Retired battery detection method and device, battery management system and storage medium

Abstract

The invention provides a decommissioned battery detection method and device, a battery management system and a storage medium, and the method comprises the steps: obtaining multi-source internal characteristic data of a to-be-detected decommissioned battery, the internal characteristic data comprising at least one of an AC impedance spectrum, pulse characteristic data and a constant-current discharge curve; inputting the multi-source internal characteristic data into a battery health state evaluation model, and obtaining a health state evaluation result of the retired battery output by the battery health state evaluation model; wherein the battery health state evaluation model is a model formed by fusing a physical model based on an electrochemical mechanism and a machine learning model. According to the invention, through fusion of an electrochemical mechanism and machine learning, the detection accuracy of the health state of the retired battery is improved.

Description

Technical Field

[0001] This invention relates to the field of battery technology, and specifically to a method, apparatus, battery management system, and storage medium for testing retired batteries. Background Technology

[0002] The large-scale retirement of power batteries is an inevitable result of the booming development of the new energy vehicle industry. Accurate and rapid health status assessment of retired batteries is an indispensable prerequisite for their safe and efficient reuse. However, existing retired battery testing technologies face serious challenges in terms of efficiency, accuracy, and comprehensiveness.

[0003] In related technologies, on the one hand, retired battery testing methods mainly rely on traditional external characteristic tests, such as measuring open-circuit voltage and DC internal resistance. While these methods are simple and quick, they cannot deeply reflect the complex electrochemical changes within the battery, such as active lithium loss and electrode material degradation, leading to inaccurate evaluation results and an inability to reliably predict battery performance and risks during long-term use. On the other hand, to obtain more accurate parameters, standard charge-discharge tests are typically used. This method directly measures capacity through a complete charge-discharge cycle, offering high accuracy, but it is extremely time-consuming, usually requiring several hours or even longer. This inefficiency cannot meet the practical needs of large-scale, industrialized sorting of retired batteries, becoming a bottleneck restricting the development of the cascade utilization industry.

[0004] Therefore, there is an urgent need in this field for a method for detecting retired batteries that can balance efficiency and accuracy, so as to achieve rapid, accurate and reliable diagnosis of battery health status. Summary of the Invention

[0005] In view of this, it is necessary to provide a method, apparatus, battery management system and storage medium for detecting retired batteries, so as to solve the technical problems of low detection speed and accuracy of health status detection of retired batteries in the prior art.

[0006] To address the aforementioned technical problems, in a first aspect, the present invention provides a method for detecting decommissioned batteries, the method comprising: Acquire multi-source internal characteristic data of the retired battery to be tested, wherein the internal characteristic data includes at least one of AC impedance spectrum, pulse characteristic data, and constant current discharge curve; The multi-source internal characteristic data are input into the battery health status assessment model to obtain the health status assessment result of the retired battery output by the battery health status assessment model. The battery health status assessment model is a fusion of a physical model based on electrochemical mechanisms and a machine learning model.

[0007] In one possible implementation, the step of inputting the multi-source internal characteristic data into a battery health status assessment model and obtaining the health status assessment result of the retired battery output by the battery health status assessment model includes: The internal characteristic data from multiple sources are input into the physical model to obtain the first electrical performance feature extracted by the physical model; The first electrical performance feature is input into the machine learning model to perform health status assessment and prediction, and the health status assessment result is obtained.

[0008] In one possible implementation, the machine learning model includes a first machine learning model and a second machine learning model, wherein the second machine learning model includes a feature extraction sub-model and a classification sub-model; the step of inputting the multi-source intrinsic characteristic data into the battery health status assessment model and obtaining the health status assessment result of the retired battery output by the battery health status assessment model includes: The multi-source intrinsic characteristic data are input into the physical model to obtain the second electrical performance feature extracted by the physical model, and the second electrical performance feature is input into the first machine learning model to perform health status assessment and prediction to obtain the first prediction result. The third electrical performance feature extracted by the physical model is obtained from the pulse characteristic data or constant current discharge through the feature extraction sub-model in the second machine model, and the third electrical performance feature is input into the classification sub-model in the machine learning model to obtain the second prediction result; The first prediction result and the second prediction result are fused together to obtain the health status assessment result.

[0009] In one possible implementation, the physical model includes a physical rule-based verifier and a health status corrector; the step of inputting the multi-source internal characteristic data into the battery health status assessment model and obtaining the health status assessment result of the retired battery output by the battery health status assessment model includes: The machine learning model is used to process the multi-source intrinsic characteristic data to generate predicted battery parameter values; The predicted battery parameters are input into the physical rule validator in the physical model, and the validator results are output. When the verification result does not meet the physical rules, the predicted value of the battery parameter is corrected using the health status corrector to obtain the health status assessment result.

[0010] In one possible implementation, after obtaining the health status assessment result of the retired battery output by the battery health status assessment model, the method further includes: A battery parameter database is constructed based on the multi-source internal characteristic data and health status assessment results of multiple retired batteries.

[0011] In one possible implementation, after obtaining the health status assessment result of the retired battery output by the battery health status assessment model, the method further includes: Based on the health status assessment results, the retired batteries are sorted into a tiered reuse system at the package level, module level, or cell level.

[0012] In one possible implementation, after sorting the retired batteries into a cascade reuse system at the pack, module, or cell level based on the health status assessment results, the method further includes: Based on the sorting results, the topology of the battery cells or battery packs in the sorted whole pack, module or cell-level cascade reuse system is reconstructed using genetic algorithms or particle swarm optimization algorithms to form recombinant battery packs.

[0013] Secondly, the present invention also provides a retired battery testing device, the device comprising: The acquisition unit is used to acquire multi-source internal characteristic data of the retired battery to be tested, wherein the internal characteristic data includes at least one of AC impedance spectrum, pulse characteristic data, and constant current discharge curve; The detection unit is used to input the multi-source internal characteristic data into the battery health status assessment model and obtain the health status assessment result of the retired battery output by the battery health status assessment model. The battery health status assessment model is a fusion of a physical model based on electrochemical mechanisms and a machine learning model.

[0014] Thirdly, the present invention also provides a battery management system, including a memory and a processor, wherein the memory is used to store a program; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the decommissioned battery detection method described in any of the above implementations.

[0015] Fourthly, the present invention also provides a computer-readable storage medium for storing a computer-readable program or instructions, which, when executed by a processor, can implement the steps in the decommissioned battery detection method described in any of the above implementations.

[0016] The beneficial effects of this invention are: The retired battery detection method provided by this invention acquires multi-source internal characteristic data of the retired battery to be tested. This internal characteristic data includes at least one of AC impedance spectroscopy, pulse characteristic data, and constant current discharge curves. This provides a multi-source dataset that comprehensively characterizes the internal health state of the battery from different dimensions, and the acquisition process does not require a complete charge-discharge cycle. This achieves rapid acquisition of internal characteristic data that can quickly reflect the internal state process of the battery, shortening the detection time. The multi-source internal characteristic data is input into a battery health state assessment model to obtain the health state assessment result of the retired battery output by the model. The battery health state assessment model is a fusion of a physical model based on electrochemical mechanisms and a machine learning model. Since the physical model is used to explain and predict the electrochemical behavior of the battery, its prediction results are constrained by physical laws, avoiding prediction errors that may occur in purely data-driven models. The machine learning model is used to analyze and process the acquired multi-source internal characteristic data, achieving rapid and accurate assessment of the battery health state of retired batteries. This invention, through the fusion of electrochemical mechanisms and machine learning, improves the detection accuracy of the health state of retired batteries and reduces human error. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic flowchart of an embodiment of the retired battery testing method provided by the present invention; Figure 2 This is a schematic diagram of an embodiment of the retired battery testing device provided by the present invention; Figure 3 This is a schematic diagram of an embodiment of the battery management system provided by the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0020] In the description of the embodiments of the present invention, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0021] The terms "first," "second," etc., used in the embodiments of this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a technical feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature.

[0022] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0023] This invention provides a method, apparatus, battery management system, and storage medium for testing retired batteries, which are described below.

[0024] Figure 1 This is a schematic flowchart of an embodiment of the retired battery testing method provided by the present invention. The executing entity in this embodiment is an electrical device or a control module within the electrical device. The control module can be a Battery Management System (BMS), a Vehicle Control Unit (VCU), etc. This embodiment uses a BMS as an example for detailed explanation. The retired battery testing method includes: S101. Obtain multi-source internal characteristic data of the retired battery to be tested, wherein the internal characteristic data includes at least one of AC impedance spectrum, pulse characteristic data, and constant current discharge curve.

[0025] Internal characteristic data refers to parameters directly related to the microscopic electrochemical processes inside the battery, reflecting the intrinsic causes and accuracy of battery aging, such as electrode material structure, active lithium-ion content, and interfacial reaction impedance. Multi-source internal characteristic data refers to internal characteristic data from multiple data sources. In this embodiment, the internal characteristic data includes at least one of AC impedance spectroscopy, pulse characteristic data, and constant current discharge curves.

[0026] Specifically, an AC impedance spectroscopy analyzer can be used to apply a small-amplitude sinusoidal excitation signal with a frequency from 10kHz to 0.1Hz when the retired battery to be tested is in a preset percentage state of charge, such as around 50% state of charge. The AC impedance spectrum of the retired battery can be measured and recorded. This AC impedance spectrum can sensitively reflect the internal ohmic resistance, charge transfer process and ion diffusion behavior of the battery.

[0027] Using a pulse characteristic tester, a short-duration (e.g., 10 seconds) high-current (e.g., 1C rate) discharge pulse is applied to the battery, followed by resting. The dynamic response of the terminal voltage of the retired battery during the pulse and the resting relaxation process is recorded, i.e., pulse characteristic data. This pulse characteristic data is used to analyze the polarization characteristics and dynamic internal resistance of the retired battery.

[0028] Using a constant current discharge tester, the battery is subjected to constant current discharge (e.g., 0.5C rate) under a specific temperature-controlled environment until the cutoff voltage is reached, and the constant current discharge curve of the entire process is recorded. This constant current discharge curve is used as a direct basis for calculating the actual capacity of retired batteries.

[0029] Understandably, this embodiment obtains a multi-source dataset that comprehensively characterizes the internal health state of the battery from different dimensions by using multi-source internal characteristic data of the retired battery to be tested. Simultaneously, the AC impedance spectroscopy in this embodiment applies a series of frequency perturbation signals and completes the measurement within a few minutes; the pulse characteristic means that a pulse followed by relaxation typically completes within a few minutes; the constant current discharge curve can collect curve segments within a certain initial time window without requiring a complete charge-discharge cycle, achieving rapid acquisition of internal characteristic data that quickly reflects the internal state process of the battery, significantly shortening the detection time and accelerating the detection speed of retired batteries.

[0030] S102. Input the multi-source internal characteristic data into the battery health status assessment model to obtain the health status assessment result of the retired battery output by the battery health status assessment model; wherein, the battery health status assessment model is a model after fusing a physical model based on electrochemical mechanism and a machine learning model.

[0031] The health status assessment results include parameters such as the capacity, internal resistance, and health status percentage of the retired battery.

[0032] A physical model based on electrochemical mechanisms is a mathematical equation or circuit model established based on the actual electrochemical reactions and physical processes occurring inside the battery. For example, this physical model can be an equivalent circuit model of electrochemical impedance spectroscopy, used to fit the equivalent circuit component parameters from the AC impedance spectrum. The equivalent circuit component parameters include at least one of ohmic internal resistance, charge transfer impedance, and Warburg impedance. The physical model is interpretable, meaning that every variable and parameter in the physical model has a clear physical meaning.

[0033] Machine learning models refer to regression or classification models obtained by training retired batteries on their electrical performance characteristics and health status using machine learning algorithms such as deep learning or support vector machines.

[0034] The battery health status assessment model is obtained by training the electrochemical mechanism physical model as physical prior knowledge of a specific network layer in a machine learning algorithm.

[0035] The battery health status assessment model is a fusion of a physical model and a machine learning model. This fusion is achieved through at least one of the following methods: extracting physically meaningful feature parameters from multi-source internal characteristic data and using these parameters as input features for training and prediction in the machine learning model; weighting or nonlinearly fusing the outputs of the electrochemical mechanism physical model and the machine learning model at the decision layer to train and predict the machine learning model; incorporating physical constraints from the physical model as part of the loss function of the machine learning model during training to ensure that the predictions conform to the electrochemical mechanism. These physical constraints include: a negative correlation between battery capacity and internal resistance, or a mapping relationship between battery health status and pulse voltage relaxation time characteristics; and constructing the physical model as prior physical knowledge for a specific network layer in the machine learning model, where the electrochemical impedance spectroscopy equivalent circuit model is encoded as a differentiable layer of the neural network, and the parameters of its circuit components are learned and updated through network training.

[0036] Specifically, multi-source internal characteristic data is used as the battery health status assessment model, and the output of the battery health status assessment model is used as the health status assessment result of the retired battery. Since the battery health status assessment model integrates a physical model based on electrochemical mechanisms and a machine learning model, the physical model is used to explain and predict the electrochemical behavior of the battery, so that the prediction results of the physical model are constrained by physical laws, avoiding the prediction errors that may occur in purely data-driven models that violate common sense. The machine learning model is used to analyze and process the collected multi-source internal characteristic data, realizing the rapid and accurate assessment of the battery health status of retired batteries. By integrating electrochemical mechanisms and machine learning, the accuracy and reliability of detection are improved, and human error is reduced.

[0037] In summary, the retired battery detection method provided by this invention acquires multi-source internal characteristic data of the retired battery to be tested. This internal characteristic data includes at least one of AC impedance spectroscopy, pulse characteristic data, and constant current discharge curves. This provides a comprehensive multi-source dataset characterizing the internal health state of the battery from different dimensions, and the acquisition process does not require a complete charge-discharge cycle. This achieves rapid acquisition of internal characteristic data that quickly reflects the internal state process of the battery, shortening the detection time. The multi-source internal characteristic data is input into a battery health status assessment model to obtain the health status assessment result of the retired battery output by the model. The battery health status assessment model is a fusion of a physical model based on electrochemical mechanisms and a machine learning model. Since the physical model is used to explain and predict the electrochemical behavior of the battery, its prediction results are constrained by physical laws, avoiding the common-sense prediction errors that may occur with purely data-driven models. The machine learning model is used to analyze and process the acquired multi-source internal characteristic data, achieving rapid and accurate assessment of the battery health status of the retired battery. Through the fusion of electrochemical mechanisms and machine learning, the detection accuracy of the retired battery's health status is improved, and human error is reduced.

[0038] In some embodiments of the present invention, step S102 includes: S201. Input the multi-source internal characteristic data into the physical model to obtain the first electrical performance feature extracted by the physical model; S202. Input the first electrical performance feature into the machine learning model to perform health status assessment and prediction, and obtain the health status assessment result.

[0039] The machine learning model can be an SVR model, trained using a large amount of historical battery data. The input to the training data consists of electrical performance characteristics extracted from retired battery samples using the same physical model, while the output is the actual capacity and internal resistance of the samples, accurately measured through standard experiments.

[0040] Specifically, a physical model can be used to process the multi-source intrinsic characteristic data to obtain a first electrical performance feature. This first electrical performance feature is then used as input to a machine learning model, and the output of the machine learning model is the health status assessment result. Understandably, because a physical model extracts electrical performance features with clear electrochemical significance from the intrinsic characteristic data, the input to the machine learning model becomes interpretable feature parameters, enhancing the interpretability and reliability of the health status assessment results and improving detection accuracy and efficiency.

[0041] In one specific implementation, an equivalent circuit model can be selected as the physical model, such as [R]. e (R mt C ct (W)] model. Where R e R represents ohmic resistance. mt With C ct The parallel circuit represents the charge transfer process, and W represents the Warburg impedance characterizing ion diffusion. A nonlinear least-squares fitting algorithm is used to fit the measured AC impedance spectrum data with the theoretical spectrum of the selected equivalent circuit model. This fitting process is automatically completed by dedicated software or a built-in algorithm, ultimately outputting a set of accurate first electrical performance characteristics, including: ohmic internal resistance R. e Charge transfer impedance R mt Double-layer capacitance C ct The Warburg coefficient W, and other parameters, all have clear electrochemical and physical meanings, constituting the electrical performance characteristic vector. The vector corresponding to the first electrical performance characteristic (R...) e R mt C ct The model takes W as input to the machine learning model and outputs a predicted value for the state of health (SOH), capacity, and estimated value for the DC internal resistance of the retired battery.

[0042] In some embodiments of the present invention, the machine learning model includes a first machine learning model and a second machine learning model, wherein the second machine learning model includes a feature extraction sub-model and a classification sub-model; step S102 includes: S301. Input the multi-source internal characteristic data into the physical model to obtain the second electrical performance feature extracted by the physical model, and input the second electrical performance feature into the first machine learning model to perform health status assessment and prediction to obtain the first prediction result; S302. Using the feature extraction sub-model in the second machine model, obtain the third electrical performance feature extracted by the physical model from the pulse characteristic data or constant current discharge, and input the third electrical performance feature into the classification sub-model in the machine learning model to obtain the second prediction result; S303. The first prediction result and the second prediction result are fused to obtain the health status assessment result.

[0043] In this model, the first machine learning model is cascaded with the physical model, while the feature extraction sub-model and the classification sub-model are cascaded in the second machine model.

[0044] Specifically, the first prediction result in this embodiment can be calculated in the same way as the health assessment result calculated in steps S201-S202, and its determination method will not be repeated here.

[0045] The pulse characteristic data and / or constant current discharge curve are input into the second machine learning model, which is a composite model containing a feature extraction sub-model and a classification sub-model. The feature extraction sub-model, such as a one-dimensional convolutional neural network, extracts deep and abstract third electrical performance features from the pulse voltage response curve or constant current discharge curve. Then, the third electrical performance features are input into the classification sub-model, such as a fully connected neural network, which analyzes the data and outputs a second prediction result. The second prediction result is then fused using a weighted average method, that is, a weight is assigned to the first prediction result and the second prediction result to calculate a weighted average as the final health status assessment result.

[0046] Understandably, this embodiment constructs two complementary and independent evaluation pathways. The first pathway is based on solid electrochemical principles, exhibiting strong anti-interference capabilities and high reliability; the second pathway can extract deep features from complex curves, resulting in a final evaluation result that is both interpretable and highly accurate. Furthermore, due to learning from different data sources and model architectures, the model has stronger generalization capabilities and can more accurately handle retired batteries from different sources with different aging modes, thereby improving the overall efficiency and accuracy of health status assessment.

[0047] In some embodiments of the present invention, the physical model includes a physical rule-based verifier and a health status corrector; step S102 includes: S401. The machine learning model is used to process the multi-source intrinsic characteristic data to generate predicted battery parameter values; S402. Input the predicted battery parameter values ​​into the physical rule validator in the physical model and output the verification results; S403. When the verification result does not meet the physical rules, the predicted value of the battery parameters is corrected using the health status corrector to obtain the health status assessment result.

[0048] The machine learning model can be a pre-trained deep learning neural network that analyzes the input data and outputs a set of preliminary battery parameter predictions, including predicted capacity, predicted internal resistance, and predicted health status.

[0049] The physical rule validator is a decision logic module with built-in electrochemical common-sense rules. After receiving battery parameter predictions from a machine learning model, it checks whether these predictions conform to basic physical rules. In this embodiment, the rule could be a negative correlation constraint between battery capacity and internal resistance; or a quantitative mapping relationship between battery health status and pulse voltage relaxation time characteristics. The validator performs the following operations: it checks the input predicted capacity C and predicted internal resistance R. A capacity-internal resistance critical relationship curve is preset. The validator determines whether the point (C, R) lies within this critical curve. The output verification result can be a Boolean logic value (yes / no) or a confidence score, indicating whether the current prediction satisfies the physical rules.

[0050] The health status corrector is a rule-based or algorithm-based correction module that is activated when the validator issues a "not satisfied" signal. Its correction strategy can be a lookup table approach: a (capacity, internal resistance) matching lookup table based on historical data is pre-stored. When the machine learning model outputs an abnormal combination of (high capacity, high internal resistance), the corrector retrieves the most probable reasonable capacity value historically associated with that internal resistance value from the lookup table, and replaces the original predicted capacity with this value.

[0051] Specifically, if the verification result meets the physical rules, the original predicted value of the machine learning model is directly output; if the verification result does not meet the physical rules, the predicted battery parameters are corrected using a health status corrector to obtain the health status assessment result. Understandably, this embodiment uses a validator to filter the output of the purely data-driven model, effectively filtering and identifying erroneous predictions that violate basic electrochemical laws, avoiding unreasonable health status assessment results, and improving the accuracy of the health status assessment results through calibration.

[0052] In some embodiments of the present invention, after step S102, the method further includes: S501. Based on the multi-source internal characteristic data and health status assessment results of multiple retired batteries, a battery parameter database is constructed.

[0053] Specifically, a structured and searchable battery parameter database can be constructed based on the multi-source internal characteristic data of multiple retired batteries and their corresponding health status assessment results. This database can be used to store and manage the collected multi-source data, and it can support rapid data retrieval and analysis, and provide data support for the training of battery health status assessment models and / or battery health status assessment.

[0054] In some embodiments of the present invention, after step S102, the method further includes: S601. Based on the health status assessment results, the retired batteries are sorted into a tiered reuse system at the package level, module level, or cell level.

[0055] Specifically, after obtaining the health status assessment results (including capacity, internal resistance, and SOH) of each retired battery cell, the retired batteries are sorted into a tiered reuse system at the whole pack level, module level, or cell level. Based on the test results, the retired batteries are sorted to form a three-level tiered reuse system of whole pack, module, and cell, realizing the refined and maximized utilization of retired battery resources.

[0056] In some embodiments of the present invention, after step S601, the method further includes: S701. Based on the sorting results, use a genetic algorithm or particle swarm optimization algorithm to perform topological reconstruction on the battery cells or battery packs in the sorted whole pack, module or cell-level cascade reuse system to form a recombinant battery pack.

[0057] Specifically, from the sorted results, key parameters of the multiple cells to be reconstructed are obtained, such as the precise capacity, DC internal resistance, and open-circuit voltage of each cell. The topology reconstruction problem is modeled as an optimization problem, with optimization objectives and constraints set, and solved using a genetic algorithm or particle swarm optimization algorithm. Based on the solution results, an optimal topology connection graph is generated, and the cells are connected by laser welding or other methods to form a highly consistent reconstructed battery pack. Understandably, topology reconstruction using a genetic algorithm or particle swarm optimization algorithm can minimize the differences between parallel branches and improve the consistency of the reconstructed battery pack.

[0058] To better implement the decommissioned battery detection method in this embodiment of the invention, based on the decommissioned battery detection method, correspondingly, as follows: Figure 2 As shown, this embodiment of the invention also provides a decommissioned battery testing device, the decommissioned battery testing device 200 comprising: The acquisition unit 201 is used to acquire multi-source internal characteristic data of the retired battery to be tested, wherein the internal characteristic data includes at least one of AC impedance spectrum, pulse characteristic data, and constant current discharge curve; The detection unit 202 is used to input the multi-source internal characteristic data into the battery health status assessment model and obtain the health status assessment result of the retired battery output by the battery health status assessment model. The battery health status assessment model is a fusion of a physical model based on electrochemical mechanisms and a machine learning model.

[0059] The retired battery testing device 200 provided in the above embodiments can realize the technical solutions described in the above retired battery testing method embodiments. The specific implementation principles of each module or unit can be found in the corresponding content in the above retired battery testing method embodiments, and will not be repeated here.

[0060] like Figure 3 As shown, the present invention also provides a battery management system 300. The battery management system 300 includes a processor 301, a memory 302, and a display 303. Figure 3 Only some of the components of the battery management system 300 are shown; however, it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0061] In some embodiments, processor 301 may be a central processing unit (CPU), microprocessor, or other data processing chip, used to run program code stored in memory 302 or process data, such as the decommissioned battery detection method of the present invention.

[0062] In some embodiments, processor 301 may be a single server or a group of servers. The server group may be centralized or distributed. In some embodiments, processor 301 may be local or remote. In some embodiments, processor 301 may be implemented on a cloud platform. In one embodiment, the cloud platform may include a private cloud, public cloud, hybrid cloud, community cloud, distributed cloud, intranet, multi-cloud, etc., or any combination thereof.

[0063] In some embodiments, memory 302 may be an internal storage unit of the battery management system 300, such as a hard disk or memory of the battery management system 300. In other embodiments, memory 302 may also be an external storage device of the battery management system 300, such as a pluggable hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., provided on the battery management system 300.

[0064] Furthermore, the memory 302 may include both internal storage units of the battery management system 300 and external storage devices. The memory 302 is used to store application software and various types of data for which the battery management system 300 is installed.

[0065] In some embodiments, display 303 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. Display 303 is used to display information from the battery management system 300 and to display a visual user interface. The components 301-303 of the battery management system 300 communicate with each other via a system bus.

[0066] In one embodiment, when the processor 301 executes the retired battery detection program in the memory 302, the following steps can be implemented: Acquire multi-source internal characteristic data of the retired battery to be tested, wherein the internal characteristic data includes at least one of AC impedance spectrum, pulse characteristic data, and constant current discharge curve; The multi-source internal characteristic data are input into the battery health status assessment model to obtain the health status assessment result of the retired battery output by the battery health status assessment model. The battery health status assessment model is a fusion of a physical model based on electrochemical mechanisms and a machine learning model.

[0067] It should be understood that when the processor 301 executes the retired battery detection program in the memory 302, in addition to the functions mentioned above, it can also perform other functions, as detailed in the description of the corresponding method embodiments above.

[0068] Furthermore, this embodiment of the invention does not specifically limit the type of battery management system 300 mentioned. The battery management system 300 can be a portable battery management system for mobile phones, tablets, personal digital assistants (PDAs), wearable devices, laptops, etc. Exemplary embodiments of the portable battery management system include, but are not limited to, portable battery management systems running iOS, Android, Microsoft, or other operating systems. The aforementioned portable battery management system can also be other portable battery management systems, such as laptops with touch-sensitive surfaces (e.g., touch panels). It should also be understood that in some other embodiments of the invention, the battery management system 300 may not be a portable battery management system, but rather a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0069] Accordingly, this application also provides a computer-readable storage medium for storing a computer-readable program or instruction. When the program or instruction is executed by a processor, it can implement the steps or functions of the decommissioned battery detection method provided in the above-described method embodiments.

[0070] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.), and the computer program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0071] The above provides a detailed description of the retired battery testing method, apparatus, battery management system, and storage medium provided by the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, those skilled in the art will recognize that there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for testing decommissioned batteries, characterized in that, The method includes: Acquire multi-source internal characteristic data of the retired battery to be tested, wherein the internal characteristic data includes at least one of AC impedance spectrum, pulse characteristic data, and constant current discharge curve; The multi-source internal characteristic data are input into the battery health status assessment model to obtain the health status assessment result of the retired battery output by the battery health status assessment model. The battery health status assessment model is a fusion of a physical model based on electrochemical mechanisms and a machine learning model.

2. The method for testing decommissioned batteries according to claim 1, characterized in that, The step of inputting the multi-source internal characteristic data into the battery health status assessment model and obtaining the health status assessment result of the retired battery output by the battery health status assessment model includes: The internal characteristic data from multiple sources are input into the physical model to obtain the first electrical performance feature extracted by the physical model; The first electrical performance feature is input into the machine learning model to perform health status assessment and prediction, and the health status assessment result is obtained.

3. The method for testing decommissioned batteries according to claim 1, characterized in that, The machine learning model includes a first machine learning model and a second machine learning model, the second machine learning model including a feature extraction sub-model and a classification sub-model; the step of inputting the multi-source intrinsic characteristic data into the battery health status assessment model and obtaining the health status assessment result of the retired battery output by the battery health status assessment model includes: The multi-source intrinsic characteristic data are input into the physical model to obtain the second electrical performance feature extracted by the physical model, and the second electrical performance feature is input into the first machine learning model to perform health status assessment and prediction to obtain the first prediction result. The third electrical performance feature extracted by the physical model is obtained from the pulse characteristic data or constant current discharge through the feature extraction sub-model in the second machine model, and the third electrical performance feature is input into the classification sub-model in the machine learning model to obtain the second prediction result; The first prediction result and the second prediction result are fused together to obtain the health status assessment result.

4. The method for testing decommissioned batteries according to claim 2, characterized in that, The physical model includes a physical rule-based verifier and a health status corrector; the step of inputting the multi-source internal characteristic data into the battery health status assessment model and obtaining the health status assessment result of the retired battery output by the battery health status assessment model includes: The machine learning model is used to process the multi-source intrinsic characteristic data to generate predicted battery parameter values; The predicted battery parameters are input into the physical rule validator in the physical model, and the validator results are output. When the verification result does not meet the physical rules, the predicted value of the battery parameter is corrected using the health status corrector to obtain the health status assessment result.

5. The method for testing decommissioned batteries according to claim 1, characterized in that, After obtaining the health status assessment result of the retired battery output by the battery health status assessment model, the method further includes: A battery parameter database is constructed based on the multi-source internal characteristic data and health status assessment results of multiple retired batteries.

6. The method for testing decommissioned batteries according to claim 1, characterized in that, After obtaining the health status assessment result of the retired battery output by the battery health status assessment model, the method further includes: Based on the health status assessment results, the retired batteries are sorted into a tiered reuse system at the package level, module level, or cell level.

7. The method for testing decommissioned batteries according to claim 6, characterized in that, After sorting the retired batteries into a cascade reuse system at the pack level, module level, or cell level based on the health status assessment results, the system further includes: Based on the sorting results, the topology of the battery cells or battery packs in the sorted whole pack, module or cell-level cascade reuse system is reconstructed using genetic algorithms or particle swarm optimization algorithms to form recombinant battery packs.

8. A device for testing decommissioned batteries, characterized in that, The device includes: The acquisition unit is used to acquire multi-source internal characteristic data of the retired battery to be tested, wherein the internal characteristic data includes at least one of AC impedance spectrum, pulse characteristic data, and constant current discharge curve; The detection unit is used to input the multi-source internal characteristic data into the battery health status assessment model and obtain the health status assessment result of the retired battery output by the battery health status assessment model. The battery health status assessment model is a fusion of a physical model based on electrochemical mechanisms and a machine learning model.

9. A battery management system, characterized in that, It includes a memory and a processor, wherein the memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the decommissioned battery detection method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Used to store computer-readable programs or instructions, which, when executed by a processor, can perform the steps in the decommissioned battery detection method according to any one of claims 1 to 7.

Images