BMS-based battery energy storage safety management system

Abstract

Disclosed is a BMS-based battery energy storage safety management system. The system comprises a BMS module, a control center module, an evaluation analysis module, a coupling analysis module, a prediction module, and a data storage module. The BMS module collects monitoring data of a battery under different operation states and then, according to the monitoring data, performs initial monitoring of a real-time health state of the battery. The evaluation analysis module, according to a change in data indicators in the monitoring data, evaluates and analyzes all fault analysis models in the BMS module to obtain an evaluation analysis result. Then the coupling analysis module combines the evaluation analysis result with the data indicators of the monitoring data to retrieve multidimensional spatio-temporal data, and constructs a coupling field model to perform coupling analysis on battery operation state. The prediction module then constructs a fault prediction model by using a coupling data set as training data, accurately locates coupling faults of different coupling surfaces, and implements comprehensive analysis and location of coupling faults.

Description

Battery energy storage safety management system based on BMS TECHNICAL FIELD

[0001] The present application relates to the technical field of battery energy storage, in particular to a battery energy storage safety management system based on BMS. BACKGROUND

[0002] With the rapid development of computer intelligence technology, the diagnosis of power battery faults of electric vehicles is being transferred from the battery monomer level to the entire battery system; in the prior art, the positioning and analysis of battery faults by the BMS battery management system mainly measure and analyze the battery temperature and voltage, and determine whether the voltage and temperature of the battery are abnormal by establishing a corresponding fault analysis model, but this battery fault positioning method which analyzes from a certain angle can only accurately locate a single fault, and cannot well solve the problem of multiple fault coupling, cannot accurately describe and analyze the battery safety problem, and the scalability of the BMS management system is limited; under the actual application scene requirement, the models established for fault prediction are not the same, and the requirements for data dimensions and time scales are also not the same. TECHNICAL PROBLEM

[0003] In order to analyze and locate the coupling faults and the internal faults of the battery in real time, the prediction accuracy of the coupling faults is improved by fusing multi-dimensional data analysis, so that in complex use scenarios, accurate real-time fault prediction of the battery can be made, and the present application provides a battery energy storage safety management system based on BMS. TECHNICAL SOLUTION

[0004] In order to solve the above problems, that is, to solve the problems mentioned in the background art, the present application provides a battery energy storage safety management system based on BMS, which optimizes the fault analysis model constructed by the BMS module when a coupling fault occurs under the coordinated analysis of various modules, the evaluation and analysis module scores the fault analysis model suitable for different changes according to the changes of multiple data detection indexes, and then analyzes the coupling coordination degree between the model constructed by the data index set, the coupling analysis module merges the monitoring data and the multi-dimensional spatio-temporal data set according to the evaluation and analysis result, and establishes a coupling field model for analysis, obtains a coupling surface with equal energy exchange rate, and then the prediction module predicts the coupling surface for faults, the analysis and management range of the BMS is greatly expanded by connecting the BMS module with the control center module, the analysis range of the monitoring data is expanded to multi-dimensional spatio-temporal data, the accuracy of the coupling fault analysis and prediction is improved, the coupling analysis module is used to more comprehensively analyze the coupling faults, the precise positioning of the coupling faults is realized, and the efficiency of the battery energy storage safety management is improved.

[0005] The application provides a battery energy storage safety management system based on BMS, which comprises a BMS module, a control center module, an evaluation and analysis module, a coupling analysis module, a prediction module and a data storage module.

[0006] The BMS module collects monitoring data of the battery under different operating states, sends the monitoring data to the data storage module for data storage, and constructs different fault analysis models through the monitoring data to analyze and locate different types of battery faults.

[0007] The battery energy storage safety management system is controlled, and the obtained multi-dimensional space-time data is stored in the data storage module.

[0008] The evaluation and analysis module evaluates and analyzes all the fault analysis models in the BMS module according to the changes of the data indicators in the monitoring data to obtain evaluation and analysis results, and sends the evaluation and analysis results to the coupling analysis module.

[0009] The coupling analysis module combines the evaluation and analysis results with the data indicators of the monitoring data to retrieve external data of the multi-dimensional space-time data, and the external data and the monitoring data form a coupling data set. The coupling analysis module constructs a coupling field model of an electric field, a stress field and an energy field to analyze the operating state of the battery. First, the coupling data set is used to set the space-time data grid of the electric field, the stress field and the energy field. Then, the coupling field model is used to analyze the data in the space-time data grid to obtain different coupling surfaces.

[0010] The prediction module constructs a fault prediction model using the coupling data set as training data, and uses a deep learning model to learn the data set to accurately locate the coupling faults of different coupling surfaces.

[0011] Specifically, the coupling analysis module retrieves external data in the multi-dimensional space-time data set according to the coupling coordination degree and the data indicator set in the evaluation and analysis results, and constructs a coupling field model to analyze the coupling data set,

[0012] In the coupling field, the coupling analysis process between the electric field, the stress field and the energy field is as follows:

[0013] ;

[0014] ;

[0015] wherein, is the diffusion heat value, is the energy transmission vector, is the electric field energy conversion value, is the heat generation rate, is the specific heat capacity of combustion, is the collision initial energy value, is the rate, is the coupling error, is the stress field energy conversion value, is the loss rate, the coupling analysis module carries out coupling analysis on the environment scene corresponding to the battery operating state to obtain the energy exchange rate, and data points with the same energy exchange rate are recorded as a coupling surface, and the energy exchange rates of different coupling surfaces are different.

[0016] In particular, the evaluation analysis module carries out real-time evaluation analysis on the data indicators that change in the monitoring data, and different data indicator combinations corresponding to different fault types change when a coupling fault occurs. The data indicator combination that changes when a coupling fault occurs is recorded as , is the data indicator value, is the subscript of the data indicator, is the number of data indicators that change due to the coupling fault, and different subsets are obtained by extracting elements in set B according to different fault analysis models in the BMS module is the number of fault analysis models, and the evaluation analysis module evaluates the analysis process of different fault analysis models to obtain the corresponding evaluation value, and the calculation formula is:

[0017] ;

[0018] wherein is the number of overlapping data indicators of the jth fault analysis model and the data indicator combination , is the weight proportion of the overlapping data indicators in the fault analysis model, is the change rate of the overlapping data indicators, and are the maximum change rate and the minimum change rate of the jth overlapping data indicator, which are obtained by analyzing the historical monitoring data through the fault analysis model. The number of fault analysis models that overlap between the data indicator combination B and the data indicator combination B is

[0019] , ,and the evaluation analysis module calculates the coupling degree of the fault analysis model according to evaluation values , and the coupling degree calculation formula is as follows:

[0020] ;

[0021] The calculation formula of the coupling coordination degree is as follows:

[0022] ;

[0023] ;

[0024] wherein is a coupling degree of k failure analysis models, is a coordination index, is a weight, a coupling coordination degree is a coordination degree between the failure analysis models corresponding to the data indicator set B, when the greater, the better the coordination degree between different failure analysis models corresponding to the data indicator set, the evaluation analysis module will send the evaluation analysis result including the coupling coordination degree to the coupling analysis module.

[0025] The prediction module extracts features of the coupling data set using a neural network algorithm, and constructs a neural network prediction model to divide the external data and the monitoring data on the coupling surface into a target domain Y and an edge domain X for learning.

[0026] In the prediction analysis process, the learning area migrates from the edge domain X to the target domain Y, and the weight update formula of the classifier is as follows:

[0027] ;

[0028] wherein, the variable represents a predicted value, and the variable represents a true value, is a classifier, indicates a learning rate, and when the classifier is used, the weight of the classifier represents the error size between the predicted value and the true value, and the prediction function is:

[0029] ;

[0030] The prediction function is used for model fusion, wherein is a symbol function, , and is a weight coefficient of the classifier, w is a prediction function of the monitoring classifier, and Π is a mapping function.

[0031] The coupling analysis module retrieves external data in the multi-dimensional spatio-temporal data according to the coupling coordination degree and the data indicator set, the external data is data obtained in addition to data monitoring of the BMS module, and the correlation coefficient of the multi-dimensional spatio-temporal data and the data indicator is calculated, when the correlation coefficient value is greater than the coupling coordination degree, the data retrieval is completed, and the external data related to the data indicator is obtained.

[0032] The monitoring data of the BMS module includes historical data and real-time data of battery monitoring, and the multi-dimensional space-time data includes a cloud data set, and the BMS module establishes a single fault analysis model according to the combination of different data indicators of battery monitoring.

[0033] The data acquisition module of the BMS module acquires monitoring data under different operating states of the battery, and the data indicators in the monitoring data include single battery voltage, battery pole temperature, battery loop current, battery pack terminal voltage and battery system insulation resistance. Advantages

[0034] 1、In the present application, the evaluation and analysis module of the system evaluates and analyzes the fault analysis model in the BMS module, different data indicators change when different battery fault types occur, and the BMS module realizes positioning analysis and management of faults by constructing different fault analysis models, but a single monitoring model cannot analyze the coupling relationship between multiple battery faults, the evaluation and analysis module first evaluates the analysis process of different fault analysis models according to the set of changed data indicators to obtain an evaluation value, then analyzes the coupling coordination degree in different BMS modules, and sends the evaluation and analysis result to the coupling analysis module, the evaluation and analysis module analyzes and evaluates the coupling degree of the coupling fault through the analysis process of the BMS module, and then the coupling analysis module performs coupling analysis, thereby improving the accuracy of coupling analysis.

[0035] 2、The coupling analysis module calls the multi-dimensional space-time data set according to the coupling coordination degree in the evaluation and analysis result and the monitoring data set to obtain a coupling data set, realizes coupling monitoring between battery faults based on big data analysis, and performs coupling analysis on energy exchange in the environment scene where the battery is located according to the coupling data set, divides the data points of the energy exchange rate in the environment where the battery is located to obtain different coupling surfaces, expands the data analysis range while improving the precision of model analysis, and then the prediction module performs fault analysis on different coupling surfaces in combination with the analysis process of the coupling analysis module. BRIEF DESCRIPTION OF DRAWINGS

[0036] Fig. 1 is a whole flowchart of the present application;

[0037] Fig. 2 is a whole module diagram of the present application;

[0038] Fig. 3 is an evaluation and analysis module of the present application;

[0039] Fig. 4 is a coupling analysis module of the present application. Embodiment of the present application

[0040] The preferred embodiments of the present application will be described below with reference to Figs. 1 to 4. Those skilled in the art should understand that these embodiments are only used to explain the technical principles of the present application and are not intended to limit the protection scope of the present application.

[0041] As a key technology to support the development of smart grid and energy transformation, the safety management of battery energy storage has always been the focus of attention. In particular, in new energy vehicles, the safety hazards and problems such as spontaneous combustion of battery energy vehicles are increasingly prominent. There are multiple complex chemical and physical reactions in the process of battery energy storage, which can easily cause battery thermal runaway and further cause energy storage safety accidents. Battery system failures are mainly divided into mechanical failures and electrical failures. The main cause of mechanical failure is that the vehicle is hit or squeezed during driving. After the vehicle is hit or squeezed, the battery pack will deform, causing the electrolyte and gas inside the battery to overflow, the battery to bulge, the internal resistance to increase, and the heat inside the battery to increase rapidly. If the heat is not effectively dissipated, it is easy to cause an accident. When the battery system has electrical faults such as overcharging, over-discharging, and excessive output power, it will cause the battery pack temperature to rise, causing the electrolyte and gas inside the battery to overflow, fire, explosion, and other accidents. Due to the numerous factors affecting battery safety, it is particularly urgent to effectively monitor and evaluate the management of battery energy storage systems.

[0042] BMS (Battery Management System), battery management system, which is one of the core subsystems for monitoring the state of energy storage batteries, is mainly for intelligent management and maintenance of each battery unit to prevent overcharging and over-discharging of the battery, prolong the service life of the battery, and monitor the state of the battery. Responsible for monitoring the running state of each battery in the battery energy storage unit to ensure safe and reliable operation of the energy storage unit. BMS can monitor and collect the state parameters of energy storage batteries in real time, analyze and calculate related state parameters as needed, obtain more system state evaluation parameters, and realize effective control of energy storage batteries according to specific protection control strategies to ensure safe and reliable operation of the entire battery energy storage unit. At the same time, BMS can interact with external other devices (PCS, EMS, fire extinguishing system, etc.) through its own communication interface, analog / digital input interface, form a linkage control of each subsystem in the entire energy storage power station, and ensure safe, reliable, and efficient grid-connected operation of the power station. However, the scalability of BMS battery management is limited. In order to improve the real-time monitoring efficiency of battery energy storage, a fault diagnosis and prediction system for the entire battery energy storage safety management system is established from the perspective of big data analysis.

[0043] As shown in FIG. 1 to FIG. 4, the application proposes a battery energy storage safety management system based on BMS, which comprises a BMS module, a control center module, an evaluation analysis module, a coupling analysis module, a prediction module and a data storage module; on the basis of massive operation data, data features are extracted, a time series model and a fusion model of faults and features are established, the relationship between features and faults is obtained by training the model with a large amount of data, and finally the faults are predicted based on the relationship.

[0044] The BMS module collects monitoring data of the battery under different operating states and sends the monitoring data to the data storage module for data storage; the functions of the battery management system (BMS) include battery parameter detection, battery state establishment, online diagnosis, battery safety protection and alarm, charge control, battery consistency control, thermal management function, network function and information storage; the battery parameter detection includes total voltage, total current, single cell voltage detection, temperature detection, insulation detection, collision detection, impedance detection, smoke detection, etc.

[0045] The control center module is the highest control end, controls the battery energy storage safety management system, and stores the obtained multi-dimensional space-time data to the data storage module; the BMS module comprises a data acquisition module and a battery management module, the BMS module monitors the real-time health status of the battery, and constructs different fault analysis models through monitoring data to analyze and position different types of battery faults;

[0046] In the battery management process of the BMS module, the battery fault diagnosis algorithm establishes corresponding battery fault analysis models through different data indexes, such as voltage anomaly detection, internal resistance estimation, temperature anomaly detection, SOC (State of Charge) estimation, SOH (State of Health) estimation, etc.; in the temperature detection, the temperature of different battery monomers is collected by using a temperature sensor, and the temperature anomaly is detected; the monitoring data of the temperature is compared with the temperature threshold to realize the temperature anomaly diagnosis. For fault diagnosis, the BMS module monitors the faults of the battery from different aspects, but when the battery operates in a complex environment, a single aspect cannot accurately analyze the coupling faults, and there is a coupling relationship between different faults, so it is important to realize the algorithm optimization of the BMS module in different scenes.

[0047] But in order to solve the problem that it is difficult to diagnose the coupling phenomenon of multiple faults and make accurate prediction of faults in actual use scenarios with complex and changeable working conditions, first of all, the evaluation and analysis module evaluates and analyzes all fault analysis models in the BMS module according to the changes of data indicators in the monitoring data to obtain evaluation and analysis results, and sends the evaluation and analysis results to the coupling analysis module; different faults occur, the monitored data indicators in the BMS module are different, when there is a coupling fault in the battery management system, the coupling fault is a fault affected by multiple influencing factors, and the evaluation and analysis module comprehensively analyzes the mutual influencing factors in the coupling fault analysis;

[0048] The BMS module monitors the charging by monitoring indicators such as voltage, current and resistance when the battery is charging and discharging, but this is only an internal factor affecting the battery charging. When charging in an extremely low temperature environment, the charging speed of the battery is slow, and the endurance time is shortened, so the charging state should be comprehensively analyzed in the process of coupling of different factors.

[0049] When the data indicators change, the BMS module performs initial monitoring and fault analysis on the data indicators and the state of the battery. The change of the data indicators is complex, and the BMS module only predicts the fault from a single angle, which ignores the coupling relationship between faults;

[0050] The coupling analysis module combines the evaluation and analysis results with the data indicators of the monitoring data to retrieve external data of multi-dimensional space-time data. The external data and the monitoring data form a coupling data set, the coupling analysis module constructs a coupling field model of electric field, stress field and energy field to analyze the coupling of the battery operating state. First, set the space-time data grid of the electric field, stress field and energy field according to the coupling data set, and then analyze the data in the space-time data grid by using the coupling field model to obtain different coupling surfaces;

[0051] The prediction module constructs a fault prediction model with the coupling data set as training data, uses a deep learning model to learn the data set, accurately locates the coupling faults of different coupling surfaces, and based on a large number of researches on the boundary, performance characteristics, internal principles and fault expansion of the fault occurrence, uses data mining to visualize the relationship between data, extracts features, uses machine learning to establish a model to analyze the relationship between the fault and the features, and diagnoses the battery fault.

[0052] The coupling analysis module calls external data in the multidimensional data set according to the coupling coordination degree and the data index set in the evaluation analysis result, and constructs a coupling field model to perform coupling analysis on the coupling data set. In the charging process, the state of charge of the battery is related to not only the values of the voltage, current and other data indexes inside the battery, but also the temperature in the environmental scene. The multidimensional space-time data include the environmental temperature data and weather data of the region where the battery is located. The coupling field is constructed based on the diffusion of energy. In the charging process of the battery, the battery is affected by the sum of the energy field and the electric field. The battery safety management reduces the safety hazard of the battery and avoids fire during the charging process. In the coupling field, the coupling analysis process between the electric field, the stress field and the energy field is as follows:

[0053]

[0054]

[0055] wherein, is a diffusion heat value, is an energy transmission vector, is an electric field energy conversion value, is a heat generation rate, is a specific heat capacity of combustion, is a collision initial energy value, is a rate, is a coupling error, is a stress field energy conversion value, is a loss rate, the coupling field model is mapped into a multidimensional space, the energy conversion and diffusion process in the coupling field are dynamically analyzed, the energy exchange process in the coupling field is classified according to the coupling relationship, different coupling surfaces are obtained, the coupling surfaces are composed of different data points in the multidimensional space, the coupling surface dynamically fluctuates in different states, the coupling analysis module performs coupling analysis on the environmental scene corresponding to the battery operating state to obtain an energy exchange rate, data points with the same energy exchange rate are recorded as a coupling surface, and the exchange rates of different coupling surfaces are different.

[0056] In the coupling analysis module, the analysis is performed from the energy conversion angle. When a mechanical failure occurs, the mechanical energy caused by stress is converted into potential energy and internal energy. The internal energy is converted into heat energy by the combustion of fuel inside the battery. The temperature change is realized through energy diffusion. The energy conversion occurs between the temperatures at different positions inside the battery and the environmental temperature outside the battery. When the heat dissipation is abnormal, the energy in the overall coupling field is transferred.

[0057] The evaluation analysis module performs real-time evaluation analysis on the data indexes that change in the monitoring data. When different fault types occur, the changed data index combination is different. The changed data index combination when the coupling fault occurs is recorded as​​ is the subscript of the data index, and different subsets are obtained by extracting elements in set B according to different fault analysis models in the BMS module is the number of fault analysis models, the evaluation analysis module evaluates the analysis process of different fault analysis models to obtain the corresponding evaluation value, and the calculation formula is:

[0058]

[0059] wherein is the number of data indexes that coincide with the combination of the jth fault analysis model and the data index group, is the weight proportion of the data index in the fault analysis model, is the change rate of the data index, and are the maximum change rate and the minimum change rate of the jth data index when a fault occurs, which are obtained by analyzing historical monitoring data through the fault analysis model. The number of fault analysis models that coincide between the data index combination B is

[0060] The evaluation analysis module calculates the coupling degree of the fault analysis model according to the evaluation values, and the coupling degree calculation formula is as follows:

[0061]

[0062] The calculation formula of the coupling coordination degree is as follows:

[0063]

[0064]

[0065] wherein is the coupling degree of the kth fault analysis model, is the coordination index, is the weight, and the coupling coordination degree is the coordination degree between the fault analysis models corresponding to the data index set B, when is larger, the coordination degree between different fault analysis models corresponding to the data index set is better, and the evaluation analysis module sends the evaluation analysis result including the coupling coordination degree to the coupling analysis module.

[0066] ​​​​​​​​​​The prediction module extracts features of the coupled data set by using a neural network algorithm, and constructs a neural network prediction model, divides the external data and the monitoring data on the coupling surface into a target domain Y and an edge domain X for learning, and through the initial management of the battery energy storage by the BMS module, single fault diagnosis and state monitoring can be realized, when the battery is damaged by external collision, etc., the battery operating environment changes sharply, multiple fault types and complex causes cause the data of multiple monitoring data indicators to change, forming a coupled fault, the prediction module improves the prediction accuracy through the prediction analysis of the coupling surface data, and through the evaluation analysis module and the coupling analysis module, the battery fault analysis and prediction are more comprehensive, in the prediction process, the data points on the coupling surface are related to the external data and the detection data, the prediction module uses two classifiers to perform transfer learning on different regions, and constructs the transfer between the region corresponding to the external data and the region corresponding to the detection data.

[0067] In the prediction analysis process, the learning region migrates from the edge domain X to the target domain Y, and the weight updating formula of the classifier is as follows:

[0068] ;

[0069] Wherein, the variable represents the predicted value, the variable represents the true value, is the classifier, and the learning rate is represented by When using the classifier , the weight of the classifier represents the error size between the predicted value and the true value, and the prediction function is:

[0070] ;

[0071] The prediction function is used for model fusion, wherein is a symbol function, , and is the weight coefficient of the classifier, w is the prediction function of the monitoring classifier, and Π is the mapping function.

[0072] The coupling analysis module retrieves external data in the multi-dimensional space-time data according to the coupling coordination degree and the data index set, the external data is data obtained in addition to the data monitoring of the BMS module, and the correlation coefficient of the multi-dimensional space-time data and the data index is calculated, when the correlation coefficient value is greater than the coupling coordination degree, the data retrieval is completed, and the external data related to the data index is obtained, the coupling coordination degree is used as the standard for evaluating the coupling degree of the fault, and the greater the coupling degree, the more complex the data set required for fault analysis.

[0073] The monitoring data of the BMS module includes historical data and real-time data of battery monitoring, and multi-dimensional space-time data includes a cloud data set, and the BMS module establishes a single fault analysis model according to the combination of different data indexes of battery monitoring.

[0074] The BMS module performs fault analysis according to the change values of different data indexes, for example, the state establishment of the battery: SOC battery remaining capacity (nuclear power state), SOH battery health state (health state), SOF (function state), and SOP (power state).

[0075] ;

[0076] 。

[0077] The data acquisition module of the BMS module acquires monitoring data under different operating states of the battery, and the data indexes in the monitoring data include single battery voltage, battery pole temperature, battery loop current, battery pack terminal voltage, and battery system insulation resistance. The monitoring data of the BMS module includes single battery voltage, battery pole temperature, battery loop current, battery pack terminal voltage, and battery system insulation resistance, necessary analysis and calculation are performed on related state parameters to obtain state evaluation parameters of the system, and effective control is realized on the energy storage battery body according to a specific protection control strategy, so that safe and reliable operation of the battery energy storage is ensured.

[0078] In specific use, the system includes a BMS module, a control center module, an evaluation analysis

[0079] The module, the coupling analysis module, the prediction module and the data storage module, the BMS module includes the data acquisition module and the battery management module, the BMS module monitors the real-time health state of the battery, and different fault analysis models are constructed by monitoring data to analyze different types of battery faults, the data acquisition module of the BMS module collects the monitoring data of the battery under different operating conditions, and sends the monitoring data to the data storage module for data storage, the control center module is the highest control end, controls the battery energy storage safety management system, the evaluation and analysis module evaluates and analyzes all the fault analysis models in the BMS module according to the change of the data index in the monitoring data to obtain the evaluation and analysis result; first, the evaluation value of the different fault analysis model analysis process is obtained according to the changed data index set, then the coupling coordination degree in the different BMS modules is analyzed, and the evaluation and analysis result is sent to the coupling analysis module, the coupling degree of the coupling fault is analyzed and evaluated by the BMS module analysis process, and then the coupling analysis module is coupled; the coupling analysis module combines the evaluation and analysis result and the data index of the monitoring data to retrieve external data of multi-dimensional space-time data, the external data and the monitoring data form a coupling data set, the coupling analysis module constructs the coupling field model of the electric field, the stress field and the energy field to analyze the coupling of the battery operating state, first, the space-time data grid of the electric field, the stress field and the energy field is set according to the coupling data set, then the coupling field model is used to analyze the data in the space-time data grid to obtain different coupling surfaces; then the prediction module uses the coupling data set as training data to construct a fault prediction model, uses a deep learning model to learn the data set, accurately locates the coupling fault of different coupling surfaces, and realizes comprehensive analysis and positioning of the coupling fault.

[0080] The technical scheme of the application has been described in connection with the preferred embodiments shown in the drawings, but those skilled in the art can easily understand that the protection scope of the application is obviously not limited to these specific embodiments. Those skilled in the art can make equivalent changes or replacements to related technical features without departing from the principles of the application, and the technical scheme after the changes or replacements will fall within the protection scope of the application.

Claims

1. A battery energy storage safety management system based on BMS, characterized in that: The system comprises a BMS module, a control center module, an evaluation analysis module, a coupling analysis module, a prediction module and a data storage module. The BMS module collects monitoring data of the battery under different operating states, sends the monitoring data to the data storage module for data storage, and constructs different fault analysis models through the monitoring data to analyze and locate different types of battery faults. The control center module controls the battery energy storage safety management system and stores the obtained multi-dimensional spatio-temporal data in the data storage module. The evaluation analysis module evaluates and analyzes all fault analysis models in the BMS module according to the changes of data indicators in the monitoring data to obtain evaluation analysis results, and sends the evaluation analysis results to the coupling analysis module. The coupling analysis module retrieves external data of the multi-dimensional spatio-temporal data set according to the coupling coordination degree and the data indicator set in the evaluation analysis results, and constructs a coupling field model to analyze the coupling data set. In the coupling field, the coupling analysis process between the electric field, the stress field and the energy field is as follows: The loss rate is the energy exchange rate of the battery operating state corresponding to the environmental scene, and the data points with the same energy exchange rate are recorded as a coupling surface. The energy exchange rates of different coupling surfaces are different.

2. A battery energy storage safety management system based on BMS according to claim 1, characterized in that: The maximum and minimum change rates of the jth coincident data indicator are obtained by analyzing the historical monitoring data through the fault analysis model. ； ； wherein, to diffuse the heat value, for the energy transfer vector, for the electric field energy conversion value, for the heat production rate, for the specific heat capacity at constant pressure, collision initial energy value, for rate, to couple errors, for the stress field energy conversion value, The calculation formula of is as follows:

3. The battery energy storage safety management system based on BMS of claim 1, wherein, The evaluation analysis module performs real-time evaluation analysis on the changed data indicators in the monitoring data, and different combinations of the changed data indicators correspond to different fault types. When a coupling fault occurs, the combination of the changed data indicators is denoted as , is a data indicator value, is a data indicator subscript, is the number of data indicators changed by the coupling fault, and different subsets are obtained according to different fault analysis models in the BMS module , is the number of fault analysis models, and the evaluation analysis module evaluates the analysis process of different fault analysis models to obtain a corresponding evaluation value, and the calculation formula is: ； wherein For the first Fault analysis model and data indicator combination the number of coincident data indicators, To determine the weight proportion of the coincidence data index in the fault analysis model, to indicate the rate of change of the data point, With The larger the coupling coordination degree is, the better the coordination between different fault analysis models corresponding to the data indicator set is. The evaluation analysis module sends the evaluation analysis results including the coupling coordination degree to the coupling analysis module.

4. The battery energy storage safety management system based on BMS according to claim 3, characterized in that, The number of fault analysis models coinciding with the data index combination B is , , The evaluation analysis module further calculates the coupling degree of the fault analysis models according to evaluation values , and the coupling degree calculation formula is as follows: ; Coupling coordination degree The prediction module extracts the features of the coupling data set using a neural network algorithm, and constructs a neural network prediction model to divide the external data and the monitoring data on the coupling surface into a target domain Y and an edge domain X for learning. ; ; wherein a coupling degree for k failure analysis models, To coordinate the indices, weight, coupling coordination degree The coordination degree between the fault analysis models corresponding to the data index set B is: In the prediction analysis process, the learning area migrates from the edge domain X to the target domain Y, and the weight update formula of the classifier is as follows:

5. The battery energy storage safety management system based on BMS of claim 1, wherein, When the weight of the classifier represents the error size between the prediction value and the true value, the prediction function is:

6. The battery energy storage safety management system based on BMS according to claim 5, wherein, The weight coefficient of the classifier is w, the prediction function of the monitoring classifier is w, and the mapping function is Π. ; wherein the variables represents the predicted value, variable represents the true value, For the classifier, denotes the learning rate, and ​ ; Model fusion is performed using a prediction function, wherein = 1 if x > 0 、 ​ 7. The battery energy storage safety management system based on BMS of claim 1, wherein, The coupling analysis module calls external data in the multi-dimensional space-time data according to the coupling coordination degree and the data index set, the external data is data obtained in addition to data monitoring of the BMS module, and a correlation coefficient of the multi-dimensional space-time data and the data index is calculated, when the correlation coefficient value is greater than the coupling coordination degree, data calling is completed, and external data related to the data index is obtained.

8. The battery energy storage safety management system based on BMS of claim 1, wherein, The monitoring data of the BMS module includes historical data and real-time data of battery monitoring, and the multi-dimensional space-time data includes a cloud data set, and the BMS module establishes a single fault analysis model according to combination of different data indexes of battery monitoring.

9. The battery energy storage safety management system based on BMS of claim 1, wherein, The data acquisition module of the BMS module acquires monitoring data under different operating states of the battery, and data indexes in the monitoring data include cell voltage, battery pole temperature, battery loop current, battery pack terminal voltage and battery system insulation resistance.

Images