A battery health prediction method, system, storage medium and program product

Abstract

A battery health prediction method, system, storage medium and program product, in which the driving data of the target battery pack is obtained; a microscopic electrochemical stress model is established based on instantaneous operating condition data, a mesoscopic stress evolution model is constructed based on single trip data, and a macroscopic stress accumulation model is generated based on long-term use data; the microscopic electrochemical stress model, the mesoscopic stress evolution model and the macroscopic stress accumulation model are integrated to form an electrochemical stress field; a stress trigger model is constructed based on the electrochemical stress field; a battery performance compensation equation is established according to the output result of the material response function; the current driving data is input into the stress trigger model to obtain the real-time stress field distribution state; the stress field distribution state is input into the battery performance compensation equation to obtain the updated driving strategy constraint condition; the battery health state is predicted according to the updated driving strategy constraint condition combined with the historical stress accumulation data. This application improves the accuracy of battery health prediction.

Description

Technical Field

[0001] The present application belongs to the field of electrical digital data processing, and in particular, relates to a battery health prediction method, system, storage medium and program product. Background Art

[0002] With the rapid development of new energy vehicles, the health status of batteries as core power components directly affects the service life and safety performance of vehicles. Traditional battery health detection methods mainly rely on static test data in laboratory environments, which cannot accurately reflect the dynamic health status changes of batteries during actual road driving. In addition, there is a large deviation between the detection results and the actual usage scenarios, and it cannot provide users with effective battery usage recommendations.

[0003] In related technologies, the current health of the target battery pack can be obtained, combined with the user's driving route and driving habits to generate driving data, and the trained battery health prediction model can be used to predict the future health of the battery, and provide users with health loss prediction reports and optimized driving suggestions. This method uses a recursive neural network to extract and predict features of time series data, realizing dynamic evaluation of the battery health status.

[0004] However, when modeling the battery health status, this method regards driving habits as independent external input variables and ignores the feedback regulation of driving habits on the internal aging mechanism of the battery. This one-way causal relationship assumption makes it difficult for the model to capture the dynamic coupling effect between driving behavior and battery performance, resulting in functional degradation in long-term predictions. The prediction results will gradually deviate from the actual state over time, and the accuracy of battery health predictions will gradually decrease. Summary of the invention

[0005] The present application provides a battery health prediction method, system, storage medium and program product for improving the accuracy of battery health prediction.

[0006] In a first aspect, the present application provides a battery health prediction method, which obtains driving data of a target battery pack at different times, wherein the driving data includes micro-scale instantaneous operating condition data, meso-scale single trip data, and macro-scale long-term use data, wherein the instantaneous operating condition data includes instantaneous current, voltage, and temperature of the target battery pack, the single trip data includes charge and discharge depth and driving conditions, and the long-term use data includes historical cycle times and capacity changes;

[0007] A microscopic electrochemical stress model is established based on instantaneous operating data, a mesoscopic stress evolution model is constructed based on single-trip data, and a macroscopic stress accumulation model is generated based on long-term use data;

[0008] The microscopic electrochemical stress model, the macroscopic stress evolution model and the macroscopic stress accumulation model are integrated to form an electrochemical stress field;

[0009] A stress trigger model is constructed based on the electrochemical stress field. The stress trigger model includes a stress threshold function and a material response function. The stress threshold function is used to characterize the stress field changes caused by different driving behaviors, and the material response function is used to describe the evolution law of the material microstructure caused by the stress field changes.

[0010] A battery performance compensation equation is established based on the output result of the material response function. The battery performance compensation equation is a function that can adjust the driving strategy constraint conditions. The driving strategy constraint conditions include the maximum charge and discharge rate, the operating temperature range, and the charge and discharge depth limit.

[0011] Input the current driving data into the stress trigger model to obtain the real-time stress field distribution state;

[0012] Input the stress field distribution state into the battery performance compensation equation to obtain the updated driving strategy constraint conditions;

[0013] The battery health status is predicted based on the updated driving strategy constraints combined with historical stress accumulation data.

[0014] By adopting the above technical solution, three levels of stress models, micro, meso and macro, are constructed by acquiring driving data at different time scales, and an electrochemical stress field is integrated to fully reflect the stress change law of the battery during use. The stress trigger model constructed based on this stress field contains stress threshold function and material response function, which can accurately describe the relationship between driving behavior and material microstructure evolution. The stress field state is converted into driving strategy constraints through the battery performance compensation equation, and the mapping of stress state to usage strategy is realized, so that the system can capture key information of battery performance degradation in time, dynamically adjust the usage strategy, and reduce the risk of accelerated degradation caused by unreasonable use. Taking into account the stress response mechanism at the material level, the prediction results have stronger physical meaning and interpretability, and improve the accuracy of battery health prediction.

[0015] In conjunction with some embodiments of the first aspect, in some embodiments, a microscopic electrochemical stress model is established based on instantaneous operating condition data, specifically including:

[0016] Calculate ion concentration gradient and potential distribution based on instantaneous current and voltage;

[0017] The diffusion stress field equation is established based on the ion concentration gradient, and the electrochemical stress field equation is established based on the potential distribution;

[0018] The diffusion stress field equation is coupled with the electrochemical stress field equation to obtain a microscopic electrochemical stress model.

[0019] By adopting the above technical solution, the diffusion stress field equation and the electrochemical stress field equation are established by calculating the ion concentration gradient and the potential distribution, and the two equations are coupled to obtain a microscopic electrochemical stress model, which can accurately reflect the stress distribution state inside the battery. Since the model is based on the microscopic electrochemical mechanism, it can directly reflect the stress evolution process at the material level, reducing the information loss that may be caused by relying solely on macroscopic parameters. The physical rationality of the model is improved, making the stress prediction results closer to the actual state inside the battery, providing more reliable data support for subsequent health status assessments.

[0020] In conjunction with some embodiments of the first aspect, in some embodiments, constructing a stress trigger model based on the electrochemical stress field specifically includes:

[0021] Calculate the principal stress components and shear stress components in the electrochemical stress field;

[0022] The material structure evolution function is established based on the principal stress components; the interface stress transfer function is constructed based on the shear stress components;

[0023] The material structure evolution function and the interface stress transfer function are combined to form a stress trigger model.

[0024] By adopting the above technical solution, by calculating the principal stress component and shear stress component in the electrochemical stress field, the material structure evolution function and the interface stress transfer function are established respectively, and the two functions are combined to form a stress trigger model, which can fully reflect the evolution law of material properties caused by changes in the stress field. By modeling the effects of principal stress and shear stress separately, the mechanism of action of different types of stress on material properties can be more accurately described, improving the model's ability to characterize the material degradation process, enabling the system to more accurately predict the performance degradation caused by stress changes, and improving the accuracy of health status assessment.

[0025] In conjunction with some embodiments of the first aspect, in some embodiments, the battery performance compensation equation is: In the equation, is the battery performance compensation coefficient, is the initial compensation coefficient, is the comprehensive damage function, is the time variable, is the temperature correction coefficient, is the integration variable.

[0026] By adopting the above technical solution, the cumulative effect of performance degradation is reflected in the form of an exponential function. The comprehensive damage function in the equation includes the coupling effect of various stress factors, and the temperature correction coefficient reflects the regulatory effect of ambient temperature on performance. This mathematical model design takes into account both the cumulative characteristics of performance degradation and the immediate impact of temperature, and can accurately describe the change of battery performance over time and temperature. Since the output of the compensation equation is limited to between 0 and 1, excessive fluctuations in the compensation coefficient are avoided, ensuring the rationality of the compensation result. This mathematical expression based on physical meaning improves the accuracy of the compensation effect and makes the adjustment of driving strategy more reasonable and reliable.

[0027] In combination with some embodiments of the first aspect, in some embodiments, the battery health state is predicted based on the updated driving strategy constraint conditions combined with the historical stress accumulation data, specifically including:

[0028] Convert the updated driving strategy constraints into a time series feature data matrix, and convert the historical stress accumulation data into a structural feature data matrix;

[0029] Perform weighted sum operation on the time series feature data matrix and the structural feature data matrix to obtain a fused feature vector;

[0030] Calculate the principal components of the fused feature vectors to obtain the eigenvalue sequence after dimensionality reduction;

[0031] Calculate the health status probability distribution value according to the characteristic value sequence;

[0032] Compare the health status probability distribution value with the preset health status classification standard to obtain the health level determination result;

[0033] Calculate the estimated remaining life span based on the health level determination result combined with the updated driving strategy constraints;

[0034] The health level determination results and remaining life estimation are combined to predict the battery health status.

[0035] By adopting the above technical solution, the updated driving strategy constraints are converted into a time series feature data matrix, the historical stress accumulation data is converted into a structural feature data matrix, the two matrices are weighted summed to obtain a fused feature vector, the principal component of the fused feature vector is calculated to obtain the eigenvalue sequence after dimension reduction, the health state probability distribution value is calculated according to the eigenvalue sequence and compared with the preset health state classification standard to obtain the health level determination result, the remaining life estimation value is calculated in combination with the updated driving strategy constraints, and finally the technical solution of predicting the battery health state by combining the health level determination result and the remaining life estimation value is realized. The effective fusion of time series features and structural features is realized, the key feature information is extracted through dimensionality reduction processing, and the interference of redundant features on the prediction accuracy is reduced. The health state determination method based on probability distribution improves the reliability of the prediction results and reduces the possible misjudgment caused by single indicator judgment. The prediction method combining health level determination and remaining life estimation not only gives a qualitative evaluation of the current health state of the battery, but also provides a quantitative life prediction result, making the prediction result more comprehensive and accurate, and improving the accuracy and reliability of the prediction.

[0036] In combination with some embodiments of the first aspect, in some embodiments, after predicting the battery health state according to the updated driving strategy constraint combined with the historical stress accumulation data, the method further includes:

[0037] Calculate the optimal operating range based on the battery health status;

[0038] Generate regional division boundary conditions based on the optimal working range;

[0039] Power allocation is optimized under boundary conditions to obtain power output reference values ​​and power allocation weight coefficients for each time period.

[0040] By adopting the above technical solution, by calculating the optimal working range according to the battery health status, generating area division boundary conditions based on the optimal working range, and optimizing power allocation under the boundary conditions to obtain the power output reference value and power allocation weight coefficient for each time period, the battery is always operated in the working range that best suits its current health status, reducing the impact of unreasonable working conditions on battery life.

[0041] In combination with some embodiments of the first aspect, in some embodiments, power allocation optimization is performed under boundary conditions to obtain a power output reference value and a power allocation weight coefficient for each time period, specifically including:

[0042] Calculate the health status decay rate of each time period, and determine the weight coefficient of each time period according to the health status decay rate;

[0043] Calculate the power upper limit threshold of each time period based on the weight coefficient;

[0044] Combining the weight coefficient and the power upper limit threshold to generate a power allocation sequence;

[0045] The power reference value and allocation weight of each time period are output according to the power allocation sequence.

[0046] By adopting the above technical scheme, the weight coefficient is determined by calculating the health state decay rate of each time period, the power upper limit threshold is calculated based on the weight coefficient, and the power allocation sequence is generated by combining the weight coefficient and the power upper limit threshold. The power reference value and allocation weight of each time period are output according to the power allocation sequence. A quantitative relationship between the health state decay rate and power allocation is established, making the power allocation more accurate and reasonable.

[0047] In a second aspect, an embodiment of the present application provides a battery health prediction system, which includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and one or more processors call the computer instructions to enable the system to execute the method described in the first aspect and any possible implementation method of the first aspect.

[0048] In a third aspect, an embodiment of the present application provides a computer-readable storage medium, comprising instructions, which, when executed on a system, causes the system to execute the method described in the first aspect and any possible implementation of the first aspect.

[0049] In a fourth aspect, an embodiment of the present application provides a computer program product. When the computer program product runs on a system, the system executes the method described in any possible implementation manner in the first aspect.

[0050] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0051] 1. The present application provides a battery health prediction method, which constructs stress models at the micro, meso and macro levels by acquiring driving data at different time scales, and integrates them to form an electrochemical stress field, which can fully reflect the stress change law of the battery during use. The stress trigger model constructed based on this stress field includes a stress threshold function and a material response function, which can accurately describe the relationship between driving behavior and the evolution of the microstructure of the material. The stress field state is converted into a driving strategy constraint through the battery performance compensation equation, and the mapping of stress state to usage strategy is realized, so that the system can capture the key information of battery performance degradation in time, dynamically adjust the usage strategy, and reduce the risk of accelerated degradation caused by unreasonable use. Taking into account the stress response mechanism at the material level, the prediction results have stronger physical meaning and interpretability, and improve the accuracy of battery health prediction.

[0052] 2. The present application provides a battery health prediction method, which reflects the cumulative effect of performance degradation in the form of an exponential function. The comprehensive damage function in the equation includes the coupling effect of various stress factors, and the temperature correction coefficient reflects the regulatory effect of ambient temperature on performance. This mathematical model design not only takes into account the cumulative characteristics of performance degradation, but also includes the immediate impact of temperature, and can accurately describe the change pattern of battery performance over time and temperature. Since the output of the compensation equation is limited to between 0 and 1, excessive fluctuations in the compensation coefficient are avoided, ensuring the rationality of the compensation result. This mathematical expression based on physical meaning improves the accuracy of the compensation effect and makes the adjustment of driving strategy more reasonable and reliable.

[0053] 3. The present application provides a battery health prediction method, which converts the updated driving strategy constraints into a time series feature data matrix, converts the historical stress accumulation data into a structural feature data matrix, performs a weighted sum operation on the two matrices to obtain a fused feature vector, calculates the principal component of the fused feature vector to obtain a reduced eigenvalue sequence, calculates the health state probability distribution value according to the eigenvalue sequence and compares it with the preset health state classification standard to obtain a health level determination result, calculates the remaining life estimate value in combination with the updated driving strategy constraints, and finally predicts the battery health state by combining the health level determination result and the remaining life estimate value. The technical solution realizes the effective fusion of time series features and structural features, extracts key feature information through dimensionality reduction processing, and reduces the interference of redundant features on prediction accuracy. The health state determination method based on probability distribution improves the reliability of the prediction results and reduces the possible misjudgment caused by single indicator judgment. The prediction method that combines health level determination and remaining life estimation not only gives a qualitative evaluation of the current health state of the battery, but also provides a quantitative life prediction result, making the prediction result more comprehensive and accurate, and improving the accuracy and reliability of the prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is a flow chart of a battery health prediction method in an embodiment of the present application.

[0055] Figure 2 It is another flowchart of a battery health prediction method in an embodiment of the present application.

[0056] Figure 3 It is a schematic diagram of the physical device structure of a battery health prediction system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0057] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to be used as limitations to the present application. As used in the specification and appended claims of the present application, the singular expressions "one", "a kind of", "said", "above", "the" and "this" are intended to also include plural expressions, unless there is a clear indication to the contrary in the context. It should also be understood that the term "and / or" used in the present application refers to any or all possible combinations comprising one or more listed items.

[0058] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as suggesting or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, unless otherwise specified, "plurality" means two or more.

[0059] The following uses an embodiment and combines Figure 1 , a battery health prediction method in an embodiment of the present application is described:

[0060] See also Figure 1 , is a flow chart of a battery health prediction method in an embodiment of the present application.

[0061] S101, obtaining driving data of the target battery pack at different times, establishing a microscopic electrochemical stress model based on instantaneous operating condition data, constructing a mesoscopic stress evolution model based on single trip data, and generating a macroscopic stress accumulation model based on long-term use data;

[0062] The system obtains the driving data of the target battery pack at different times. The driving data includes micro-scale instantaneous operating condition data, meso-scale single trip data and macro-scale long-term use data. The instantaneous operating condition data includes the instantaneous current, voltage and temperature of the target battery pack, the single trip data includes the charge and discharge depth and driving conditions, and the long-term use data includes the historical number of cycles and capacity changes. A micro-electrochemical stress model is established based on the instantaneous operating condition data, a meso-stress evolution model is constructed based on the single trip data, and a macro-stress accumulation model is generated based on the long-term use data.

[0063] Among them, a microscopic electrochemical stress model is established based on the instantaneous operating condition data, specifically including: calculating the ion concentration gradient and potential distribution according to the instantaneous current and voltage; establishing the diffusion stress field equation based on the ion concentration gradient, and establishing the electrochemical stress field equation based on the potential distribution; coupling the diffusion stress field equation with the electrochemical stress field equation to obtain the microscopic electrochemical stress model.

[0064] This step is the data basis for battery health prediction. The system needs to comprehensively collect various driving condition parameters of the battery pack during actual use, covering the microscopic instantaneous state, the mesoscopic single-trip characteristics and the macroscopic long-term use history. This multi-scale data fusion helps to depict the overall picture of battery stress evolution. During data collection, the system can measure instantaneous quantities such as current, voltage, and temperature in real time through on-board sensors, obtain single-trip information such as charge and discharge depth and driving conditions through the driving recorder, and read long-term data such as cycle number and capacity change from the battery management system (BMS).

[0065] Based on the collected data, the system can construct stress models that describe different physical and chemical processes. The microscopic electrochemical stress model focuses on the ion diffusion and potential distribution inside the battery, and uses partial differential equations to describe the stress field caused by concentration gradients and electric field changes. The mesoscopic stress evolution model focuses on stress fluctuations within a single trip, and analyzes the effects of charge and discharge depth and operating conditions on material fatigue and interface failure. The macroscopic stress accumulation model focuses on the long-term aging of the battery, and describes the effects of cycle times and capacity attenuation on the health status of the battery. The system uses multi-physics field coupling simulation, machine learning and other methods to establish a cross-scale stress evolution model.

[0066] S102, integrating the microscopic electrochemical stress model, the macroscopic stress evolution model and the macroscopic stress accumulation model to form an electrochemical stress field;

[0067] This step aims to integrate stress models of different scales and construct a unified electrochemical stress field representation. The electrochemical stress field is a bridge to describe the interaction of multiple physical fields inside the battery, the evolution of material structure, and performance degradation. Through the construction of the stress field, the system can quantitatively analyze the battery health status under different driving conditions and predict future performance trends.

[0068] During the model integration process, the system needs to deal with the coupling relationship between different scales and different physical fields. The microscopic electrochemical stress model emphasizes the local stress caused by ion diffusion and charge transfer, the mesoscopic stress evolution model describes the material fatigue accumulation during the charging and discharging process, and the macroscopic stress accumulation model considers the capacity attenuation caused by long-term cycling. The system can adopt a multi-scale modeling method to nest and bridge different models in the time and space dimensions. For example, the microscopic model is used as the boundary condition of the mesoscopic model, and the mesoscopic model is embedded in the macroscopic model for iterative solution. At the same time, machine learning methods are used to establish a mapping relationship between different models to achieve real-time updating and prediction of the stress field.

[0069] In the construction of electrochemical stress fields, the computational complexity and solution efficiency of the model are key challenges. Solving complex multi-physics field coupling equations usually requires a lot of computing resources and time. To this end, the system can introduce technologies such as reduced-order models and proxy models to improve computational efficiency while ensuring accuracy. By performing feature decomposition and sensitivity analysis on the original model, key modes and dominant factors are extracted to construct a simplified proxy model. At the same time, the system can also adopt strategies such as parallel computing and distributed optimization to make full use of multi-core processors and cluster resources to accelerate the solution of stress fields.

[0070] S103, constructing a stress trigger model based on electrochemical stress field;

[0071] The system builds a stress trigger model based on the electrochemical stress field. The stress trigger model includes a stress threshold function and a material response function. The stress threshold function is used to characterize the stress field changes caused by different driving behaviors, and the material response function is used to describe the evolution law of the material microstructure caused by the stress field changes. Specifically: calculate the principal stress component and shear stress component in the electrochemical stress field; establish the material structure evolution function based on the principal stress component; construct the interface stress transfer function based on the shear stress component; combine the material structure evolution function and the interface stress transfer function to form a stress trigger model.

[0072] Based on the electrochemical stress field, this step further constructs a stress trigger model for characterizing the material structure evolution and performance degradation. The stress trigger model consists of two parts: the stress threshold function and the material response function, which respectively characterize the stress field changes caused by driving conditions and the material degradation laws caused by the stress field changes. Through the stress trigger model, the system can monitor the battery health status in real time and warn of possible performance degradation and safety risks.

[0073] In the construction of the stress trigger model, the system needs to quantify the key stress components in the electrochemical stress field. Through methods such as tensor decomposition, the system can extract the principal stress and shear stress components to characterize the anisotropy of the material and the interface failure characteristics. Based on the principal stress components, the system constructs the material structure evolution function to describe the microstructural changes such as lattice distortion and phase change induced by stress. At the same time, the interface stress transfer function is constructed based on the shear stress component to characterize the stress distribution and transfer law at the interface between the current collector and the active material, the electrolyte and the diaphragm. The material structure evolution function and the interface stress transfer function are coupled to form a complete stress trigger model.

[0074] S104, establishing a battery performance compensation equation according to the output result of the material response function;

[0075] The system establishes a battery performance compensation equation based on the output results of the material response function. The battery performance compensation equation is a function that can adjust the driving strategy constraints. The driving strategy constraints include the maximum charge and discharge rate, the operating temperature range, and the charge and discharge depth limit. The battery performance compensation equation is: In the equation, is the battery performance compensation coefficient, is the initial compensation coefficient, is the comprehensive damage function, is the time variable, is the temperature correction coefficient, is the integration variable.

[0076] This step uses the material response function in the stress trigger model to establish a battery performance compensation equation. The compensation equation achieves adaptive optimization of battery performance by adjusting the driving strategy constraints. The system comprehensively considers the evolution of material structure and the cumulative effect of stress, and dynamically adjusts constraint parameters such as charge and discharge rate, operating temperature range, and charge and discharge depth to extend the battery life while ensuring battery safety and reliability.

[0077] In establishing the compensation equation, the system needs to strike a balance between driving performance and battery life. Overly conservative constraint settings will limit the vehicle's power and range, while overly aggressive operating conditions may exacerbate battery aging. To this end, when designing the compensation equation, the system can introduce multi-objective optimization theories, such as Pareto optimality and weighting methods, to find the optimal balance between driving performance and battery life. At the same time, considering the complexity and variability of actual working conditions, the system can also use adaptive control, reinforcement learning and other methods to achieve online adjustment and optimization of the compensation equation, thereby improving the robustness and intelligence of battery management.

[0078] S105, inputting current driving data into a stress trigger model to obtain a real-time stress field distribution state;

[0079] This step inputs the real-time collected driving data into the constructed stress trigger model to obtain the stress field distribution state inside the battery. The stress field distribution reflects the stress characteristics and material evolution trend of the battery under the current working conditions, and is an important basis for evaluating the health status of the battery and warning of performance degradation.

[0080] In solving the stress field distribution, the system needs to efficiently process massive real-time data streams. The current, voltage, temperature and other signals of each cell in the battery pack are continuously generated at a frequency of milliseconds, which places strict requirements on the timeliness of calculation. To this end, the system can adopt technologies such as incremental learning and stream data processing to improve the real-time performance of stress field solution while ensuring the calculation accuracy. Through the incremental learning method, the system can quickly integrate new data on the basis of the original model to achieve dynamic update of stress field distribution.

[0081] S106, inputting the stress field distribution state into the battery performance compensation equation to obtain updated driving strategy constraint conditions;

[0082] This step uses the real-time stress field distribution state as input, substitutes it into the battery performance compensation equation, and dynamically adjusts the driving strategy constraints. The battery compensation equation is based on the material response characteristics, quantifies the impact of stress field changes on performance indicators such as capacity and power, and gives the change in constraints required for performance recovery. Based on the calculation results of the compensation equation, the system adaptively updates constraint parameters such as the maximum charge and discharge rate, operating temperature range, and charge and discharge depth limit, extending battery life while taking into account vehicle driving performance.

[0083] In the updating of driving strategy constraints, the system needs to balance immediacy and robustness. The adjustment of strategy constraints should be based on the long-term stress accumulation effect to avoid fluctuations in driving experience caused by too frequent changes. At the same time, the setting of constraints must also take into account the uncertainty of actual working conditions, such as sudden changes in ambient temperature and sudden changes in road slope, to ensure the adaptability and robustness of the strategy. To this end, the system can introduce methods such as model predictive control (MPC) and robust optimization to perform multi-step prediction and optimization based on the compensation equation to obtain a constraint update strategy that takes into account both long-term and short-term effects. In addition, the system can also integrate technologies such as driver behavior pattern recognition and road condition information perception to personalize and situationally adjust driving strategy constraints, thereby improving the intelligence and humanization of battery management.

[0084] For the adaptability problems that may be caused by the update of driving strategy constraints, the system can adopt a progressive adjustment strategy to gradually transition the constraints to the target value within a certain period to reduce the fluctuation of the driving experience. At the same time, through the human-computer interaction interface, the system can provide the driver with an explanation of the strategy adjustment to enhance the transparency and acceptability of the adjustment process. In addition, the system can also establish a driver feedback mechanism to collect the driver's subjective evaluation of the strategy adjustment effect, and combine the objective battery performance to continuously optimize and improve the compensation equation and constraint update strategy.

[0085] S107. Predict the battery health status based on the updated driving strategy constraints and historical stress accumulation data.

[0086] The system predicts the battery health status based on the updated driving strategy constraints and historical stress accumulation data.

[0087] This step predicts the battery's health status based on the updated driving strategy constraints and historical stress accumulation data. The battery health status is usually characterized by indicators such as remaining available capacity (RUL) and peak power, reflecting the battery's available energy and power output capabilities in the future. Through health status prediction, the system can perceive the battery performance degradation trend in advance, optimize energy management and maintenance strategies, and maximize the battery's use value.

[0088] In the health status prediction, the system needs to establish a correlation model between degradation mechanisms such as long-term capacity attenuation and internal resistance increase and historical stress accumulation. Traditional empirical formulas, such as the Arrhenius equation and the Eyring equation, are simple and practical, but it is difficult to describe the capacity attenuation behavior under complex working conditions. To this end, the system can use machine learning methods to explore the deep correlation between stress accumulation and capacity attenuation. Based on models such as the long short-term memory (LSTM) neural network and the temporal convolutional network (TCN), the system can establish a nonlinear mapping of temporal degradation characteristics and stress history to predict the capacity attenuation curve and RUL. At the same time, the deep learning model can also incorporate influencing factors such as material properties and ambient temperature to improve the multi-field coupling characteristics and robustness of the prediction.

[0089] How to implement this step specifically will be described in detail below in conjunction with another embodiment and will not be repeated here.

[0090] In the above embodiment, by acquiring driving data at different time scales, stress models at the micro, meso and macro levels are constructed, and an electrochemical stress field is integrated to fully reflect the stress change law of the battery during use. The stress trigger model constructed based on this stress field includes a stress threshold function and a material response function, which can accurately describe the relationship between driving behavior and the evolution of the material microstructure. The stress field state is converted into a driving strategy constraint through the battery performance compensation equation, and the mapping of stress state to usage strategy is realized, so that the system can capture the key information of battery performance degradation in time, dynamically adjust the usage strategy, and reduce the risk of accelerated degradation caused by unreasonable use. Taking into account the stress response mechanism at the material level, the prediction results have stronger physical meaning and interpretability, and improve the accuracy of battery health prediction.

[0091] In step S107 of the above embodiment, the system predicts the battery health status based on the updated driving strategy constraints combined with the historical stress accumulation data. Figure 2 , describe this step in detail:

[0092] See also Figure 2 , is another flow chart of a battery health prediction method in an embodiment of the present application.

[0093] S201, converting the updated driving strategy constraint condition into a time series characteristic data matrix, and converting the historical stress accumulation data into a structural characteristic data matrix;

[0094] In this step, the system first converts the updated driving strategy constraints into a time-series feature data matrix. Driving strategy constraints usually include a series of time-related parameters, such as speed limit, acceleration limit, charge and discharge power limit, etc., which may vary in different time periods. Converting these time-related constraints into a time-series feature data matrix can better characterize the changing characteristics of the driving strategy in the time dimension. At the same time, the system also converts the historical stress accumulation data into a structural feature data matrix. During use, the battery will be affected by various stress factors, such as temperature, current, voltage, etc. These stresses will cause certain physical or chemical changes inside the battery, thereby affecting the health of the battery. Converting the historical stress accumulation data into a structural feature data matrix can reflect the evolution of the internal structure of the battery.

[0095] In specific implementation, the system can use a variety of data conversion technologies to convert driving strategy constraints and historical stress accumulation data into feature matrices respectively. For driving strategy constraints, the system can extract the constraint parameters in each time period in a time series manner to form a time series feature vector, and the feature vectors of multiple time periods form a time series feature matrix. For historical stress accumulation data, the system can integrate the data corresponding to different stress factors to construct a multi-dimensional feature matrix that characterizes the changes in the internal structure of the battery. During the conversion process, the system can also pre-process the original data, such as denoising and normalization, to improve the accuracy of subsequent analysis.

[0096] S202, performing a weighted sum operation on the time series feature data matrix and the structural feature data matrix to obtain a fused feature vector;

[0097] The purpose of this step is to fuse the timing features and structural features to obtain a feature representation that comprehensively reflects the battery health status. The system first assigns certain weights to the timing feature data matrix and the structural feature data matrix respectively, and then performs a weighted summation operation to obtain the fused feature vector. The weights here can reflect the importance of different features to the battery health status, and can be set based on prior knowledge or data analysis results. Through the weighted summation method, the role of important features can be highlighted while retaining the original feature information, and the redundancy and interference between features can also be reduced.

[0098] In specific implementation, the system can adopt a variety of strategies to determine the weights of features. A simple method is to directly give fixed weight values ​​based on expert experience, such as giving a weight of 0.6 to the timing feature matrix and a weight of 0.4 to the structural feature matrix. Another more flexible method is to adaptively learn weight values ​​through machine learning algorithms, such as using weighted least squares, support vector machines and other algorithms to optimize weight parameters based on training data to maximize the correlation between the fused feature vector and the battery health status. In addition to weighted summation, the system can also explore other feature fusion methods, such as feature splicing, feature mapping, etc., to better mine the correlation information between timing features and structural features.

[0099] S203, calculating the principal component of the fused feature vector to obtain a eigenvalue sequence after dimensionality reduction;

[0100] In this step, the system performs principal component analysis (PCA) on the fused feature vectors to extract their main feature components and obtain the eigenvalue sequence after dimensionality reduction. Principal component analysis is a commonly used unsupervised dimensionality reduction method that projects the original high-dimensional feature data into a low-dimensional subspace through linear transformation, maximizing the variance of the projected data, that is, retaining the most discriminative and representative information in the original data. The purpose of doing this is to reduce the complexity of subsequent calculations on the one hand, and to filter out some noise and redundant information on the other hand, and improve the robustness of feature representation.

[0101] In the specific implementation, the system first centralizes the fused eigenvector, that is, subtracts its mean so that the feature data is centered on the zero point. Then, the covariance matrix of the eigenvector is calculated, and the eigenvalue decomposition is performed on it to obtain the eigenvalue and eigenvector. The eigenvalue represents the variance of the corresponding principal component, and the eigenvector represents the direction of the principal component. According to the size of the eigenvalue, the system can select the eigenvectors corresponding to the first few largest eigenvalues ​​as the principal components, project the original eigenvectors onto these principal components, and obtain the eigenvalue sequence after dimensionality reduction. The number of principal components can be determined based on the cumulative variance contribution rate, such as selecting the first few principal components with a cumulative variance contribution rate of 90%.

[0102] S204, calculating the health status probability distribution value according to the characteristic value sequence;

[0103] The purpose of this step is to use the eigenvalue sequence after dimensionality reduction to estimate the probability distribution of the battery health status. The battery health status can usually be divided into several discrete levels, such as healthy, sub-healthy, slightly degraded, moderately degraded, severely degraded, etc. Given a eigenvalue sequence, the system needs to predict the probability value of each health level. This is actually a multi-classification problem that can be solved using a probabilistic statistical model or a machine learning method.

[0104] In specific implementation, the system can use a method based on the Gaussian mixture model (GMM) to model the probability distribution of health status. Assume that each health level corresponds to a Gaussian component, and the eigenvalue sequence is a mixture distribution of these Gaussian components. The system can use the expectation maximization (EM) algorithm to estimate the parameters of the GMM, including the mean, covariance matrix, and mixing coefficient of each Gaussian component, through iterative optimization. Then, for a given eigenvalue sequence, the posterior probability of belonging to each health level is calculated based on the learned GMM parameters as an estimate of the probability distribution of the health state. In addition to GMM, the system can also use other probability models, such as hidden Markov models (HMMs), conditional random fields (CRFs), etc., to consider the temporal dependency between eigenvalues ​​to more accurately characterize the dynamic evolution of the health state.

[0105] Another implementation method is to directly use the training data to build a health status classifier. The system can use the feature value sequence as input and the corresponding health level as a label, and use supervised learning algorithms to train the classification model, such as logistic regression, support vector machine, decision tree, neural network, etc. The trained classifier can directly predict the health status to which the new feature value sequence belongs, and output the confidence or probability value of each health level. Compared with GMM, this method can more flexibly introduce nonlinear decision boundaries and capture the complex relationship between feature values ​​and health status. The system can select the appropriate classification algorithm according to the specific application scenario and data characteristics, and adjust the model's hyperparameters, such as regularization coefficient, kernel function type, etc., to achieve the best classification performance.

[0106] S205, comparing the health status probability distribution value with the preset health status classification standard to obtain a health level determination result;

[0107] In this step, the system needs to determine the health level of the battery based on the health status probability distribution value. The preset health status classification standard is to divide the battery health status into several levels based on expert knowledge or industry experience, such as healthy, sub-healthy, slightly degraded, moderately degraded, severely degraded, etc. Each level corresponds to a probability threshold range. The system compares the health status probability distribution value with these threshold ranges to determine its health level.

[0108] In specific implementation, the system can adopt simple decision rules, such as selecting the health level with the largest probability value as the judgment result. For example, if the health status probability distribution value is [0.1, 0.2, 0.4, 0.2, 0.1], and the corresponding health levels are [healthy, sub-healthy, slightly degraded, moderately degraded, severely degraded], the system will determine that the battery is in a slightly degraded state. Another more robust approach is to set a confidence threshold, and only when the probability value of a health level exceeds the threshold, it will be used as a judgment result. This can avoid erroneous judgments caused by probability estimation errors. For example, if the confidence threshold is set to 0.6, the above probability distribution cannot determine the health level, and further analysis or testing is required.

[0109] In actual applications, the health status classification standards may vary depending on factors such as application scenarios and battery types. In order to improve the adaptability of health level determination, the system can adopt an adaptive classification strategy to dynamically adjust the probability threshold range based on actual operating data. For example, a clustering algorithm can be used to analyze a large amount of historical data to automatically discover the clustered areas of probability distribution under different health states, thereby determining a reasonable threshold boundary. In addition, considering that the evolution of the battery health state has a certain degree of continuity and smoothness, a sudden change in the determination result may mean a misjudgment or anomaly. To this end, the system can introduce a time series smoothing or filtering mechanism to correct the determination results, such as using moving average, Kalman filtering and other methods to reduce the volatility of a single judgment.

[0110] S206, calculating the estimated remaining lifespan according to the health level determination result combined with the updated driving strategy constraint condition;

[0111] In this step, the system needs to estimate the remaining life of the battery, that is, how long the battery can continue to be used in its current health state. This requires comprehensive consideration of the battery health level determination results and the updated driving strategy constraints. The health level reflects the current performance level and degree of degradation of the battery, while the driving strategy constraints affect the future usage intensity and stress level of the battery. The two together determine the battery life consumption rate.

[0112] In specific implementation, the system can adopt a physical model-based approach to establish a mathematical relationship between health level, driving strategy and remaining life according to the equivalent circuit model and aging mechanism of the battery. For example, the Arrhenius equation can be used to describe the accelerated effect of temperature on battery aging, and the Palmgren-Miner linear cumulative damage theory can be used to estimate the loss of cycle life, and then combined with the capacity decay value corresponding to the current health level, the remaining use time of the battery under a given driving strategy can be predicted. The advantage of this method is that it has clear physical meaning and strong interpretability, but the disadvantage is that it requires relatively precise prior knowledge of battery materials and structures, and it is difficult to obtain relevant parameters.

[0113] Another implementation method is based on a data-driven approach, which uses a large amount of historical operating data to establish a mapping relationship between health level, driving strategy and remaining life through a machine learning algorithm. For example, the system can select a batch of batteries with similar health levels, track and record their life under different driving strategies, extract key characteristic parameters of the driving strategy (such as charge and discharge rate, operating temperature, etc.), and then use regression models (such as multivariate linear regression, support vector regression, etc.) to fit the functional relationship between health level, driving strategy characteristics and remaining life. When predicting, you only need to input the current battery health level and driving strategy characteristics into the trained model to get an estimate of the remaining life. Compared with the physical model, this data-driven method does not require a deep understanding of the internal mechanism of the battery, but requires a large amount of historical data as training samples, and the generalization ability of the model may be limited by the quality and distribution of the training data.

[0114] Considering that the battery life prediction itself has a large uncertainty, a single estimated value may not fully reflect the credibility of the prediction result. To this end, the system can give a confidence interval for the remaining life, indicating the probability that the true life falls within this interval under a certain confidence level. The width of the confidence interval can reflect the uncertainty of the prediction, and the confidence level can reflect the reliability requirements of the prediction. The system can flexibly set the confidence level and interval width according to the specific application scenario and decision-making risk, and balance the prediction accuracy and robustness. In addition, since the driving strategy may fluctuate and deviate in actual use, in order to improve the robustness of the life prediction, the system can also introduce an adaptive correction mechanism to dynamically correct the parameters of the remaining life estimation model according to the real-time operating status of the battery and the accumulated life data, and respond to changes in driving strategies in a timely manner.

[0115] S207: Predict the battery health status based on the health level determination result and the remaining life estimation value.

[0116] The purpose of this step is to obtain the comprehensive health status assessment results of the battery in the current and future, and provide a basis for decision-making and maintenance strategy optimization. The system needs to integrate and analyze the previously obtained health level determination results and the remaining life estimation value to give a health status prediction that takes into account both static performance and dynamic trends.

[0117] In the specific implementation, the system can design a weighted scoring function to quantify the health level and remaining life, and then perform weighted summation according to a certain weight coefficient to obtain a comprehensive health index. For example, the health level can be assigned a score of 5 to 1 from "healthy" to "severely deteriorated", and the remaining life can be converted into a score of 1 to 5 as a percentage of the design life, and then multiplied by the weight coefficients W1 and W2 respectively, and the sum is obtained to obtain a comprehensive health index of 0 to 10 points. The weight coefficient can be determined according to the relative importance of the health level and remaining life to the battery performance, or it can be adaptively learned from the data through a machine learning algorithm. The higher the comprehensive health index, the better the health of the battery. The system can further divide the comprehensive health index into several levels, such as excellent, good, general, poor, extremely poor, etc., for intuitive understanding and decision-making applications.

[0118] In addition to the weighted scoring method, the system can also use other information fusion strategies, such as DS evidence theory, fuzzy logic reasoning, etc., to treat the health level and remaining life as two different sources of evidence or fuzzy propositions, and obtain the probability distribution or membership function of the comprehensive health status through evidence combination or membership synthesis operation. This method can better handle the uncertainty and conflict between the health level and the remaining life, and obtain a more reasonable fusion result. For example, when the health level is "sub-healthy" but the remaining life prediction value is very high, the weighted scoring method may obtain a medium comprehensive health index, while the DS evidence theory may believe that the battery is more likely to be in a "healthy" state.

[0119] In the above embodiment, by converting the updated driving strategy constraints into a time series feature data matrix, converting the historical stress accumulation data into a structural feature data matrix, performing a weighted sum operation on the two matrices to obtain a fused feature vector, calculating the principal component of the fused feature vector to obtain a reduced eigenvalue sequence, calculating the health state probability distribution value according to the eigenvalue sequence and comparing it with the preset health state classification standard to obtain a health level determination result, combining the updated driving strategy constraints to calculate the remaining life estimate, and finally combining the health level determination result and the remaining life estimate to predict the battery health state. The technical solution realizes the effective fusion of time series features and structural features, extracts key feature information through dimensionality reduction processing, and reduces the interference of redundant features on prediction accuracy. The health state determination method based on probability distribution improves the reliability of the prediction results and reduces the possible misjudgment caused by single indicator judgment. The prediction method combining health level determination and remaining life estimation not only gives a qualitative evaluation of the current health state of the battery, but also provides a quantitative life prediction result, making the prediction result more comprehensive and accurate, and improving the accuracy and reliability of the prediction.

[0120] Further, in another embodiment, the system calculates the optimal operating range according to the battery health state;

[0121] Generate regional division boundary conditions based on the optimal working range;

[0122] Power allocation is optimized under boundary conditions to obtain power output reference values ​​and power allocation weight coefficients for each time period. Specifically, the health state decay rate of each time period is calculated, and the weight coefficient of each time period is determined according to the health state decay rate;

[0123] Calculate the power upper limit threshold of each time period based on the weight coefficient;

[0124] Combining the weight coefficient and the power upper limit threshold to generate a power allocation sequence;

[0125] The power reference value and allocation weight of each time period are output according to the power allocation sequence.

[0126] In this embodiment, the system first determines the optimal operating voltage range and current range of the battery by analyzing the current health status of the battery to form an optimal working area; then divides the working area into three sub-areas: high-efficiency area, balance area and protection area according to the characteristics of the working area, and sets corresponding boundary constraints for each sub-area; then calculates the health status decay rate of each time period through continuous monitoring, and converts the decay rate into a time period weight coefficient, and assigns a smaller weight to the time period with a larger decay rate; then calculates the power upper limit threshold of each time period in combination with the weight coefficient, and the time period with a larger weight allows a larger power output; based on the calculated weight coefficient and power threshold, generates a complete power timing allocation sequence; finally, determines the actual power reference value and the corresponding allocation weight for each time period, and realizes adaptive power optimization allocation based on health status.

[0127] In the above embodiment, a technical solution is provided by calculating the optimal working range according to the battery health status, generating area division boundary conditions based on the optimal working range, and optimizing power allocation under the boundary conditions to obtain power output reference values ​​for each time period and power allocation weight coefficients. This allows the battery to always operate in the working range that best suits its current health status, thereby reducing the impact of unreasonable working conditions on battery life.

[0128] The following describes the system in the embodiment of the present invention from the perspective of hardware processing. Figure 3 , which is a schematic diagram of the physical device structure of a battery health prediction system provided in an embodiment of the present application.

[0129] It should be noted that Figure 3 The structure of the system shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0130] like Figure 3 As shown, the system includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 302 or the program loaded from the storage part 308 to the random access memory (RAM) 303, such as executing the method in the above embodiment. In RAM 303, various programs and data required for system operation are also stored. CPU 301, ROM 302 and RAM 303 are connected to each other through a bus 304. Input / output (I / O) interface 305 is also connected to bus 304.

[0131] The following components are connected to the I / O interface 305: an input section 306 including a camera, an infrared sensor, etc.; an output section 307 including a liquid crystal display (LCD) and a speaker, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 310 as needed so that a computer program read therefrom is installed into the storage section 308 as needed.

[0132] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 309, and / or installed from a removable medium 311. When the computer program is executed by the central processing unit (CPU) 301, various functions defined in the present invention are performed.

[0133] It should be noted that the computer-readable medium shown in the embodiment of the present invention may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, device or device. In the present invention, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable computer program. Such a propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing.

[0134] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Among them, each box in the flowchart or block diagram can represent a module, a program segment, or a part of the code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0135] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the system described in the above embodiment; or may exist independently without being assembled into the system. The above storage medium carries one or more computer programs, and when the above one or more computer programs are executed by a processor of a system, the system implements the method provided in the above embodiment.

[0136] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0137] As used in the above embodiments, the term "when..." may be interpreted to mean "if..." or "after..." or "in response to determining..." or "in response to detecting...", depending on the context. Similarly, the phrases "upon determining..." or "if (the stated condition or event) is detected" may be interpreted to mean "if determining..." or "in response to determining..." or "upon detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)", depending on the context.

[0138] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from a website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrated. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state hard disk), etc.

[0139] Those skilled in the art can understand that to implement all or part of the processes in the above-mentioned embodiments, the processes can be completed by computer programs to instruct related hardware, and the programs can be stored in computer-readable storage media. When the programs are executed, they can include the processes of the above-mentioned method embodiments. The aforementioned storage media include: ROM or random access memory RAM, magnetic disk or optical disk and other media that can store program codes.

Claims

1. A battery health prediction method, characterized in that: include: Acquire driving data of the target battery pack at different times, the driving data including micro-scale instantaneous operating condition data, meso-scale single trip data, and macro-scale long-term use data, the instantaneous operating condition data including instantaneous current, voltage, and temperature of the target battery pack, the single trip data including charge and discharge depth and driving conditions, and the long-term use data including historical cycle times and capacity changes; Establishing a microscopic electrochemical stress model based on the instantaneous operating condition data, constructing a mesoscopic stress evolution model based on the single trip data, and generating a macroscopic stress accumulation model based on the long-term use data; Integrating the microscopic electrochemical stress model, the mesoscopic stress evolution model and the macroscopic stress accumulation model to form an electrochemical stress field; A stress trigger model is constructed based on the electrochemical stress field, wherein the stress trigger model includes a stress threshold function and a material response function, wherein the stress threshold function is used to characterize stress field changes caused by different driving behaviors, and the material response function is used to describe the evolution law of the material microstructure caused by the stress field changes; Establishing a battery performance compensation equation according to the output result of the material response function, wherein the battery performance compensation equation is a function capable of adjusting driving strategy constraints, wherein the driving strategy constraints include a maximum charge and discharge rate, an operating temperature range, and a charge and discharge depth limit; Inputting current driving data into the stress trigger model to obtain a real-time stress field distribution state; Inputting the stress field distribution state into the battery performance compensation equation to obtain updated driving strategy constraint conditions; The battery health status is predicted based on the updated driving strategy constraints combined with historical stress accumulation data.

2. The method according to claim 1, characterized in that: The establishing of a microscopic electrochemical stress model based on the instantaneous operating condition data specifically includes: Calculating ion concentration gradient and potential distribution according to the instantaneous current and the voltage; Establishing a diffusion stress field equation based on the ion concentration gradient, and establishing an electrochemical stress field equation based on the potential distribution; The diffusion stress field equation and the electrochemical stress field equation are coupled and calculated to obtain a microscopic electrochemical stress model.

3. The method according to claim 1, characterized in that The constructing of a stress trigger model based on the electrochemical stress field specifically includes: calculating principal stress components and shear stress components in the electrochemical stress field; Establishing a material structure evolution function based on the principal stress components; constructing an interface stress transfer function based on the shear stress components; The material structure evolution function and the interface stress transfer function are combined to form a stress trigger model.

4. The method according to claim 1, characterized in that The battery performance compensation equation is: In the equation, is the battery performance compensation coefficient, is the initial compensation coefficient, is the comprehensive damage function, is a time variable, is the temperature correction factor, is the integration variable.

5. The method according to claim 1, characterized in that The method of predicting the battery health status according to the updated driving strategy constraint conditions combined with historical stress accumulation data specifically includes: Converting the updated driving strategy constraint condition into a time series characteristic data matrix, and converting the historical stress accumulation data into a structural characteristic data matrix; Performing a weighted sum operation on the time series feature data matrix and the structural feature data matrix to obtain a fused feature vector; Calculate the principal component of the fused feature vector to obtain a eigenvalue sequence after dimension reduction; Calculate the health status probability distribution value according to the characteristic value sequence; Comparing the health status probability distribution value with a preset health status grading standard to obtain a health grade determination result; Calculate the estimated remaining life according to the health level determination result and the updated driving strategy constraint condition; The health status of the battery is predicted by combining the health level determination result and the remaining life estimation value.

6. The method according to claim 1, characterized in that After predicting the battery health state according to the updated driving strategy constraint conditions combined with historical stress accumulation data, the method further includes: Calculating an optimal operating range according to the battery health status; generating a region division boundary condition based on the optimal working interval; Power allocation is optimized under the boundary conditions to obtain power output reference values ​​and power allocation weight coefficients for each time period.

7. The method according to claim 6, characterized in that The power allocation optimization is performed under the boundary conditions to obtain the power output reference value and the power allocation weight coefficient for each time period, specifically including: Calculating the health state decay rate of each time period, and determining the weight coefficient of each time period according to the health state decay rate; Calculate the power upper limit threshold of each time period based on the weight coefficient; generating a power allocation sequence by combining the weight coefficient and the power upper limit threshold; The power reference value and allocation weight of each time period are output according to the power allocation sequence.

8. A battery health prediction system, characterized in that: The system comprises: One or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to cause the system to execute the method as described in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that: When the instructions are executed on a system, the system is caused to execute the method according to any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product is run on a system, the system is caused to execute the method according to any one of claims 1 to 7.

Images