Safety detection method, system and storage medium for energy storage system

Abstract

This application relates to the field of data processing technology, and discloses a safety detection method, system, and storage medium for energy storage systems. The method includes: collecting voltage, current, temperature, and internal resistance data of the energy storage system to form a basic monitoring dataset; generating a preprocessed data sequence through data preprocessing; mapping the preprocessed data sequence to a multi-dimensional feature space to obtain a target feature set; establishing a safety detection model using the target feature set; generating a risk assessment report based on the safety detection model; and analyzing system performance and outputting a maintenance plan based on the assessment report. This application addresses the technical problem of intelligent analysis and accurate fault diagnosis of multi-parameter data in energy storage systems. By establishing a multi-level analysis framework including feature extraction, intelligent diagnosis, and fault early warning functions, and combining deep learning algorithms to comprehensively evaluate the system status, intelligent monitoring and early warning of the safety status of energy storage systems are achieved.

Description

Technical Field

[0001] This application relates to the field of data processing, and in particular to a safety detection method, system and storage medium for an energy storage system. Background Technology

[0002] With the rapid development of the new energy industry, energy storage systems are increasingly widely used in areas such as power peak shaving, renewable energy grid connection, and distributed generation. As a key device for energy storage and conversion, the safety and reliability of energy storage systems directly affect the stable operation of the entire energy system. Currently, the safety inspection of energy storage systems mainly relies on regular inspections and offline testing, with methods including electrical parameter measurement, thermal imaging scanning, and insulation performance checks. Meanwhile, some advanced online monitoring systems are gradually being put into use. These systems can collect real-time operating data from energy storage devices and perform basic status monitoring and alarms.

[0003] However, existing safety inspection technologies for energy storage systems have several shortcomings. First, traditional manual inspection methods suffer from long inspection cycles, low efficiency, and a high risk of missed detections, making it difficult to promptly identify potential safety hazards in the system. Second, while online monitoring systems can collect data in real time, their ability to identify fault characteristics under complex operating conditions is limited due to a lack of effective data analysis and processing methods. Furthermore, existing detection methods often analyze individual parameters in isolation, failing to effectively handle the complex relationships between parameters in energy storage systems, especially when faced with multi-parameter coupling and nonlinear variations, making it difficult to accurately determine the actual safety status of the system. In addition, existing technologies are weak in fault early warning and trend prediction, failing to provide effective decision support for preventative maintenance of the system. Summary of the Invention

[0004] This application provides a safety detection method, system, and storage medium for energy storage systems, addressing the technical problem of intelligent analysis and accurate fault diagnosis of multi-parameter data in energy storage systems. By establishing a multi-level analysis framework including feature extraction, intelligent diagnosis, and fault early warning functions, and combining deep learning algorithms to comprehensively evaluate the system status, intelligent monitoring and early warning of the safety status of energy storage systems are achieved.

[0005] Firstly, this application provides a safety detection method for an energy storage system. The method includes: monitoring and analyzing the operating parameters of the energy storage system; establishing a monitoring parameter system based on the multi-parameter, nonlinear, and time-varying characteristics of the energy storage system; collecting voltage, current, temperature, and internal resistance data for the monitoring parameter system to obtain a basic monitoring dataset; establishing a data preprocessing model based on the basic monitoring dataset; performing data cleaning and standardization on the basic monitoring dataset; extracting time series features and statistical features using feature engineering methods to generate a preprocessed data sequence; and constructing a feature space model based on the preprocessed data sequence. Parameters are mapped to a multi-dimensional feature space according to their type. Key feature indicators are selected through a feature evaluation mechanism to form a target feature set. An intelligent diagnostic model is established based on the target feature set. Correlation analysis is performed on the electrical, thermal, and state parameters in the target feature set. The model parameters are optimized through cross-validation to obtain a safety detection model. Fault diagnosis rules are set according to the safety detection model. Real-time operating data of the energy storage system is input into the safety detection model and compared with preset safety thresholds to generate a fault risk assessment report. A health assessment system is established based on the fault risk assessment report. Trend analysis of system performance is performed in conjunction with historical operating data to output maintenance decision-making schemes.

[0006] Secondly, this application provides a safety detection system for an energy storage system, the safety detection system for the energy storage system comprising:

[0007] The data acquisition module is used to monitor and analyze the operating parameters of the energy storage system. Based on the multi-parameter, nonlinear and time-varying characteristics of the energy storage system, a monitoring parameter system is established. Voltage, current, temperature and internal resistance data are collected for the monitoring parameter system to obtain the basic monitoring dataset.

[0008] The extraction module is used to establish a data preprocessing model based on the basic monitoring dataset, perform data cleaning and standardization on the basic monitoring dataset, extract time series features and statistical features through feature engineering methods, and generate preprocessed data sequences.

[0009] The mapping module is used to construct a feature space model based on the preprocessed data sequence, map the preprocessed data sequence to a multidimensional feature space according to parameter type, and filter key feature indicators through a feature evaluation mechanism to form a target feature set.

[0010] The verification module is used to establish an intelligent diagnostic model based on the target feature set, perform correlation analysis on the electrical parameters, thermal parameters and state parameters in the target feature set, optimize the model parameters through cross-validation, and obtain a safety detection model.

[0011] The input module is used to set fault diagnosis rules according to the safety detection model, input the real-time operation data of the energy storage system into the safety detection model, compare it with the preset safety threshold, and generate a fault risk assessment report.

[0012] The analysis module is used to establish a health assessment system based on the fault risk assessment report, perform trend analysis on system performance in combination with historical operating data, and output maintenance decision-making schemes.

[0013] A third aspect of this application provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the aforementioned safety detection method for an energy storage system.

[0014] The technical solution provided in this application achieves comprehensive monitoring and intelligent diagnosis of the safety status of energy storage systems by establishing a data acquisition, processing, analysis, and decision-making chain. In the data acquisition stage, a monitoring parameter system is established based on the multi-parameter, nonlinear, and time-varying characteristics of energy storage systems, enabling accurate acquisition of key parameters such as voltage, current, temperature, and internal resistance, thus solving the problems of incomplete parameter acquisition and insufficient sampling accuracy in traditional methods. In the data preprocessing stage, a data preprocessing model is established to clean and standardize the basic monitoring dataset, and time-series and statistical features are extracted using feature engineering methods, significantly improving data quality and usability. Regarding feature space construction, a parameter type mapping and feature evaluation mechanism is adopted to achieve data dimensionality reduction and key feature extraction, providing high-quality feature input for subsequent analysis. This invention employs a deep learning network structure including a feature input layer, parameter mapping layer, feature extraction layer, fault diagnosis layer, and output layer. This artificial intelligence model design, specifically optimized for the characteristics of energy storage systems, enables accurate diagnosis of energy storage systems under complex operating conditions. The design of each layer of the model fully considers the professional characteristics of energy storage systems. For example, the feature extraction layer uses a parallel neural network structure to extract time-series features, statistical features, and correlation features respectively. The fault diagnosis layer designs a specific loss function based on the professional knowledge of energy storage systems, significantly improving the accuracy and reliability of fault identification. By setting dynamic fault diagnosis rules and combining real-time operating data with preset safety thresholds, a real-time risk assessment mechanism is established. Finally, by establishing a health assessment system and combining historical operating data to conduct trend analysis of system performance, closed-loop management from fault diagnosis to maintenance decision-making is achieved, providing reliable technical support for the safe operation and preventive maintenance of energy storage systems. The overall solution effectively solves the technical challenges of data diversity, fault complexity, and real-time diagnosis in the safety detection of energy storage systems through multi-level data processing and multi-level analysis architecture, realizing the intelligent and automated safety detection of energy storage systems. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of one embodiment of the safety detection method for an energy storage system in this application.

[0017] Figure 2 A schematic diagram of the network structure of the security detection model in this application embodiment.

[0018] Figure 3 This is a schematic diagram of one embodiment of the safety detection system for the energy storage system in this application. Detailed Implementation

[0019] This application provides a safety detection method, system, and storage medium for an energy storage system. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0020] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 One embodiment of the safety detection method for energy storage systems in this application includes:

[0021] Step S101: Monitor and analyze the operating parameters of the energy storage system. Based on the multi-parameter, nonlinear and time-varying nature of the energy storage system, establish a monitoring parameter system. Collect voltage, current, temperature and internal resistance data for the monitoring parameter system to obtain the basic monitoring dataset.

[0022] Step S102: Establish a data preprocessing model based on the basic monitoring dataset, perform data cleaning and standardization on the basic monitoring dataset, extract time series features and statistical features through feature engineering methods, and generate preprocessed data sequences.

[0023] Step S103: Construct a feature space model based on the preprocessed data sequence, map the preprocessed data sequence to a multidimensional feature space according to parameter type, and screen key feature indicators through a feature evaluation mechanism to form a target feature set;

[0024] Step S104: Establish an intelligent diagnostic model based on the target feature set, perform correlation analysis on the electrical parameters, thermal parameters and state parameters in the target feature set, optimize the model parameters through cross-validation, and obtain a safety detection model;

[0025] Step S105: Set fault diagnosis rules according to the safety detection model, input the real-time operation data of the energy storage system into the safety detection model, compare it with the preset safety threshold, and generate a fault risk assessment report.

[0026] Step S106: Establish a health assessment system based on the fault risk assessment report, conduct trend analysis on system performance in conjunction with historical operating data, and output maintenance decision-making plans.

[0027] It is understood that the executing entity of this application can be a safety monitoring system for an energy storage system, or it can be a terminal or a server; no specific limitation is made here. This application's embodiment uses a server as an example for illustration.

[0028] Specifically, when collecting operating parameters of an energy storage system through a sensor network, multi-dimensional data, including voltage, current, temperature, and internal resistance, are collected to address the system's multi-parameter characteristics. Given the nonlinear characteristics of the energy storage system, the collected data must consider the coupling relationships between parameters, such as the mutual influence between voltage and temperature. For time-varying characteristics, the sampling frequency needs to be dynamically adjusted according to the parameter change rate; voltage and current sampling frequencies are typically in the millisecond range, while temperature sampling frequencies are in the second range. Phase calibration is particularly important during data acquisition; precise control of the sampling timing ensures the time alignment of data for different parameters. The collected raw data undergoes preliminary processing through data aggregation nodes to form a basic monitoring dataset.

[0029] In the data preprocessing stage, the basic monitoring dataset is cleaned to remove outliers caused by sensor malfunctions or communication interference. For spikes in voltage data, they are identified and corrected by calculating the rate of change of adjacent data points. Missing values ​​in temperature data are supplemented through interpolation of adjacent measurement points. Standardization processes unify parameters of different dimensions to a comparable scale; voltage data is typically normalized to the [0,1] interval, and temperature data is converted into relative changes. In feature engineering, when extracting time-series features, the rate of change, fluctuation amplitude, and periodicity of parameters are calculated. Statistical features include descriptive indicators such as mean, standard deviation, and skewness, generating a preprocessed data sequence containing multidimensional features. The feature space construction stage focuses on solving data dimensionality reduction and feature selection problems. When mapping the preprocessed data sequence to a multidimensional feature space, the physical relationships between parameters are considered, such as voltage-temperature coupling features and current-internal resistance correlation features. Feature evaluation uses an information gain-based screening method to calculate the contribution of each feature to the system state judgment. By setting contribution thresholds, a target feature set is formed. The construction of the intelligent diagnostic model is based on this target feature set. The model comprises a network structure with a feature input layer, a parameter mapping layer, a feature extraction layer, a fault diagnosis layer, and an output layer. The feature input layer receives electrical, thermal, and state parameters, while the parameter mapping layer performs spatial transformation and scale adjustment. The feature extraction layer contains parallel processing units that extract temporal, statistical, and correlation features. The fault diagnosis layer performs pattern recognition based on the extracted features, and the output layer generates evaluation results. Cross-validation is used to optimize model parameters and improve diagnostic accuracy.

[0030] The development of fault diagnosis rules is based on a safety detection model. Real-time operational data from the energy storage system is input into the model, and after feature extraction and pattern recognition, it is compared with preset safety thresholds. Threshold settings consider the physical limits and historical statistical characteristics of parameters, such as voltage over-limit thresholds and temperature alarm thresholds. The comparison results generate a fault risk assessment report, including information such as fault type, risk level, and development trend. The establishment of a health assessment system integrates fault risk assessment results and historical operational data. By analyzing fault types, development trends, and impact levels, a fault-performance mapping relationship is established. Based on performance degradation patterns in historical data, lifespan prediction and maintenance planning are performed. System performance trend analysis, based on the rate of change and degradation trajectory of performance indicators, outputs maintenance decision plans that include maintenance project priorities, implementation timing, and resource requirements.

[0031] For example, during operation, voltage sensors detected fluctuations in the voltage of individual cells in the battery pack of an energy storage power station. Data preprocessing revealed that these fluctuations exhibited periodicity and were correlated with temperature changes. Feature space analysis showed that the time lag between voltage fluctuations and temperature changes was approximately 10 minutes, indicating that temperature changes were a potential cause of the voltage fluctuations. Based on this characteristic, the diagnostic model identified an anomaly in the temperature control system. The risk assessment report indicated that this anomaly could lead to accelerated battery performance degradation. Health assessment, based on historical data analysis, showed that similar faults, if not addressed promptly, would result in a 15% reduction in battery capacity within 2-3 months. The maintenance decision plan recommended overhauling the temperature control system within one week and increasing the frequency of temperature monitoring.

[0032] In this embodiment, a comprehensive monitoring and intelligent diagnosis of the safety status of the energy storage system is achieved by establishing a data acquisition, processing, analysis, and decision-making chain. In the data acquisition stage, a monitoring parameter system is established based on the multi-parameter, nonlinear, and time-varying characteristics of the energy storage system, enabling accurate acquisition of key parameters such as voltage, current, temperature, and internal resistance, thus solving the problems of incomplete parameter acquisition and insufficient sampling accuracy in traditional methods. In the data preprocessing stage, a data preprocessing model is established to clean and standardize the basic monitoring dataset, and feature engineering methods are combined to extract time-series and statistical features, significantly improving data quality and usability. Regarding feature space construction, a parameter type mapping and feature evaluation mechanism is adopted to achieve data dimensionality reduction and key feature extraction, providing high-quality feature input for subsequent analysis. This invention employs a deep learning network structure including a feature input layer, parameter mapping layer, feature extraction layer, fault diagnosis layer, and output layer. This artificial intelligence model design, specifically optimized for the characteristics of energy storage systems, enables accurate diagnosis of energy storage systems under complex operating conditions. The design of each layer of the model fully considers the professional characteristics of energy storage systems. For example, the feature extraction layer uses a parallel neural network structure to extract time-series features, statistical features, and correlation features respectively. The fault diagnosis layer designs a specific loss function based on the professional knowledge of energy storage systems, significantly improving the accuracy and reliability of fault identification. By setting dynamic fault diagnosis rules and combining real-time operating data with preset safety thresholds, a real-time risk assessment mechanism is established. Finally, by establishing a health assessment system and combining historical operating data to conduct trend analysis of system performance, closed-loop management from fault diagnosis to maintenance decision-making is achieved, providing reliable technical support for the safe operation and preventive maintenance of energy storage systems. The overall solution effectively solves the technical challenges of data diversity, fault complexity, and real-time diagnosis in the safety detection of energy storage systems through multi-level data processing and multi-level analysis architecture, realizing the intelligent and automated safety detection of energy storage systems.

[0033] In one specific embodiment, the process of performing step S101 may specifically include the following steps:

[0034] (1) Obtain the basic operating data of the energy storage system, perform cross-analysis on the voltage and current data in the basic operating data, and output electrical characteristic parameters;

[0035] (2) Correlation analysis is performed between electrical characteristic parameters and temperature data, and the electrical characteristic parameters are corrected according to the changing trend of temperature data to form a temperature correction coefficient;

[0036] (3) Data fusion is performed on temperature correction coefficient and internal resistance data, and the degree of mutual influence between parameters is determined through data correlation analysis to generate parameter correlation matrix;

[0037] (4) Classify the parameter correlation matrix according to the multi-parameter, nonlinear and time-varying nature of the energy storage system, perform time series analysis on the classified data, and establish a monitoring parameter system.

[0038] (5) Set data acquisition rules according to the monitoring parameter system, and sample voltage, current, temperature and internal resistance data in stages according to the acquisition rules to generate a graded data sequence;

[0039] (6) Perform data verification and outlier processing on the hierarchical data sequence, and correct the abnormal data points through data compensation methods to obtain the basic monitoring dataset.

[0040] Specifically, starting with the acquisition of basic operational data, the focus is on the cross-analysis of voltage and current data by collecting basic operational data from the energy storage system. This cross-analysis, through resistivity calculation, power factor analysis, and harmonic content assessment, forms a set of electrical characteristic parameters reflecting the operating status of the energy storage system. These parameters include electrical impedance characteristics, power characteristics, and power quality indicators.

[0041] After obtaining the electrical characteristic parameters, the impact of temperature on the energy storage system performance needs to be considered. The correlation analysis between temperature and electrical characteristic parameters uses the following calculation formula:

[0042]

[0043] Among them, M tc P represents the temperature correction factor. i ΔT represents the value of the i-th electrical characteristic parameter. i α represents the temperature deviation at the corresponding parameter point. i β represents the temperature sensitivity coefficient. i γ represents the temperature decay factor, γ represents the temperature change rate weight, and ω represents the temperature dynamic response coefficient. This represents the rate of temperature change, and n represents the number of electrical characteristic parameters.

[0044] The fusion of temperature correction coefficient and internal resistance data employs a weighted average method to calculate the degree of mutual influence between parameters. The temperature correction coefficient is applied to correct the internal resistance data, and then correlation analysis is used to establish the influence relationships between parameters, generating a parameter correlation matrix. Each element in the parameter correlation matrix represents the intensity of influence between different parameters. The data classification of the parameter correlation matrix follows the three basic characteristics of energy storage systems: multi-parameter nature, nonlinearity, and time-varying nature. Multi-parameter nature is reflected in the grouping of different types of parameters; nonlinearity is reflected through the mapping of nonlinear relationships between parameters; and time-varying nature is reflected through the calculation of parameter change rates. The classified data undergoes time-series analysis, including trend analysis, periodic analysis, and abrupt change detection, thereby establishing a monitoring parameter system.

[0045] Based on the data acquisition rules set by the monitoring parameter system, a hierarchical sampling strategy is adopted. High-frequency sampling is used for rapidly changing parameters such as voltage and current, while low-frequency sampling is used for relatively slowly changing parameters such as temperature and internal resistance. The selection of sampling frequency matches the time-varying characteristics of the parameters, forming a multi-level sampling sequence. The acquired hierarchical data sequences are validated, including numerical range checks, rate of change checks, and data continuity verification. For detected abnormal data points, data compensation methods are used for correction. Data compensation methods include linear interpolation, spline interpolation, or data correction based on physical models, resulting in a reliable basic monitoring dataset.

[0046] Taking a battery bank in an energy storage power station as an example, voltage and current data of the battery bank are collected during data processing. Cross-analysis reveals that the phase difference between voltage and current changes with temperature. This phenomenon is input into a temperature correction model to calculate the influence coefficient of temperature on electrical parameters. Combined with battery internal resistance data, a correlation model between parameters is established. Data classification shows that the impact of temperature changes on battery performance has a significant lag effect. Based on this, a differentiated data sampling strategy is developed, increasing the sampling frequency during periods of significant temperature change. Outlier detection identifies voltage fluctuations caused by drastic temperature changes, and after data compensation, more accurate basic data for performance evaluation is obtained.

[0047] In one specific embodiment, the process of performing step S102 may specifically include the following steps:

[0048] (1) Construct a data quality assessment system for the basic monitoring dataset, perform data phase analysis on voltage, current, temperature and internal resistance data, identify phase shifts in the data acquisition process, and form a phase correction sequence;

[0049] (2) The phase correction sequence is calibrated with the basic monitoring dataset, and the impact of data acquisition delay is eliminated by the phase compensation method, and the time series calibration data is output.

[0050] (3) Perform wavelet decomposition on the noise components in the time-series calibration data, extract data features of different frequency bands, separate high-frequency noise from effective signals, and generate signal decomposition results;

[0051] (4) Reconstruct the data based on the signal decomposition results, and combine the effective signals of different frequency bands according to their energy contribution to construct the signal reconstruction data;

[0052] (5) The signal reconstruction data is segmented according to the working state of the energy storage system, and the characteristic parameters of each working state are extracted, including charging characteristics, discharging characteristics and static characteristics, to generate a state characteristic sequence.

[0053] (6) Perform data fusion processing on the state feature sequence, combine time series features with state features, extract the dynamic relationship between data through feature correlation analysis, and generate a preprocessed data sequence.

[0054] Specifically, data quality assessment is conducted on the basic monitoring dataset, with a focus on phase shift issues during data acquisition. Voltage, current, temperature, and internal resistance data exhibit phase differences during acquisition due to factors such as sensor response time and signal transmission delay. Data phase analysis employs correlation analysis to calculate the time delay relationship between different parameters, determines the phase shift by comparing parameter time series, and generates a phase correction sequence containing phase deviation information for each parameter.

[0055] The data calibration process for the phase correction sequence employs phase compensation technology to align the data in time based on the physical correlation between parameters. For rapidly changing voltage and current data, the compensation method dynamically adjusts based on the rate of parameter change; for slowly changing temperature and internal resistance data, linear interpolation is used for time synchronization. Phase compensation eliminates the impact of data acquisition delay, forming time-series calibration data. Time-series calibration data often contains noise components of various frequencies, requiring signal processing using wavelet decomposition. Wavelet decomposition breaks down the original signal into sub-signals of different frequency bands. High-frequency components mainly contain measurement noise and interference signals, while low-frequency components contain the basic trend of parameter changes. By setting frequency thresholds, high-frequency noise and valid signals are distinguished, generating multi-level signal decomposition results.

[0056] During data reconstruction, an energy contribution analysis method is employed for signals in different frequency bands. The energy proportion of each frequency band is calculated; high-energy regions typically correspond to the main changes in parameters, while low-energy regions reflect secondary changes. Weighting coefficients are set based on energy contribution, and effective signals from different frequency bands are weighted and combined to construct reconstructed signal data reflecting the true changes in parameters. The operational status analysis of the reconstructed signal data is based on the operating characteristics of the energy storage system. According to the variation patterns of voltage and current, the data is divided into three operational states: charging, discharging, and resting. Charging characteristics include parameters such as the charging voltage curve and the rate of change of charging current; discharging characteristics include indicators such as the depth of discharge and the discharge rate; resting characteristics reflect the system's self-discharge characteristics and temperature variation patterns. Feature parameters are extracted for each operational state to form a state feature sequence.

[0057] The fusion of state feature sequences and time series features employs a multidimensional feature analysis method. Time series features reflect the changing trends and periodic characteristics of parameters, while state features describe performance indicators under different operating conditions. Through feature correlation analysis, dynamic correlations between parameters are established, including the order of parameter changes, the intensity of their impact, and their duration, generating a preprocessed data sequence containing complete feature information.

[0058] Taking a lithium battery pack in an energy storage power station as an example, a significant acquisition delay was found in the voltage and temperature data during data processing. Phase analysis determined that the temperature data lagged behind the voltage data by approximately 2 seconds, and a phase correction value was generated accordingly. After phase compensation, the correspondence between temperature and voltage changes became more accurate. Wavelet decomposition was performed on the calibrated data to separate high-frequency interference caused by grid fluctuations. In data reconstruction, the energy contribution during the charging process mainly came from mid-frequency signals, corresponding to the main charging characteristics of the battery. Operating state analysis showed that during the fast charging phase, the rate of temperature rise and the charging current exhibited a significant positive correlation. Through feature fusion analysis, a dynamic correlation model between charging current, temperature change, and battery performance was established.

[0059] In one specific embodiment, the process of executing step S103 may specifically include the following steps:

[0060] (1) Construct a dynamic feature evaluation system for the preprocessed data sequence, extract the mutation points in the time series data based on the operating condition switching characteristics of the energy storage system, and generate the operating condition switching sequence.

[0061] (2) Divide the data into segments for the operating condition transition sequence, extract the operating condition features from the voltage data, current data, temperature data and internal resistance data in each segment, and construct the operating condition feature space;

[0062] (3) Perform energy distribution analysis on the data in the working condition feature space, calculate the energy proportion of each parameter under different working conditions, and output the energy feature matrix;

[0063] (4) Based on the energy feature matrix, feature mapping is performed to project the parameters of the high-energy region onto the multi-dimensional feature space to form a key feature mapping map;

[0064] (5) Perform data coupling analysis on the key feature mapping map, and generate a parameter coupling matrix by calculating the phase relationship and amplitude relationship between parameters;

[0065] (6) The coupling degree values ​​in the parameter coupling matrix are correlated with the working condition transformation sequence to extract feature parameters with significant working condition identification capabilities and form a target feature set.

[0066] Specifically, when performing dynamic feature evaluation on preprocessed data sequences, the focus is on the operating condition switching characteristics of the energy storage system, mainly including key nodes such as charge / discharge switching, load changes, and operating mode transitions. By analyzing the parameter change rate in the time-series data, when the parameter change rate exceeds a set threshold, it is marked as an operating condition switching point, thus generating an operating condition switching sequence reflecting the system's operating state transitions. Based on the generated operating condition switching sequence, the data is segmented. For each operating condition segment, features of voltage, current, temperature, and internal resistance data are extracted. Operating condition feature extraction includes statistical features (mean, standard deviation), time-domain features (rise time, settling time), and frequency-domain features (dominant frequency components, harmonic content), which together constitute the operating condition feature space.

[0067] When performing energy distribution analysis on data in the operating condition characteristic space, the following calculation method is used:

[0068]

[0069] Among them, E dist λ represents the characteristic value of energy distribution. k W represents the weighting coefficient for the k-th working condition. i D represents the weight of the i-th parameter. ki R represents the standardized value of the i-th parameter under the k-th operating condition. j H represents the j-th reference value. j ξ represents the corresponding adjustment factor. k The value represents the attenuation coefficient under operating conditions, m represents the number of operating conditions, n represents the number of parameters, and p represents the number of reference values.

[0070] The data coupling degree analysis of the key feature map is performed using the following formula:

[0071]

[0072] Among them, C coup Represents the coupling coefficient. V represents the cross-correlation coefficient between parameters x and y. x and V y S represents the magnitude of the parameter. x and S y θ represents the standard deviation of the parameter. x and θ y The phase angle, ψ, represents the parameter. xy q represents the phase difference sensitivity, and q represents the total number of parameters.

[0073] After the coupling degree analysis is completed, the parameter coupling matrix is ​​correlated with the operating condition transition sequence. By calculating the variation characteristics of parameter coupling degree under different operating conditions, characteristic parameters that are sensitive to changes in operating conditions are identified, forming a target feature set.

[0074] Taking a battery pack in an energy storage power station as an example, during the charging and discharging process, dynamic feature evaluation identifies the charging and discharging switching points. Near the switching points, the rates of change of voltage and current increase significantly, while temperature changes show a certain lag. Feature extraction of data segments before and after the switching reveals a clear correlation between voltage rise and temperature changes during charging, while internal resistance changes are closely related to current magnitude during discharging. Energy distribution analysis determines the importance of each parameter under different operating conditions; for example, under fast charging conditions, current and temperature account for a higher proportion of energy. Feature mapping and coupling analysis show that voltage-temperature coupling increases significantly towards the end of charging, while current-internal resistance coupling is stronger in the early stages of discharging.

[0075] In one specific embodiment, the process of executing step S104 may specifically include the following steps:

[0076] (1) Group the target feature set according to functional attributes, calculate the data correlation strength between electrical parameters, thermal parameters and state parameters, and generate a parameter correlation matrix;

[0077] (2) Introduce time delay analysis for the parameter correlation matrix, calculate the response delay characteristics between different parameters, and establish a parameter time delay mapping table;

[0078] (3) Combine the time delay data and correlation data in the parameter time delay mapping table to construct the parameter influence link, quantify the causal relationship in the link, and output the parameter causal relationship diagram;

[0079] (4) Perform hierarchical analysis on the parameter causal relationship diagram, identify the direct and indirect influence paths between parameters, and construct a hierarchical influence network;

[0080] (5) Establish parameter interaction rules through hierarchical influence network, match the parameter change patterns in historical fault data with hierarchical influence network, and generate fault feature library;

[0081] (6) Data verification is performed based on the fault feature library. The parameter interaction rules are verified and corrected through cross-validation to obtain a safety detection model. The safety detection model includes a feature input layer, a parameter mapping layer, a feature extraction layer, a fault diagnosis layer, and an output layer.

[0082] Specifically, based on functional attributes, the data is divided into three main categories: electrical parameters (including voltage and current), thermal parameters (including temperature), and state parameters (including internal resistance). The correlation strength between these parameters is then calculated. Cross-correlation analysis is used to assess the changing relationships between different parameters and generate a parameter correlation matrix.

[0083] The time delay analysis uses the following calculation method:

[0084]

[0085] Among them, T lag κ represents the time-delay eigenvalue of the parameter. ij t represents the correlation coefficient between parameters i and j. i and t j η represents the response time of parameters i and j, respectively. ij ζ represents the response decay factor, ζ represents the time delay sensitivity coefficient, and ρ represents the response decay factor. ij This represents the ratio of the rate of change of the parameters, where n represents the number of parameters.

[0086] The causal relationship between parameters and the link is quantified using the following formula:

[0087]

[0088] Among them, C rel μ represents the strength of the causal relationship. k This represents the weight of the k-th link node. Indicates the change in the causal parameter. μ represents the change in the resulting parameter. k Indicates link sensitivity, σ k Represents the time decay coefficient, Δt k This represents the node time interval, and m represents the number of link nodes.

[0089] Based on the aforementioned mathematical model, time-delay data and correlation data are combined to construct parameter influence chains reflecting the relationships between parameters. By quantifying the strength of causal relationships between different parameters, a parameter causal relationship graph is formed, clearly showing the transmission paths of influence between parameters. During the hierarchical analysis, direct and indirect influence paths between parameters are identified. Direct influence paths represent the direct interaction between parameters, such as temperature changes directly caused by current changes; indirect influence paths reflect the influence between parameters through intermediate parameters, such as voltage changes affecting temperature through current changes. These influence paths collectively constitute a hierarchical influence network.

[0090] In the process of matching the hierarchical influence network with historical fault data, the parameter change patterns in the historical fault data are extracted and matched with the influence paths in the network to establish fault-parameter association rules, forming a fault feature library containing multiple fault modes.

[0091] Cross-validation is employed to verify and correct parameter interaction rules, constructing a five-layer safety detection model: the feature input layer receives raw parameter data; the parameter mapping layer performs data standardization and mapping transformation; the feature extraction layer contains multiple parallel processing units to extract time-series features, statistical features, and correlation features; the fault diagnosis layer performs fault mode identification based on features; and the output layer generates diagnostic results. Taking a power battery pack in an energy storage power station as an example, when a battery thermal runaway fault occurs, an abnormal increase in battery temperature is observed. Time-delay analysis reveals that voltage fluctuations, increased current, and decreased internal resistance occur before the temperature anomaly. The parameter influence chain shows that the increased current leads to a decrease in internal resistance, which in turn accelerates the temperature rise, forming a positive feedback effect. This positive feedback loop is identified as a typical characteristic of thermal runaway faults in the hierarchical influence network. Based on this characteristic, the safety detection model can identify potential thermal runaway risks early when anomalies occur in the current-internal resistance-temperature chain.

[0092] In one specific embodiment, the process of executing step S105 may specifically include the following steps:

[0093] (1) Extract key monitoring point combinations from the feature input layer in the safety detection model, input voltage, current, temperature and internal resistance data into the feature input layer, calculate the rate of change and phase relationship of the data through the parameter mapping layer, and generate a dynamic early warning matrix;

[0094] (2) The dynamic early warning matrix input parameter mapping layer is spatially transformed, and the inflection point of parameter change trend is identified through data distribution characteristics. Boundary division is carried out for different risk levels to form a fault classification standard.

[0095] (3) Input the fault classification standard into the feature extraction layer, extract the time-series features, statistical features and correlation features through the parallel feature processing unit, construct the response link between parameters, and output the fault propagation path diagram;

[0096] (4) Perform feature matching on the fault diagnosis layer for the fault propagation path map, calculate the similarity between the real-time operation data of the energy storage system and the historical fault modes, and generate the fault source identification result.

[0097] (5) Perform risk quantification analysis on the fault source identification results at the fault diagnosis layer, calculate the risk level quantification index based on the difference between the fault development trend and the system safety boundary;

[0098] (6) The risk level quantitative indicators are analyzed in multiple dimensions at the output layer. Combined with the fault type identification results, development trend prediction and impact assessment, a fault risk assessment report is generated.

[0099] Specifically, if Figure 2 The diagram shown is a network structure diagram of the security detection model in this application embodiment. The structure includes five main layers: the feature input layer receives raw data of voltage, current, temperature and internal resistance; the parameter mapping layer performs spatial transformation and data standardization processing; the feature extraction layer includes three parallel processing units: time-series features, statistical features and correlation features, which are responsible for feature extraction in different dimensions; the fault diagnosis layer performs feature matching and risk quantification analysis; and the output layer generates a risk assessment report.

[0100] The feature input layer receives collected voltage, current, temperature, and internal resistance data, which are then preliminarily processed by the parameter mapping layer. During this process, the rate of change of each parameter is calculated, including the rate of change of voltage dV / dt, current dI / dt, temperature dT / dt, and internal resistance dR / dt. Simultaneously, the phase relationship between parameters is analyzed to determine the sequence and temporal differences of parameter changes, forming a dynamic early warning matrix reflecting the dynamic characteristics of parameter changes. In the parameter mapping layer, the dynamic early warning matrix undergoes spatial transformation. This process analyzes the distribution characteristics of the data, including statistics such as mean, variance, and kurtosis, to identify inflection points in the parameter change trends. Inflection points typically represent significant changes in the system state, and risk levels are delineated based on these inflection points. The delineation process considers the magnitude, rate, and duration of parameter changes, forming a fault classification standard that includes multiple risk levels.

[0101] After the fault classification criteria are input into the feature extraction layer, multi-dimensional analysis is performed through parallel feature processing units. The temporal feature processing unit extracts the time-domain features of the parameters, such as trends, periodicity, and abrupt changes; the statistical feature processing unit calculates the statistical characteristics of the parameters, including distribution characteristics and correlation indices; and the correlation feature processing unit analyzes the mutual influence relationships between parameters. These features collectively construct the response links between parameters, describing the propagation path of the fault in the system and outputting a fault propagation path diagram. In the fault diagnosis layer, feature matching analysis is performed on the fault propagation path diagram. Real-time operating data of the energy storage system is compared with typical fault modes recorded in the historical fault database, and a similarity index is calculated. The similarity calculation considers the temporal, amplitude, and correlation characteristics of parameter changes, and determines the most matching fault type through a comprehensive score, generating an identification result that includes the fault source location and fault type.

[0102] The fault diagnosis layer is also responsible for quantifying the risk of fault source identification results. By calculating the difference between the fault development trend and the system's preset safety boundaries, the severity and rate of development of the fault are assessed. This process comprehensively considers the degree of parameter exceeding limits, the rate of change, and the cumulative impact, outputting quantitative indicators reflecting the level of fault risk. At the output layer, the risk level quantitative indicators undergo multi-dimensional comprehensive analysis. Combined with the fault type identification results, the typical characteristics and scope of impact of the fault are analyzed; based on development trend prediction, the direction and speed of fault evolution are assessed; and through impact degree assessment, the impact of the fault on the overall system performance is determined. These analytical results are integrated to form a detailed fault risk assessment report.

[0103] Taking a lithium iron phosphate battery pack in an energy storage power station as an example, the working process of the safety detection model is as follows: When the feature input layer receives the monitoring data of the battery pack, it finds that the temperature of a certain single cell has risen abnormally. The parameter mapping layer analysis shows that the temperature change rate exceeds the normal range and forms an abnormal phase relationship with the voltage and current parameters. The feature extraction layer, through parallel processing, finds a significant correlation between the temperature anomaly and the decrease in internal resistance and the increase in current. In the fault diagnosis layer, this parameter change characteristic highly matches the typical characteristics of battery thermal runaway. Risk quantification analysis shows that the temperature rise rate has approached the safety boundary, and there is a risk of thermal diffusion. The assessment report points out that this is a typical early sign of thermal runaway and provides specific handling recommendations.

[0104] In one specific embodiment, the process of executing step S106 may specifically include the following steps:

[0105] (1) Perform spatiotemporal decoupling analysis on the fault type, development trend and impact data in the fault risk assessment report, associate the development characteristics of different types of faults with the performance degradation law, and generate a fault-performance mapping matrix.

[0106] (2) Perform health status analysis on the fault-performance mapping matrix, and construct a performance degradation index sequence by calculating the degradation rate and performance decay curve of each parameter;

[0107] (3) The performance degradation index sequence is correlated with historical operating data to extract the performance change characteristics of the energy storage system under different operating conditions and output the operating condition-performance correlation map.

[0108] (4) Life characteristic analysis is performed based on the working condition-performance correlation map. By calculating the correlation between the degradation trajectory of key performance parameters and the remaining life, a life prediction curve is formed.

[0109] (5) Establish maintenance timing assessment criteria based on the life prediction curve, and generate a maintenance priority ranking by combining system operation constraints and maintenance cost factors;

[0110] (6) Conduct a comprehensive analysis of maintenance priority ranking and performance degradation trend, make overall plans for the urgency of maintenance projects and resource allocation, and output maintenance decision-making schemes.

[0111] Specifically, spatiotemporal decoupling analysis analyzes fault type, development trend, and impact data in both time and spatial dimensions. Time-dimensional analysis focuses on the temporal characteristics of fault development, while spatial-dimensional analysis focuses on the propagation characteristics of faults within the system. By establishing a mapping relationship between the development characteristics of different fault types and their corresponding performance degradation patterns, a mapping matrix reflecting the fault-performance correspondence is generated.

[0112] The following calculation method is used when performing health status analysis on the fault-performance mapping matrix:

[0113]

[0114] Among them, D idx χ² represents the performance degradation index. i F represents the weighting coefficient of the i-th performance parameter. i Represents the fault characteristic value, ω i Indicates degradation sensitivity, υ i P represents the time decay factor. i Let P0 represent the current performance value, P0 represent the initial performance value, t represent the running time, and n represent the number of performance parameters. This formula calculates a performance degradation index sequence reflecting the system's health status. The correlation analysis between the performance degradation index sequence and historical operating data focuses on the performance change characteristics under different operating conditions. By extracting the performance parameter change patterns under each operating condition, a correspondence between operating conditions and performance degradation is established, forming an operating condition-performance correlation graph. This graph illustrates the degree and characteristics of the impact of different operating conditions on system performance.

[0115] Lifespan characteristic analysis is based on the operating condition-performance correlation graph. By analyzing the degradation trajectory of key performance parameters, the health status of the system is assessed, and the remaining lifespan is calculated. Lifespan prediction considers the performance degradation rate, operating condition influencing factors, and historical data statistical patterns, and plots a prediction curve reflecting the system's remaining lifespan. Maintenance timing assessment criteria are established based on the lifespan prediction curve. The assessment criteria comprehensively consider system operating constraints (such as operating load requirements, environmental limitations, etc.) and maintenance cost factors (including maintenance costs, downtime losses, etc.). Through multi-factor comprehensive evaluation, different maintenance items are prioritized to form a maintenance execution sequence.

[0116] This involves a comprehensive analysis of maintenance priorities and performance degradation trends. By assessing the urgency of maintenance projects and the resources required, a reasonable maintenance plan is developed, ultimately outputting a maintenance decision plan that includes specific implementation schemes.

[0117] Taking the maintenance of a battery pack in an energy storage power station as an example: Spatiotemporal decoupling analysis revealed that some individual cells in the battery pack experienced accelerated capacity decay, which was more pronounced under high-temperature conditions. Health status analysis showed that the performance degradation index of the affected batteries continued to rise, with the decay rate exceeding normal levels. Correlation analysis revealed that high-temperature conditions and depth of charge / discharge jointly affected battery performance degradation. Lifetime characteristic analysis predicted that if the current usage pattern is maintained, some batteries will reach the replacement threshold in the short term. Considering both replacement costs and performance requirements, the maintenance decision recommends replacing aging batteries in batches and adjusting the charge / discharge strategy to slow down the degradation rate of other batteries.

[0118] In one specific embodiment, the process of performing spatiotemporal decoupling analysis on the fault type, development trend, and impact level data in the fault risk assessment report may specifically include the following steps:

[0119] (1) Perform time dimension analysis on the parameter data in the fault risk assessment report, associate fault type with time sequence characteristics, extract the fault occurrence time, development cycle and evolution rate, and form a fault time sequence characteristic table.

[0120] (2) Decompose the fault timing feature table into spatial dimensions, quantitatively describe the propagation path and impact range of the fault development trend among different system components, and generate a fault spatial distribution map.

[0121] (3) Cross-analyze the fault spatial distribution map with the impact data, and output the fault impact weight matrix by calculating the impact intensity and attenuation law of the fault at different spatial locations.

[0122] (4) Perform performance correlation analysis on the fault impact weight matrix, quantify the degree of impact of faults on various performance indicators of the energy storage system, and construct a performance degradation factor sequence.

[0123] (5) Match the performance degradation factor sequence with historical failure cases, and output a failure-performance correspondence table by identifying the performance degradation pattern under similar failure modes.

[0124] (6) Data fusion is performed on the fault-performance correspondence table to comprehensively evaluate the performance change characteristics caused by different types of faults and generate a fault-performance mapping matrix.

[0125] Specifically, time-series data is extracted from the fault risk assessment report, and timeline analysis is established for each fault type. The time-series analysis focuses on the fault's occurrence time (determined by the first appearance of parameter anomalies), development cycle (the time interval from the appearance of fault symptoms to the full manifestation of the fault), and evolution rate (the rate at which parameters deviate from normal values). These time-series characteristics are compiled and recorded in a fault time-series characteristic table. The spatial dimension decomposition process focuses on the propagation characteristics of the fault among different components of the energy storage system. Spatial mapping is performed on the data in the fault time-series characteristic table to track the process of the fault spreading from its occurrence location to the surrounding area. By analyzing the physical connections and energy transfer paths between different components, the propagation direction and diffusion speed of the fault are determined. Quantitative descriptions, including propagation distance, diffusion area, and depth of influence, are used to generate a fault spatial distribution map reflecting the spatial distribution characteristics of the fault.

[0126] Cross-analysis of fault spatial distribution maps and impact data is conducted from multiple dimensions. The impact intensity of the fault at various spatial locations is calculated, considering the attenuation of fault energy during propagation. The calculation of impact intensity is based on factors such as parameter deviations, component distances, and propagation impedance, while the attenuation pattern reflects the weakening trend of fault impact with increasing distance. These analytical results are organized into a fault impact weight matrix, quantitatively describing the degree of fault impact on components at different locations. In the performance correlation analysis phase, a correspondence is established between the fault impact weight matrix and the performance indicators of the energy storage system. The impact of faults on core performance indicators such as capacity, efficiency, and lifetime is analyzed and quantitatively calculated. By establishing a mapping relationship between fault impact and performance degradation, a performance degradation factor sequence reflecting the degree of degradation of each performance indicator is constructed. This sequence records the degradation rate and trend characteristics of different performance indicators.

[0127] The matching process between performance degradation factor sequences and historical failure cases employs pattern recognition methods. By comparing the performance degradation characteristics caused by the current failure with the degradation patterns recorded in historical cases, similar failure types and development patterns are identified. The matching process considers multiple feature dimensions such as degradation rate, degradation pattern, and impact degree, outputting a failure-performance correspondence table containing detailed relationships. The data fusion process systematically integrates the information in the failure-performance correspondence table. For different types of failures, the performance change characteristics they cause are analyzed, including direct and indirect impacts. By comprehensively evaluating the impact of failures on various performance indicators, a failure-performance mapping relationship is established, generating a failure-performance mapping matrix.

[0128] Taking the battery pack fault analysis of an energy storage power station as an example, when an abnormal temperature was detected in the battery pack, time-series analysis revealed that the temperature rise first occurred in a single cell and then spread to the surrounding area. Spatial analysis showed that the fault started from this single cell and propagated to adjacent cells along the heat conduction path, exhibiting a circular diffusion characteristic. Cross-analysis showed that the temperature of cells closer to the fault source rose faster, while the impact on distant cells weakened with increasing distance. Performance analysis revealed that the temperature abnormality led to a decrease in battery capacity, and the rate of capacity degradation was positively correlated with the duration and peak value of the temperature abnormality. Comparison with historical cases showed that this temperature abnormality-capacity degradation pattern highly matched typical thermal runaway faults. The fault-performance mapping matrix clearly demonstrated how the temperature abnormality affected various performance indicators of the entire battery pack through thermal diffusion.

[0129] The safety detection method for the energy storage system in the embodiments of this application has been described above. The safety detection system for the energy storage system in the embodiments of this application is described below. Please refer to [link / reference]. Figure 3 One embodiment of the safety detection system for the energy storage system in this application includes:

[0130] The data acquisition module is used to monitor and analyze the operating parameters of the energy storage system. Based on the multi-parameter, nonlinear and time-varying characteristics of the energy storage system, a monitoring parameter system is established. Voltage, current, temperature and internal resistance data are collected for the monitoring parameter system to obtain the basic monitoring dataset.

[0131] The extraction module is used to establish a data preprocessing model based on the basic monitoring dataset, perform data cleaning and standardization on the basic monitoring dataset, extract time series features and statistical features through feature engineering methods, and generate preprocessed data sequences.

[0132] The mapping module is used to construct a feature space model based on the preprocessed data sequence, map the preprocessed data sequence to a multidimensional feature space according to parameter type, and filter key feature indicators through a feature evaluation mechanism to form a target feature set.

[0133] The verification module is used to establish an intelligent diagnostic model based on the target feature set, perform correlation analysis on the electrical parameters, thermal parameters and state parameters in the target feature set, optimize the model parameters through cross-validation, and obtain a safety detection model.

[0134] The input module is used to set fault diagnosis rules according to the safety detection model, input the real-time operation data of the energy storage system into the safety detection model, compare it with the preset safety threshold, and generate a fault risk assessment report.

[0135] The analysis module is used to establish a health assessment system based on the fault risk assessment report, perform trend analysis on system performance in combination with historical operating data, and output maintenance decision-making schemes.

[0136] Through the collaborative efforts of the aforementioned components, and by establishing a data acquisition, processing, analysis, and decision-making chain, comprehensive monitoring and intelligent diagnosis of the safety status of energy storage systems are achieved. In the data acquisition stage, a monitoring parameter system is established based on the multi-parameter, nonlinear, and time-varying characteristics of energy storage systems, enabling accurate acquisition of key parameters such as voltage, current, temperature, and internal resistance, thus solving the problems of incomplete parameter acquisition and insufficient sampling accuracy in traditional methods. In the data preprocessing stage, a data preprocessing model is established to clean and standardize the basic monitoring dataset, and feature engineering methods are combined to extract time-series and statistical features, significantly improving data quality and usability. Regarding feature space construction, a parameter type mapping and feature evaluation mechanism is adopted to achieve data dimensionality reduction and key feature extraction, providing high-quality feature input for subsequent analysis. This invention employs a deep learning network structure comprising a feature input layer, parameter mapping layer, feature extraction layer, fault diagnosis layer, and output layer. This artificial intelligence model design, specifically optimized for the characteristics of energy storage systems, enables accurate diagnosis of energy storage systems under complex operating conditions. The design of each layer of the model fully considers the professional characteristics of energy storage systems. For example, the feature extraction layer uses a parallel neural network structure to extract time-series features, statistical features, and correlation features respectively. The fault diagnosis layer designs a specific loss function based on the professional knowledge of energy storage systems, significantly improving the accuracy and reliability of fault identification. By setting dynamic fault diagnosis rules and combining real-time operating data with preset safety thresholds, a real-time risk assessment mechanism is established. Finally, by establishing a health assessment system and combining historical operating data to conduct trend analysis of system performance, closed-loop management from fault diagnosis to maintenance decision-making is achieved, providing reliable technical support for the safe operation and preventive maintenance of energy storage systems. The overall solution effectively solves the technical challenges of data diversity, fault complexity, and real-time diagnosis in the safety detection of energy storage systems through multi-level data processing and multi-level analysis architecture, realizing the intelligent and automated safety detection of energy storage systems.

[0137] This application also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the steps of the safety detection method for the energy storage system.

[0138] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the system and units described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0139] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0140] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A safety testing method for an energy storage system, characterized in that, The safety testing method for the energy storage system includes: The operating parameters of the energy storage system are monitored and analyzed. A monitoring parameter system is established based on the multi-parameter, nonlinear and time-varying characteristics of the energy storage system. Voltage, current, temperature and internal resistance data are collected for the monitoring parameter system to obtain the basic monitoring dataset. A data preprocessing model is established based on the basic monitoring dataset. Data cleaning and standardization are performed on the dataset. Time series and statistical features are extracted using feature engineering methods to generate a preprocessed data sequence. This includes: constructing a data quality assessment system for the basic monitoring dataset; performing phase analysis on voltage, current, temperature, and internal resistance data to identify phase shifts during data acquisition and form a phase correction sequence; calibrating the phase correction sequence with the basic monitoring dataset; eliminating the impact of data acquisition delays using phase compensation methods to output time-series calibration data; and further processing the noise components in the time-series calibration data. Wavelet decomposition is performed to extract data features from different frequency bands, separating high-frequency noise from effective signals to generate signal decomposition results. Data reconstruction is then performed on these results, weighting and combining effective signals from different frequency bands according to their energy contribution to construct reconstructed signal data. This reconstructed signal data is then segmented according to the operating states of the energy storage system, extracting feature parameters for each operating state, including charging, discharging, and resting characteristics, to generate a state feature sequence. Finally, data fusion processing is performed on this state feature sequence, combining time-series features with state features. Feature correlation analysis is used to extract the dynamic relationships between data, generating a preprocessed data sequence. A feature space model is constructed based on the preprocessed data sequence. The preprocessed data sequence is mapped to a multidimensional feature space according to parameter type. Key feature indicators are screened through a feature evaluation mechanism to form a target feature set. An intelligent diagnostic model is established based on the target feature set. Correlation analysis is performed on the electrical parameters, thermal parameters and state parameters in the target feature set. The model parameters are optimized through cross-validation to obtain a safety detection model. Based on the safety detection model, fault diagnosis rules are set, real-time operating data of the energy storage system is input into the safety detection model, compared with preset safety thresholds, and a fault risk assessment report is generated. A health assessment system is established based on the aforementioned fault risk assessment report. By combining historical operating data, trend analysis of system performance is conducted, and maintenance decision-making plans are output.

2. The safety testing method for an energy storage system according to claim 1, characterized in that, The monitoring and analysis of the operating parameters of the energy storage system involves establishing a monitoring parameter system based on the system's multi-parameter, nonlinear, and time-varying characteristics. Voltage, current, temperature, and internal resistance data are collected for this monitoring parameter system to obtain a basic monitoring dataset, including: Acquire basic operating data of the energy storage system, perform cross-analysis on the voltage and current data in the basic operating data, and output electrical characteristic parameters; The electrical characteristic parameters are correlated with temperature data, and the electrical characteristic parameters are corrected according to the changing trend of temperature data to form a temperature correction coefficient. The temperature correction coefficient and internal resistance data are fused, and the degree of mutual influence between parameters is determined through data correlation analysis to generate a parameter correlation matrix. The parameter correlation matrix is ​​classified according to the multi-parameter, nonlinear and time-varying nature of the energy storage system. Time series analysis is performed on the classified data to establish a monitoring parameter system. According to the monitoring parameter system, data acquisition rules are set, and voltage, current, temperature and internal resistance data are sampled in stages according to the acquisition rules to generate a graded data sequence. The hierarchical data sequence is subjected to data verification and outlier processing. Abnormal data points are corrected through data compensation methods to obtain the basic monitoring dataset.

3. The safety testing method for an energy storage system according to claim 1, characterized in that, The step involves constructing a feature space model based on the preprocessed data sequence, mapping the preprocessed data sequence to a multidimensional feature space according to parameter types, and selecting key feature indicators through a feature evaluation mechanism to form a target feature set, including: A dynamic feature evaluation system is constructed for the preprocessed data sequence, and abrupt change points in the time series data are extracted based on the operating condition switching characteristics of the energy storage system to generate an operating condition switching sequence; The operating condition transition sequence is divided into data segments, and operating condition features are extracted from the voltage data, current data, temperature data and internal resistance data in each segment to construct an operating condition feature space. Energy distribution analysis is performed on the data in the working condition feature space to calculate the energy proportion of each parameter under different working conditions and output the energy feature matrix. Based on the energy feature matrix, feature mapping is performed to project the parameters of the high-energy region onto a multi-dimensional feature space to form a key feature mapping map. Data coupling analysis is performed on the key feature mapping map, and a parameter coupling matrix is ​​generated by calculating the phase relationship and amplitude relationship between parameters; The coupling degree values ​​in the parameter coupling matrix are correlated with the working condition transformation sequence to extract feature parameters with significant working condition identification capabilities, forming a target feature set.

4. The safety detection method for an energy storage system according to claim 1, characterized in that, The intelligent diagnostic model established based on the target feature set involves performing correlation analysis on the electrical, thermal, and state parameters within the target feature set, optimizing the model parameters through cross-validation, and obtaining a safety detection model, including: The target feature set is grouped according to functional attributes, and the data correlation strength between electrical parameters, thermal parameters and state parameters is calculated to generate a parameter correlation matrix; Time delay analysis is introduced for the parameter correlation matrix to calculate the response delay characteristics between different parameters and establish a parameter time delay mapping table; The time delay data and correlation data in the parameter time delay mapping table are combined to construct the parameter influence link, and the causal relationship in the link is quantified to output the parameter causal relationship graph. A hierarchical analysis is performed on the parameter causal relationship graph to identify the direct and indirect influence paths between parameters and to construct a hierarchical influence network. By establishing parameter interaction rules through the hierarchical influence network, the parameter change patterns in historical fault data are matched with the hierarchical influence network to generate a fault feature library. Data verification is performed based on the fault feature library, and the parameter interaction rules are verified and corrected through cross-validation to obtain a safety detection model. The safety detection model includes a feature input layer, a parameter mapping layer, a feature extraction layer, a fault diagnosis layer, and an output layer.

5. The safety testing method for an energy storage system according to claim 4, characterized in that, The step involves setting fault diagnosis rules based on the safety detection model, inputting real-time operating data of the energy storage system into the safety detection model, comparing it with preset safety thresholds, and generating a fault risk assessment report, including: The key monitoring point combination is extracted from the feature input layer of the safety detection model. Voltage, current, temperature and internal resistance data are input into the feature input layer. The rate of change and phase relationship of the data are calculated through the parameter mapping layer to generate a dynamic early warning matrix. The dynamic early warning matrix is ​​input into the parameter mapping layer for spatial transformation. The inflection point of the parameter change trend is identified through data distribution characteristics. Boundaries are divided for different risk levels to form a fault classification standard. The fault classification criteria are input into the feature extraction layer, and the time-series features, statistical features and correlation features are extracted by the parallel feature processing unit to construct the response link between parameters and output the fault propagation path diagram. Feature matching is performed on the fault diagnosis layer for the fault propagation path diagram. The similarity between the real-time operation data of the energy storage system and the historical fault modes is calculated to generate the fault source identification result. The fault source identification results are subjected to risk quantification analysis at the fault diagnosis layer. Based on the difference between the fault development trend and the system safety boundary, the risk level quantification index is output. The aforementioned risk level quantification indicators are comprehensively analyzed from multiple dimensions at the output layer. Combined with the fault type identification results, development trend prediction, and impact assessment, a fault risk assessment report is generated.

6. The safety detection method for an energy storage system according to claim 1, characterized in that, The aforementioned health assessment system, established based on the fault risk assessment report, performs trend analysis on system performance using historical operating data, and outputs maintenance decision-making plans, including: Spatiotemporal decoupling analysis was performed on the fault type, development trend and impact data in the fault risk assessment report to correlate the development characteristics of different types of faults with the performance degradation law and generate a fault-performance mapping matrix. A health status analysis is performed on the fault-performance mapping matrix, and a performance degradation index sequence is constructed by calculating the degradation rate and performance decay curve of each parameter. The performance degradation index sequence is correlated with historical operating data to extract the performance change characteristics of the energy storage system under different operating conditions, and an operating condition-performance correlation map is output. Based on the aforementioned operating condition-performance correlation map, a lifespan characteristic analysis is performed. By calculating the correlation between the degradation trajectory of key performance parameters and the remaining lifespan, a lifespan prediction curve is formed. A maintenance timing assessment criterion is established based on the life prediction curve, and a maintenance priority ranking is generated by combining system operation constraints and maintenance cost factors. By comprehensively analyzing the maintenance priority ranking and performance degradation trend, the urgency of maintenance projects and resource allocation are planned in a coordinated manner, and maintenance decision-making schemes are output.

7. The safety detection method for an energy storage system according to claim 6, characterized in that, The method involves performing spatiotemporal decoupling analysis on the fault type, development trend, and impact data in the fault risk assessment report, correlated with the development characteristics of different fault types and performance degradation patterns, and generated a fault-performance mapping matrix, including: The parameter data in the fault risk assessment report are analyzed in terms of time dimension. The fault type is associated with the time sequence characteristics, and the fault occurrence time, development cycle and evolution rate are extracted to form a fault time sequence characteristic table. The fault timing feature table is decomposed into spatial dimensions to quantitatively describe the propagation path and impact range of the fault development trend among different system components, and a fault spatial distribution map is generated. By cross-analyzing the fault spatial distribution map and the impact degree data, the impact intensity and attenuation law of the fault at different spatial locations are calculated, and the fault impact weight matrix is ​​output. A performance correlation analysis was performed on the fault impact weight matrix to quantify the degree of impact of faults on various performance indicators of the energy storage system and to construct a performance degradation factor sequence. The performance degradation factor sequence is matched with historical failure cases to identify the performance degradation patterns under similar failure modes and output a failure-performance correspondence table. Data fusion is performed on the fault-performance correspondence table to comprehensively evaluate the performance change characteristics caused by different types of faults and generate a fault-performance mapping matrix.

8. A safety detection system for an energy storage system, used to implement the safety detection method for an energy storage system as described in any one of claims 1-7, characterized in that, The safety detection system of the energy storage system includes: The data acquisition module is used to monitor and analyze the operating parameters of the energy storage system. Based on the multi-parameter, nonlinear and time-varying characteristics of the energy storage system, a monitoring parameter system is established. Voltage, current, temperature and internal resistance data are collected for the monitoring parameter system to obtain the basic monitoring dataset. The extraction module is used to establish a data preprocessing model based on the basic monitoring dataset, perform data cleaning and standardization on the basic monitoring dataset, extract time series features and statistical features through feature engineering methods, and generate a preprocessed data sequence. This includes: constructing a data quality assessment system for the basic monitoring dataset; performing data phase analysis on voltage, current, temperature, and internal resistance data to identify phase shifts during data acquisition and form a phase correction sequence; calibrating the phase correction sequence with the basic monitoring dataset; eliminating the impact of data acquisition delays through phase compensation methods; and outputting time-series calibration data; and removing noise from the time-series calibration data. The acoustic components are decomposed using wavelet decomposition to extract data features from different frequency bands, separating high-frequency noise from effective signals to generate signal decomposition results. Data reconstruction is then performed on these results, weighting and combining effective signals from different frequency bands according to their energy contribution to construct reconstructed signal data. This reconstructed signal data is then segmented according to the operating states of the energy storage system, extracting characteristic parameters for each operating state, including charging, discharging, and resting characteristics, to generate a state feature sequence. Finally, the state feature sequence undergoes data fusion processing, combining time-series features with state features, and using feature correlation analysis to extract dynamic relationships between data, generating a preprocessed data sequence. The mapping module is used to construct a feature space model based on the preprocessed data sequence, map the preprocessed data sequence to a multidimensional feature space according to parameter type, and filter key feature indicators through a feature evaluation mechanism to form a target feature set. The verification module is used to establish an intelligent diagnostic model based on the target feature set, perform correlation analysis on the electrical parameters, thermal parameters and state parameters in the target feature set, optimize the model parameters through cross-validation, and obtain a safety detection model. The input module is used to set fault diagnosis rules according to the safety detection model, input the real-time operation data of the energy storage system into the safety detection model, compare it with the preset safety threshold, and generate a fault risk assessment report. The analysis module is used to establish a health assessment system based on the fault risk assessment report, perform trend analysis on system performance in combination with historical operating data, and output maintenance decision-making schemes.

9. A computer-readable storage medium storing instructions thereon, characterized in that, When the instruction is executed by the processor, it implements the safety detection method for the energy storage system as described in any one of claims 1-7.

Images