Modelica power battery early warning method and device based on monitoring data

Abstract

The application relates to the technical field of power batteries, in particular to a Modelica power battery early warning method and device based on monitoring data. The method comprises the following steps: constructing a power battery model based on a Modelica language, wherein the power battery model comprises an electro-thermal coupling submodel, an aging submodel and a thermal runaway submodel; generating a dynamic correction coefficient matrix according to real-time monitoring data of a power battery, and correcting coefficients in the power battery model according to the dynamic correction coefficient matrix; the dynamic correction coefficient matrix comprises a plurality of dynamically changing correction coefficients; inputting real-time monitoring data into the power battery model to obtain performance parameters of the power battery predicted by the power battery model; and analyzing the performance parameters by using an early warning model to obtain whether early warning is needed and an early warning type, so that state prediction and early warning of the power battery can be realized under the condition of limited data.

Description

Technical Field

[0001] The present application relates to the field of power battery technology, and in particular to a Modelica power battery early warning method and device based on monitoring data. Background Art

[0002] Power battery early warning refers to the early detection and prediction of potential failures and safety hazards of power batteries through monitoring the performance degradation of batteries during the aging process and real-time analysis of parameters such as heat and pressure under critical conditions of thermal runaway.

[0003] Traditional methods are mainly based on physical system analysis:

[0004] 1. Temperature sensor monitoring: Temperature sensors installed inside or outside the battery monitor battery temperature changes in real time. When the temperature exceeds a set safety threshold, an alert is triggered or protective measures are taken. While simple, this method can only detect temperature increases after they have already occurred and cannot predict problems in advance.

[0005] 2. Voltage and current monitoring: The battery management system (BMS) monitors the battery's voltage and current in real time. By analyzing abnormal voltage fluctuations or overcharge and over-discharge, the BMS can determine the battery's health and issue warnings. However, this method is slow to predict battery aging and thermal runaway, typically only detecting faults that have already occurred.

[0006] 3. Overvoltage / overcurrent protection circuits: Traditionally, battery packs are equipped with overvoltage and overcurrent protection circuits. When overcharge, overdischarge, or a short circuit is detected, the circuits automatically cut off power to prevent further damage or danger. This is a protective measure, but it only responds immediately after a fault occurs and cannot predict risks in advance.

[0007] Traditional methods mainly rely on post-response mechanisms or external physical protection measures, lack accurate monitoring and prediction of the complex chemical reactions and aging processes inside the battery, have limited early warning capabilities, and often cannot provide sufficient safety protection before a failure occurs.

[0008] In view of this, this application is filed. Summary of the Invention

[0009] The purpose of this application is to provide a Modelica power battery early warning method and device based on monitoring data, so as to realize power battery status prediction and early warning under limited data conditions.

[0010] In order to achieve the above objectives, this application adopts the following technical solutions:

[0011] In a first aspect, the present application provides a Modelica power battery early warning method based on monitoring data, comprising:

[0012] Constructing a power battery model based on the Modelica language, wherein the power battery model includes an electric-thermal coupling sub-model, an aging sub-model, and a thermal runaway sub-model;

[0013] Generating a dynamic correction coefficient matrix according to real-time monitoring data of the power battery, and correcting the coefficients in the power battery model according to the dynamic correction coefficient matrix; the dynamic correction coefficient matrix includes a plurality of dynamically changing correction coefficients;

[0014] Inputting the real-time monitoring data into the power battery model to obtain the performance parameters of the power battery predicted by the power battery model;

[0015] The performance parameters are analyzed using an early warning model to determine whether an early warning should be issued and the type of early warning.

[0016] In a second aspect, the present application provides an electronic device, comprising:

[0017] at least one processor, and a memory communicatively coupled to the at least one processor;

[0018] The memory stores instructions that can be executed by at least one of the processors, and the instructions are executed by at least one of the processors so that the at least one processor can execute the above-mentioned Modelica power battery early warning method based on monitoring data.

[0019] Compared with the prior art, the present invention has the following advantages:

[0020] This application uses the Modelica model to construct an electrothermal coupling sub-model of the power battery model, as well as a thermal runaway and aging sub-model; through the limited data of the monitoring platform, a correction matrix is ​​formed to perform real-time corrections on the main parameters of the power battery model (such as voltage, current, heat generation, etc.), ultimately achieving power battery status prediction and early warning with limited data. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the specific implementation methods of the present application or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the specific implementation methods or the description of the prior art. Obviously, the drawings described below are some implementation methods of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0022] Figure 1This is a flow chart of a Modelica power battery early warning method based on monitoring data provided by this application;

[0023] Figure 2 It is a structural diagram of the electronic device provided in this application. DETAILED DESCRIPTION

[0024] The following description of exemplary embodiments of the present application is made in conjunction with the accompanying drawings, including various details of the embodiments of the present application to facilitate understanding. These details should be considered as merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present application. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0025] The present application is further described in detail below with reference to the embodiments.

[0026] Figure 1 This is a flowchart of a monitoring data-based Modelica power battery early warning method provided in this embodiment. This method can be executed by a computer program and integrated into an electronic device, such as an electronic control unit (ECU), a telematics box (T-box), or the cloud. This embodiment uses a monitoring data-based Modelica power battery early warning method integrated into the ECU, T-box, or cloud to predict power battery failures and provide timely warnings.

[0027] like Figure 1 As shown, this embodiment provides a Modelica power battery early warning method based on monitoring data, including the following steps:

[0028] S110 , constructing a power battery model based on the Modelica language, wherein the power battery model includes an electric-thermal coupling sub-model, an aging sub-model, and a thermal runaway sub-model.

[0029] The power battery model, built using the object-oriented Modelica language, calculates the battery's electrical and thermal characteristics, simulating battery aging and thermal runaway. The model simulates the battery's electrical, thermal, aging, and thermal runaway characteristics based on real-time monitoring data and a dynamic correction coefficient matrix. The electrothermal coupling submodel utilizes common industry practices, such as equivalent circuit models, electrochemical models, and the Bernadi heat generation model. The aging submodel utilizes a semi-mechanistic, semi-empirical model based on calendar life and cycle life. The thermal runaway submodel employs a semi-mechanistic, semi-empirical model tailored to specific conditions, such as puncture, high temperature, and overcharge.

[0030] For example, the aging sub-model uses the calendar life model and the cycle life model based on the Arrhenius formula:

[0031] ;

[0032] in, represents calendar life, A refers to the pre-factor (related to factors such as battery materials and structure, which comprehensively reflects the inherent aging characteristics of the battery without considering the influence of temperature. Different battery systems have different values ​​and can be obtained by fitting a large amount of experimental data), and e is the natural base number. Ea is the activation energy (it represents the energy barrier that needs to be overcome for the aging reaction to occur inside the battery, and its value depends on the type of chemical reaction of the battery, such as electrolyte decomposition, side reactions between electrode materials and electrolytes, etc., and the unit is usually J / mol), R is the molar gas constant, T is the absolute temperature, k is the coefficient related to voltage (its value reflects the extent of the influence of storage voltage on calendar life, which can be determined by fitting experimental data, and the unit is usually V-1), V is voltage, H is humidity, O is oxygen concentration, m and n are coefficients related to humidity and oxygen concentration, respectively (they reflect the sensitivity of humidity and oxygen concentration to battery aging, which also need to be determined through experiments).

[0033] ;

[0034] in, Indicates cycle life. B(C) is a function related to the charge and discharge rate. Generally, as the charge and discharge rate increases, the B(C) value decreases. Its specific form can be obtained by fitting the cycle life test data under different charge and discharge rates. e is the natural base number. Eb It is the activation energy related to cycle aging. Its physical meaning is similar to the activation energy in the calendar life model, but it is aimed at the chemical reaction during the charge and discharge cycle. The unit is J / mol). R is the molar gas constant, T is the absolute temperature, q It is the depth of discharge correlation coefficient, which indicates the degree of influence of deep discharge on cycle life. It is determined by experimental fitting and the unit is usually 1. DOD It is the depth of discharge. fs Is the correction factor of the staged charging strategy. When a reasonable staged charging strategy is adopted fs Greater than 1. r is a comprehensive coefficient that represents the impact of electrode material durability and electrolyte stability on cycle life. The r value is related to factors such as the structural stability of the electrode material and the antioxidant properties of the electrolyte. It can be obtained by fitting battery cycle life test data of different materials and electrolyte systems.

[0035] The key formulas of the thermal runaway sub-model include the overall equation for the side reactions of thermal runaway in power batteries:

[0036] ;

[0037] Where: P is the total power of the side reaction; Psei is the power of the decomposition reaction of the solid electrolyte interface film (i.e., SEI film); Pneg is the power of the reaction between the negative electrode and the electrolyte; Ppos is the power of the reaction between the positive electrode and the electrolyte; Pele is the power of the self-decomposition reaction of the electrolyte. The calculation process of the above-mentioned various powers can be found in the mature formulas in the industry and will not be repeated here.

[0038] The thermal runaway submodel is based on the thermal runaway mechanism of power batteries. By inputting the heat generated by the battery during discharge, self-heating, or short-circuiting, it simulates the triggering and propagation of thermal runaway, enabling prediction and analysis of heat diffusion paths and critical states. The electrothermal coupling submodel, aging submodel, and thermal runaway submodel are coupled together for calculation. Specifically, the aging submodel and thermal runaway submodel are coupled. The decay characteristics calculated by the aging submodel (such as SEI film growth, electrode material degradation, and lithium deposition) all influence the calculations of the thermal runaway submodel. For example, as the SEI film thickens, the internal resistance increases, which affects the battery power in the thermal runaway submodel. The thermal runaway submodel influences the battery aging submodel, and the two are coupled in multiple ways. For example, the temperature affects the aging rate: the temperature calculated by the thermal runaway submodel is one of the main drivers of battery aging. Aging sub-models typically include electrochemical characteristics (such as SEI growth, lithium deposition, and active material loss), while thermal runaway sub-models focus on temperature changes. At excessively high temperatures, lithium deposition is exacerbated, potentially leading to lithium dendrite growth and increasing short-circuit risk. Side reactions induced by thermal runaway (such as redox reactions) can alter the aging pathway.

[0039] In one specific embodiment, the electrothermal coupling sub-model calculates the total heat generation rate of the battery under different operating conditions, including ohmic heat, polarization heat, and electrochemical reaction heat, by inputting current, voltage, ambient temperature, thermal parameters (such as heat capacity, thermal conductivity, and convective heat transfer coefficient), and internal resistance parameters. The output instantaneous heat generation rate and temperature field distribution serve as inputs for the thermal runaway sub-model and the aging sub-model. Driven by temperature, SOC range, current density, and historical heat generation data, the aging sub-model simulates processes such as SEI film growth, lithium deposition, and active material loss. It outputs capacity decay rate, internal resistance growth rate, and thermal stability degradation indicators, which are then fed back to correct the physical parameters of the electrothermal coupling sub-model, such as the first polarization resistance, first polarization capacitance, and effective capacity, forming a self-consistent electro-thermal-aging closed loop. The thermal runaway coupling sub-model determines whether a critical thermal reaction process is triggered based on the temperature field from the electrothermal coupling sub-model and the stability attenuation factor from the aging sub-model. By calculating the hotspot growth rate, spontaneous reaction heat, thermal diffusion rate, etc., it outputs the thermal risk factor and safety status assessment value. At the same time, it reversely adjusts the thermal coupling boundary conditions and aging acceleration factors to construct a power battery model that can be used for thermal safety prediction.

[0040] S120 : Generate a dynamic correction coefficient matrix according to the real-time monitoring data of the power battery, and correct the coefficients in the power battery model according to the dynamic correction coefficient matrix.

[0041] Since the power battery model constructed by the S110 is based on ideal parameters, the power battery will be affected by various factors during actual vehicle use. Therefore, the coefficients in the power battery model need to be corrected based on the actual power battery monitoring data to make the power battery model more representative of the vehicle's actual use process.

[0042] Real-time monitoring data for power batteries includes primary and secondary data. Primary data includes at least battery pack voltage and current, single cell voltage and temperature, battery SOC, battery pack ambient temperature and humidity, and charge and discharge rates. Secondary data includes at least vehicle distance traveled, location information, battery system thermal management system information, and the driver's driving habits.

[0043] This embodiment uses a correction model to generate a dynamic correction coefficient matrix based on real-time monitoring data. "Dynamic" here means that the correction coefficient matrix changes as the monitoring data changes. This dynamic correction coefficient matrix includes multiple correction coefficients used to modify the coefficients in the power battery model. These correction coefficients are multiplied by the corresponding values ​​in the power battery model to correct the model parameters.

[0044] Optionally, the correction algorithm includes the following five cases:

[0045] 1. According to the environmental conditions of the power battery, generate capacity attenuation correction factor, battery ambient temperature coefficient, road condition correction factor, aging correction factor, mileage correction factor and thermal runaway correction factor.

[0046] Specifically, combining positioning information can determine the geographic range in which the vehicle operates. Based on this geographic range, a capacity decay coefficient, battery ambient temperature coefficient, aging correction factor, and thermal runaway correction factor are generated. For example, if a vehicle operates at high latitudes for an extended period, battery degradation will be accelerated. Therefore, the capacity decay correction coefficient is set to be greater than 1. The capacity decay correction coefficient is multiplied by the capacity decay rate in the power battery model to correct the capacity decay rate. If the vehicle operates in high-temperature areas for an extended period, the battery ambient temperature will factor in battery degradation and thermal runaway. Therefore, the battery ambient temperature coefficient is set to be greater than 1. The capacity decay rate is corrected by multiplying the battery ambient temperature coefficient by the capacity decay rate. Accordingly, the aging correction factor and thermal runaway correction factor will also differ from those in other regions. The aging correction factors include at least 1) an internal resistance growth coefficient, which reflects the change in internal resistance with aging and is used to correct the internal resistance growth rate in the power battery model; and 2) a deep discharge correction coefficient, which adjusts for the impact of deep discharge on life and is used to correct the degree of impact of deep discharge on life in the power battery model. The thermal runaway correction factor includes at least: 1) a critical temperature correction factor, which reflects the change in the temperature threshold that triggers thermal runaway and is used to multiply the critical temperature to correct the critical temperature; 2) a heat generation correction factor, which describes the change in the heat generation rate under unit energy output and is used to multiply the heat generation coefficient to correct the heat generation coefficient; 3) a cooling efficiency coefficient, which reflects the impact of the performance of the thermal management system on the risk of thermal runaway of the power battery and is used to multiply the cooling efficiency to correct the cooling efficiency.

[0047] The road condition correction factor measures the impact of different road conditions on power battery energy consumption and is used to correct battery energy consumption. In complex road conditions (for example), the power battery energy consumption is higher, so the road condition correction factor is greater than 1. The road condition correction factor is multiplied by the battery energy consumption to correct energy consumption. In high-speed road conditions, the power battery energy consumption is lower, so the road condition correction factor is less than 1. The road condition correction factor is multiplied by the battery energy consumption to correct energy consumption.

[0048] The mileage correction factor measures the cumulative effect of long-term driving on battery aging. Long cumulative mileage accelerates battery aging, significantly increasing the aging acceleration factor. In this case, the mileage correction factor is greater than 1, and the mileage correction factor is multiplied by the aging acceleration factor to obtain the corrected aging acceleration factor. Short cumulative mileage does not accelerate battery aging, and the aging acceleration factor decreases. In this case, the mileage correction factor is less than 1, and the mileage correction factor is multiplied by the aging acceleration factor to obtain the corrected aging acceleration factor.

[0049] 2. According to the driving habits of the driver of the vehicle equipped with the power battery, a driving habit correction coefficient is generated, including: a battery life attenuation correction coefficient, a capacity attenuation correction coefficient and a speed correction coefficient.

[0050] Among them, the battery life attenuation coefficient is a quantitative index for measuring the decline of battery performance over time or the number of uses, and the battery life attenuation correction coefficient is used to correct the battery life attenuation coefficient.

[0051] Firstly, by feature extraction and analysis of vehicle operation data (such as acceleration, deceleration, speed, fuel consumption, etc.), positioning data (such as road type, slope), and inertial sensor data (such as acceleration, angular velocity), the classification of driving habits is obtained. According to the behavior mode, the driving habits can be divided into aggressive type (frequent sudden acceleration / sudden braking, high-speed driving), conservative type (smooth driving, low-speed driving), and standard type (between aggressive and conservative). Different driving habits will generate different driving habit correction coefficients, which will have different effects on the power battery model. For example, the aggressive driving habit (frequent sudden acceleration and sudden braking) will cause the battery to frequently experience high-rate charging and discharging, increase the electrochemical stress, and accelerate the capacity attenuation, so the battery life attenuation correction coefficient and the capacity attenuation correction coefficient are both large. The conservative driving habit usually keeps the battery in a stable working state, which helps to delay aging, so the battery life attenuation correction coefficient and the capacity attenuation correction coefficient are both 1. The standard driving habit has an impact on the battery between aggressive and conservative, but it may still put pressure on the thermal management system and battery life due to high-speed driving or occasional high-power demand, so the values of the battery life attenuation correction coefficient and the capacity attenuation correction coefficient can be set according to the actual situation.

[0052] The speed correction coefficient is used to represent the influence of vehicle speed on battery temperature. If the driver often drives at high speed, it will speed up the rate of battery temperature rise, so the speed correction coefficient is greater than 1, and the speed correction coefficient is multiplied by the battery temperature rise rate to get the corrected battery temperature rise rate. On the contrary, if the driver often drives at low speed, it will reduce the rate of battery temperature rise, so the speed correction coefficient is less than 1, and the speed correction coefficient is multiplied by the battery temperature rise rate to get the corrected battery temperature rise rate.

[0053] 3. If the charge and discharge rate of the power battery is greater than the set value, a physical parameter correction coefficient of the power battery is generated. If the power battery is discharged and charged at a high rate, the physical parameters of the power battery itself will also be corrected online, and the physical parameter correction coefficient will be multiplied by the corresponding physical parameter to correct the physical parameter. Optionally, the physical parameters include: 1) battery internal resistance, which represents the transient response impedance of the battery; 2) first polarization resistance, which forms the first polarization link with the first polarization capacitance, reflecting the rapid polarization behavior in the battery (such as electrochemical reaction at the electrode interface); 3) first polarization capacitance, which represents the capacitance effect paired with the first polarization resistance; 4) second polarization resistance, which forms the second polarization link with the second polarization capacitance, reflecting the slow polarization behavior in the battery (such as diffusion process, charge migration, etc.); 5) second polarization capacitance, which represents the slowly changing capacitance effect paired with the second polarization resistance, corresponding to the slower electrochemical process inside the battery; 6) open circuit voltage.

[0054] It should be noted that the power battery model contains numerous parameters, many of which require correction. This embodiment describes only some of the correction coefficients. The specific values ​​of the correction coefficients need to be set based on actual conditions. For example, the specific values ​​of the correction coefficients can be determined through regression analysis, principal component analysis, discrete Fourier transform algorithm, clustering algorithm, curve fitting, and other methods.

[0055] For example, real-time monitoring data from power batteries is clustered according to their environment. Within each cluster are real-time monitoring data from the same environment. Based on this data, the impact of battery capacity decay, changes in ambient temperature, road conditions, and mileage on various battery parameters (i.e., correlation analysis) is analyzed. Parameters with high correlation are selected as affected battery parameters. The degree of impact is determined based on actual test data, thereby determining the capacity decay correction factor, battery ambient temperature correction factor, road condition correction factor, mileage correction factor, aging correction factor, and thermal runaway correction factor for each specific environment.

[0056] For example, if the charge / discharge rate of a power battery exceeds a set value, an online identification method is used to determine the physical parameter correction coefficient. First, the resistance-capacitance (RC) model in the power battery model is converted into a discrete state space representation. In power battery modeling, the resistance-capacitance (RC) model is a simplified model widely used to characterize the dynamic voltage response of the battery, often used in state estimation, thermal management control, and digital twins. Converting the RC model into a discrete state space representation facilitates integration with filtering algorithms, enabling online identification of physical parameters. Taking a first-order RC model as an example, state variables may include the voltage across the capacitor and the battery state of charge. The input is current, and the output is terminal voltage. The current physical parameters are then identified using an extended Kalman filter or recursive least squares method. For example, to dynamically update model parameters such as resistance and capacitance, an extended Kalman filter (EKF) or recursive least squares (RLS) method can be introduced for online parameter identification. Taking the EKF as an example, the parameters to be identified and the system state can be combined and expanded into an "augmented state vector." At each moment, error correction is performed based on the input current and the actual measured terminal voltage to estimate the impedance parameters at that moment. When a system experiences thermal decay or aging, parameters such as resistance will change accordingly, and the EKF can effectively capture their time-varying characteristics. After identifying the current and calibrated physical parameters, a physical parameter correction coefficient is derived based on these two parameters. The calibrated physical parameters are the parameters when the power battery's charge and discharge rate is less than or equal to the set value. The correction coefficient can be the ratio of the current and calibrated physical parameters. Therefore, to correct the coefficients in the power battery model, simply multiply the calibrated physical parameters by the correction coefficient. This coefficient can be fed back into the thermal runaway or aging sub-model to correct the heat-generated power or internal resistance power consumption in real time. Similar logic can be extended to the dynamic identification and correction of capacitance parameters or other parameters, thereby constructing a battery lifecycle simulation model with self-updating capabilities.

[0057] 4. Optionally, a multi-factor coupling correction coefficient is generated based on the coupling influence between the real-time monitoring data. The multi-factor coupling correction coefficient is used to take into account the coupling influence of factors such as ambient temperature, charge and discharge rate, SOC, SOH, aging degree, etc., and reflects the coordinated changes in the conditions for thermal runaway under the joint action of these factors. The multi-factor coupling correction coefficient is used to correct the trigger threshold parameters (such as starting temperature, critical temperature rise rate, reaction rate constant) in the thermal runaway sub-model. In a specific embodiment, the multi-factor coupling correction coefficient is usually established through multi-factor sensitivity analysis or experimental / data-driven methods. For example, a multivariable nonlinear function can be constructed, the function value is the multi-factor coupling correction coefficient, and the independent variable is each real-time monitoring data. The function is fitted using experimental data or simulation results, such as using a neural network, response surface method, polynomial regression, etc.

[0058] 5. Optionally, a time dimension correction factor is generated based on the battery operating time / usage cycle in real-time monitoring data. This time dimension correction factor characterizes trends in battery performance degradation and thermal stability during long-term operation, correcting time-dependent parameters in the model to reflect the impact of battery aging on thermal safety. The time dimension correction factor is used to correct for changes in heat generation factors (such as internal resistance and electrochemical reaction heat source terms) or life-related functions (such as timing function parameters in the thermal runaway probability model) caused by battery aging. The value of the time dimension correction factor can be calibrated based on the battery's operating time. In one specific embodiment, the time dimension correction factor is primarily used to reflect the impact of time / usage cycle on the thermal runaway model. For example, based on measured / accelerated aging test data, the time dimension correction factor of the power battery model is fitted to the trend of heat generation factors over time, such as a curve showing the measured internal resistance increasing with the number of cycles. The measured internal resistance is then divided by the calibrated internal resistance to obtain the time dimension correction factor for the current battery operating time / usage cycle.

[0059] S130: Input the real-time monitoring data into a power battery model to obtain performance parameters of the power battery predicted by the power battery model.

[0060] At the start of the power battery model's operation, the electrothermal coupling sub-model couples the power battery voltage based on the initial state of charge (SOC), initial operating current, and initial ambient temperature. The initial voltage, current, SOC, and temperature are then fed into the aging and thermal runaway sub-models for corresponding calculations. For example, the aging sub-model calculates the current state of health (SOH), while the thermal runaway sub-model corrects the battery's heat generation. After the initial calculations are complete, the power battery model continuously receives ambient temperature and current parameters, which are then internally combined with a dynamic correction coefficient matrix to correct the power battery's physical parameters, such as resistance and capacitance, as well as the capacity decay coefficient, internal resistance growth, and life decay coefficient (see above for details). Simultaneously, the correction model continuously reads real-time monitoring data to dynamically adjust the correction coefficients. These correction coefficients are fed into the power battery model in real time for online correction, continuously iterating the model's parameters. Performance parameters calculated by the power battery model include, but are not limited to, battery voltage, battery charge and discharge current, battery temperature, battery state of charge (SOC), and battery state of health (SOH). Among them, the battery SOH needs to be calculated based on battery parameters, such as battery voltage, current and SOC, which will affect the battery SOH.

[0061] S140: Analyze the performance parameters using an early warning model to determine whether an early warning should be issued and the type of early warning.

[0062] Optionally, the early warning model includes: a neural network model, a gradient early warning model, a data threshold model, and a frequency domain model. An appropriate early warning model is selected based on performance parameters.

[0063] The performance parameters are input into the neural network model to obtain the output of the neural network model, indicating whether an early warning is issued and the type of early warning.

[0064] Performance parameters are input into the gradient warning model, which calculates the gradient change of the performance parameters. If the gradient change exceeds a set threshold, a warning is issued and the warning type is determined. For example, if the performance parameter is the battery ambient temperature, the gradient warning model calculates the gradient of the battery ambient temperature and identifies vehicles with large temperature changes. The warning type is battery pack thermal balancing failure or localized overtemperature.

[0065] The performance parameters are input into the data threshold model, which processes the performance parameters and compares them with the set threshold. If the set threshold is exceeded, an early warning is issued and the warning type is determined. For example, if the performance parameter is the battery ambient temperature, the battery ambient temperature is compared with the temperature threshold through the data threshold model. If the temperature threshold is exceeded, the early warning type is a thermal management anomaly. For example, the insulation resistance output by the power battery model is obtained, and the sudden drop value of the insulation resistance is calculated using the data threshold model. If the sudden drop value is greater than the threshold, it is considered that the insulation resistance value is decreasing too quickly, and a battery external short circuit fault or charging pile abnormality has occurred.

[0066] The performance parameters are input into a frequency-domain model, which then performs a Fourier transform or wavelet transform on performance parameters such as battery voltage, charge / discharge current, and temperature. This extracts the frequency characteristics of these parameters and identifies abnormal frequency components or energy distribution changes to determine whether the power battery is experiencing potential failures or thermal runaway risks. This method is suitable for detecting potential hazards such as periodic fluctuations and small disturbances that are difficult to identify in the time domain.

[0067] The above-mentioned different early warning models can monitor different performance parameters.

[0068] Optionally, before using the early warning model to analyze the performance parameters and determine whether to issue an early warning and the type of early warning, it also includes: correcting the parameters in the early warning model based on real-time monitoring data, specifically, correcting the parameters in the neural network, and correcting the thresholds in the gradient early warning model, data threshold model and frequency domain model.

[0069] For example, the ambient temperature at the current location is determined based on the location information in real-time monitoring data. If the ambient temperature at the current location exceeds a set threshold (e.g., 40 degrees Celsius), indicating a high temperature environment, the risk of thermal runaway increases. To provide a timely warning, the temperature threshold in the data threshold model is lowered, thereby detecting thermal management anomalies in advance. If the warning model is a neural network model, it must be pre-trained using training samples. The training samples include: the actual performance parameters of the power battery and labels (including no warning, warning, and warning type). Warning types include but are not limited to internal cell short circuit, poor cell consistency, external battery short circuit, charging pile anomaly, and thermal management anomaly.

[0070] In actual operation, since there are multiple warning models, and the corresponding warning types are obtained based on different performance parameters and different model algorithms. This embodiment adjusts the external weights of the warning models according to the actual monitoring data to reflect the importance of different warning types. Optionally, after obtaining whether to issue a warning and the warning type, it also includes: obtaining the external weights of each warning model according to the real-time monitoring data; determining the type of thermal runaway warning based on the external weights of the warning models and the outputs of each warning model. For example, according to the real-time monitoring data, if it is found that the vehicle has been running in a high-temperature area for a long time, the external weights of the following warning models will be increased:

[0071] 1) A thermal runaway risk model estimates thermal diffusivity using the flash diffusion method based on battery performance parameters. Risk levels are categorized based on severity, time (duration of thermal runaway), and probability (frequency of occurrence). For example, high severity + high probability corresponds to extremely high risk. High temperatures are a typical trigger for thermal runaway, emphasizing the thermal diffusion and superposition mechanisms.

[0072] 2) The insulation resistance fluctuation model calculates the insulation resistance of the battery based on the battery's SOC, temperature, voltage, and current, and determines the magnitude of the insulation resistance change over time. High temperatures can accelerate insulation aging, causing insulation performance degradation.

[0073] 3) The cell consistency fluctuation model is used to calculate the voltage and current consistency of each cell based on battery SOC and voltage data. High temperatures may cause inconsistent thermal responses of different cells, exacerbating differences between cells.

[0074] 4) The temperature gradual change process model is used to combine multi-source data such as ambient temperature, battery SOC, current, voltage, and capacity to analyze the slow temperature increase trend caused by heat accumulation under long-term high-temperature conditions, identify potential thermal runaway risks in advance, and provide trend warnings for the abnormal temperature evolution process. Long-term high temperature will cause slow temperature accumulation, requiring early warning of thermal trends.

[0075] 5) A large temperature fluctuation model is used to calculate the temperature fluctuations over time based on the battery temperature. Abnormal thermal management control strategies under high temperature fluctuations can easily lead to safety hazards.

[0076] 6. The cell voltage slow-change model is used to identify voltage response hysteresis, flatness, or atypical fluctuations caused by high temperature, aging, or abnormal operating conditions based on the battery SOC, voltage, and ambient temperature, thereby detecting potential cell anomalies or thermal runaway risks in advance. High temperatures affect the electrochemical processes of the cell, which may manifest as slow or abnormal voltage changes. This model is required for dynamic trend identification and anomaly judgment. High temperatures affect the electrochemical processes of the cell, which may manifest as slow or abnormal voltage changes.

[0077] For example, if real-time monitoring data reveals that the vehicle's mileage exceeds a set value, the battery aging risk is high and the battery life is severely reduced. In this case, the external weights of the following early warning models for aging and life will increase:

[0078] 1) The significant capacity degradation model analyzes the consistency of capacity between cells and whether the overall capacity is rapidly decaying or experiencing abnormal jumps based on the SOC, capacity, SOH and other data of the battery pack and each cell. It identifies sudden capacity drops caused by manufacturing defects, accelerated aging or abnormal operating conditions, and is used to assess whether the available energy of the battery system is below the safety threshold, thereby achieving rapid early warning of significant performance degradation or failure risks.

[0079] 2) A single cell capacity degradation model analyzes the capacity degradation trends and inconsistencies of certain cells relative to other cells based on the SOC, capacity, and SOH data of each cell in the battery pack, identifying localized aging, cell imbalance, or potential failure risks. This model can be used to monitor the presence of "short cells" in the battery system, proactively identifying system performance bottlenecks caused by cell degradation, and ensuring overall pack energy utilization and operational safety.

[0080] 3) Cell consistency trend model: Based on the SOC, voltage, capacity, SOH and other data of each cell in the battery pack, it analyzes the consistency change trend of key parameters between cells and identifies whether there is a gradually increasing imbalance. This model is used to monitor the dynamic process of the difference between cells over time and determine whether there is a potential thermal runaway cause or performance bottleneck.

[0081] 4) The cell capacity slow-change model is used to analyze and detect the slow attenuation of capacity under normal use conditions based on data such as SOC, voltage change, and capacity change. By analyzing the subtle change trends in charge and discharge capacity, it can identify early performance degradation signals and evaluate changes in battery life status and health level.

[0082] After the above analysis, the warning types output by the warning model are sorted in descending order of external weights, and the warning types in the top N (natural numbers) positions are fed back to the warning platform.

[0083] Based on the above embodiments, this application has the following technical effects:

[0084] 1. This application uses the Modelica model to implement the electric-thermal coupling of the battery model. Combined with the thermal runaway and aging models, a matrix of correction coefficients is formed based on limited monitoring data. The main parameters of the power battery model (such as voltage, current, and heat generation) are iterated and corrected, ultimately achieving battery early warning and prediction with limited data.

[0085] 2. The revised model can further improve the accuracy of thermal runaway prediction and aging characteristics simulation, and enhance the credibility of simulation.

[0086] 3. This application uses limited monitoring data as input, combined with a power battery model to calculate and predict battery performance parameters. Combined with the built-in prediction model, it provides intelligent early warning of thermal runaway and aging trends, thereby improving the safety and life management capabilities of the power battery system.

[0087] like Figure 2 As shown, this embodiment provides an electronic device, including:

[0088] at least one processor; and

[0089] a memory communicatively connected to at least one processor; wherein,

[0090] The memory stores instructions executable by at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the above method. The at least one processor in the electronic device is capable of performing the above method, thereby having at least the same advantages as the above method.

[0091] Optionally, the electronic device also includes interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. The various components are connected to each other using different buses and can be installed on a common motherboard or installed in other ways as needed. The processor can process instructions executed in the electronic device, including instructions stored in or on the memory to display graphical information of a GUI (Graphical User Interface) on an external input / output device (such as a display device coupled to the interface). In other embodiments, if necessary, multiple processors can be used with multiple memories, and / or multiple buses can be used with multiple memories. Similarly, multiple electronic devices can be connected (for example, as a server array, a group of blade servers, or a multi-processor system), with each device providing part of the necessary operations. Figure 2 A processor 301 is taken as an example.

[0092] Memory 302, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the monitoring data-based Modelica power battery early warning method in the embodiments of this application. Processor 301 executes the software programs, instructions, and modules stored in memory 302 to execute various functional applications and data processing of the device, thereby implementing the aforementioned monitoring data-based Modelica power battery early warning method.

[0093] The memory 302 may primarily include a program storage area and a data storage area. The program storage area may store an operating system and at least one application required for a function; the data storage area may store data created based on the use of the terminal, etc. Furthermore, the memory 302 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 302 may further include a memory remotely located relative to the processor 301, and these remote memories may be connected to the device via a network. Examples of the aforementioned networks include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0094] The electronic device may further include: an input device 303 and an output device 304. The processor 301, the memory 302, the input device 303 and the output device 304 may be connected via a bus or other means. Figure 2 The bus connection is taken as an example.

[0095] The input device 303 can receive input digital or character information, and the output device 304 can include a display device, an auxiliary lighting device (e.g., an LED), a tactile feedback device (e.g., a vibration motor), etc. The display device may include, but is not limited to, a liquid crystal display (LCD), a light-emitting diode (LED) display, and a plasma display. In some embodiments, the display device may be a touch screen.

[0096] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this application can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in this application can be achieved. This is not a limitation herein.

[0097] The above specific embodiments do not limit the scope of protection of this application. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application shall be included within the scope of protection of this application.

Claims

1. A Modelica power battery early warning method based on monitoring data, characterized in that: include: Constructing a power battery model based on the Modelica language, wherein the power battery model includes an electric-thermal coupling sub-model, an aging sub-model, and a thermal runaway sub-model; generating a dynamic correction coefficient matrix based on real-time monitoring data of the power battery, and correcting coefficients in the power battery model based on the dynamic correction coefficient matrix; the dynamic correction coefficient matrix includes a plurality of dynamically changing correction coefficients; the correction coefficients are to be multiplied by corresponding values ​​in the power battery model to achieve the purpose of correcting model parameters; Inputting the real-time monitoring data into the power battery model to obtain the performance parameters of the power battery predicted by the power battery model; Analyze the performance parameters using an early warning model to determine whether to issue an early warning and the type of early warning; Among them, the dynamic correction coefficient matrix is ​​generated according to the real-time monitoring data of the power battery, including: Generate capacity attenuation correction factor, battery ambient temperature factor, road condition correction factor, mileage correction factor, aging correction factor and thermal runaway correction factor according to the environment of the power battery; Generate a driving habit correction factor based on the driving habits of the driver of the vehicle carrying the power battery, the driving habit correction factor including a battery life attenuation correction factor, a capacity attenuation correction factor, and a speed correction factor; If the charge and discharge rate of the power battery is greater than a set value, generating a physical parameter correction coefficient of the power battery; generating a multi-factor coupling correction coefficient according to the coupling influence between the real-time monitoring data; A time dimension correction coefficient is generated according to the battery operation time or usage cycle in the real-time monitoring data.

2. The Modelica power battery early warning method based on monitoring data according to claim 1, characterized in that: Generating a physical parameter correction coefficient of the power battery includes: Convert the resistance and capacitance model in the power battery model into a discrete state space form; Identify current physical parameters using extended Kalman filtering or recursive least squares method; Obtaining a physical parameter correction coefficient based on the current physical parameters and the calibrated physical parameters; The calibrated physical parameters are parameters when the charge and discharge rate of the power battery is less than or equal to the set value.

3. The Modelica power battery early warning method based on monitoring data according to claim 1, characterized in that: The performance parameters include: battery voltage, battery charge and discharge current, battery temperature, battery charge state and battery health state.

4. The Modelica power battery early warning method based on monitoring data according to claim 3 is characterized in that: The early warning model includes: a neural network model, a gradient early warning model, a data threshold model and a frequency domain model.

5. The Modelica power battery early warning method based on monitoring data according to claim 4 is characterized in that: Before analyzing the performance parameters using the early warning model to determine whether to issue an early warning and the type of early warning, the following steps are also included: The parameters in the early warning model are modified according to the real-time monitoring data.

6. The Modelica power battery early warning method based on monitoring data according to claim 5, characterized in that: After obtaining whether an early warning is issued and the type of early warning, the following is also included: Obtaining external weights of each early warning model based on the real-time monitoring data; The type of thermal runaway warning is determined according to the external weights of the warning models and the outputs of each warning model.

7. An electronic device, characterized in that: include: at least one processor, and a memory communicatively coupled to the at least one processor; The memory stores instructions that can be executed by at least one of the processors, and the instructions are executed by at least one of the processors so that at least one of the processors can execute the Modelica power battery early warning method based on monitoring data according to any one of claims 1 to 6.

Images