium, and large, and the fuzzy set of the sampling interval adjustment factor includes low, medium, and high;

[0090] A fuzzy rule base is established according to the fuzzy set, the fuzzy language value and its membership degree are used as input items, and fuzzy reasoning is performed according to the rule base, and the fuzzy output variable is obtained by using the Mamdani reasoning mechanism to obtain a lossless priority value;

[0091] Defuzzifying the lossless priority value to obtain a sharp quantization value, and making a judgment based on a preset lossless priority threshold. If the sharp quantization value is greater than or equal to the preset lossless priority threshold, the lossless compression method is preferentially used, otherwise the lossy compression method is used;

[0092] Based on the judgment result, a compression control strategy adjustment signal is obtained, and the corresponding lossless or lossy compression strategy is selected for the data collected by the energy storage system;

[0093] Collect feedback parameters after the round of compression to construct a compression effect scoring function to obtain a compression strategy performance score value; the feedback parameters include actual compression rate, actual reconstruction error, and actual compression delay;

[0094] If the score value is lower than the preset threshold for several consecutive cycles, the score value is fed back to the fuzzy inference system to trigger the optimization of the fuzzy rule base to achieve adaptive evolution of the compression strategy decision capability.

[0095] The specific calculation formula of the compression effect scoring function is as follows:

[0096]

[0097] Where, is the compression strategy performance score, is the target compression ratio, is the actual compression ratio, is the maximum acceptable error, is the actual reconstruction error, is the maximum acceptable compression processing delay, is the actual compression delay, 、 、 is the weight coefficient.

[0098] It should be noted that .

[0099] Furthermore, the introduction of a genetic algorithm to optimize the fuzzy rule base to achieve adaptive adjustment of the compression strategy can improve the accuracy of compression strategy judgments. Through the evolution of the genetic algorithm driven by the scoring function, the fuzzy rules are gradually approached to the high-performance range, effectively improving the accuracy and matching degree of the judgment on whether to give priority to lossless compression. Faced with the complex and dynamically changing operating conditions of the energy storage system, the genetic algorithm can quickly search for the optimal rule combination and avoid the problem of artificially preset rules failing under some extreme working conditions. Using real-time compression execution feedback as the fitness function, the fuzzy rule base is continuously adjusted to construct a closed-loop evolution chain with compression feedback-driven closed-loop optimization capabilities.

[0100] The optimization of the trigger fuzzy rule base is specifically as follows:

[0101] Based on genetic algorithm, fuzzy rules are encoded as genetic individuals in the form of rule vectors. Each fuzzy rule includes the matching relationship between the fuzzy language value of the input parameter and the output strategy priority.

[0102] Using the compression effect scoring function as the fitness function of the genetic algorithm, an iterative evolution operation is performed on the individuals, the evolution operation including selecting a rule combination with a higher scoring value as a parent individual based on a roulette wheel method or an elite retention strategy, performing a crossover operation on the selected parent generation using a two-point crossover or uniform crossover strategy to exchange rule genes to generate new rule individuals, and a mutation operation for randomly perturbing the fuzzy set labels or output strategy labels in the individuals to increase population diversity;

[0103] After iterating for a preset number of evolutionary rounds, the set of fuzzy rule individuals with the largest fitness value is selected as the updated fuzzy rule base and replaces the rules in the current reasoning system to achieve adaptive evolution of the fuzzy reasoning system;

[0104] Repeated iterative evolutionary operations form a closed-loop fuzzy rule optimization process driven by compression feedback, which improves the environmental adaptability and judgment accuracy of the compression strategy determination system.

[0105] The fuzzy rule base is specifically:

[0106] If the error tolerance is extremely small and the sampling interval adjustment factor is low, the compression control strategy is to give priority to lossless compression;

[0107] If the error tolerance is small and the compression rate is low, the compression control strategy tends to adopt lossless compression;

[0108] If the error tolerance is large, the sampling interval adjustment factor is high, and the compression rate is high, the compression control strategy is to give priority to lossy compression;

[0109] If the error tolerance is medium, the sampling interval adjustment factor is medium, and the compression rate is high, the compression control strategy is acceptable lossy compression;

[0110] If the error tolerance is extremely small and the compression rate is medium or high, the compression control strategy must use lossless compression;

[0111] If the error tolerance is medium and the sampling interval adjustment factor is low, the compression control strategy is moderately biased towards lossless compression;

[0112] If the error tolerance is large, the sampling interval adjustment factor is medium or high, but the compression rate is low, the compression strategy is to use lossy compression as appropriate;

[0113] If the error tolerance is small and the sampling interval adjustment factor is high, the compression control strategy is to select a lossy compression method that takes into account both accuracy and compression rate.

[0114] It should be noted that the fuzzy rule base does not output absolute conclusions; it only provides a tendency. Therefore, a preset threshold must be used to "hard-judge" this tendency into a specific compression method. Through the adaptive evolution of the fuzzy rule base, particularly the introduction of a genetic algorithm for optimization, the system can adjust the compression control strategy in real time based on feedback from the compression effect scoring function. This enables the system to automatically optimize the compression strategy and maintain efficient performance under the dynamically changing operating conditions of the energy storage system without manual intervention. Optimizing the fuzzy rule base through a genetic algorithm effectively improves the accuracy of compression strategy determination and avoids the limitations of rule settings, especially under extreme operating conditions. The genetic algorithm can quickly search and find the optimal rule combination, enabling fine-tuning of the compression strategy and improving system stability and performance. The use of fuzzy inference mechanisms and Mamdani inference rules can handle nonlinear relationships between parameters such as compression rate, error tolerance, and sampling interval, avoiding the shortcomings of traditional control methods in handling complex inputs. The determination of lossless priority can more accurately reflect actual system requirements and improve decision accuracy. Because fuzzy control methods inherently tolerate uncertainty and noise, the system can effectively address various uncertainties in energy storage system operation, such as real-time compression errors and environmental changes, providing a more robust control strategy. By accurately selecting between lossless and lossy compression strategies, the system can select the most appropriate compression method based on actual operating conditions, optimizing the use of computing resources and storage space, avoiding over-compression or over-retention of data, and improving the overall performance of the energy storage system.

[0115] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0116] The above embodiments may be implemented in whole or in part through software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0117] Those skilled in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0118] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0119] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0120] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for collecting and compressing energy storage detailed data based on batch-stream fusion, characterized in that: include: Obtaining the operating status parameters of the energy storage system and performing state perception and hierarchical mapping of the energy storage system to obtain a state level; The state perception vector is obtained through differential processing, its evolution trend is extracted to generate the state evolution rate vector, and the state potential function is constructed based on the principal component analysis method; The state level is obtained by performing state perception and state classification mapping according to the state potential function; State potential function, specifically: Where, is the state potential function value, is the state of charge at time The evolution rate of Estimates of health status at time points The evolution rate of is the cell temperature at time point The evolution rate of 、 、 are weight coefficients respectively; A Markov decision process model based on state levels is constructed, and deep reinforcement learning methods are used for training and optimization to generate a compression control policy function. The compression control parameters are adaptively generated according to different state levels through the policy function. Perform fuzzy processing on compression control parameters, dynamically select and adjust compression strategy based on fuzzy inference rule base, and generate compression control strategy adjustment signal; Based on the compression control strategy, the signal is adjusted to perform the corresponding data compression operation, and the feedback parameters after the compression execution are collected for optimization; Constructing a compression effect scoring function based on feedback parameters to obtain a compression strategy performance score value, wherein the feedback parameters include actual compression rate, actual reconstruction error, and actual compression delay; If the score value is lower than the preset threshold for several consecutive cycles, the score value will be fed back to the fuzzy inference system to trigger the optimization of the fuzzy rule base; Compression effect scoring function, specifically: Where, is the compression strategy performance score, is the target compression ratio, is the actual compression ratio, is the maximum acceptable error, is the actual reconstruction error, is the maximum acceptable compression processing delay, is the actual compression delay, 、 、 is the weight coefficient.

2. The energy storage detailed data acquisition and compression control method based on batch-stream fusion according to claim 1 is characterized in that: The state perception vector is obtained by differential processing, and its evolution trend is extracted to generate the state evolution rate vector, which is specifically: Obtain the energy storage system operating state parameters at the current moment, construct a state vector, and perform differential processing on the state vectors at adjacent time points to obtain a state perception vector; According to the state perception vector, a sliding time window sequence is constructed based on the current moment and the preset window length. For each type of state perception vector in the window sequence, its evolution trend in the time dimension is extracted and a state evolution rate vector is formed.

3. The energy storage detailed data acquisition and compression control method based on batch-stream fusion according to claim 2 is characterized in that: The evolution trend of each state perception vector in the window sequence in the time dimension is extracted and a state evolution rate vector is formed, which is specifically: A state perception vector sequence within a sliding time window is obtained, from which a state variable sequence including at least the state of charge, health state estimation value, and battery cell temperature is extracted; time trend fitting is performed on each variable in the state variable sequence based on the least squares method to generate a corresponding fitting model; a fitting slope is extracted from the fitting model to construct a state evolution rate vector that characterizes the dynamic evolution trend of the state variables.

4. The energy storage detailed data acquisition and compression control method based on batch-stream fusion according to claim 1 is characterized in that: The state level is obtained by performing state perception and state classification mapping according to the state potential function, specifically: Obtain and sort the state potential function sequence within a preset time period; Generate a set of candidate partition points based on the sorted state potential function sequence, where each candidate partition point is obtained by the median of the adjacent potential function values; For each candidate partition point, the sequence is divided into two subsets, and the entropy value of each subset is calculated; Calculate the information gain of each candidate partition point based on the entropy value and the number of samples, and determine the optimal partition point according to the principle of maximizing the information gain; The state potential function sequence is graded using the optimal division point, and a state grade classification result of the energy storage system is output.

5. The energy storage detailed data acquisition and compression control method based on batch-stream fusion according to claim 4 is characterized in that: The steps of using the optimal dividing point to classify the state potential function sequence and outputting the state classification result of the energy storage system are as follows: (1) Obtain the state potential function sequence and iteratively determine the classification threshold set based on the principle of maximizing information gain; (2) By recursively partitioning, the information gain of the candidate partition points is calculated level by level, and the partition point with the largest information gain is selected as the classification threshold of the current level; (3) Repeat step (2) until the preset number of grading levels is reached, and a complete set of state potential function grading thresholds is generated; (4) Performing level mapping on the real-time state potential function value according to the grading threshold set, and outputting the state level of the energy storage system.

6. The energy storage detailed data acquisition and compression control method based on batch-stream fusion according to claim 1 is characterized in that: The deep reinforcement learning method is used for training optimization to generate a compression control strategy function, and the compression control parameters are adaptively generated according to different state levels through the strategy function, specifically: The Markov decision process model is iteratively trained using a deep deterministic policy gradient algorithm, and interactive optimization of policy evaluation and policy improvement is achieved through a dual neural network architecture until the cumulative discounted reward value of the policy function converges to a preset threshold; The trained strategy function is deployed as the compression control strategy mapping function, and a dynamic mapping relationship between the state level and the compression parameter group is established to obtain the compression control parameters corresponding to the state level of the current energy storage system. The mapping relationship is implemented through an improved radial basis function neural network, and parameter boundary constraints are set to ensure the feasibility of the compression operation.

7. The energy storage detailed data acquisition and compression control method based on batch-stream fusion according to claim 6 is characterized in that: The fuzzy processing of the compression control parameters is performed, and the compression strategy is dynamically selected and adjusted based on the fuzzy inference rule base to generate a compression control strategy adjustment signal, specifically: Performing fuzzification processing on the compression control parameters, mapping continuous input values ​​into fuzzy language values ​​using triangular membership functions to form fuzzy sets; A fuzzy rule base is established according to the fuzzy set, the fuzzy language value and its membership degree are used as input items, and fuzzy reasoning is performed according to the rule base, and the fuzzy output variable is obtained by using the Mamdani reasoning mechanism to obtain a lossless priority value; Defuzzifying the lossless priority value to obtain a clear quantized value, and making a determination based on a preset lossless priority threshold; According to the judgment result, a compression control strategy adjustment signal is obtained, and the corresponding lossless or lossy compression strategy selection is performed on the data collected by the energy storage system.

8. The energy storage detailed data acquisition and compression control method based on batch-stream fusion according to claim 1 is characterized in that: The optimization of the trigger fuzzy rule base is specifically as follows: Based on genetic algorithm, fuzzy rules are encoded as genetic individuals in the form of rule vectors. Each fuzzy rule includes the matching relationship between the fuzzy language value of the input parameter and the output strategy priority. Using the compression effect scoring function as the fitness function of the genetic algorithm, performing an iterative evolution operation on the individual; After iterating for a preset number of evolutionary rounds, the set of fuzzy rule individuals with the largest fitness value is selected as the updated fuzzy rule base and replaces the rules in the current reasoning system; Repeat the iterative optimization operation to form a closed-loop fuzzy rule optimization process driven by compressed feedback to optimize the fuzzy rule base.

Images