A method, system, and electronic device for assessing the health status of retired batteries.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-05-29
- Publication Date
- 2026-08-14
AI Technical Summary
[0005]有鉴于此，有必要提供一种退役电池健康状态评估方法、系统及电子设备，用以解决现有技术中存在的模型泛化性差、精度不足的技术问题
[0020]本发明的有益效果是：本发明提供的退役电池健康状态评估方法，首先获取退役电池的电化学机理特征和运行数据特征，通过模态对齐算法将电化学机理特征和运行数据特征映射至统一特征空间并最小化其分布差异，实现了机理知识与数据驱动的深层融合，有效解决了多源异构特征维度不统一、分布不一致导致的融合失真问题，提升了后续健康状态估计的一致性与稳定性。进一步地，基于机理模型置信度与有效数据量动态计算机理模态、数据模态权重，当机理模型可靠时权重向机理侧倾斜，当数据充足时权重向数据侧倾斜，避免了固定权重融合在工况变化时性能下降的问题，无需人工干预，对未知类型退役电池仍具备强泛化性。
Smart Images
Figure CN122568285A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of battery monitoring technology, specifically to a method, system, and electronic device for assessing the health status of retired batteries. Background Technology
[0002] Retired power batteries are characterized by large batch dispersion, complex aging mechanisms, and scarce available labeled data. Accurate state of health (SOH) assessment is a crucial step in battery reuse and remanufacturing. Currently, retired battery SOH assessment mainly falls into two technical routes: electrochemical mechanism assessment and data-driven assessment. Mechanism assessment relies on physical models such as equivalent circuit models and electrochemical models to calculate the SOH by identifying internal battery parameters, possessing physical interpretability. Data-driven assessment relies on operational data such as voltage, temperature, and internal resistance, using machine learning and deep learning to fit battery aging patterns, offering strong adaptability. With the diversification of retired battery types and the increasing complexity of scenarios, a single assessment route is insufficient to meet the engineering requirements of high precision and strong generalization. Therefore, the fusion of mechanism and data assessment has become the mainstream research direction in the industry.
[0003] Existing technologies for assessing the health status of retired batteries have significant drawbacks: First, most employ purely data-driven deep learning models, relying on extensive labeled charge-discharge data for offline training. While these models offer high accuracy for battery types present in the training set, they become inaccurate when faced with retired batteries not found in the training set. The highly discrete nature of retired batteries and their significantly different aging behaviors compared to fresh batteries or those under standard operating conditions lead to assessment errors. Second, traditional equivalent circuit and electrochemical models are based on idealized assumptions, making it difficult to accurately characterize the complex nonlinear aging behavior within retired batteries. This is especially true when battery aging is uneven and individual variations are large, significantly increasing the parameter identification error of mechanistic models and reducing the accuracy of health status estimation.
[0004] In summary, existing technologies have not yet achieved a deep integration of electrochemical mechanism modes and operational data modes, and cannot simultaneously take into account evaluation accuracy, generalization ability, and adaptation to the discreteness and scarcity of labels of retired batteries. Summary of the Invention
[0005] In view of this, it is necessary to provide a method, system and electronic device for assessing the health status of retired batteries, in order to solve the technical problems of poor model generalization and insufficient accuracy in the existing technology.
[0006] To address the aforementioned technical problems, in a first aspect, the present invention provides a method for assessing the health status of retired batteries, comprising: To obtain the electrochemical mechanism characteristics and operational data characteristics of retired batteries; Based on the modal alignment algorithm, the electrochemical mechanism features and the operational data features are mapped to a unified feature space, and the distribution difference between the electrochemical mechanism features and the operational data features in the unified feature space is minimized to obtain the aligned electrochemical mechanism features and operational data features. Based on the aligned electrochemical mechanism characteristics and operational data characteristics, respectively, the estimated values of the electrochemical mechanism mode health status and the estimated values of the data mode health status are calculated; Based on the confidence level of the mechanism model and the amount of effective data, the weights of the electrochemical mechanism mode and the weights of the operating data mode are dynamically calculated. The health status assessment value is obtained by weighted fusion of the electrochemical mechanism mode weights, the operating data mode weights, the electrochemical mechanism mode health status estimate, and the operating data mode health status estimate.
[0007] In one possible implementation, the electrochemical mechanism features include SEI film impedance, polarization resistance, and lithium-ion concentration, and the operating data features include voltage curves, temperature sequences, and internal resistance variations.
[0008] In one possible implementation, the modality alignment algorithm includes: The electrochemical mechanism features are mapped to the hidden feature space using a mechanism feature extractor to obtain the first hidden feature. The running data features are mapped to the same hidden feature space by a data feature extractor to obtain the second hidden feature. Based on the domain discriminator, source modality discrimination is performed on the first hidden feature and the second hidden feature, and adversarial migration loss is calculated; Based on the health status regressor, predict the health status from the first hidden feature and the second hidden feature respectively, and calculate the regression task loss; The total loss function is the weighted difference between minimizing the regression task loss and the adversarial transfer loss. The mechanism feature extractor, data feature extractor, domain discriminator and health status regressor are trained together.
[0009] In one possible implementation, the formula for calculating the adversarial migration loss is:
[0010] In the formula, To combat migration losses, For mechanism feature extractor, For data feature extractor, For domain discriminator, Characterized by electrochemical mechanisms, For running data characteristics.
[0011] In one possible implementation, the electrochemical mechanism modal health state estimate and the operational data modal health state estimate are calculated based on the aligned electrochemical mechanism characteristics and operational data characteristics, respectively, including: The aligned electrochemical mechanism features are input into the health state regressor, which outputs an estimated value of the electrochemical mechanism mode health state. The aligned operational data features are input into the health status regressor, and the health status regressor outputs the operational data modal health status estimate. The health state regressor is a pre-trained deep neural network, and its parameters are shared when calculating the electrochemical mechanism modal health state estimate and the operational data modal health state estimate.
[0012] In one possible implementation, the formulas for calculating the electrochemical mechanism modal health state estimate and the operational data modal health state estimate are as follows:
[0013]
[0014] In the formula, For electrochemical mechanism mode weights, To run the data modal weights, For the confidence level of the mechanistic model, For effective data volume, , This is the preset sensitivity coefficient.
[0015] In one possible implementation, the formula for calculating the confidence level of the mechanistic model is:
[0016] In the formula, This represents the standard deviation of the output voltage prediction error of the mechanistic model.
[0017] In one possible implementation, the method further includes a small-sample cold start: When the retired battery to be evaluated is a new type of battery that does not appear in the training set, the parameters of the pre-trained mechanism feature extractor, data feature extractor, and domain discriminator are fixed, and the health status regressor is fine-tuned using the charge-discharge cycle data of the new type of battery.
[0018] On the other hand, the present invention also provides a health status assessment system for retired batteries, comprising: The data acquisition module is used to electrically connect to the retired battery, and to collect voltage, current and temperature signals in real time during the charging and discharging process of the retired battery, and convert the voltage, current and temperature signals into digital voltage data, current data and temperature data; The processing module, electrically connected to the data acquisition module, is used to receive the voltage data, current data, and temperature data, and specifically includes: The feature acquisition module is used to acquire the electrochemical mechanism characteristics and operational data characteristics of retired batteries; The feature alignment module is used to map the electrochemical mechanism features and the operational data features to a unified feature space based on the modal alignment algorithm, and to minimize the distribution difference between the electrochemical mechanism features and the operational data features in the unified feature space, so as to obtain the aligned electrochemical mechanism features and operational data features. The health status calculation module is used to calculate the estimated values of the electrochemical mechanism mode health status and the estimated values of the operational data mode health status based on the aligned electrochemical mechanism characteristics and operational data characteristics, respectively. The weighted fusion module is used to dynamically calculate the electrochemical mechanism mode weights and operational data mode weights based on the confidence level of the mechanism model and the amount of effective data. Based on the electrochemical mechanism mode weights, operational data mode weights, electrochemical mechanism mode health status estimates, and operational data mode health status estimates, a weighted fusion is performed to obtain a health status assessment value. The display module, electrically connected to the processing module, is used to output and display the health status assessment value.
[0019] Secondly, the present invention also provides an electronic device, comprising: Data acquisition unit, used to acquire voltage, current and temperature data of retired batteries; A display used to show health status assessment values; Memory, used to store programs; The processor, coupled to the data acquisition unit, the display, and the memory, is used to execute the program stored in the memory to implement the steps in the retired battery health status assessment method described in any of the above implementations.
[0020] The beneficial effects of this invention are as follows: The method for assessing the health status of retired batteries provided by this invention first obtains the electrochemical mechanism characteristics and operational data characteristics of the retired batteries. Then, through a modal alignment algorithm, the electrochemical mechanism characteristics and operational data characteristics are mapped to a unified feature space and their distribution differences are minimized. This achieves a deep fusion of mechanistic knowledge and data-driven approaches, effectively solving the fusion distortion problem caused by inconsistent dimensions and distributions of multi-source heterogeneous features, and improving the consistency and stability of subsequent health status estimation. Furthermore, based on the confidence level of the mechanism model and the amount of effective data, the weights of the mechanism modes and data modes are dynamically calculated. When the mechanism model is reliable, the weights tilt towards the mechanism side; when the data is sufficient, the weights tilt towards the data side. This avoids the performance degradation problem of fixed-weight fusion when operating conditions change, requires no manual intervention, and still has strong generalization ability for unknown types of retired batteries. Attached Figure Description
[0021] 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.
[0022] Figure 1 A schematic flowchart of an embodiment of the method for assessing the health status of retired batteries provided by the present invention; Figure 2 This is a schematic flowchart of an embodiment of the modal alignment algorithm provided by the present invention; Figure 3 Provided by the present invention Figure 1 A schematic diagram of an embodiment of S103; Figure 4 A schematic diagram of an embodiment of the retired battery health status assessment system provided by the present invention; Figure 5 A schematic diagram of an embodiment of the electronic device provided by the present invention. Detailed Implementation
[0023] 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.
[0024] 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.
[0025] 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.
[0026] 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.
[0027] Before demonstrating the embodiments, the following terms will be explained.
[0028] Electrochemical mechanism characteristics: These are physical parameters that characterize the internal electrochemical aging state of retired batteries. They are identified by electrochemical mechanism models and directly reflect the physical nature of battery degradation. These parameters include SEI film impedance, polarization resistance, lithium-ion concentration, ohmic internal resistance, and maximum usable capacity.
[0029] Operational data characteristics: These refer to the external operating parameters collected from retired batteries under conditions such as charging, discharging, resting, and pulse operation. These parameters are extracted through data processing and reflect changes in the battery's macroscopic performance, including voltage curves, temperature sequences, internal resistance changes, charge / discharge rates, voltage relaxation, and temperature rise rates.
[0030] Mechanism model confidence level: refers to a quantitative evaluation index of the reliability of the output results of the current electrochemical mechanism model, which is calculated based on the error between the predicted value of the battery terminal voltage by the mechanism model and the actual measured value.
[0031] Effective data volume: refers to the number of valid, abnormal, and modelable charge-discharge cycles or time-series data points used for assessing the health status of retired batteries, excluding invalid samples such as short circuits, over-temperature, and missing data.
[0032] This invention provides a method, system, and electronic device for assessing the health status of retired batteries, which will be described below.
[0033] Figure 1This is a schematic flowchart of an embodiment of the retired battery health status assessment method provided by the present invention. The executing entity of the method of the present invention can be a retired battery health status assessment system, specifically executed by an electronic device, host computer, edge computing unit, or processor of battery testing equipment with data processing and model calculation capabilities. It is mainly applied to scenarios of retired power battery recycling, sorting, cascade utilization, and remanufacturing. Figure 1 As shown, the methods for assessing the health status of retired batteries include: S101. Obtain the electrochemical mechanism characteristics and operational data characteristics of retired batteries.
[0034] S102. Based on the modal alignment algorithm, the electrochemical mechanism features and the operational data features are mapped to a unified feature space, and the distribution difference between the electrochemical mechanism features and the operational data features in the unified feature space is minimized to obtain the aligned electrochemical mechanism features and operational data features.
[0035] S103. Based on the aligned electrochemical mechanism characteristics and operating data characteristics, respectively, calculate the estimated values of the physical mode health status and the estimated values of the data mode health status.
[0036] S104. Based on the confidence level of the mechanism model and the amount of effective data, dynamically calculate the electrochemical mechanism mode weight and the operating data mode weight. Based on the electrochemical mechanism mode weight, the operating data mode weight, the electrochemical mechanism mode health status estimate, and the operating data mode health status estimate, perform weighted fusion to obtain the health status assessment value.
[0037] In summary, the method for assessing the health status of retired batteries provided in this invention first acquires the electrochemical mechanism characteristics and operational data characteristics of the retired batteries. Then, a mode alignment algorithm maps these characteristics to a unified feature space and minimizes their distribution differences, achieving a deep fusion of mechanistic knowledge and data-driven approaches. This effectively solves the fusion distortion problem caused by inconsistent dimensions and distributions of multi-source heterogeneous features, improving the consistency and stability of subsequent health status estimation. Furthermore, based on the confidence level of the mechanistic model and the amount of effective data, the weights of the mechanistic modes and data modes are dynamically calculated. When the mechanistic model is reliable, the weights tilt towards the mechanistic side; when the data is sufficient, the weights tilt towards the data side. This avoids the performance degradation problem of fixed-weight fusion under changing operating conditions, requires no manual intervention, and still possesses strong generalization ability for retired batteries of unknown types.
[0038] In some embodiments of the present invention, the electrochemical mechanism features include SEI film impedance, polarization resistance and lithium ion concentration, and the operating data features include voltage curves, temperature sequences and internal resistance changes.
[0039] It should be noted that, in this embodiment of the invention, constant current charge-discharge, pulse excitation, or static relaxation excitation conditions are applied to retired batteries to collect the timing signals of battery terminal voltage, operating current, and surface temperature. Based on existing second-order RC equivalent circuit models or single-particle electrochemical models, the timing signals are fitted and identified using the least squares method, Kalman filtering algorithm, or recursive parameter identification method to obtain electrochemical mechanism characteristics such as SEI film impedance, polarization resistance, lithium ion concentration, ohmic internal resistance, and polarization capacitance parameters.
[0040] It should also be noted that, in this embodiment of the invention, the original signals of the battery during charging, discharging and resting processes are collected in real time by voltage sensors, current sensors, temperature sensors and internal resistance testers. Then, the original signals are preprocessed by denoising, filtering, outlier removal and normalization to obtain time-series data such as voltage curves, temperature sequences, internal resistance change curves and charging and discharging capacity. Finally, based on the time-series data, operating data features such as slope, variance, rate of change, extreme values and relaxation offset are extracted.
[0041] In some embodiments of the present invention, such as Figure 2 As shown, the modal alignment algorithm in step S102 includes: S201. The electrochemical mechanism features are mapped to the hidden feature space by the mechanism feature extractor to obtain the first hidden feature; S202. The running data features are mapped to the same hidden feature space by a data feature extractor to obtain the second hidden feature. S203. Based on the domain discriminator, perform source mode discrimination on the first hidden feature and the second hidden feature, and calculate the adversarial migration loss; S204. Based on the health status regressor, predict the health status from the first hidden feature and the second hidden feature respectively, and calculate the regression task loss; S205. Using the weighted difference between minimizing the regression task loss and the adversarial transfer loss as the total loss function, jointly train the mechanism feature extractor, data feature extractor, domain discriminator, and health status regressor.
[0042] It should be noted that both the mechanism feature extractor and the data feature extractor are multi-layer fully connected neural networks, and their output dimensions are the same, being the dimension of the hidden feature space. The domain discriminator is a binary classification fully connected network used to distinguish whether the input hidden features come from the mechanism branch or the data branch. The health status regressor is a fully connected regression network used to predict health status values from the hidden features.
[0043] In some embodiments of the present invention, the formula for calculating the adversarial migration loss is as follows:
[0044] In the formula, For mechanism feature extractor, For data feature extractor, For domain discriminator, Characterized by electrochemical mechanisms, For running data characteristics.
[0045] It should be understood that the adversarial migration loss is used to measure the distributional difference between mechanistic features and data features. The smaller the loss, the closer the bimodal features are, thus making the fusion more reliable and the evaluation more stable.
[0046] In some embodiments of the present invention, the formula for calculating the total loss function is as follows:
[0047] In the formula, To combat migration losses, Health status assessment task loss (MAE loss); To combat migration losses, This is the balance coefficient, with a value ranging from 0.1 to 0.5.
[0048] The embodiments of the present invention use adversarial training to make the distribution of dual-modal features as close as possible in a unified space, thereby achieving deep fusion, so that the final health status assessment of retired batteries has the advantages of both mechanism interpretability and data fitting.
[0049] In some embodiments of the present invention, such as Figure 3 As shown, step S103 includes: S301. Input the aligned electrochemical mechanism features into the health state regressor, and the health state regressor outputs the electrochemical mechanism mode health state estimate. S302. Input the aligned running data features into the health status regressor, and the health status regressor outputs the running data modal health status estimate. The health state regressor is a pre-trained deep neural network, and its parameters are shared when calculating the electrochemical mechanism modal health state estimate and the operational data modal health state estimate.
[0050] It should be understood that the parameters of the health state regressor are shared when calculating the electrochemical mechanism modal health state estimate and the operational data modal health state estimate, which can reduce the number of model parameters, reduce computational complexity and overfitting risk. At the same time, by learning a universal mapping from a unified hidden space to the health state through the regressor, it ensures that the dual-path outputs are on the same standard scale, avoids numerical deviations caused by different networks, and enhances the stability of the evaluation results after fusion.
[0051] In some embodiments of the present invention, the calculation formulas for the electrochemical mechanism modal health state estimate and the operational data modal health state estimate are as follows:
[0052]
[0053] In the formula, For electrochemical mechanism mode weights, To run the data modal weights, For the confidence level of the mechanistic model, For effective data volume, , This is the preset sensitivity coefficient.
[0054] It should be understood that the effective data volume The effective data volume is the number of valid and complete charge-discharge cycles retained after data cleaning. First, time-series data of voltage, current, temperature, and internal resistance throughout the entire charge-discharge process of retired batteries are collected. Then, abnormal samples such as overvoltage, overtemperature, sudden current changes, and missing data are removed, noise is filtered, and drift is corrected. Finally, the total number of valid, complete charge-discharge cycles suitable for modeling is calculated, which is the effective data volume. Effective data volume The larger the value, the more abundant the data modality sample and the more stable the data-driven estimation results, thus giving higher weight to the running data modality in adaptive weighted fusion.
[0055] In some embodiments of the present invention, the formula for calculating the confidence level of the mechanism model is as follows:
[0056] In the formula, This represents the standard deviation of the output voltage prediction error of the mechanistic model.
[0057] It should be noted that when the confidence level of the mechanistic model When the value approaches 1, the mode weights of the electrochemical mechanism It is also close to 1, at which point the focus is on mechanism estimation; when the effective data volume is... When the value is large, the operating data modal weights As the number of variables increases, the focus shifts to data estimation, and the exponential function ensures the confidence level of the weighted stochastic model. and effective data volume Smooth changes and avoid abrupt changes caused by threshold switching.
[0058] Compared to traditional fixed-weight fusion, this invention can adjust the fusion results in real time based on current operating conditions and data quality, making the fusion results more reliable. For example, when abnormal battery conditions lead to increased prediction errors in the mechanistic model, the confidence level of the mechanistic model will be adjusted. Automatic reduction and weight shift to the data side avoid interference from erroneous mechanism estimation; when only a small amount of charge / discharge data is available, the effective data volume is [not specified]. The weight is relatively small, and the weight shifts towards the mechanism side.
[0059] In some embodiments of the present invention, the method further includes a small-sample cold start: When the retired battery to be evaluated is a new type of battery that does not appear in the training set, the parameters of the pre-trained mechanism feature extractor, data feature extractor, and domain discriminator are fixed, and the health status regressor is fine-tuned using the charge-discharge cycle data of the new type of battery.
[0060] It should be noted that the embodiments of the present invention use small sample cold start, which does not require retraining of the entire model and only requires 1-3 cycles to complete the adaptation, thereby significantly reducing data dependence, shortening the evaluation cycle, and meeting the needs of rapid on-site detection of retired batteries.
[0061] To better implement the retired battery health status assessment method in this embodiment of the invention, based on the retired battery health status assessment method, correspondingly, as follows: Figure 4 As shown, this embodiment of the invention also provides a retired battery health status assessment system. The retired battery health status assessment system 400 includes: Data acquisition module 401 is used to electrically connect to the retired battery, and to acquire voltage signals, current signals and temperature signals in real time during the charging and discharging process of the retired battery, and convert the voltage signals, current signals and temperature signals into digital voltage data, current data and temperature data; Processing module 402, electrically connected to data acquisition module 401, is used to receive voltage data, current data, and temperature data, specifically including: The feature acquisition module is used to acquire the electrochemical mechanism characteristics and operational data characteristics of retired batteries; The feature alignment module is used to map the electrochemical mechanism features and the operational data features to a unified feature space based on the modal alignment algorithm, and to minimize the distribution difference between the electrochemical mechanism features and the operational data features in the unified feature space, so as to obtain the aligned electrochemical mechanism features and operational data features. The health status calculation module is used to calculate the estimated values of the electrochemical mechanism mode health status and the estimated values of the operational data mode health status based on the aligned electrochemical mechanism characteristics and operational data characteristics, respectively. The weighted fusion module is used to dynamically calculate the electrochemical mechanism mode weights and operational data mode weights based on the confidence level of the mechanism model and the amount of effective data. Based on the electrochemical mechanism mode weights, operational data mode weights, electrochemical mechanism mode health status estimates, and operational data mode health status estimates, a weighted fusion is performed to obtain a health status assessment value. Display module 403, electrically connected to processing module 402, is used to output and display the health status assessment value.
[0062] The retired battery health status assessment system 400 provided in the above embodiments can realize the technical solutions described in the above retired battery health status assessment method embodiments. The specific implementation principles of each module or unit can be found in the corresponding content in the above retired battery health status assessment method embodiments, and will not be repeated here.
[0063] like Figure 5 As shown, the present invention also provides an electronic device 500. The electronic device 500 includes a data acquisition unit 501, a display 502, a memory 503, and a processor 504. Figure 5 Only some components of the electronic device 500 are shown, but it should be understood that it is not required to implement all the components shown, and more or fewer components may be implemented instead.
[0064] In some embodiments, processor 504 may be a central processing unit (CPU), microprocessor, or other data processing chip, used to run program code stored in memory 503 or process data, such as the retired battery health status assessment method of the present invention.
[0065] In some embodiments, processor 504 may be a single server or a group of servers. The server group may be centralized or distributed. In some embodiments, processor 504 may be local or remote. In some embodiments, processor 504 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.
[0066] In some embodiments, memory 503 may be an internal storage unit of electronic device 500, such as a hard disk or memory of electronic device 500. In other embodiments, memory 503 may also be an external storage device of electronic device 500, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. equipped on electronic device 500.
[0067] Furthermore, the memory 503 may include both internal storage units of the electronic device 500 and external storage devices. The memory 503 is used to store application software and various types of data installed on the electronic device 500.
[0068] In some embodiments, display 502 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. Display 502 is used to display information from electronic device 500 and to display a visual user interface. Components 501-504 of electronic device 500 communicate with each other via a system bus.
[0069] In one embodiment, when the processor 504 executes the retired battery health status assessment program in the memory 503, the following steps can be implemented: To obtain the electrochemical mechanism characteristics and operational data characteristics of retired batteries; Based on the modal alignment algorithm, the electrochemical mechanism features and the operational data features are mapped to a unified feature space, and the distribution difference between the electrochemical mechanism features and the operational data features in the unified feature space is minimized to obtain the aligned electrochemical mechanism features and operational data features. Based on the aligned electrochemical mechanism characteristics and operational data characteristics, respectively, calculate the electrochemical mechanism modal health status estimate and the operational data modal health status estimate; Based on the confidence level of the mechanism model and the amount of effective data, the weights of the electrochemical mechanism mode and the weights of the operating data mode are dynamically calculated. The health status assessment value is obtained by weighted fusion of the electrochemical mechanism mode weights, the operating data mode weights, the electrochemical mechanism mode health status estimate, and the operating data mode health status estimate.
[0070] It should be understood that when the processor 504 executes the retired battery health status assessment program in the memory 503, in addition to the functions mentioned above, it can also perform other functions, as detailed in the description of the corresponding method embodiments above.
[0071] 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.
[0072] The above provides a detailed description of the method, apparatus, electronic device, and storage medium for assessing the health status of retired batteries provided by 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, 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 this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.
Claims
1. A method for assessing the health status of retired batteries, characterized in that, include: To obtain the electrochemical mechanism characteristics and operational data characteristics of retired batteries; Based on the modal alignment algorithm, the electrochemical mechanism features and the operational data features are mapped to a unified feature space, and the distribution difference between the electrochemical mechanism features and the operational data features in the unified feature space is minimized to obtain the aligned electrochemical mechanism features and operational data features. Based on the aligned electrochemical mechanism characteristics and operational data characteristics, respectively, calculate the electrochemical mechanism modal health status estimate and the operational data modal health status estimate; Based on the confidence level of the mechanism model and the amount of effective data, the weights of the electrochemical mechanism mode and the weights of the operating data mode are dynamically calculated. The health status assessment value is obtained by weighted fusion of the electrochemical mechanism mode weights, the operating data mode weights, the electrochemical mechanism mode health status estimate, and the operating data mode health status estimate.
2. The method for assessing the health status of decommissioned batteries according to claim 1, characterized in that, The electrochemical mechanism features include SEI film impedance, polarization resistance, and lithium ion concentration, while the operational data features include voltage curves, temperature sequences, and internal resistance variations.
3. The method for assessing the health status of decommissioned batteries according to claim 1, characterized in that, The modality alignment algorithm includes: The electrochemical mechanism features are mapped to the hidden feature space using a mechanism feature extractor to obtain the first hidden feature. The running data features are mapped to the same hidden feature space by a data feature extractor to obtain the second hidden feature. Based on the domain discriminator, source modality discrimination is performed on the first hidden feature and the second hidden feature, and adversarial migration loss is calculated; Based on the health status regressor, predict the health status from the first hidden feature and the second hidden feature respectively, and calculate the regression task loss; The total loss function is the weighted difference between minimizing the regression task loss and the adversarial transfer loss. The mechanism feature extractor, data feature extractor, domain discriminator and health status regressor are trained together.
4. The method for assessing the health status of decommissioned batteries according to claim 3, characterized in that, The formula for calculating the adversarial migration loss is as follows: In the formula, To combat migration losses, For mechanism feature extractor, For data feature extractor, For domain discriminator, Characterized by electrochemical mechanisms, For running data characteristics.
5. The method for assessing the health status of decommissioned batteries according to claim 4, characterized in that, Based on the aligned electrochemical mechanism characteristics and operational data characteristics, respectively, calculate the estimated values of the electrochemical mechanism modal health state and the estimated values of the operational data modal health state, including: The aligned electrochemical mechanism features are input into the health state regressor, which outputs an estimated value of the electrochemical mechanism mode health state. The aligned operational data features are input into the health status regressor, and the health status regressor outputs the operational data modal health status estimate. The health state regressor is a pre-trained deep neural network, and its parameters are shared when calculating the electrochemical mechanism modal health state estimate and the operational data modal health state estimate.
6. The method for assessing the health status of decommissioned batteries according to claim 5, characterized in that, The calculation formulas for the electrochemical mechanism modal health status estimate and the operational data modal health status estimate are as follows: In the formula, For electrochemical mechanism mode weights, To run the data modal weights, For the confidence level of the mechanistic model, For effective data volume, , This is the preset sensitivity coefficient.
7. The method for assessing the health status of decommissioned batteries according to claim 6, characterized in that, The formula for calculating the confidence level of the mechanistic model is as follows: In the formula, This represents the standard deviation of the output voltage prediction error of the mechanistic model.
8. The method for assessing the health status of decommissioned batteries according to claim 1, characterized in that, The method also includes small-sample cold start: When the retired battery to be evaluated is a new type of battery that does not appear in the training set, the parameters of the pre-trained mechanism feature extractor, data feature extractor, and domain discriminator are fixed, and the health status regressor is fine-tuned using the charge-discharge cycle data of the new type of battery.
9. A health status assessment system for retired batteries, characterized in that, include: The data acquisition module is used to electrically connect to the retired battery, and to collect voltage, current and temperature signals in real time during the charging and discharging process of the retired battery, and convert the voltage, current and temperature signals into digital voltage data, current data and temperature data; The processing module, electrically connected to the data acquisition module, is used to receive the voltage data, current data, and temperature data, and specifically includes: The feature acquisition module is used to acquire the electrochemical mechanism characteristics and operational data characteristics of retired batteries; The feature alignment module is used to map the electrochemical mechanism features and the operational data features to a unified feature space based on the modal alignment algorithm, and to minimize the distribution difference between the electrochemical mechanism features and the operational data features in the unified feature space, so as to obtain the aligned mechanism features and data features. The health status calculation module is used to calculate the estimated values of the electrochemical mechanism mode health status and the estimated values of the operational data mode health status based on the aligned electrochemical mechanism characteristics and operational data characteristics, respectively. The weighted fusion module is used to dynamically calculate the electrochemical mechanism mode weights and operational data mode weights based on the confidence level of the mechanism model and the amount of effective data. Based on the electrochemical mechanism mode weights, operational data mode weights, electrochemical mechanism mode health status estimates, and operational data mode health status estimates, a weighted fusion is performed to obtain a health status assessment value. The display module, electrically connected to the processing module, is used to output and display the health status assessment value.
10. An electronic device, characterized in that, include: Data acquisition unit, used to acquire voltage, current and temperature data of retired batteries; A display used to show health status assessment values; Memory, used to store programs; The processor, coupled to the data acquisition unit, the display, and the memory, is used to execute the program stored in the memory to implement the steps in the method for assessing the health status of retired batteries according to any one of claims 1 to 8.