Operation data analysis method and system based on intelligent energy storage system

Abstract

The invention discloses an operation data analysis method and system based on an intelligent energy storage system, and relates to the technical field of data analysis and management, and the method comprises the steps: collecting original data, carrying out the preprocessing, extracting a feature index, and generating a confidence feature vector; based on the confidence feature vector, calculating a comprehensive confidence score to judge a credible state variable, then generating a trust factor, then performing correction based on the trust factor, obtaining a comprehensive trust factor, and performing grade division to obtain a trust result; based on the trust result, obtaining an average trust factor value of each grade, calculating an initial weight and a fuzzy membership weight of each grade, fusing the initial weight and the fuzzy membership weight to obtain a comprehensive weight, and combining the average trust factor value to generate overall trust strength for data analysis to obtain an analysis result; according to the method, the accuracy and the self-adaptability of energy storage operation data analysis can be effectively improved, and the real-time monitoring and intelligent decision-making capabilities of the system are remarkably enhanced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data analysis and management, and particularly relates to a running data analysis method and system based on an intelligent energy storage system. BACKGROUND

[0002] With the rapid development of new energy power generation and distributed energy storage systems, energy storage technology has become an important foundation for supporting energy structure transformation and intelligent dispatch of power systems. In recent years, energy storage units such as lithium batteries, super capacitors and flow batteries have been widely used in smart grids, microgrids and renewable energy grid-connected scenarios. Intelligent energy storage systems not only bear the functions of energy storage and release, but also have multi-dimensional characteristics such as running state perception, autonomous scheduling and health management. In order to realize real-time monitoring and data security analysis of the running state of the energy storage system, the current industry usually adopts multi-source sensing collection and cloud analysis technology, combines time series prediction models, anomaly detection algorithms and state estimation methods to evaluate the system running data. However, most of the traditional running data analysis is based on fixed threshold or statistical average model, which is difficult to adapt to the dynamic changes of the running state of the energy storage system. Especially in the background of multiple sensing nodes, complex operating conditions and frequent data drift, the existing methods often cannot accurately reflect the data reliability, and it is also difficult to realize dynamic trust evaluation and hierarchical judgment of the system running state, thereby reducing the accuracy and stability of the energy storage system running data analysis. SUMMARY

[0003] In view of the above existing problems, the present application is proposed.

[0004] Therefore, the present application provides a running data analysis method and system based on an intelligent energy storage system, which solves the problem that the existing method often cannot accurately reflect the data reliability, and it is also difficult to realize dynamic trust evaluation and hierarchical judgment of the system running state.

[0005] To solve the above technical problems, the present application provides the following technical solutions: In a first aspect, the present application provides a running data analysis method based on an intelligent energy storage system, which comprises, collecting raw data for preprocessing, extracting feature indicators and generating a confidence feature vector; The raw data includes voltage, current data, temperature data and power output data; Based on the confidence feature vector, the comprehensive confidence score is calculated to determine the confidence state variable, and then the trust factor is generated. Based on the trust factor, the comprehensive trust factor is obtained for hierarchical division, and the trust result is obtained; Based on the trust result, the average trust factor value of each level is obtained, and the initial weight and fuzzy membership weight of each level are calculated and fused to obtain a comprehensive weight, and the average trust factor value is combined to generate a whole trust strength for data analysis to obtain an analysis result; Based on the analysis result, the database is used for storage.

[0006] As a preferred scheme of the operation data analysis method based on the intelligent energy storage system, after the confidence feature vector is used to calculate the comprehensive confidence score for the determination of the trusted state variable, a trust factor is generated, and the trust factor is used for correction to obtain a comprehensive trust factor for level division to obtain a trust result, including the following steps: Based on the confidence feature vector, the weight coefficient of each data in the confidence feature vector is set through historical data correlation analysis, and the comprehensive confidence score of each confidence feature vector is calculated using the weighted sum formula, and the determination of the trusted state variable is performed to obtain a determination result, including trusted and untrusted; The total number of trusted times and the total number of comprehensive confidence scores in the determination result are counted, and the ratio between the total number of trusted times and the total number of comprehensive confidence scores is defined as a basic probability, and the maximum likelihood estimation is used to obtain the likelihood probability under the trusted and untrusted states; The likelihood probabilities under the trusted and untrusted states are summed, and the sum probability is taken as an evidence value, and the posterior probability is calculated based on the basic probability, the likelihood probability and the evidence value, and is defined as the trust factor; After the mean value formula is used to obtain the mean value of the standardized data in each window, the standardized deviation between the standardized data and the mean value is calculated; The detection threshold is set using historical regression analysis, and when the standardized deviation is greater than or equal to the detection threshold, it is determined that the standardized data in the current window has abnormal drift, otherwise it is determined that the standardized data in the current window is normal; Based on the abnormal drift of the standardized data, the corresponding trust factor is queried for correction in the form of exponential decay to obtain a corrected trust factor; The corrected trust factor and the trust factor are fused to generate a comprehensive trust factor; The level threshold is set according to the statistical distribution of the comprehensive trust factor And , and The level threshold is used to divide the trust level of the corrected trust factor at the moment , to obtain a trust result including levels 1, 2 and 3.

[0007] As a preferred scheme of the operation data analysis method based on the intelligent energy storage system, wherein: based on the trust result, the average trust factor value of each level is obtained, and the initial weight and fuzzy membership weight of each level are calculated for fusion to obtain the comprehensive weight, including the following steps: The average trust factor value of each level is obtained using the mean formula, and the initial weight of each level is calculated according to the average trust factor value; The measurable indicators are set using the empirical rule, including the energy balance degree, signal delay and fluctuation sensitivity, the importance ratio between the measurable indicators is obtained by using the fuzzy analytic hierarchy process combined with the measurable indicators, and the fuzzy judgment matrix is constructed for normalization to obtain the fuzzy membership weight of each indicator; The initial weight and the fuzzy membership weight are fused to obtain the comprehensive weight of the level.

[0008] As a preferred scheme of the operation data analysis method based on the intelligent energy storage system, wherein: the overall trust strength is generated based on the average trust factor value for data analysis to obtain the analysis result, including the following steps: Based on the comprehensive weight, the average trust factor value of each level is weighted and aggregated to obtain the overall trust strength ; The quantization threshold is set, and when the overall trust strength is greater than or equal to the quantization threshold, it indicates that the current intelligent energy storage system is running normally, and the data is reliable, otherwise it indicates that the current intelligent energy storage system is running abnormally, and the data is unreliable, and the mobile terminal sends a reminder information to the analysis personnel.

[0009] As a preferred scheme of the operation data analysis method based on the intelligent energy storage system, wherein: based on the analysis result, the database is used for storage, and the overall trust strength and the quantization threshold are stored through the database, and a time stamp and a unique ID are added to each data during the storage process.

[0010] As a preferred scheme of the operation data analysis method based on the intelligent energy storage system, wherein: the acquisition of the original data for pretreatment refers to installing voltage and current sensors at the battery module end, installing thermocouple temperature sensors on the surface of the battery module end, and installing power sensors on the AC side of the bidirectional converter; The voltage and current data of the battery module end are collected in real time through the voltage and current sensors, the temperature data of the battery cell on the battery module are collected in real time through the thermocouple temperature sensor, and the power output data of the battery module are collected in real time through the power sensor; All data are denoised, abnormal value screened out and standardized to obtain standardized data.

[0011] As a preferred scheme of the operation data analysis method based on the intelligent energy storage system, wherein: the step of extracting the feature indicators and generating the confidence feature vector comprises the following steps: Based on the standardized data, the sliding window technology is used for division, and the effective sampling point proportion of each window is calculated as the continuity feature of the standardized data of each sensor in the window. After the difference operation is performed on the standardized data in the current time window and the previous time window, the transient fluctuation degree is calculated to measure the change intensity of the data between adjacent time points. After the window is slid forward, the average absolute deviation of the data in the current time window and the past time window is calculated, and the consistency score is mapped, and then the mutual information between the data of the current sensor and the adjacent sensor is calculated to obtain the relative correlation indicator. The continuity feature, the transient fluctuation degree, the consistency score, and the relative correlation indicator at the time point are combined in the calculation order to generate the confidence feature vector of each sensor at the time point .

[0012] In a second aspect, the present application provides an operation data analysis system based on an intelligent energy storage system, comprising, a collection and extraction module for collecting raw data for preprocessing, extracting feature indicators, and generating a confidence feature vector; a determination and division module for calculating a comprehensive confidence score based on the confidence feature vector to determine the confidence state variable, generating a trust factor, and then modifying the trust factor based on the trust factor to obtain a comprehensive trust factor for grade division to obtain a trust result; a fusion analysis module for obtaining the average trust factor value of each grade based on the trust result, calculating the initial weight and fuzzy membership weight of each grade for fusion to obtain a comprehensive weight, combining the average trust factor value to generate the overall confidence strength for data analysis to obtain an analysis result; a storage module for storing the analysis result in a database.

[0013] In a third aspect, the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, any step of the operation data analysis method based on the intelligent energy storage system according to the first aspect of the present application is implemented.

[0014] In a fourth aspect, the present application provides a computer readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, any step of the operation data analysis method based on the intelligent energy storage system according to the first aspect of the present application is implemented.

[0015] The application has the beneficial effects that: the application realizes quantitative description of the credibility of multi-source operation data by constructing a confidence feature vector, automatically setting weights for each feature data, and calculating a comprehensive confidence score, and establishes a probability mechanism of the energy storage node in the credible and non-credible states by using Gaussian likelihood estimation and Bayesian posterior probability to calculate a trust factor, so that dynamic trust updating of the operation data can be realized, and the accuracy of abnormal detection and state identification during data analysis is significantly improved, and secondly, the accuracy and adaptability of the energy storage operation data analysis can be effectively improved by correcting the trust factor in an exponential decay manner, introducing an information entropy and a fuzzy analytic hierarchy process fusion weighting mechanism, and weighting and aggregating the average trust factor value by a level comprehensive weight, and the real-time monitoring and intelligent decision-making capability of the system is significantly enhanced. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the application, the drawings needed to be used in the embodiment description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative effort on the basis of these drawings.

[0017] Fig. 1 The flowchart of the operation data analysis method based on the intelligent energy storage system in embodiment 1.

[0018] Fig. 2 The structure diagram of the operation data analysis system based on the intelligent energy storage system in embodiment 1.

[0019] Fig. 3 The flowchart of the trust level division in embodiment 1. DETAILED DESCRIPTION

[0020] In order to make the above-mentioned purposes, features and advantages of the application more apparent and easy to understand, the specific embodiments of the application will be described in detail below with reference to the drawings of the specification.

[0021] In the following description, many specific details are set forth in order to provide a thorough understanding of the application, but the application can also be practiced without the other ways different from those described herein, and those skilled in the art can make similar generalizations without departing from the connotation of the application, therefore the application is not limited by the specific embodiments disclosed below.

[0022] Secondly, the "one embodiment" or "embodiment" referred to herein means that the specific features, structures or characteristics can be included in at least one implementation of the application. "In one embodiment" appearing in different places in the specification does not mean the same embodiment, nor is it an independent or alternative embodiment that excludes other embodiments.

[0023] Embodiment 1, refer to Figs. 1-3 For the first embodiment of the present application, the embodiment provides a method for analyzing operation data based on an intelligent energy storage system, comprising the following steps: S1, collecting raw data for pretreatment, extracting feature indicators and generating a confidence feature vector; Specifically, collecting raw data for pretreatment means installing voltage and current sensors at the end of the battery module, installing thermocouple temperature sensors on the surface of the battery module, and installing power sensors on the AC side of the bidirectional converter; Real-time collection of voltage and current data at the end of the battery module through voltage and current sensors, real-time collection of temperature data of the battery cell on the battery module through thermocouple temperature sensors, and real-time collection of power output data of the battery module through power sensors; All data are denoised, outliers are screened out and standardized to obtain standardized data.

[0024] By arranging voltage, current, temperature and power sensors at the end of the battery module, multi-dimensional data acquisition from the energy side, the heat side and the power side is realized. Compared with the traditional scheme of monitoring only voltage or current signals, this multi-source collaborative collection method can capture more implicit information about electrochemical and thermal coupling effects. This not only realizes comprehensive perception of the running state of the system, but also provides high-resolution data support for subsequent trust calculation, health assessment and intelligent scheduling. In the data collection stage, denoising, outlier screening and standardization are introduced, so that the present application significantly reduces measurement noise, improves signal-to-noise ratio (SNR) without destroying data timing characteristics, and eliminates analysis bias caused by unit difference and numerical scale inconsistency of different measurement channels, so that the subsequent confidence feature vector construction and probability modeling of the system have a unified input basis.

[0025] Further, the feature indicators are extracted and the confidence feature vector is generated, comprising the following steps: Based on the standardized data, use the sliding window technology to divide and count the effective sampling point proportion of each sensor standardized data in each window as the continuity feature; After difference operation of the standardized data in the current time window and the previous time window, the transient fluctuation degree is calculated to measure the change degree between adjacent time points, and the formula is: In the formula, represents the transient fluctuation degree of the i-th sensor at time t, represents the standardized data value of the i-th sensor at time t, represents the standardized data value of the i-th sensor at time t,​​​​ Each sensor at time Standardized data values ​​at that time Represents a very small positive number. Represents absolute value; After sliding the window forward, the average absolute deviation of the data in the current time window and the past time windows is calculated and mapped to a consistency score. Then, the mutual information between the current sensor and the adjacent sensor data is calculated to obtain the relative correlation index. The formula for calculating the average absolute deviation of data within the current time window and the past time windows is as follows: In the formula, Indicates the first Each sensor at time Consistency score at the time Indicates the length of the sliding window. Indicates the first One sensor in Standardized data values ​​from a point in time; The formula for calculating the mutual information between the current sensor data and the data from adjacent sensors is as follows: In the formula, Indicates the first Each sensor at time The relative correlation index at time, Indicates the first The sensor and the first The amount of mutual information between standardized data from individual sensors. Indicates belonging to, Indicates the first A collection of sensor domains; Time The continuity characteristics, transient fluctuations, consistency scores, and relative correlation indices of each sensor are combined in the order of calculation to generate the value of each sensor at time t. The confidence feature vector.

[0026] By statistically analyzing the proportion of effective sampling points of standardized sensor data within a sliding window, a continuity feature is obtained. This feature quantifies the temporal stability of node signals, reflecting whether the sensor experiences prolonged signal interruptions, packet loss, or communication anomalies. This not only provides a fundamental temporal integrity metric for the trust vector but also enables the system to quickly identify the reliability of the data source when signals are missing. Furthermore, through the normalized calculation of transient fluctuations, the system can capture abrupt changes in voltage, current, or SOC signals in real time, thereby achieving early anomaly detection. Compared to traditional threshold-based differential detection, this method significantly reduces the false trigger rate to sporadic noise while maintaining sensitivity. Moreover, the acquisition of consistency scores enables adaptive smooth evaluation in the time domain, allowing the identification of gradual anomalies or slow drift behavior of sensor nodes, thus improving the system's ability to detect latent degradation or early faults. Secondly, the relative correlation index enables the system to judge spatial consistency. When the mutual information between a sensor's behavior and its neighboring sensors significantly decreases, it can be immediately identified as a local anomaly source, realizing the transformation from "single-point anomaly detection" to "system-wide correlation identification." This design overcomes the limitation of existing technologies where trust assessment is confined to the time series level, achieving trust modeling in both spatiotemporal dimensions. The confidence feature vector, a multidimensional representation of the node's operational state, serves as the fundamental input for subsequent confidence assessment and trust factor generation.

[0027] S2. Based on the confidence feature vector, calculate the comprehensive confidence score to determine the trustworthy state variable, generate the trust factor, and then make corrections based on the trust factor to obtain the comprehensive trust factor for level classification, and obtain the trust result. Specifically, based on the confidence feature vector, a comprehensive confidence score is calculated to determine the trustworthiness of the state variable, a trust factor is generated, and then the trust factor is used for correction to obtain a comprehensive trust factor for level classification, thus obtaining the trust result. This includes the following steps: Based on confidence feature vectors, a weight coefficient is assigned to each data point in the confidence feature vector through correlation analysis of historical data. Then, a weighted summation formula is used to calculate the comprehensive confidence score of each confidence feature vector. Finally, the credible state variables are determined to obtain the judgment result, including credible and uncredible. The subsequent determination of the reliable state variables refers to setting a defined threshold using rules of thumb. By combining the comprehensive confidence score, the time value of each node is generated. The reliable state variable is given by the formula: In the formula, Indicates the first Each sensor at time The reliable state variable at that time, Indicates the first Each sensor at time the comprehensive confidence score at time t; The ratio between the total number of trusted times and the total number of comprehensive confidence scores is defined as the base probability, and the likelihood probabilities in the trusted and untrusted states are obtained using maximum likelihood estimation, and the formula is: In the formula, represents the likelihood probability of the comprehensive confidence score under the given trusted state condition, represents a normalization constant term, represents twice the value of pi, represents the dimension of the confidence feature vector, represents an exponential function, represents the confidence feature vector of the i th sensor at time t, represents the mean value of the i th sensor under the trusted state variable (which can be obtained by statistics), represents the covariance matrix (which is generated by calculating the difference vector between each confidence feature vector and the mean vector); The likelihood probabilities in the trusted and untrusted states are summed, and the sum probability is taken as the evidence value. Based on the base probability, the likelihood probability, and the evidence value, the posterior probability is calculated, which is defined as the trust factor, and the formula is: In the formula, represents the posterior probability of the confidence feature vector under the given trusted state condition, represents the base probability under the trusted condition, represents the evidence value of the confidence feature vector, represents the trust factor of the i th sensor at time t, represents the trust factor of the i th sensor at time t, represents the trust factor of the i th sensor at time t, After obtaining the mean value of the standardized data in each window using the mean formula, the standardized deviation between the standardized data and the mean value is calculated, and the formula is: In the formula, represents the standardized deviation of the i th sensor at time t, represents the standardized deviation of the i th sensor at time t, represents the drift amplification coefficient (which can be set through experiments, or take a constant value of 1.0 or 1.5), represents the mean value of the standardized data of the i th sensor, represents the mean value of the standardized data of the i th sensor, represents the mean value of the standardized data of the i th sensor, represents the mean value of the standardized data of the i th sensor, ​​​a standard deviation of the standardized data of the sensor; A detection threshold is set using historical regression analysis. When the standardized deviation is greater than or equal to the detection threshold, it is determined that the standardized data in the current window has abnormal drift, otherwise it is determined that the standardized data in the current window is normal. Based on the abnormal drift of the standardized data, the corresponding trust factor is queried for correction by an exponential decay method to obtain a corrected trust factor, and the formula is: In the formula, represents the corrected trust factor, represents the base of the natural logarithm, represents the decay coefficient (which can be set by relevant field knowledge, or take the reference value 0.2 or 0.5); The corrected trust factor and the trust factor are fused to generate a comprehensive trust factor, and the formula is: In the formula, represents the comprehensive trust factor at time , and represents the weight coefficient of the corrected trust factor; The weight coefficient of the corrected trust factor is calculated according to the formula: According to the statistical distribution of the comprehensive trust factor, a level threshold is set and , and The corrected trust factor at time is divided into trust levels using the level threshold to obtain a trust result, including levels 1, 2, and 3, and the formula is: In the formula, represents the trust level of the sensor at time ; Level 1 (trusted): When the comprehensive trust factor , it means that the node data is considered to be highly trusted; Level 2 (medium trust): When , it means that the data trustworthiness of the node is medium; Level 3 (untrusted): When the comprehensive trust factor , it means that the data trustworthiness of the node is low and is considered to be untrusted.

[0028] The comprehensive confidence score is a multi-dimensional confidence metric obtained by weighted summation of confidence feature vectors, used to measure the overall credibility level of the current node's data. The credibility state variable is used to make a binary judgment of the node's state as either credible or untrustworthy. (Empirical threshold) By combining historical distribution data, the judgment result is determined not only by instantaneous values ​​but also by the influence of operational trends. This step effectively avoids misjudgments caused by occasional data fluctuations, improving the stability and robustness of the judgment. The use of maximum likelihood estimation enables this invention to shift from empirical threshold judgment to probability-driven trust inference, significantly improving the system's self-learning ability and the accuracy of trust assessment. Furthermore, the trust factor is based on posterior probability, comprehensively calculating the trustworthiness of a node under given confidence conditions by combining base probability, likelihood probability, and evidence values. This step formally completes the mapping from "observed values" to "confidence inference results," dynamically characterizing the evolution trajectory of node trust, and providing a probabilistic quantitative basis for subsequent correction and classification. Secondly, standardized deviation is used to measure drift by comparing the standardized data of nodes with the mean deviation; its calculation introduces a drift amplification factor. This enables the model to amplify and detect minute shifts. By combining historical regression analysis with threshold settings, short-term fluctuations can be distinguished from long-term anomaly drifts. This mechanism enables node-level data quality diagnosis, effectively improving the foresight of anomaly detection in energy storage systems. Secondly, the original trust factor is dynamically corrected using an exponential decay function. The decay coefficient is used. The exponential decay characteristic ensures that the larger the deviation, the stronger the correction, which makes this step an automatic suppression mechanism for trust. This method can suppress the problem of excessive trust caused by data drift, and ensure the physical consistency and numerical stability of the system's trust assessment; Third, by integrating and correcting the trust values ​​before and after the trust factor fusion, this method not only achieves adaptive weighting of the two types of trust information but also ensures that the proportion of corrected trust values ​​increases with the degree of drift. This method enables continuous and differentiable adjustment of trust factor fusion, avoiding the boundary abruptness problem caused by traditional linear weighting. Fourth, set a threshold based on the statistical distribution of the comprehensive trust factor. and The trust level of sensor data is divided into three levels. The classification results not only reflect the reliability level of the data, but also provide interpretable decision-making basis for analysis and early warning, energy scheduling, and health assessment.

[0029] S3. Based on the trust results, obtain the average trust factor value of each level, and calculate the initial weight and fuzzy membership weight of each level and fuse them to obtain the comprehensive weight. Combine the average trust factor value to generate the overall trust strength for data analysis and obtain the analysis results. Specifically, based on the trust results, the average trust factor value for each level is obtained, and the initial weight and fuzzy membership weight for each level are calculated and fused to obtain the comprehensive weight, including the following steps: The average trust factor value for each level is obtained using the mean formula. The initial weight for each level is then calculated based on the average trust factor value, using the following formula: In the formula, Indicates the first The initial weights of each level, Indicates the total number of grades. Indicates the first Average trust factor value for each level Represents the natural logarithm function. Indicates the first Average trust factor value for each level; Measurable indicators are set using rules of thumb, including energy balance, signal delay and fluctuation sensitivity. Fuzzy hierarchical analysis is used in combination with the measurable indicators to obtain the importance ratio between the measurable indicators. A fuzzy judgment matrix is ​​constructed and normalized to obtain the fuzzy membership weight of each indicator. The initial weights and fuzzy membership weights are combined to obtain the comprehensive weight of the level, as shown in the formula: In the formula, Indicates the first The overall weight of each level Indicates the first Fuzzy membership weights at each level, Indicates the first The initial weights of each level, Indicates the first Fuzzy membership weights at each level.

[0030] This method utilizes information entropy theory to map the uncertainty of trust distribution among sensor data into quantified weights. It avoids the subjectivity of traditional manual weight setting, making the weight allocation process entirely driven by system operational data. This allows the system to automatically increase the weight of a certain level of trust factor when its fluctuation is small and its credibility distribution is concentrated, thereby achieving adaptive analysis based on confidence distribution. This ensures that the energy storage system maintains the scientific and rational nature of data evaluation at different operational stages. Furthermore, by establishing a fuzzy judgment matrix of three-dimensional indicators—energy balance, signal delay, and fluctuation sensitivity—and calculating and normalizing the importance ratios of each indicator, fuzzy membership weights are obtained. This step enables the model to dynamically balance the influence weights of various performance indicators when facing multiple factors, avoiding the dominance of a single indicator in the judgment. This significantly improves the adaptability and anti-interference performance of the energy storage system under complex operating conditions, allowing operational analysis to not only rely on trust data but also comprehensively consider system performance fluctuation characteristics. Moreover, the fusion process of comprehensive weights combines statistical characteristics with expert experience characteristics, forming a multi-dimensional decision-making model that conforms to mathematical laws and possesses system perception capabilities. The comprehensive weighting accurately reflects the actual operational risks of the system at different trust levels, making the subsequent calculation of overall trust strength more discriminative and interpretable. Secondly, the calculation results of the comprehensive weighting directly participate in the generation of the subsequent overall trust strength. This transforms multi-dimensional, hierarchical trust information into a single quantitative indicator, thereby enabling health assessment and real-time decision support for intelligent energy storage systems.

[0031] Furthermore, by combining the average trust factor value, an overall trust strength is generated for data analysis, yielding analysis results including the following steps: Based on a comprehensive weighting, the average trust factor values ​​of each level are weighted and aggregated to obtain the overall trust strength. The formula is: A quantitative threshold was set based on experiments and knowledge from books in related fields, to determine the overall credibility strength. When the value is greater than or equal to the quantification threshold, it indicates that the current smart energy storage system is operating normally and the data is reliable; otherwise, it indicates that the current smart energy storage system is operating abnormally and the data is unreliable, and a reminder message will be sent to the analysts via mobile terminal.

[0032] By comprehensively weighting and aggregating the average trust factor values ​​at each level, a hierarchical fusion of trust results is achieved. Compared to the traditional averaging method, this method integrates the statistical characteristics of trust distribution with the fuzzy weights of operational performance indicators, realizing a synergistic expression of "data credibility" and "system state characteristics." This not only evolves the trust assessment of energy storage systems from a single numerical value to a multi-dimensional fusion result, but also enhances the algorithm's adaptability and applicability to multiple scenarios. Furthermore, this weighted result can serve as a core reference parameter for system scheduling, fault early warning, and health assessment, and is applicable to complex energy management platforms such as microgrids, distributed energy storage, and virtual power plants. Moreover, by comparing the overall credibility strength with a quantification threshold, system-level state identification can be achieved. When the overall credibility strength... When the overall reliability is greater than or equal to the quantization threshold, the system is considered to be operating normally and the data is reliable. When the value falls below the quantified threshold, an anomaly detection is triggered, and an alert is sent to maintenance personnel via mobile terminal. This step achieves a closed-loop transformation from "data analysis" to "operational decision-making," enhancing the system's autonomous diagnosis and intelligent operation and maintenance capabilities while avoiding the lag and subjectivity of manual monitoring.

[0033] S4. Based on the analysis results, store them in a database; Specifically, based on the analysis results, using a database for storage means storing the overall trust strength and quantification threshold in a database, and adding a timestamp and unique ID to each data point during the storage process.

[0034] By storing the overall trust strength and quantification threshold in a database, persistent and traceable management of trust analysis results can be achieved. Adding a timestamp to each record during storage enables temporal correlation and dynamic analysis of the data. Furthermore, a unique ID mechanism ensures global uniqueness and indexability of the data.

[0035] This embodiment also provides an operation data analysis system based on an intelligent energy storage system, including: The data acquisition and extraction module is used to collect raw data, preprocess it, extract feature indicators, and generate confidence feature vectors. The judgment and classification module is used to calculate the comprehensive confidence score based on the confidence feature vector to determine the trustworthy state variable, generate the trust factor, and then make corrections based on the trust factor to obtain the comprehensive trust factor for level classification and obtain the trust result. The fusion analysis module is used to obtain the average trust factor value of each level based on the trust results, calculate the initial weight and fuzzy membership weight of each level and fuse them to obtain the comprehensive weight. Combined with the average trust factor value, the overall trust strength is generated for data analysis to obtain the analysis results. The storage module is used to store the analysis results using a database.

[0036] This embodiment also provides a computer device applicable to the operation data analysis method based on an intelligent energy storage system, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the operation data analysis method based on an intelligent energy storage system as proposed in the above embodiment. The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0037] This embodiment also provides a storage medium storing a computer program. When executed by a processor, the program implements the operation data analysis method for an intelligent energy storage system as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0038] 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 method for analyzing operational data based on an intelligent energy storage system, characterized in that: include, Collect raw data, preprocess it, extract feature indicators, and generate confidence feature vectors; The raw data includes voltage, current, temperature, and power output data; Based on the confidence feature vector, a comprehensive confidence score is calculated to determine the trustworthy state variable, a trust factor is generated, and then the trust factor is used for correction to obtain a comprehensive trust factor for level classification, thus obtaining the trust result. Based on the trust results, the average trust factor value of each level is obtained, and the initial weight and fuzzy membership weight of each level are calculated and fused to obtain the comprehensive weight. Combined with the average trust factor value, the overall trust strength is generated for data analysis to obtain the analysis results. Based on the analysis results, a database is used for storage.

2. The method for analyzing operational data based on an intelligent energy storage system as described in claim 1, characterized in that: The process of calculating a comprehensive confidence score based on the confidence feature vector to determine the trustworthiness of the state variable, generating a trust factor, and then refining the result based on the trust factor to obtain a comprehensive trust factor for level classification, thereby obtaining the trust result, includes the following steps: Based on confidence feature vectors, a weight coefficient is assigned to each data point in the confidence feature vector through correlation analysis of historical data. Then, a weighted summation formula is used to calculate the comprehensive confidence score of each confidence feature vector. Finally, the credible state variables are determined to obtain the judgment result, including credible and uncredible. The ratio between the total number of credible occurrences and the total number of comprehensive confidence scores in the statistical judgment results is defined as the basic probability, and the maximum likelihood estimation is used to obtain the likelihood probability under the credible and uncredible states. The likelihood probabilities under credible and uncredible states are summed, and the summed probability is used as the evidence value. Based on the base probability, likelihood probability, and evidence value, the posterior probability is calculated and defined as the trust factor. After obtaining the mean of the standardized data within each window using the mean formula, calculate the standardized deviation between the standardized data and the mean. Historical regression analysis is used to set a detection threshold. When the standardization deviation is greater than or equal to the detection threshold, it is determined that there is abnormal drift in the standardized data in the current window; otherwise, the standardized data in the current window is considered normal. Backtracking is performed on standardized data with abnormal drift, and the corresponding trust factor is queried to correct the exponential decay method, thus obtaining the corrected trust factor. The modified trust factor is merged with the trust factor to generate a comprehensive trust factor; The level threshold is set based on the statistical distribution of the comprehensive trust factor. and ,and Use level thresholds for time The corrected trust factor is used to classify trust levels, resulting in trust results, including levels 1, 2, and 3.

3. The method for analyzing operational data based on an intelligent energy storage system as described in claim 2, characterized in that: The process of obtaining the average trust factor value for each level based on the trust results, and then calculating and fusing the initial weight and fuzzy membership weight for each level to obtain the comprehensive weight, includes the following steps: The average trust factor value for each level is obtained using the mean formula, and the initial weight for each level is calculated based on the average trust factor value. Measurable indicators are set using rules of thumb, including energy balance, signal delay and fluctuation sensitivity. Fuzzy hierarchical analysis is used in combination with the measurable indicators to obtain the importance ratio between the measurable indicators. A fuzzy judgment matrix is ​​constructed and normalized to obtain the fuzzy membership weight of each indicator. The initial weights and fuzzy membership weights are combined to obtain the comprehensive weights of the levels.

4. The method for analyzing operational data based on an intelligent energy storage system as described in claim 3, characterized in that: The process of combining the average trust factor value to generate overall trust strength for data analysis, and obtaining analysis results, includes the following steps: Based on a comprehensive weighting, the average trust factor values ​​of each level are weighted and aggregated to obtain the overall trust strength. ; Set a quantization threshold when the overall credibility strength When the value is greater than or equal to the quantification threshold, it indicates that the current smart energy storage system is operating normally and the data is reliable; otherwise, it indicates that the current smart energy storage system is operating abnormally and the data is unreliable, and a reminder message will be sent to the analysts via mobile terminal.

5. The method for analyzing operational data based on an intelligent energy storage system as described in claim 4, characterized in that: The phrase "based on the analysis results, use a database for storage" refers to storing the overall trust strength and quantification threshold in a database, and adding a timestamp and unique ID to each data point during the storage process.

6. The method for analyzing operational data based on an intelligent energy storage system as described in claim 5, characterized in that: The preprocessing of the collected raw data refers to installing voltage and current sensors at the battery module end, installing thermocouple temperature sensors on the surface of the battery module end, and installing power sensors on the AC side of the bidirectional converter. Voltage and current data of the battery module are collected in real time through voltage and current sensors, temperature data of the battery cells on the battery module are collected in real time through thermocouple temperature sensors, and power output data of the battery module is collected in real time through power sensors. All data are denoised, outlier removed, and standardized to obtain standardized data.

7. The method for analyzing operational data based on an intelligent energy storage system as described in claim 6, characterized in that: The extraction of feature indicators and generation of confidence feature vectors includes the following steps: Based on standardized data, a sliding window technique is used to divide the data, and the proportion of effective sampling points of standardized data from each sensor within each window is used as a continuity feature. After performing a difference calculation on the standardized data within the current time window and the previous time window, the transient volatility is calculated to measure the degree of drastic change in data between adjacent time points. After sliding the window forward, the average absolute deviation of the data in the current time window and the past time windows is calculated and mapped to a consistency score. Then, the mutual information between the current sensor and the adjacent sensor data is calculated to obtain the relative correlation index. Time The continuity characteristics, transient fluctuations, consistency scores, and relative correlation indices of each sensor are combined in the order of calculation to generate the value of each sensor at time t. The confidence feature vector.

8. A system for analyzing operational data of an intelligent energy storage system, based on the method for analyzing operational data of an intelligent energy storage system as described in any one of claims 1 to 7, characterized in that: include, The data acquisition and extraction module is used to collect raw data, preprocess it, extract feature indicators, and generate confidence feature vectors. The judgment and classification module is used to calculate the comprehensive confidence score based on the confidence feature vector to determine the trustworthy state variable, generate the trust factor, and then make corrections based on the trust factor to obtain the comprehensive trust factor for level classification and obtain the trust result. The fusion analysis module is used to obtain the average trust factor value of each level based on the trust results, calculate the initial weight and fuzzy membership weight of each level and fuse them to obtain the comprehensive weight. Combined with the average trust factor value, the overall trust strength is generated for data analysis to obtain the analysis results. The storage module is used to store the analysis results using a database.

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 operation data analysis method based on the intelligent 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 operation data analysis method based on the intelligent energy storage system as described in any one of claims 1 to 7.

Images