Fault prediction method, system and equipment based on multi-physics field coupling, medium and product

Abstract

The invention relates to the technical field of electric power systems, and discloses a fault prediction method, system, device, medium and product based on multi-physics field coupling, the method comprises the following steps: inputting current operation data of a battery module into a multi-physics field coupling dynamics model, determining current multi-physics field distribution of the battery module, and predicting the fault of the battery module through the multi-physics field coupling dynamics model; therefore, a complex coupling evolution mechanism in the battery is revealed from the perspective of multi-physics field coupling, the current estimation state of the battery module is estimated by adopting extended Kalman filtering so as to correct the current state of the battery module, and the estimation state of the battery module at a future preset moment is predicted through a risk trend prediction model based on LSTM so as to correct the current state of the battery module. Therefore, the fault state of the battery module is predicted through the estimated state of the battery module, the fault state of the battery module is predicted through the multi-dimensional physical quantity, the battery fault can be accurately predicted comprehensively and timely, and the false alarm rate and the missing report rate of the battery fault are reduced.

Description

Technical Field

[0001] This invention relates to the field of power system technology, and in particular to a fault prediction method, system, device, medium and product based on multi-physics coupling. Background Technology

[0002] With the rapid development of energy storage systems, the safety and reliability of lithium-ion batteries have become core challenges for the industry. Batteries may experience thermal runaway, internal short circuits, structural fatigue, and gas leaks under complex operating conditions. The evolution of these failures involves the coupling effects of multiple physical fields, including thermal, electrical, mechanical, chemical, and acoustic factors. For example, mechanical abuse (such as impact and compression) or electrical abuse (such as overcharging) can trigger internal short circuits, leading to a chain reaction of exothermic reactions and ultimately thermal runaway or even fire and explosion. Currently, various standards impose stringent requirements on the thermal safety and overall safety of battery systems, necessitating the development of high-precision fault prediction and health management technologies.

[0003] Traditional battery fault prediction methods often focus on fault signals in a single dimension, making it difficult to predict battery faults comprehensively and in a timely manner, resulting in high false alarm and false negative rates. Summary of the Invention

[0004] In view of this, in order to solve the above-mentioned technical problems, the present invention provides a fault prediction method, system, device, medium and product based on multi-physics coupling.

[0005] The first aspect of this invention provides a fault prediction method based on multi-physics coupling, comprising:

[0006] Based on the pre-constructed multiphysics coupling dynamics model of the battery module and combined with the current operating data of the battery module, the current multiphysics distribution of the battery module is determined; wherein, the multiphysics coupling dynamics model is obtained by coupling multiple physical fields including electrothermal, stress, structural acoustics and gas dynamics.

[0007] Based on the current multiphysics distribution, an extended Kalman filter is used to estimate the current estimated state of the battery module.

[0008] Based on a pre-trained LSTM risk trend prediction model, combined with the current estimated state of the battery module, the estimated state of the battery module at a future preset time is predicted.

[0009] The fault state of the battery module is predicted based on the estimated state of the predicted battery module.

[0010] Preferably, the method further includes:

[0011] Based on the geometric structure of the battery module, construct a geometric model of the battery module;

[0012] Based on the geometric model of the battery module, simulation models of electrothermal field, stress field, structural acoustic field, and gas dynamic acoustic field are constructed respectively.

[0013] The electrothermal field simulation model, the stress field simulation model, the structural acoustic field simulation model, and the gas dynamic acoustic field simulation model are coupled to obtain the coupled dynamic model of the multiphysics field.

[0014] Preferably, the current multiphysics field distribution includes stress field distribution, temperature field distribution, acoustic energy field distribution, and gas concentration field distribution;

[0015] The step of estimating the current estimated state of the battery module using an extended Kalman filter based on the current multiphysics distribution includes:

[0016] Based on the current multiphysics field distribution and the measured multiphysics field data of the battery module obtained by the sensor, the state feature vector of the battery module is determined.

[0017] Based on the extended Kalman filter algorithm and combined with the state feature vector of the battery module, the temperature state, stress state, acoustic energy density state and gas concentration of the battery module are determined as the current estimated state.

[0018] Preferably, determining the state feature vector of the battery module based on the current multiphysics distribution and the measured multiphysics data of the battery module obtained by the sensors includes:

[0019] Physical field features are extracted from the stress field distribution, the temperature field distribution, the acoustic energy field distribution, and the gas dynamic field distribution, respectively; wherein, the physical field features include stress gradient, temperature rise rate, acoustic energy density, and gas escape concentration rate;

[0020] Acquire multi-physics field measured data of the battery module by the sensor, and determine the multi-physics field deviation based on the residual between the multi-physics field measured data and the current multi-physics field distribution;

[0021] The state feature vector of the battery module is determined based on the stress gradient, the temperature rise rate, the acoustic energy density, the gas escape concentration rate, and the multiphysics field deviation.

[0022] Preferably, predicting the fault state of the battery module based on the estimated state of the predicted battery module includes:

[0023] A dimensionless transformation is performed on the predicted estimated state of the battery module to obtain multiple dimensionless transformed estimated state features.

[0024] The estimated state features after multiple dimensionless transformations are weighted to obtain the fault risk index of the battery module.

[0025] Compare the magnitude of the fault risk index with the threshold values ​​of the fault risk index ranges corresponding to multiple preset warning levels, and determine the preset warning level corresponding to the fault risk index based on the comparison results.

[0026] The fault status of the battery module is determined based on the preset warning level corresponding to the fault risk indicator.

[0027] Preferably, the method further includes:

[0028] Based on the temperature field distribution and the gas dynamic field distribution, the temperature change rate and gas concentration change rate of the battery module are determined;

[0029] The thermal runaway risk index of the battery module is determined based on the temperature change rate and gas concentration change rate of the battery module.

[0030] If the thermal runaway risk index continuously exceeds the preset thermal runaway risk threshold within a preset time period, the battery module is determined to have a thermal runaway risk.

[0031] Secondly, the present invention also provides a fault prediction system based on multi-physics coupling, comprising:

[0032] The physics field determination module is used to determine the current multiphysics field distribution of the battery module based on the pre-constructed multiphysics field coupled dynamics model of the battery module and the current operating data of the battery module; wherein, the multiphysics field coupled dynamics model is obtained by coupling multiple physical fields including electrothermal, stress, structural acoustics and gas dynamics;

[0033] The state estimation module is used to estimate the current estimated state of the battery module based on the current multiphysics distribution using an extended Kalman filter.

[0034] The state prediction module is used to predict the estimated state of the battery module at a future preset time based on a pre-trained LSTM risk trend prediction model and the current estimated state of the battery module.

[0035] The fault state prediction module is used to predict the fault state of the battery module based on the estimated state of the predicted battery module.

[0036] Thirdly, the present invention also provides an electronic device, the electronic device including a memory and a processor, the memory storing a computer program, the computer program being executed by the processor causing the processor to perform the steps of the fault prediction method based on multiphysics coupling as described in the first aspect.

[0037] Fourthly, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed, implements the steps of the fault prediction method based on multiphysics coupling as described in the first aspect.

[0038] Fifthly, the present invention also provides a computer program product comprising a computer program stored on a non-transitory computer-readable storage medium, the computer program comprising program instructions, wherein when the program instructions are executed by a computer, the computer performs the steps of the fault prediction method based on multiphysics coupling as described in the first aspect.

[0039] As can be seen from the above technical solutions, this invention determines the current multiphysics distribution of the battery module by inputting the current operating data of the battery module into a multiphysics coupling dynamics model. This reveals the complex coupling evolution mechanism inside the battery from the perspective of multiphysics coupling. Extended Kalman filtering is used to estimate the current estimated state of the battery module to correct the current state of the battery module. Furthermore, an LSTM-based risk trend prediction model is used to predict the estimated state of the battery module at a future preset time. Thus, the fault state of the battery module is predicted based on the estimated state of the battery module. By using multidimensional physical quantities to predict the fault state of the battery module, it is possible to comprehensively and timely make accurate predictions of battery faults, thereby reducing the false alarm rate and false negative rate of battery faults. Attached Figure Description

[0040] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 An application environment diagram for a fault prediction method based on multi-physics coupling provided in an embodiment of the present invention;

[0042] Figure 2 A flowchart illustrating a fault prediction method based on multi-physics coupling provided in an embodiment of the present invention;

[0043] Figure 3A schematic diagram of a fault prediction system based on multi-physics coupling provided in an embodiment of the present invention;

[0044] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0045] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0046] Batteries, operating in complex and variable environments, are susceptible to various potential failures, including thermal runaway, internal short circuits, structural fatigue, and gas leakage. The evolution of these failures involves the complex coupling of multiple physical fields, including thermal, electrical, mechanical, chemical, and acoustic factors. For example, mechanical abuse (such as external impact or compression) or electrical abuse (such as overcharging) can trigger internal short circuits, leading to a chain reaction of exothermic reactions that ultimately result in thermal runaway, and in severe cases, even fire or explosion. Currently, relevant domestic and international standards impose extremely stringent requirements on the thermal safety performance and overall pack safety of battery systems. Therefore, there is an urgent need to develop high-precision, high-reliability fault prediction and health management technologies to improve the safety and reliability of battery use.

[0047] Traditional battery fault prediction methods typically focus only on single-dimensional fault signals, such as monitoring only voltage or temperature changes, while ignoring the interactions between multiple physical fields. This makes it difficult to provide comprehensive and timely predictions of battery faults. This limitation leads to high false alarm and false negative rates in practical applications, failing to meet the development requirements of high-reliability battery management systems.

[0048] Therefore, embodiments of this application provide a fault prediction method based on multiphysics coupling, which can be applied to, for example... Figure 1In the application environment shown, terminal 101 communicates with server 102 via a network. A data storage system can store the data that server 102 needs to process. The data storage system can be integrated onto server 102 or placed on a cloud or other network server. Terminal 101 or server 102 determines the current multiphysics distribution of the battery module based on a pre-built coupled dynamics model of the multiphysics field, combined with the current operating data of the battery module. The coupled dynamics model of the multiphysics field is obtained by coupling multiple physical fields including electrothermal, stress, structural acoustics, and gas dynamics. Based on the current multiphysics distribution, an extended Kalman filter is used to estimate the current estimated state of the battery module. Based on a pre-trained LSTM risk trend prediction model, combined with the current estimated state of the battery module, the estimated state of the battery module at a future preset time is predicted. Finally, the fault state of the battery module is predicted based on the predicted estimated state of the battery module.

[0049] Terminal 101 can be, but is not limited to, various personal computers, laptops, smartphones, and tablets.

[0050] Server 102 can be a standalone physical server, a server cluster or distributed system consisting of multiple physical servers, or a cloud server that provides cloud computing services.

[0051] like Figure 2 As shown, this application provides a fault prediction method based on multiphysics coupling, which is applied to... Figure 1 Taking terminal 101 or server 102 as an example, the explanation includes the following steps S1 to S4. Wherein:

[0052] Step S1: Based on the pre-constructed multi-physics coupled dynamics model of the battery module and combined with the current operating data of the battery module, determine the current multi-physics distribution of the battery module; wherein, the multi-physics coupled dynamics model is obtained by coupling multiple physical fields including electrothermal, stress, structural acoustics and gas dynamics.

[0053] The coupled dynamics model is derived from the multi-physics response characteristics of the battery module by coupling multiple physical fields, including electrothermal, stress, structural acoustics, and gas dynamics. Specifically, the construction process of the coupled dynamics model includes: constructing a geometric model of the battery module based on its geometry; constructing simulation models for the electrothermal field, stress field, structural acoustic field, and gas dynamic acoustic field based on the geometric model of the battery module; and coupling these simulation models to obtain the multi-physics coupled dynamics model.

[0054] Among them, the electrothermal coupling process takes into account the Joule heat generated when the current passes through the electrode material, as well as the reverse effect of temperature change on conductivity, ion mobility and electrochemical reaction rate; the stress field captures the mechanical deformation caused by the volume expansion and contraction induced by lithium ion insertion / extraction during charge and discharge cycles; the structural acoustic field reflects the micro-vibration characteristics of the battery casing under the combined action of thermal stress and electrochemical stress; and the gas dynamic field quantifies the dynamic disturbance of gas generation behavior on the internal pressure distribution and heat conduction path.

[0055] Among these factors, the geometric parameters of the battery module are key to the simulation accuracy. These include the overall dimensions of the module (length, width, and height), the arrangement and stacking of the cells, the width and depth of the heat dissipation channels, and the distribution and density of various connectors (such as busbars and mounting brackets). Specifically, the overall dimensions of the module are fundamental parameters determining the spatial resolution of the electrothermal and stress fields, directly affecting the accuracy of identifying temperature gradients and stress concentration areas. The arrangement and stacking of the cells determine the current path distribution and heat diffusion direction, thus affecting the risk of localized overheating. The width and depth of the heat dissipation channels jointly regulate the thermal resistance characteristics, while the position and density of the busbars and mounting brackets are directly related to structural stiffness and vibration modal frequencies.

[0056] These structural parameters directly determine the spatial coupling strength of multiple physical fields such as electricity, heat, and force, and directly affect the reasonable setting of boundary conditions for each field equation. In finite element modeling, to balance computational efficiency and numerical accuracy, a non-uniform meshing strategy is typically adopted: local mesh refinement is performed at the cell-to-cell contact interface, current collection regions, and heat source concentration regions, while a sparser mesh is appropriately used in other regions where physical quantities change gradually. This differentiated meshing method effectively captures the drastic changes in high-gradient temperature fields and stress concentration regions, thereby significantly improving the accuracy of numerical solutions in key areas and the overall reliability of the model simulation.

[0057] By constructing a system of ordinary differential equations, the evolution of each physical quantity over time is precisely characterized, revealing the complex interaction mechanisms in the occurrence and development of faults. Among these mechanisms, during the charging and discharging process of the battery module, current flow and electrochemical reactions inevitably generate heat, leading to an increase in internal temperature. This heat generation primarily includes the Ohmic thermal effect (i.e.,...) The heat generated (including heat from heating) and electrochemical side reactions, and the time-varying dynamics of the electrothermal field are described by the following heat balance equation:

[0058]

[0059] In the formula, This represents the equivalent heat capacity of the module. This refers to the instantaneous temperature inside the module. It is the charging and discharging current. This refers to the battery's internal resistance. This represents the heat released by the electrochemical side reaction. The convective heat transfer coefficient is... This represents the surface area of ​​the battery module. This refers to the ambient temperature. The equation comprehensively considers the balance between heat generation and dissipation, reflecting real-time changes in the battery's thermal state.

[0060] During charge-discharge cycles, lithium ions repeatedly insert and extract between the positive and negative electrode materials. This electrochemical process induces periodic expansion and contraction of the electrode material's lattice volume. Simultaneously, Joule heating generated by current passing through internal resistance and the thermal effects inherent in the electrochemical reactions cause fluctuations in the battery's internal temperature, leading to uneven thermal deformation due to differences in the coefficients of thermal expansion of different materials. This electrode deformation caused by ion migration coupled with the thermal expansion effect due to temperature changes induces a complex distribution of mechanical stress within the battery module. Over time, this stress accumulates and evolves dynamically, potentially causing problems such as interface delamination, active material breakage, or current collector fatigue. The combined effect of thermal and electrochemical stress further exacerbates material structural degradation, significantly increasing the risk of mechanical damage, capacity decay, and reduced battery life.

[0061] The governing equations of the stress field are:

[0062]

[0063] in, It is instantaneous stress. The Young's modulus of the material. For total strain, The coefficient of thermal expansion is This is the initial reference temperature. The sensitivity coefficient for volume expansion caused by lithium-ion intercalation. This model describes the microstructural changes caused by lithium-ion concentration. It helps analyze the coupling mechanism between electrochemical processes and mechanical responses, revealing the impact of stress accumulation on battery performance and lifespan.

[0064] Among them, the volume expansion sensitivity coefficient is a key parameter that reflects the linear relationship between the lithium-ion insertion depth and the volume change of the electrode material. Its value directly depends on the crystal structure stability and micropore distribution characteristics of the positive and negative electrode active materials.

[0065] When microscopic structural damage or microcrack propagation occurs inside a battery, stress release or material fracture will radiate acoustic signals in specific frequency bands. These signals are often concentrated in the low-frequency range because their longer wavelengths and stronger penetrating power can reflect early changes in the mechanical state inside the battery. The frequency characteristics, energy distribution, and time-frequency properties exhibited by low-frequency structural acoustic waves are important bases for early fault detection and health status diagnosis of batteries. High-sensitivity acoustic sensors can effectively collect these fault-related acoustic emission signals, providing a reliable data foundation for subsequent signal analysis and fault identification. In a structural acoustic field, the trend of acoustic energy density variation can reflect the development process of structural anomalies. The acoustic energy density index is defined as follows:

[0066]

[0067] in, The frequency domain sound pressure amplitude, and The low-frequency range of interest (typically below 100kHz) was defined. ω is the angular frequency.

[0068] If the internal temperature of a battery rises abnormally or materials decompose, a certain amount of flammable gas will be released, significantly increasing the overall safety risk of the system. This gas release process can essentially be considered a temperature-driven chemical kinetic reaction, with a clear exponential dependence on temperature. The gas dynamic acoustic field formed during the gas release process can typically be quantitatively described using Arrhenius-type equations.

[0069]

[0070] in, Indicates gas concentration. The frequency factor of the reaction, For activation energy, Let be the ideal gas constant. This is an empirical function that reflects the influence of the battery's state of charge (SOC) and state of health (SOH) on the gas release rate. This model effectively captures the dynamic changes in gas concentration with temperature and battery state.

[0071] Among them, the frequency factor is a key parameter characterizing the intrinsic rate of reaction, and its value is closely related to the chemical bond energy of the electrode material, the interfacial reactivity, and the electrolyte composition; within different temperature ranges... The value of needs to be determined by fitting gas release experimental data under multiple operating conditions to ensure that the Arrhenius model has engineering applicability in a wide temperature range of -20℃ to 85℃.

[0072] Activation energy is a core physical property parameter characterizing the height of the reaction energy barrier. Its value directly reflects the energy threshold that needs to be overcome for electrolyte decomposition or electrode material phase transition. In typical lithium-ion battery systems, the activation energy range of key side reactions such as SEI film growth, electrolyte oxidation, and positive electrode oxygen release is concentrated in the range of 60–120 kJ / mol. This range has an inherent thermodynamic consistency with the measured thermal runaway initiation temperature (about 130 °C).

[0073] The ideal gas constant, with a value of 8.314 J / (mol·K), is a universal physical constant that serves as a bridge between temperature and energy units in the Arrhenius equation. It ensures the dimensional consistency of the reaction rate expression in the International System of Units (SI) and makes the ratio of activation energy to thermodynamic temperature dimensionless, thereby guaranteeing the theoretical rigor and engineering portability of the model across different temperature scales. f(SOC, SOH) comprehensively characterizes the coupled regulatory effect of the battery's electrochemical state on gas release: high SOC exacerbates the oxidation tendency of the cathode.

[0074] The current operating data of the battery module includes several key parameters such as initial temperature, stress, gas concentration, acoustic energy density, and current. These data comprehensively reflect the actual state and dynamic changes of the battery during operation. By inputting the current operating data of the battery module into a highly accurate coupled dynamic model in real time, the distribution of multiple physical fields can be calculated efficiently and accurately, thereby realizing the dynamic simulation and real-time monitoring of the complex physical processes inside the battery.

[0075] Step S2: Based on the current multiphysics distribution, use extended Kalman filtering to estimate the current estimated state of the battery module.

[0076] Among them, the Extended Kalman Filter (EKF) is a state estimation algorithm specifically designed for handling nonlinear systems. Based on the current multiphysics distribution, this algorithm fully considers the nonlinear characteristics of the system's state variables and the interference of external noise. By dynamically updating the state estimate and error covariance, it achieves an accurate estimate of the current state of the battery module. The EKF effectively improves the accuracy and robustness of the state estimation by performing local linearization on the nonlinear system, enabling it to maintain good tracking performance even under complex operating conditions.

[0077] Step S3: Based on the pre-trained LSTM risk trend prediction model and combined with the current estimated state of the battery module, predict the estimated state of the battery module at a future preset time.

[0078] Long Short-Term Memory (LSTM) neural networks are recurrent neural network structures specifically designed for time-series data modeling. To effectively capture the time-series dependencies of system state characteristics, this application introduces an LSTM neural network to train a risk trend prediction model. The training process of the risk trend prediction model is as follows: using the historical estimated state of the battery module as input and the future multi-step state evolution trajectory as labels, end-to-end fitting is achieved by minimizing the prediction error to obtain the optimal convergence of the model parameters. After training, the risk trend prediction model receives the current estimated state as input and predicts the future state evolution trajectory, i.e.:

[0079]

[0080] In the formula, The length of the historical time window. To predict the step size, This describes the operation process of a long short-term memory neural network. It is a sequence of state features within a historical window. This is the sequence of state features within the prediction window.

[0081] Step S4: Predict the fault state of the battery module based on the estimated state of the predicted battery module.

[0082] Accurate prediction of battery module status allows for further analysis and assessment of its health and potential risks. By using the estimated state information of the predicted battery module, we can deduce whether the module is experiencing a fault, including potential performance degradation or abnormal behavior. This process not only helps identify current problems but also enables systematic quantitative analysis and judgment of potential battery failure modes at an early stage. This provides a scientific basis for taking preventative maintenance measures, effectively extending battery life and improving the overall system's safety and reliability.

[0083] It should be noted that, in this embodiment, the current operating data of the battery module is input into a multiphysics coupling dynamics model to determine the current multiphysics distribution of the battery module. This reveals the complex coupling evolution mechanism inside the battery from the perspective of multiphysics coupling. An extended Kalman filter is used to estimate the current estimated state of the battery module to correct the current state of the battery module. Furthermore, an LSTM-based risk trend prediction model is used to predict the estimated state of the battery module at a future preset time. Thus, the fault state of the battery module is predicted based on the estimated state of the battery module. By using multidimensional physical quantities to predict the fault state of the battery module, it is possible to comprehensively and timely make accurate predictions of battery faults, thereby reducing the false alarm rate and false negative rate of battery faults.

[0084] In some embodiments, the current multiphysics distribution includes stress field distribution, temperature field distribution, acoustic energy field distribution, and gas concentration field distribution. In this case, based on the current multiphysics distribution, an extended Kalman filter is used to estimate the current estimated state of the battery module, including: determining the state feature vector of the battery module based on the current multiphysics distribution and the measured multiphysics data of the battery module obtained by the sensor; and determining the temperature state, stress state, acoustic energy density state, and gas concentration of the battery module as the current estimated state based on the extended Kalman filter algorithm and the state feature vector of the battery module.

[0085] Sensors are a key component of the monitoring system, including various types such as stress sensors, temperature sensors, acoustic sensors, and gas sensors, forming a multi-source heterogeneous sensor array. This system can acquire multi-dimensional physical response signals generated by the battery module during operation in real time, such as stress changes, temperature fluctuations, acoustic vibrations, and gas releases. These raw signals undergo spatiotemporal alignment and noise suppression processing to eliminate the influence of timestamp inconsistencies and interference factors, ultimately yielding high-quality multiphysics field measurement data, specifically covering key indicators such as temperature distribution, stress state, gas concentration changes, and acoustic energy density. The spatiotemporal alignment process involves mapping sensor data from different sampling frequencies and installation locations to the same spatiotemporal reference frame through timestamp synchronization and spatial interpolation algorithms. Noise suppression employs a combination of adaptive wavelet thresholding and robust principal component analysis to effectively remove environmental interference and measurement outliers, ensuring the accuracy and stability of subsequent state estimation.

[0086] Subsequently, by constructing a multi-dimensional feature mapping relationship under a unified spatiotemporal benchmark, high-dimensional data from different physical fields (including electrical, thermal, and mechanical fields) are efficiently fused and collaboratively analyzed with multi-source measured signals, thereby achieving comprehensive, high-precision, and dynamic real-time monitoring of the battery's operating status.

[0087] Specifically, based on the current multiphysics field distribution and the measured multiphysics field data of the battery module obtained by the sensors, the state feature vector of the battery module is determined, including: extracting physical field features from the stress field distribution, temperature field distribution, acoustic energy field distribution, and gas dynamic field distribution respectively; wherein, the physical field features include stress gradient, temperature rise rate, acoustic energy density, and gas escape concentration rate; acquiring the measured multiphysics field data of the battery module obtained by the sensors, and determining the multiphysics field deviation based on the residual between the measured multiphysics field data and the current multiphysics field distribution; and determining the state feature vector of the battery module based on the stress gradient, temperature rise rate, acoustic energy density, gas escape concentration rate, and multiphysics field deviation.

[0088] The stress gradient, as the rate of change of stress over time in a stress field distribution, is a crucial physical quantity for measuring how quickly local stress in a material or structure changes over time. It not only reflects the potential risk of structural instability in a local area but also demonstrates the drastic nature of stress changes in both spatial and temporal dimensions. A higher stress gradient typically indicates a rapid change in the stress state of that region, potentially leading to fatigue damage, plastic deformation, or crack propagation. Therefore, it holds significant importance in the safety assessment and life prediction of engineering structures. The stress gradient is:

[0089]

[0090] In the formula, For stress gradient, Let be the stress at time t.

[0091] The rate of temperature rise is an important indicator reflecting the dynamic change of thermal state. It describes how quickly the temperature changes per unit time. The rate of temperature rise is:

[0092]

[0093] In the formula, Let t be the temperature at time t.

[0094] Acoustic energy density is a key physical quantity used to reveal the acoustic state inside a structure. It achieves an effective quantitative assessment of the degree of structural anomalies by analyzing the spatial distribution of the acoustic energy field, especially by integrating low-frequency energy. Acoustic energy density is the instantaneous acoustic energy density within a region volume V. Integrating (r,t) and taking the average, we get:

[0095]

[0096] In the formula, This represents the sound energy density.

[0097] The gas escape concentration rate is a key indicator for assessing the risk of gas release, directly reflecting the speed at which gas escapes from its source and its diffusion within the environment. This indicator is not only relevant to workplace safety management but also has significant guiding implications for environmental protection and accident prevention. By monitoring and accurately calculating the gas escape concentration rate in real time, potential hazards can be identified promptly. The gas escape concentration rate is defined as follows:

[0098]

[0099] In the formula, The initial concentration at the start of gas escape. The concentration rate of gas escape.

[0100] The multiphysics bias is determined by the residual between the measured multiphysics data and the current multiphysics distribution.

[0101]

[0102] In the formula, For multiphysics bias, k is the number of physics fields involved in the fusion (in this application, the value is 4). The weight of the residual of the i-th physical field. For a single physical field residual; where:

[0103]

[0104] In the formula, and These are the measured data and the current distribution prediction value of the i-th physical field, respectively.

[0105] The state feature vector of the battery module is composed of stress gradient, temperature rise rate, acoustic energy density, gas escape concentration rate, and multi-physics field deviation:

[0106]

[0107] In the formula, This is the state feature vector of the battery module.

[0108] The state feature vector of the battery module is set as the state variable of the extended Kalman filter. At the same time, the actual measurement data of each state feature vector of the battery module is used as the observation vector. The actual observation data and the prediction results of the multiphysics distribution are fused by the update formula of the extended Kalman filter algorithm.

[0109] This process can output multiple key state parameters of the battery module in real time, including temperature state, stress state, acoustic energy density state, and gas concentration state, and use these data as the optimal estimates for the current moment. By continuously fusing real-time observation data from sensors with model-based prediction information, the system not only achieves dynamic adjustment and correction of various state estimates, but also significantly enhances the robustness of the entire system and the accuracy of overall state estimation. This mechanism effectively addresses the uncertainties caused by changes in the external environment and internal parameter disturbances, thus providing reliable data support for the safe operation and performance optimization of the battery system. The update formula for the extended Kalman filter is:

[0110]

[0111] In the formula, For prior state estimation, Let be the posterior state estimate at time t. Here is the Kalman gain matrix. For nonlinear observation functions, Let t be the observation vector (measured data) at time t.

[0112] In some embodiments, predicting the fault state of the battery module based on the predicted estimated state of the battery module includes: performing a dimensionless transformation on the predicted estimated state of the battery module to obtain multiple dimensionless transformed estimated state features; performing weighted processing on the multiple dimensionless transformed estimated state features to obtain a fault risk index of the battery module; comparing the magnitude relationship between the fault risk index and the fault risk index interval thresholds corresponding to multiple preset warning levels, and determining the preset warning level corresponding to the fault risk index based on the comparison results; and determining the fault state of the battery module based on the preset warning level corresponding to the fault risk index.

[0113] During data processing, the estimated state parameters for each battery module have inconsistent dimensions; for example, voltage, current, and internal resistance may use different units or numerical ranges. This discrepancy severely affects the comparability of data and the effectiveness of overall analysis. Therefore, to achieve data consistency and comparability, the estimated states predicted from each module must be standardized. Specifically, this process involves performing a dimensionless transformation on the original estimated state values ​​of each module, generating multiple corresponding standardized estimated state features. This transformation completely eliminates the scale differences caused by different dimensions, significantly improving the comparability of data in subsequent calculations, while also making the data processing more efficient and convenient. Ultimately, standardization not only enhances data quality but also provides a reliable and consistent data foundation for subsequent statistical analysis, model building, and system optimization.

[0114] By weighting the estimated state characteristics after multiple dimensionless transformations, the failure risk index of the battery module is obtained, namely:

[0115]

[0116] In the formula, As a fault risk indicator, These are the weight coefficients for each feature. To estimate the state The function after dimensionless transformation.

[0117] Based on fault risk indicators, a graded response strategy is designed as follows, with three preset warning levels: Level 1, Level 2, and Level 3. The corresponding preset warning level is Level 1, and the response strategy is continuous data logging and alarm notification. The corresponding preset warning level is Level 2 warning. The response strategy is to reduce the battery module power and slow down the operating rate. Reducing the battery module power is achieved by adjusting the BMS output command to limit the charging and discharging current amplitude; slowing down the operating rate is achieved by dynamically adjusting the SOC threshold range and the temperature control fan speed. ,and Its corresponding preset warning level is Level 3, and the response strategy is emergency power outage and activation of fire-fighting linkage measures. Among these, The rate of change of the fault risk indicator. This is the threshold for changes in the fault risk indicator, such as a value of 0.4.

[0118] exist When the value is ≤0.6, it is determined to be in normal operation, and no warning response is triggered. Only routine monitoring and data archiving are maintained.

[0119] In some embodiments, the method further includes: determining the temperature change rate and gas concentration change rate of the battery module based on the temperature field distribution and the gas dynamic field distribution; determining the thermal runaway risk index of the battery module based on the temperature change rate and gas concentration change rate of the battery module; and determining that the battery module has a thermal runaway risk if the thermal runaway risk index continuously exceeds a preset thermal runaway risk threshold within a preset time period.

[0120] Specifically, regarding the risk of thermal runaway, the thermal runaway risk index for battery modules is defined as follows:

[0121]

[0122] In the formula, As a thermal runaway risk indicator, λ is the weighting coefficient for gas concentration changes. Exceeding the preset thermal runaway risk threshold In such cases, the battery module is deemed to have a risk of thermal runaway, triggering an early warning mechanism to ensure a timely and effective emergency response. In response to latency, .

[0123] Based on the same inventive concept, this application also provides a fault prediction system based on multiphysics coupling for implementing the fault prediction method based on multiphysics coupling mentioned above.

[0124] The solution provided by this system is similar to the solution described in the above method. Therefore, the specific limitations of one or more embodiments of the fault prediction system based on multi-physics coupling provided below can be found in the limitations of the fault prediction method based on multi-physics coupling above, and will not be repeated here.

[0125] like Figure 3As shown, this application provides a fault prediction system based on multiphysics coupling, including:

[0126] The physics field determination module 100 is used to determine the current multiphysics field distribution of the battery module based on the pre-constructed multiphysics field coupled dynamics model of the battery module and the current operating data of the battery module; wherein, the multiphysics field coupled dynamics model is obtained by coupling multiple physical fields including electrothermal, stress, structural acoustics and gas dynamics.

[0127] The state estimation module 200 is used to estimate the current estimated state of the battery module based on the current multiphysics distribution using an extended Kalman filter.

[0128] The state prediction module 300 is used to predict the estimated state of the battery module at a future preset time by combining the current estimated state of the battery module with the risk trend prediction model based on a pre-trained LSTM.

[0129] The fault state prediction module 400 is used to predict the fault state of the battery module based on the estimated state of the predicted battery module.

[0130] In some embodiments, the system further includes: a physics model construction module, used for:

[0131] Based on the geometric structure of the battery module, construct a geometric model of the battery module;

[0132] Based on the geometric model of the battery module, simulation models of the electrothermal field, stress field, structural acoustic field, and gas dynamic acoustic field are constructed respectively.

[0133] The electrothermal field simulation model, stress field simulation model, structural acoustic field simulation model, and gas dynamic acoustic field simulation model are coupled to obtain a multi-physics field coupled dynamic model.

[0134] In some embodiments, the current multiphysics field distribution includes a stress field distribution, a temperature field distribution, an acoustic energy field distribution, and a gas concentration field distribution; the state estimation module 200 is used for:

[0135] Based on the current multiphysics field distribution and the measured multiphysics field data of the battery module obtained by the sensor, the state feature vector of the battery module is determined.

[0136] Based on the extended Kalman filter algorithm and combined with the state feature vector of the battery module, the temperature state, stress state, acoustic energy density state and gas concentration of the battery module are determined as the current estimated state.

[0137] In some embodiments, the state estimation module 200 is configured to:

[0138] Physical field features were extracted from the stress field distribution, temperature field distribution, acoustic energy field distribution, and gas dynamic field distribution, respectively. Among them, the physical field features include stress gradient, temperature rise rate, acoustic energy density, and gas escape concentration rate.

[0139] Acquire multi-physics field measured data of the battery module by the sensor, and determine the multi-physics field deviation based on the residual between the multi-physics field measured data and the current multi-physics field distribution;

[0140] The state characteristic vector of the battery module is determined based on stress gradient, temperature rise rate, acoustic energy density, gas escape concentration rate, and multi-physics field deviation.

[0141] In some embodiments, the fault state prediction module 400 is configured to:

[0142] A dimensionless transformation is performed on the estimated state of the predicted battery module to obtain multiple dimensionless transformation estimated state features.

[0143] By weighting the estimated state characteristics after multiple dimensionless transformations, the failure risk index of the battery module is obtained.

[0144] Compare the magnitudes of the fault risk indicators with the threshold ranges of multiple preset warning levels, and determine the preset warning levels corresponding to the fault risk indicators based on the comparison results.

[0145] The fault status of the battery module is determined based on the preset warning level corresponding to the fault risk indicator.

[0146] In some embodiments, the system further includes: a thermal runaway risk assessment module, used for:

[0147] Based on the temperature field distribution and gas dynamic field distribution, the temperature change rate and gas concentration change rate of the battery module are determined;

[0148] The thermal runaway risk indicators of the battery module are determined based on the temperature change rate and gas concentration change rate of the battery module.

[0149] If the thermal runaway risk index continuously exceeds the preset thermal runaway risk threshold within a preset time period, the battery module is determined to have a thermal runaway risk.

[0150] It should be noted that, in this embodiment, the current operating data of the battery module is input into a multiphysics coupling dynamics model to determine the current multiphysics distribution of the battery module. This reveals the complex coupling evolution mechanism inside the battery from the perspective of multiphysics coupling. An extended Kalman filter is used to estimate the current estimated state of the battery module to correct the current state of the battery module. Furthermore, an LSTM-based risk trend prediction model is used to predict the estimated state of the battery module at a future preset time. Thus, the fault state of the battery module is predicted based on the estimated state of the battery module. By using multidimensional physical quantities to predict the fault state of the battery module, it is possible to comprehensively and timely make accurate predictions of battery faults, thereby reducing the false alarm rate and false negative rate of battery faults.

[0151] like Figure 4 As shown, this application embodiment provides an electronic device. The electronic device 10 includes a memory 20 and a processor 30. The memory 20 stores a computer program. When the computer program is executed by the processor 30, the processor 30 performs the following steps:

[0152] Based on the pre-constructed multiphysics coupled dynamics model of the battery module and combined with the current operating data of the battery module, the current multiphysics distribution of the battery module is determined; wherein, the multiphysics coupled dynamics model is obtained by coupling multiple physical fields including electrothermal, stress, structural acoustics and gas dynamics.

[0153] Based on the current multiphysics distribution, the extended Kalman filter is used to estimate the current estimated state of the battery module;

[0154] Based on a pre-trained LSTM risk trend prediction model, combined with the current estimated state of the battery module, the estimated state of the battery module at a future preset time is predicted.

[0155] Predict the fault state of the battery module based on the estimated state of the predicted battery module.

[0156] In some embodiments, processor 30 also performs:

[0157] Based on the geometric structure of the battery module, construct a geometric model of the battery module;

[0158] Based on the geometric model of the battery module, simulation models of the electrothermal field, stress field, structural acoustic field, and gas dynamic acoustic field are constructed respectively.

[0159] The electrothermal field simulation model, stress field simulation model, structural acoustic field simulation model, and gas dynamic acoustic field simulation model are coupled to obtain a multi-physics field coupled dynamic model.

[0160] In some embodiments, the current multiphysics field distribution includes stress field distribution, temperature field distribution, acoustic energy field distribution, and gas concentration field distribution;

[0161] Based on the current multiphysics distribution, an extended Kalman filter is used to estimate the current estimated state of the battery module, including:

[0162] Based on the current multiphysics field distribution and the measured multiphysics field data of the battery module obtained by the sensor, the state feature vector of the battery module is determined.

[0163] Based on the extended Kalman filter algorithm and combined with the state feature vector of the battery module, the temperature state, stress state, acoustic energy density state and gas concentration of the battery module are determined as the current estimated state.

[0164] In some embodiments, the state feature vector of the battery module is determined based on the current multiphysics distribution and the measured multiphysics data of the battery module obtained by the sensor, including:

[0165] Physical field features were extracted from the stress field distribution, temperature field distribution, acoustic energy field distribution, and gas dynamic field distribution, respectively. Among them, the physical field features include stress gradient, temperature rise rate, acoustic energy density, and gas escape concentration rate.

[0166] Acquire multi-physics field measured data of the battery module by the sensor, and determine the multi-physics field deviation based on the residual between the multi-physics field measured data and the current multi-physics field distribution;

[0167] The state characteristic vector of the battery module is determined based on stress gradient, temperature rise rate, acoustic energy density, gas escape concentration rate, and multi-physics field deviation.

[0168] In some embodiments, predicting the fault state of the battery module based on the estimated state of the predicted battery module includes:

[0169] A dimensionless transformation is performed on the estimated state of the predicted battery module to obtain multiple dimensionless transformation estimated state features.

[0170] By weighting the estimated state characteristics after multiple dimensionless transformations, the failure risk index of the battery module is obtained.

[0171] Compare the magnitudes of the fault risk indicators with the threshold ranges of multiple preset warning levels, and determine the preset warning levels corresponding to the fault risk indicators based on the comparison results.

[0172] The fault status of the battery module is determined based on the preset warning level corresponding to the fault risk indicator.

[0173] In some embodiments, processor 30 also performs:

[0174] Based on the temperature field distribution and gas dynamic field distribution, the temperature change rate and gas concentration change rate of the battery module are determined;

[0175] The thermal runaway risk indicators of the battery module are determined based on the temperature change rate and gas concentration change rate of the battery module.

[0176] If the thermal runaway risk index continuously exceeds the preset thermal runaway risk threshold within a preset time period, the battery module is determined to have a thermal runaway risk.

[0177] It should be noted that, in this embodiment, the current operating data of the battery module is input into a multiphysics coupling dynamics model to determine the current multiphysics distribution of the battery module. This reveals the complex coupling evolution mechanism inside the battery from the perspective of multiphysics coupling. An extended Kalman filter is used to estimate the current estimated state of the battery module to correct the current state of the battery module. Furthermore, an LSTM-based risk trend prediction model is used to predict the estimated state of the battery module at a future preset time. Thus, the fault state of the battery module is predicted based on the estimated state of the battery module. By using multidimensional physical quantities to predict the fault state of the battery module, it is possible to comprehensively and timely make accurate predictions of battery faults, thereby reducing the false alarm rate and false negative rate of battery faults.

[0178] This application provides a computer-readable storage medium storing a computer program thereon, which, when executed, performs the following steps:

[0179] Based on the pre-constructed multiphysics coupled dynamics model of the battery module and combined with the current operating data of the battery module, the current multiphysics distribution of the battery module is determined; wherein, the multiphysics coupled dynamics model is obtained by coupling multiple physical fields including electrothermal, stress, structural acoustics and gas dynamics.

[0180] Based on the current multiphysics distribution, the extended Kalman filter is used to estimate the current estimated state of the battery module;

[0181] Based on a pre-trained LSTM risk trend prediction model, combined with the current estimated state of the battery module, the estimated state of the battery module at a future preset time is predicted.

[0182] Predict the fault state of the battery module based on the estimated state of the predicted battery module.

[0183] In some embodiments, when a computer program is executed, the following also occurs:

[0184] Based on the geometric structure of the battery module, construct a geometric model of the battery module;

[0185] Based on the geometric model of the battery module, simulation models of the electrothermal field, stress field, structural acoustic field, and gas dynamic acoustic field are constructed respectively.

[0186] The electrothermal field simulation model, stress field simulation model, structural acoustic field simulation model, and gas dynamic acoustic field simulation model are coupled to obtain a multi-physics field coupled dynamic model.

[0187] In some embodiments, the current multiphysics field distribution includes stress field distribution, temperature field distribution, acoustic energy field distribution, and gas concentration field distribution;

[0188] Based on the current multiphysics distribution, an extended Kalman filter is used to estimate the current estimated state of the battery module, including:

[0189] Based on the current multiphysics field distribution and the measured multiphysics field data of the battery module obtained by the sensor, the state feature vector of the battery module is determined.

[0190] Based on the extended Kalman filter algorithm and combined with the state feature vector of the battery module, the temperature state, stress state, acoustic energy density state and gas concentration of the battery module are determined as the current estimated state.

[0191] In some embodiments, the state feature vector of the battery module is determined based on the current multiphysics distribution and the measured multiphysics data of the battery module obtained by the sensor, including:

[0192] Physical field features were extracted from the stress field distribution, temperature field distribution, acoustic energy field distribution, and gas dynamic field distribution, respectively. Among them, the physical field features include stress gradient, temperature rise rate, acoustic energy density, and gas escape concentration rate.

[0193] Acquire multi-physics field measured data of the battery module by the sensor, and determine the multi-physics field deviation based on the residual between the multi-physics field measured data and the current multi-physics field distribution;

[0194] The state characteristic vector of the battery module is determined based on stress gradient, temperature rise rate, acoustic energy density, gas escape concentration rate, and multi-physics field deviation.

[0195] In some embodiments, predicting the fault state of the battery module based on the estimated state of the predicted battery module includes:

[0196] A dimensionless transformation is performed on the estimated state of the predicted battery module to obtain multiple dimensionless transformation estimated state features.

[0197] By weighting the estimated state characteristics after multiple dimensionless transformations, the failure risk index of the battery module is obtained.

[0198] Compare the magnitudes of the fault risk indicators with the threshold ranges of multiple preset warning levels, and determine the preset warning levels corresponding to the fault risk indicators based on the comparison results.

[0199] The fault status of the battery module is determined based on the preset warning level corresponding to the fault risk indicator.

[0200] In some embodiments, when a computer program is executed, the following also occurs:

[0201] Based on the temperature field distribution and gas dynamic field distribution, the temperature change rate and gas concentration change rate of the battery module are determined;

[0202] The thermal runaway risk indicators of the battery module are determined based on the temperature change rate and gas concentration change rate of the battery module.

[0203] If the thermal runaway risk index continuously exceeds the preset thermal runaway risk threshold within a preset time period, the battery module is determined to have a thermal runaway risk.

[0204] It should be noted that, in this embodiment, the current operating data of the battery module is input into a multiphysics coupling dynamics model to determine the current multiphysics distribution of the battery module. This reveals the complex coupling evolution mechanism inside the battery from the perspective of multiphysics coupling. An extended Kalman filter is used to estimate the current estimated state of the battery module to correct the current state of the battery module. Furthermore, an LSTM-based risk trend prediction model is used to predict the estimated state of the battery module at a future preset time. Thus, the fault state of the battery module is predicted based on the estimated state of the battery module. By using multidimensional physical quantities to predict the fault state of the battery module, it is possible to comprehensively and timely make accurate predictions of battery faults, thereby reducing the false alarm rate and false negative rate of battery faults.

[0205] This application provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions, wherein when the program instructions are executed by a computer, the computer performs the following steps:

[0206] Based on the pre-constructed multiphysics coupled dynamics model of the battery module and combined with the current operating data of the battery module, the current multiphysics distribution of the battery module is determined; wherein, the multiphysics coupled dynamics model is obtained by coupling multiple physical fields including electrothermal, stress, structural acoustics and gas dynamics.

[0207] Based on the current multiphysics distribution, the extended Kalman filter is used to estimate the current estimated state of the battery module;

[0208] Based on a pre-trained LSTM risk trend prediction model, combined with the current estimated state of the battery module, the estimated state of the battery module at a future preset time is predicted.

[0209] Predict the fault state of the battery module based on the estimated state of the predicted battery module.

[0210] In some embodiments, the computer also performs:

[0211] Based on the geometric structure of the battery module, construct a geometric model of the battery module;

[0212] Based on the geometric model of the battery module, simulation models of the electrothermal field, stress field, structural acoustic field, and gas dynamic acoustic field are constructed respectively.

[0213] The electrothermal field simulation model, stress field simulation model, structural acoustic field simulation model, and gas dynamic acoustic field simulation model are coupled to obtain a multi-physics field coupled dynamic model.

[0214] In some embodiments, the current multiphysics field distribution includes stress field distribution, temperature field distribution, acoustic energy field distribution, and gas concentration field distribution;

[0215] Based on the current multiphysics distribution, an extended Kalman filter is used to estimate the current estimated state of the battery module, including:

[0216] Based on the current multiphysics field distribution and the measured multiphysics field data of the battery module obtained by the sensor, the state feature vector of the battery module is determined.

[0217] Based on the extended Kalman filter algorithm and combined with the state feature vector of the battery module, the temperature state, stress state, acoustic energy density state and gas concentration of the battery module are determined as the current estimated state.

[0218] In some embodiments, the state feature vector of the battery module is determined based on the current multiphysics distribution and the measured multiphysics data of the battery module obtained by the sensor, including:

[0219] Physical field features were extracted from the stress field distribution, temperature field distribution, acoustic energy field distribution, and gas dynamic field distribution, respectively. Among them, the physical field features include stress gradient, temperature rise rate, acoustic energy density, and gas escape concentration rate.

[0220] Acquire multi-physics field measured data of the battery module by the sensor, and determine the multi-physics field deviation based on the residual between the multi-physics field measured data and the current multi-physics field distribution;

[0221] The state characteristic vector of the battery module is determined based on stress gradient, temperature rise rate, acoustic energy density, gas escape concentration rate, and multi-physics field deviation.

[0222] In some embodiments, predicting the fault state of the battery module based on the estimated state of the predicted battery module includes:

[0223] A dimensionless transformation is performed on the estimated state of the predicted battery module to obtain multiple dimensionless transformation estimated state features.

[0224] By weighting the estimated state characteristics after multiple dimensionless transformations, the failure risk index of the battery module is obtained.

[0225] Compare the magnitudes of the fault risk indicators with the threshold ranges of multiple preset warning levels, and determine the preset warning levels corresponding to the fault risk indicators based on the comparison results.

[0226] The fault status of the battery module is determined based on the preset warning level corresponding to the fault risk indicator.

[0227] In some embodiments, the computer also performs:

[0228] Based on the temperature field distribution and gas dynamic field distribution, the temperature change rate and gas concentration change rate of the battery module are determined;

[0229] The thermal runaway risk indicators of the battery module are determined based on the temperature change rate and gas concentration change rate of the battery module.

[0230] If the thermal runaway risk index continuously exceeds the preset thermal runaway risk threshold within a preset time period, the battery module is determined to have a thermal runaway risk.

[0231] It should be noted that, in this embodiment, the current operating data of the battery module is input into a multiphysics coupling dynamics model to determine the current multiphysics distribution of the battery module. This reveals the complex coupling evolution mechanism inside the battery from the perspective of multiphysics coupling. An extended Kalman filter is used to estimate the current estimated state of the battery module to correct the current state of the battery module. Furthermore, an LSTM-based risk trend prediction model is used to predict the estimated state of the battery module at a future preset time. Thus, the fault state of the battery module is predicted based on the estimated state of the battery module. By using multidimensional physical quantities to predict the fault state of the battery module, it is possible to comprehensively and timely make accurate predictions of battery faults, thereby reducing the false alarm rate and false negative rate of battery faults.

[0232] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, electronic devices, computer storage media, and computer program products described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0233] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention 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 of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises 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 apparatus.

[0234] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0235] In the several embodiments provided by this invention, it should be understood that the disclosed systems, electronic devices, computer storage media, computer program products, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, indirect coupling or communication connection between devices or units, and may be electrical, mechanical, or other forms.

[0236] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0237] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0238] 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 the present invention, 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 for executing all or part of the steps of the methods described in the various embodiments of the present invention through a computer device (which may be a personal computer, a server, or a network device, etc.). The aforementioned storage medium includes: USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, optical disks, and other media capable of storing program code.

[0239] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to 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 the present invention.

Claims

1. A fault prediction method based on multiphysics coupling, characterized in that, include: Based on the pre-constructed multiphysics coupling dynamics model of the battery module and combined with the current operating data of the battery module, the current multiphysics distribution of the battery module is determined; wherein, the multiphysics coupling dynamics model is obtained by coupling multiple physical fields including electrothermal, stress, structural acoustics and gas dynamics. Based on the current multiphysics distribution, an extended Kalman filter is used to estimate the current estimated state of the battery module. Based on a pre-trained LSTM risk trend prediction model, combined with the current estimated state of the battery module, the estimated state of the battery module at a future preset time is predicted. The fault state of the battery module is predicted based on the estimated state of the predicted battery module.

2. The fault prediction method based on multiphysics coupling according to claim 1, characterized in that, Also includes: Based on the geometric structure of the battery module, construct a geometric model of the battery module; Based on the geometric model of the battery module, simulation models of electrothermal field, stress field, structural acoustic field, and gas dynamic acoustic field are constructed respectively. The electrothermal field simulation model, the stress field simulation model, the structural acoustic field simulation model, and the gas dynamic acoustic field simulation model are coupled to obtain the coupled dynamic model of the multiphysics field.

3. The fault prediction method based on multiphysics coupling according to claim 1, characterized in that, The current multiphysics field distribution includes stress field distribution, temperature field distribution, acoustic energy field distribution, and gas concentration field distribution; The step of estimating the current estimated state of the battery module using an extended Kalman filter based on the current multiphysics distribution includes: Based on the current multiphysics field distribution and the measured multiphysics field data of the battery module obtained by the sensor, the state feature vector of the battery module is determined. Based on the extended Kalman filter algorithm and combined with the state feature vector of the battery module, the temperature state, stress state, acoustic energy density state and gas concentration of the battery module are determined as the current estimated state.

4. The fault prediction method based on multiphysics coupling according to claim 3, characterized in that, The step of determining the state feature vector of the battery module based on the current multiphysics distribution and the measured multiphysics data of the battery module obtained by the sensors includes: Physical field features are extracted from the stress field distribution, the temperature field distribution, the acoustic energy field distribution, and the gas dynamic field distribution, respectively; wherein, the physical field features include stress gradient, temperature rise rate, acoustic energy density, and gas escape concentration rate; Acquire multi-physics field measured data of the battery module by the sensor, and determine the multi-physics field deviation based on the residual between the multi-physics field measured data and the current multi-physics field distribution; The state feature vector of the battery module is determined based on the stress gradient, the temperature rise rate, the acoustic energy density, the gas escape concentration rate, and the multiphysics field deviation.

5. The fault prediction method based on multiphysics coupling according to claim 1, characterized in that, The step of predicting the fault state of the battery module based on the estimated state of the predicted battery module includes: A dimensionless transformation is performed on the predicted estimated state of the battery module to obtain multiple dimensionless transformed estimated state features. The estimated state features after multiple dimensionless transformations are weighted to obtain the fault risk index of the battery module. Compare the magnitude of the fault risk index with the threshold values ​​of the fault risk index ranges corresponding to multiple preset warning levels, and determine the preset warning level corresponding to the fault risk index based on the comparison results. The fault status of the battery module is determined based on the preset warning level corresponding to the fault risk indicator.

6. The fault prediction method based on multiphysics coupling according to claim 3, characterized in that, Also includes: Based on the temperature field distribution and the gas dynamic field distribution, the temperature change rate and gas concentration change rate of the battery module are determined; The thermal runaway risk index of the battery module is determined based on the temperature change rate and gas concentration change rate of the battery module. If the thermal runaway risk index continuously exceeds the preset thermal runaway risk threshold within a preset time period, the battery module is determined to have a thermal runaway risk.

7. A fault prediction system based on multiphysics coupling, characterized in that, include: The physics field determination module is used to determine the current multiphysics field distribution of the battery module based on the pre-constructed multiphysics field coupled dynamics model of the battery module and the current operating data of the battery module; wherein, the multiphysics field coupled dynamics model is obtained by coupling multiple physical fields including electrothermal, stress, structural acoustics and gas dynamics; The state estimation module is used to estimate the current estimated state of the battery module based on the current multiphysics distribution using an extended Kalman filter. The state prediction module is used to predict the estimated state of the battery module at a future preset time based on a pre-trained LSTM risk trend prediction model and the current estimated state of the battery module. The fault state prediction module is used to predict the fault state of the battery module based on the estimated state of the predicted battery module.

8. An electronic device, characterized in that, The electronic device includes a memory and a processor. The memory stores a computer program, which, when executed by the processor, causes the processor to perform the steps of the fault prediction method based on multiphysics coupling as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed, it implements the steps of the fault prediction method based on multiphysics coupling as described in any one of claims 1-6.

10. A computer program product, characterized in that, The computer program product includes a computer program stored on a non-transitory computer-readable storage medium, the computer program including program instructions, wherein when the program instructions are executed by a computer, the computer performs the steps of the fault prediction method based on multiphysics coupling as described in any one of claims 1-6.

Images