Lithium ion battery health state estimation method, electronic equipment and storage medium

Abstract

The invention provides a lithium ion battery health state estimation method, electronic equipment and a storage medium, and belongs to the technical field of battery health management. The problems that an existing weighting mode cannot adapt to changes of aging characteristics of the battery in different aging stages, the advantages of two models are difficult to give full play in each stage, precision improvement is limited, and overfitting may occur are solved. The method comprises the following steps: preprocessing charging and discharging data of a battery; constructing a multi-physical mechanism fusion model; carrying out adaptive segmentation weight optimization; and carrying out weighted fusion prediction by using an adaptive weight to realize battery SOH estimation. The method is mainly used for battery health state estimation.

Description

Technical Field

[0001] This invention belongs to the field of battery health management technology, and in particular relates to a method for estimating the health status of lithium-ion batteries, an electronic device, and a storage medium. Background Technology

[0002] Lithium-ion batteries are the preferred power batteries for new energy vehicles, and accurate estimation of their State of Health (SOH) is crucial for battery life assessment and safety management. Currently, semi-empirical model-based methods are a research hotspot, with the bi-exponential model and the Arrhenius model being the most common.

[0003] The dual-exponential model describes aging by superimposing two exponential terms, which can better capture multi-stage aging characteristics and has a strong ability to fit nonlinear trends. However, the correspondence between its parameters and specific aging mechanisms is unclear, and it may overfit limited data when there are many parameters, resulting in high prediction uncertainty when the data range is exceeded.

[0004] The Arrhenius model, derived from chemical kinetics theory, has parameters with clear electrochemical significance and strong coupling with thermoelectric characteristics such as temperature and charge / discharge rate. Therefore, the model is highly interpretable and its extrapolation capability is reliable within a reasonable temperature range. However, this model has a relatively weak ability to fit the nonlinear trends of battery changes, which may lead to larger fitting errors.

[0005] To combine the advantages of both, existing technologies employ a weighted fusion method for the two models. However, current methods typically use fixed weights for overall weighting. This rigid weighting approach cannot adapt to the changing aging characteristics of batteries at different aging stages (such as early, middle, and late stages), making it difficult to fully leverage the advantages of both models at each stage. This results in limited accuracy improvement and may lead to problems such as overfitting.

[0006] Therefore, a SOH estimation method that can adapt to the characteristics of different battery aging stages and dynamically adjust the model fusion weights is needed to achieve health state estimation with higher accuracy, stronger interpretability and better extrapolation ability. Summary of the Invention

[0007] In view of this, the present invention aims to propose a method, electronic device and storage medium for estimating the health status of lithium-ion batteries, in order to solve the problem that the existing weighted method cannot adapt to the changes in the aging characteristics of batteries at different aging stages, and it is difficult to give full play to the advantages of the two models at each stage, resulting in limited accuracy improvement and the possibility of overfitting.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: According to a first aspect of the present invention, a method for estimating the health status of a lithium-ion battery is provided, comprising the following steps: Preprocess the battery charge and discharge data; Construct a multi-physics mechanism fusion model; Adaptive segmented weight optimization; We use adaptive weights for weighted fusion prediction to achieve battery SOH estimation.

[0009] Furthermore, the preprocessing of the battery charge and discharge data specifically includes: Extract the capacity sequence as the dependent variable and the cumulative ampere-hour capacity as the independent variable; The training and validation sets are randomly divided in a 7:3 ratio, and a random seed is set to ensure the repeatability of the data split.

[0010] Furthermore, the construction of the multi-physics mechanism fusion model specifically includes establishing a double-exponential model and an Arrhenius model, and defining a weighted fusion model.

[0011] Furthermore, the double-exponential model takes the form of: in, and These are parameters related to the initial capacity; and This parameter is strongly correlated with the aging rate at different time stages of internal resistance. This represents the cumulative charge / discharge ampere-hours as the number of charge / discharge cycles increases. This represents the SOH value as the number of cycles increases.

[0012] Furthermore, the formula for the Arrhenius model is: in, This is the temperature-dependent attenuation amplitude coefficient; It is the activation energy for the aging reaction; This is the attenuation coefficient related to the charge / discharge rate; The decay acceleration index; This represents the cumulative charge / discharge ampere-hours as the number of charge / discharge cycles increases. This represents the SOH value as the number of cycles increases; This refers to the battery charge / discharge rate value. It is the gas constant; This refers to the temperature at which the battery was tested.

[0013] Furthermore, the weighted fusion model takes the following form: in, For weighted parameters; This represents the SOH value as the number of cycles increases.

[0014] Furthermore, the segmented weight optimization specifically includes: Define segment boundaries based on the number of iterations; Optimize weights independently for each segment As optimization variables, an objective function is constructed with the goal of minimizing the root mean square error (RMSE) of the segmented training data. Under the constraint condition The objective function is then solved using a constrained nonlinear optimization algorithm to obtain the optimal weight value for that segment.

[0015] Furthermore, the method of using adaptive weights for weighted fusion prediction to achieve battery SOH estimation specifically includes: Define a weight matching function to automatically match the optimization weights for the corresponding segments based on the input loop count; The optimized weight parameters are invoked, and the battery SOH is estimated using the weighted fusion model.

[0016] According to a second aspect of the present invention, an electronic device is provided, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement a lithium-ion battery health state estimation method as described above.

[0017] According to a third aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing a computer program that, when executed by a processor, implements a lithium-ion battery health state estimation method as described above.

[0018] Compared with the prior art, the beneficial effects of the present invention are: 1. Segmented dynamic weighting mechanism: This invention combines segmented iteration counts with weight optimization to replace the traditional fixed weight method, enabling the model to adapt to the characteristic changes of different stages of battery aging. 2. Constrained nonlinear optimization application: Constrained nonlinear optimization algorithm is used to solve for the weights, which balances computational efficiency and physical rationality and avoids weights going out of bounds; 3. Multi-model complementarity and fusion: The dual-exponential model captures the long-term decay trend, while the Arrhenius model introduces electrothermal characteristics such as current ratio and temperature. Through weighted fusion, complementary advantages are achieved.

[0019] 4. Lightweight real-time deployment: The optimization process involves only single-variable weights, resulting in a light computational burden, making it suitable for embedded BMS online applications.

[0020] 5. This invention provides a method for estimating the health status of lithium-ion batteries, which integrates the aging mechanism and physical degradation trend of batteries. It uses a double exponential model to capture the nonlinear trend of battery capacity decay, and uses an Arrhenius model to introduce the influence of temperature stress. It integrates electrothermal characteristic parameters including initial capacity, internal resistance, charging rate and temperature, and combines the advantages of both through an adaptive weighting strategy for different aging time segments. This achieves high-precision battery health status estimation and good interpretability. Moreover, the adaptive weighting strategy can adaptively adjust the weighting parameters according to the differences in the fitting of individual models for different batteries, making it more robust and flexible. Attached Figure Description

[0021] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 This is a flowchart of a lithium-ion battery health status estimation method according to the present invention; Figure 2 This is a graph showing the performance comparison results of a model for aging lithium manganese iron phosphate batteries. Detailed Implementation

[0022] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0023] Figure 1 This is a flowchart of a lithium-ion battery health status estimation method according to the present invention.

[0024] like Figure 1 As shown, the lithium-ion battery health status estimation method includes the following steps: In step S1, the battery charging and discharging data are preprocessed.

[0025] In some embodiments, the battery charge and discharge data are preprocessed, specifically including: Extract the capacity sequence as the dependent variable and the cumulative ampere-hour capacity as the independent variable; The training and validation sets are randomly divided in a 7:3 ratio, and a random seed is set to ensure the repeatability of the data split.

[0026] Data sequences of capacity, charging rate, and temperature were extracted. The cumulative ampere-hour capacity was obtained by processing the charge and discharge capacity and used as the independent variable. The capacity obtained from the capacity test was used as the dependent variable for the training and validation sets. The charge and discharge rate and temperature were coupled into the model as electrothermal characteristics. The training and validation sets were randomly divided in a 7:3 ratio. The random seed was set using rng(42) in MATLAB to ensure the repeatability of the data partitioning and avoid performance fluctuations caused by differences in sample allocation, thereby ensuring the consistency of the reproduction results.

[0027] In step S2, a multi-physics mechanism fusion model is constructed.

[0028] It should be noted that the construction of the multi-physics mechanism fusion model specifically includes establishing a double-exponential model and an Arrhenius model, and defining a weighted fusion model.

[0029] Based on the dual core mechanisms of battery aging (nonlinear degradation trend and temperature dependence), a dual-model collaborative architecture is constructed.

[0030] In some embodiments, the double-exponential model takes the form of: in, and For parameters related to the initial capacity, This can be the initial maximum usable capacity of the battery; and This parameter is strongly correlated with the aging rate at different time stages of internal resistance. This represents the cumulative charge / discharge ampere-hours as the number of charge / discharge cycles increases. This represents the SOH value as the number of cycles increases.

[0031] In actual execution, the initial parameters are set to Based on the prior knowledge that the initial battery capacity is close to 100% and the decay rate constant is a small negative value, the convergence of the fit is ensured.

[0032] The formula for the Arrhenius model is: in, This is the temperature-dependent attenuation amplitude coefficient, characterizing the degree of influence of temperature on the total attenuation. This is the activation energy of the aging reaction; a larger value indicates that temperature has a more significant effect on the decay rate. This is the attenuation amplitude coefficient related to charge / discharge rate. The larger the value, the more significant the effect of charge / discharge rate on the attenuation rate. The decay acceleration index characterizes the degree to which the decay rate accelerates with increasing cumulative ampere-hours; This represents the cumulative charge / discharge ampere-hours as the number of charge / discharge cycles increases. This represents the SOH value as the number of cycles increases; This refers to the battery charge / discharge rate value. This is the gas constant, with a value of 8.314 J·mol⁻¹. -1 ·K -1 ; This refers to the temperature at which the battery was tested.

[0033] In actual execution, the initial parameter settings are as follows: Based on the activation energy range (10-50 J / mol) and decay acceleration characteristics of typical lithium-ion batteries, the settings avoid fitting divergence.

[0034] The weighted fusion model takes the following form: in, For weighted parameters; This represents the SOH value as the number of cycles increases.

[0035] In actual implementation, the initial value was set to w0 = 0.5 to balance the initial contributions of the two models and accelerate convergence. The constraints are as follows: To prevent overfitting.

[0036] By using a weighted fusion model, the coupling between internal resistance change and electrothermal characteristic parameters such as temperature and charge / discharge rate is achieved, and the nonlinear curve fitting characteristics of the double exponential model enable the model to have better accuracy.

[0037] In step S3, adaptive segmented weight optimization is performed.

[0038] During the construction of the weighted fusion model, it was found that simply applying a fixed overall weight to the two models resulted in minimal accuracy improvement and overfitting. However, by adaptively fitting the weight parameters for different battery aging periods, the advantages of both capacity degradation models could be fully utilized, thereby significantly improving the accuracy of SOH estimation.

[0039] In some embodiments, the segmented weight optimization specifically includes: The segment boundaries are defined based on the number of cycles (e.g., [300, 600], [600, 1000], [1000, 1500], [1500, 2000]), corresponding to different stages of battery aging; Optimize weights independently for each segment As optimization variables, an objective function is constructed with the goal of minimizing the root mean square error (RMSE) of the segmented training data. In actual execution, the objective function is: in, This represents the total number of training data segments. and These are the predicted values ​​from the bi-exponential model and the Arrhenius model, respectively. For the first Measured SOH values ​​for each data point.

[0040] Under the constraint condition The objective function is then solved using a constrained nonlinear optimization algorithm to obtain the optimal weight value for that segment.

[0041] In actual implementation, the constrained nonlinear optimization algorithm uses the fmincon function in MATLAB, with the following constraints: This algorithm ensures the physical meaning of the weights is reasonable (non-negative and the total weights sum to 1). Its advantages lie in its ability to efficiently handle boundary constraint problems, with a convergence speed that is more than 40% faster than the simplex method. It iteratively approaches the optimal solution from within the feasible region, avoiding local optima traps. It has low computational complexity (single-variable optimization), and the time taken for a single piecewise weight solution is ≤10ms. It is also suitable for embedded deployment.

[0042] In step S4, adaptive weights are used to perform weighted fusion prediction to achieve battery SOH estimation.

[0043] In some embodiments, the use of adaptive weights for weighted fusion prediction to achieve battery SOH estimation specifically includes: Define a weight matching function to automatically match the optimization weights for the corresponding segments based on the input loop count; In actual execution, a weight matching function get_weight_for_x(x_val) is defined, which automatically matches the segment to which the input independent variable x belongs and calls the corresponding optimization weight.

[0044] If a segment lacks training data (e.g., some batteries haven't reached a high number of cycles), the globally optimized weights are automatically reused. =0.5, to ensure the continuity and rationality of the forecast.

[0045] If the number of iterations in the input data lies at the boundaries of two segments, a linear interpolation method is used to generate transition weights for the segment boundary points. The formula is as follows: in, Train weights for boundary points; Optimize the weights for the left-side segments; Optimize the weights for the right-side segment; This is the upper boundary of the left segment (i.e., the starting point of the transition interval). This is the lower boundary of the right segment (i.e., the end point of the transition interval). The target number of iterations (the number of iterations for the data points to be predicted); This represents the length of the transition interval.

[0046] By using linear interpolation to generate transition weights, boundary point processing is performed to avoid abrupt changes in boundary point estimation results.

[0047] The optimized weight parameters are invoked, and the battery SOH is estimated using the weighted fusion model. The root mean square error (RMSE) and goodness-of-fit (R²) performance metrics are then calculated. 2 .

[0048] Taking lithium manganese iron phosphate batteries as an example, the performance comparison results of the aging model of lithium manganese iron phosphate batteries are shown in the figure below. Figure 2 As shown, using a dataset of lithium manganese iron phosphate batteries, an adaptive piecewise weighted fusion model is employed. Comparing its SOH estimation performance with that of a single Arrhenius formula decay model, the training set performance is improved by approximately 30%, and the validation set performance accuracy is improved by approximately 10%, demonstrating its good SOH estimation capability. Moreover, its accuracy is significantly improved compared to a single model. This method can also effectively couple electrical and thermal characteristics such as temperature and charge / discharge current, thereby improving the interpretability of the model.

[0049] By performing SOH estimation on the battery dataset using both the fusion model and the single model, it can be seen that the adaptive segmented weighted fusion model improves the accuracy of the training set performance by about 30% compared to the single Arrhenius model, and improves the accuracy of the validation set performance by more than 10%. This indicates that the model performs very well in SOH estimation, and the accuracy is significantly improved compared to the single model. At the same time, it also achieves coupling of electrothermal characteristics such as temperature and charge / discharge current, and has strong interpretability.

[0050] The lithium-ion battery health status estimation method proposed according to embodiments of the present invention has the following beneficial effects: (1) Segmented dynamic weighting mechanism: This invention combines segmented loop count with weight optimization to replace the traditional fixed weight method, enabling the model to adapt to the characteristic changes of different stages of battery aging; (2) Constrained nonlinear optimization application: The constrained nonlinear optimization algorithm is used to solve the weights, taking into account both computational efficiency and physical rationality, and avoiding weights going out of bounds; (3) Multi-model complementarity and fusion: The dual-exponential model captures the long-term decay trend, while the Arrhenius model introduces electrothermal characteristics such as current ratio and temperature, and achieves complementary advantages through weighted fusion.

[0051] (4) Lightweight real-time deployment: The optimization process involves only single variable weights, with a light computational burden, making it suitable for embedded BMS online applications.

[0052] (5) This invention provides a method for estimating the health status of lithium-ion batteries, which integrates the aging mechanism and physical degradation trend of batteries. It uses a double exponential model to capture the nonlinear trend of battery capacity decay, uses an Arrhenius model to introduce the influence of temperature stress, integrates electrothermal characteristic parameters including initial capacity, internal resistance, charging rate and temperature, and integrates the advantages of both by using an adaptive weighting strategy for different aging time segments to achieve high-precision battery health status estimation and good interpretability. Moreover, the adaptive weighting strategy can adjust the weighting parameters adaptively for the differences in the fitting of individual models for different batteries, making it more robust and flexible.

[0053] This invention proposes an electronic device, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the lithium-ion battery health state estimation method.

[0054] This invention proposes a computer-readable storage medium for storing computer instructions, which, when executed by a processor, implement the steps of the lithium-ion battery health state estimation method.

[0055] The memory in this application embodiment can be volatile memory or non-volatile memory, or it can include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous linked dynamic random access memory (SLDRAM), and direct rambus RAM (DRRAM). It should be noted that the memory used in the methods described in this invention is intended to include, but is not limited to, these and any other suitable types of memory.

[0056] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially as a computer program product. The computer program product includes one or more computer instructions. When the computer instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., high-density digital video discs (DVDs)), or semiconductor media (e.g., solid-state drives (SSDs)).

[0057] In implementation, each step of the above method can be completed by integrated logic circuits in the processor's hardware or by instructions in software. The steps of the method disclosed in the embodiments of this application can be directly implemented by a hardware processor, or by a combination of hardware and software modules in the processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method. To avoid repetition, detailed descriptions are omitted here.

[0058] It should be noted that the processor in the embodiments of this application can be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method embodiments can be completed by the integrated logic circuitry in the processor's hardware or by instructions in software form. The processor can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly embodied as being executed by a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can be located in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory, and the processor reads the information in the memory and, in conjunction with its hardware, completes the steps of the above methods.

[0059] The above provides a detailed description of the lithium-ion battery health status estimation method proposed in this invention. Specific examples have been used to illustrate the principles and implementation methods of this invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.

Claims

1. A method for estimating the state of health of a lithium-ion battery, characterized in that, Includes the following steps: Preprocess the battery charge and discharge data; Construct a multi-physics mechanism fusion model; Adaptive segmented weight optimization; We use adaptive weights for weighted fusion prediction to achieve battery SOH estimation.

2. The method for estimating the health status of a lithium-ion battery according to claim 1, characterized in that: The preprocessing of the battery charge and discharge data specifically includes: Extract the capacity sequence as the dependent variable and the cumulative ampere-hour capacity as the independent variable; The training and validation sets are randomly divided in a 7:3 ratio, and a random seed is set to ensure the repeatability of the data split.

3. The method for estimating the state of health of a lithium-ion battery according to claim 1, characterized in that: The construction of the multi-physics mechanism fusion model specifically includes establishing a double-exponential model and an Arrhenius model, and defining a weighted fusion model.

4. The method for estimating the health status of a lithium-ion battery according to claim 3, characterized in that: The form of the double-exponential model is as follows: in, and These are parameters related to the initial capacity; and This parameter is strongly correlated with the aging rate at different time stages of internal resistance. This represents the cumulative charge / discharge ampere-hours as the number of charge / discharge cycles increases. This represents the SOH value as the number of cycles increases.

5. The method for estimating the health status of a lithium-ion battery according to claim 3, characterized in that: The formula for the Arrhenius model is: in, This is the temperature-dependent attenuation amplitude coefficient; It is the activation energy for the aging reaction; This is the attenuation coefficient related to the charge / discharge rate; The decay acceleration index; This represents the cumulative charge / discharge ampere-hours as the number of charge / discharge cycles increases. This represents the SOH value as the number of cycles increases; This refers to the battery charge / discharge rate value. It is the gas constant; This refers to the temperature at which the battery was tested.

6. The method for estimating the state of health of a lithium-ion battery according to claim 3, characterized in that: The weighted fusion model takes the following form: in, These are weighted parameters; This represents the SOH value as the number of cycles increases.

7. The method for estimating the health status of a lithium-ion battery according to claim 1, characterized in that: The segmented weight optimization specifically includes: Define segment boundaries based on the number of iterations; Optimize weights independently for each segment As optimization variables, an objective function is constructed with the goal of minimizing the root mean square error (RMSE) of the segmented training data. Under the constraint condition The objective function is then solved using a constrained nonlinear optimization algorithm to obtain the optimal weight value for that segment.

8. The method for estimating the state of health of a lithium-ion battery according to claim 1, characterized in that: The method of using adaptive weights for weighted fusion prediction to achieve battery SOH estimation specifically includes: Define a weight matching function to automatically match the optimization weights for the corresponding segments based on the input loop count; The optimized weight parameters are invoked, and the battery SOH is estimated using the weighted fusion model.

9. An electronic device, characterized in that, include: The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement a lithium-ion battery health state estimation method as described in claims 1-8.

10. A computer-readable storage medium, characterized in that, It stores a computer program that enables the computer to execute a lithium-ion battery health state estimation method as described in claims 1-8.

Images