Residual value evaluation method, device and equipment of retired battery and medium

Abstract

The invention provides a residual value evaluation method, device, equipment and medium for a decommissioned battery, and the method comprises the steps: carrying out the terminal voltage detection of the decommissioned battery when the appearance of the decommissioned battery is judged to reach a detection standard through the three-dimensional point cloud data of the decommissioned battery, so as to determine the terminal voltage of the decommissioned battery, performing an electrochemical impedance spectroscopy test on the decommissioned battery to determine a plurality of characteristic parameters of the decommissioned battery; performing dimension reduction processing on the terminal voltage and the plurality of characteristic parameters to obtain a target characteristic data set of the retired battery; and inputting the target feature data set into a pre-trained battery residual value score prediction model to determine a residual value evaluation score of the retired battery. Through the method and the device, the battery residual value evaluation score can be quickly and accurately determined so as to evaluate the health state of the retired battery.

Description

Technical Field

[0001] The present application relates to the field of battery technology, and in particular to a residual value assessment method, device, equipment and medium for retired batteries. Background Art

[0002] The booming electric vehicle industry has significantly driven rapid growth in power battery production, but it has also given rise to a new social issue: the effective management and reuse of retired power batteries. When a battery's capacity drops below 80% of its initial capacity, its performance can no longer meet the needs of vehicle operation and the battery enters retirement.

[0003] However, batteries exhibit significant inconsistencies in real-world applications due to volatile environments, severely hindering the direct reuse of retired batteries. Furthermore, automotive batteries operate under complex operating conditions, resulting in varying aging patterns, further exacerbating performance differences between batteries. Compared to new batteries, retired batteries are more sensitive to overcharge and overdischarge, posing potential safety risks such as thermal runaway and even explosion. Therefore, accurate residual value assessment of retired batteries is crucial to optimize the consistency of reassembled battery packs and ensure the safety and efficiency of reuse.

[0004] Currently, the most common methods for assessing battery residual value are capacity testing and internal resistance testing. Capacity testing is the most direct method, but it has limitations such as long test cycles and the inability to monitor in real time. Internal resistance testing is typically based on an equivalent circuit model, using parameter identification to estimate residual value. However, this model is complex to establish and solve, and suffers from strong parameter dependence and limited applicability. It also suffers from serious deficiencies in test conditions, dynamic response, and long-term performance prediction. Summary of the Invention

[0005] In view of this, the purpose of this application is to provide a method, device, equipment and medium for residual value assessment of retired batteries, which does not require long-term capacity testing or parameter identification through complex equivalent circuit models, greatly improving the efficiency of battery residual value assessment. In addition, electrochemical impedance spectroscopy testing can monitor the battery status in situ and in real time. The process does not involve deep discharge or charging of the battery and will not cause damage to the battery; it analyzes and explains the battery aging status from a mechanistic perspective, thereby improving the consistency of battery sorting. By combining terminal voltage detection technology with electrochemical impedance spectroscopy testing and using a battery residual value score prediction model, the battery residual value assessment score can be quickly and accurately determined to evaluate the health status of retired batteries.

[0006] In a first aspect, an embodiment of the present application provides a method for evaluating the residual value of retired batteries, the method comprising:

[0007] When it is determined that the appearance of the retired battery meets the inspection standard using the three-dimensional point cloud data of the retired battery, a terminal voltage test is performed on the retired battery to determine the terminal voltage of the retired battery, and an electrochemical impedance spectroscopy test is performed on the retired battery to determine multiple characteristic parameters of the retired battery;

[0008] performing dimensionality reduction processing on the terminal voltage and the plurality of characteristic parameters to obtain a target characteristic data set of the retired battery;

[0009] The target feature data set is input into a pre-trained battery residual value score prediction model to determine the residual value assessment score of the retired battery.

[0010] Furthermore, the following steps are performed to determine whether the appearance of the retired battery meets the inspection standards, including:

[0011] Preprocessing the three-dimensional point cloud data to obtain processed target point cloud data; wherein the preprocessing includes removing large-scale noise and smoothing small-scale noise;

[0012] Registering the standard point cloud data of the battery in the initial state with the target point cloud data, and comparing the registered standard point cloud data with the target point cloud data point by point to determine the offset of each point;

[0013] When the offset of each point is less than or equal to a preset threshold, it is determined that the appearance of the retired battery meets the inspection standard.

[0014] Furthermore, the electrochemical impedance spectroscopy test is performed on the retired battery to determine multiple characteristic parameters of the retired battery, including:

[0015] applying voltage excitation signals of different frequencies to the retired battery within a frequency measurement interval of the retired battery's impedance spectrum to obtain current response signals of the retired battery at the different frequencies, and calculating resistance values of the retired battery at the different frequencies based on the current response signals of the retired battery at the different frequencies to obtain impedance spectrum data of the retired battery;

[0016] The DRT algorithm is used to convert the frequency domain information of the impedance spectrum data into time domain characteristics to obtain a relaxation time distribution curve;

[0017] A plurality of characteristic peaks are determined from the relaxation time distribution curve, and characteristic information of each characteristic peak is used as a plurality of characteristic parameters of the retired battery; wherein the plurality of characteristic parameters include resistance, time constant and relaxation time distribution.

[0018] Furthermore, the performing dimensionality reduction processing on the terminal voltage and the plurality of characteristic parameters to obtain a target characteristic data set of the retired battery includes:

[0019] constructing an original feature data set using the terminal voltage and the plurality of feature parameters, and preprocessing the original feature data set to obtain a feature data set to be reduced in dimension;

[0020] Constructing a kernel matrix using a preset kernel function and the feature data set to be reduced in dimension, and performing centralization processing on the kernel matrix to obtain a centralized kernel matrix;

[0021] Performing eigendecomposition on the centralized kernel matrix to obtain multiple eigenvalues of the centralized kernel matrix and an eigenvector corresponding to each eigenvalue;

[0022] According to the size of each eigenvalue, a preset number of eigenvalues corresponding to the eigenvectors are selected as principal components to obtain a low-dimensional eigenvector matrix;

[0023] The original feature data set is projected into the low-dimensional feature vector matrix to obtain the target feature data set.

[0024] Furthermore, the battery residual value score prediction model is trained through the following steps:

[0025] Acquire sample data; wherein the sample data includes a sample feature data set and a battery residual value score sample value of a plurality of sample batteries;

[0026] The sample data is input into the original prediction model of the battery residual value score, and the original prediction model of the battery residual value score is trained using a least squares regression tree combined with a gradient boosting algorithm to obtain the battery residual value score prediction model.

[0027] Furthermore, after determining the residual value assessment score of the retired battery, the residual value assessment method further includes:

[0028] When the application field of the retired battery is the energy storage field, if it is determined that the residual value assessment score of the retired battery is greater than or equal to a first preset score, then the retired battery is considered to meet the sorting criteria;

[0029] When the application field of the retired battery is low-speed electric vehicles, if it is determined that the residual value assessment score of the retired battery is greater than or equal to a second preset score, the retired battery is considered to meet the sorting criteria; wherein, the second preset score is greater than or equal to the first preset score.

[0030] In a second aspect, an embodiment of the present application further provides a residual value assessment device for retired batteries, the residual value assessment device comprising:

[0031] a characteristic parameter acquisition module, configured to, when judging by using the three-dimensional point cloud data of the retired battery that the appearance of the retired battery meets the inspection standard, perform a terminal voltage test on the retired battery to determine the terminal voltage of the retired battery, and perform an electrochemical impedance spectroscopy test on the retired battery to determine a plurality of characteristic parameters of the retired battery;

[0032] a characteristic parameter dimensionality reduction module, configured to perform dimensionality reduction processing on the terminal voltage and a plurality of the characteristic parameters to obtain a target characteristic data set of the retired battery;

[0033] The residual value score determination module is used to input the target feature data set into a pre-trained battery residual value score prediction model to determine the residual value assessment score of the retired battery.

[0034] Furthermore, the characteristic parameter acquisition module is further configured to determine whether the appearance of the retired battery meets the inspection standard through the following steps:

[0035] Preprocessing the three-dimensional point cloud data to obtain processed target point cloud data; wherein the preprocessing includes removing large-scale noise and smoothing small-scale noise;

[0036] Registering the standard point cloud data of the battery in the initial state with the target point cloud data, and comparing the registered standard point cloud data with the target point cloud data point by point to determine the offset of each point;

[0037] When the offset of each point is less than or equal to a preset threshold, it is determined that the appearance of the retired battery meets the inspection standard.

[0038] In a third aspect, an embodiment of the present application further provides an electronic device comprising: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor and the memory communicate via the bus, and when the machine-readable instructions are executed by the processor, the steps of the residual value assessment method for retired batteries as described above are performed.

[0039] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, executes the steps of the above-mentioned method for residual value assessment of retired batteries.

[0040] Embodiments of the present application provide a method, apparatus, device, and medium for residual value assessment of retired batteries. First, when the three-dimensional point cloud data of the retired battery is used to determine that the appearance of the retired battery meets the inspection standard, the terminal voltage of the retired battery is detected, and the electrochemical impedance spectroscopy test is performed on the retired battery to determine multiple characteristic parameters of the retired battery; then, dimensionality reduction processing is performed on the terminal voltage and the multiple characteristic parameters to obtain a target feature data set of the retired battery; finally, the target feature data set is input into a pre-trained battery residual value score prediction model to determine the residual value assessment score of the retired battery.

[0041] This application obtains characteristic parameters by performing terminal voltage detection and electrochemical impedance spectroscopy testing on retired batteries, performs dimensionality reduction on the characteristic parameters, and inputs the reduced-dimensionality data set into a battery residual value score prediction model to determine the residual value assessment score of retired batteries. This application does not require long-term capacity testing, nor does it require parameter identification through complex equivalent circuit models, which greatly improves the efficiency of battery residual value assessment. In addition, electrochemical impedance spectroscopy testing can monitor the battery status in situ and in real time. The process does not involve deep discharge or charging of the battery and will not damage the battery. It analyzes and explains the battery aging status from a mechanistic perspective, thereby improving the consistency of battery sorting. By combining terminal voltage detection technology with electrochemical impedance spectroscopy testing and using a battery residual value score prediction model, the battery residual value assessment score can be quickly and accurately determined to evaluate the health status of retired batteries.

[0042] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without creative work.

[0044] Figure 1 A flowchart of a method for evaluating the residual value of retired batteries provided in an embodiment of the present application;

[0045] Figure 2 This is one of the structural schematic diagrams of a residual value assessment device for retired batteries provided in an embodiment of the present application;

[0046] Figure 3 This is a second structural diagram of a residual value assessment device for retired batteries provided in an embodiment of the present application;

[0047] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application for which protection is claimed, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, each other embodiment obtained by those skilled in the art without making creative work falls within the scope of protection of the present application.

[0049] First, the application scenarios to which this application is applicable are introduced. This application can be applied in the field of battery technology.

[0050] The booming electric vehicle industry has significantly driven rapid growth in power battery production, but it has also given rise to a new social issue: the effective management and reuse of retired power batteries. When a battery's capacity drops below 80% of its initial capacity, its performance can no longer meet the needs of vehicle operation and the battery enters retirement.

[0051] However, batteries exhibit significant inconsistencies in real-world applications due to volatile environments, severely hindering the direct reuse of retired batteries. Furthermore, automotive batteries operate under complex operating conditions, resulting in varying aging patterns, further exacerbating performance differences between batteries. Compared to new batteries, retired batteries are more sensitive to overcharge and overdischarge, posing potential safety risks such as thermal runaway and even explosion. Therefore, accurate residual value assessment of retired batteries is crucial to optimize the consistency of reassembled battery packs and ensure the safety and efficiency of reuse.

[0052] Research has found that the most common methods for assessing battery residual value are capacity testing and internal resistance testing. Capacity testing is the most direct method, but it has limitations such as long test cycles and the inability to monitor in real time. Internal resistance testing is typically based on an equivalent circuit model, using parameter identification to estimate residual value. However, this model is complex to establish and solve, and suffers from strong parameter dependence and limited applicability. It also suffers from serious deficiencies in test conditions, dynamic response, and long-term performance prediction.

[0053] Based on this, an embodiment of the present application provides a residual value assessment method for retired batteries. By combining terminal voltage detection technology with electrochemical impedance spectroscopy testing and utilizing a battery residual value score prediction model, the battery residual value assessment score can be quickly and accurately determined to evaluate the health status of retired batteries.

[0054] See also Figure 1 , Figure 1 This is a flow chart of a method for evaluating the residual value of retired batteries provided in an embodiment of the present application. Figure 1 As shown in , the residual value assessment method provided in the embodiment of the present application includes:

[0055] S101: When it is determined that the appearance of a retired battery meets a test standard using the three-dimensional point cloud data of the retired battery, a terminal voltage test is performed on the retired battery to determine the terminal voltage of the retired battery, and an electrochemical impedance spectroscopy test is performed on the retired battery to determine multiple characteristic parameters of the retired battery.

[0056] Regarding step S101 above, in a specific implementation, first, three-dimensional point cloud data of the retired battery is acquired. The three-dimensional point cloud data is then used to determine whether the retired battery's appearance meets inspection standards. Here, the inspection standard is that the retired battery's appearance is free of scratches, bulges, and other issues. The three-dimensional point cloud data is collected to obtain physical property information of the retired battery, enabling rapid capture of detailed three-dimensional topographical data of the retired battery to screen for retired batteries that are intact and free of damage, cracks, or bloating. For example, the three-dimensional point cloud data of the retired battery can be acquired using a structured light sensor or lidar, which is not specifically limited in this application. When the three-dimensional point cloud data of the retired battery determines that the retired battery's appearance meets the inspection standards, the retired battery is subjected to a terminal voltage measurement to determine the terminal voltage, and an electrochemical impedance spectroscopy test is performed on the retired battery to determine multiple characteristic parameters of the retired battery. By measuring the terminal voltage of the retired battery and performing the electrochemical impedance spectroscopy test on the retired battery, key features are extracted for subsequent evaluation. For example, when performing the terminal voltage test on the retired battery, the retired battery is discharged at room temperature to reduce the terminal voltage to a lower voltage limit. The retired batteries are then left to rest for a period of time to eliminate internal polarization. After this rest period, the terminal voltage U of the retired batteries is measured as a sorting feature for residual value assessment.

[0057] As an optional implementation, determining whether the appearance of the retired battery meets the inspection standard is performed through the following steps, including:

[0058] A: Preprocess the three-dimensional point cloud data to obtain processed target point cloud data.

[0059] The preprocessing includes removing large-scale noise and smoothing small-scale noise.

[0060] In the specific implementation of step A above, the 3D point cloud data of the retired batteries is processed to remove large-scale noise and smooth small-scale noise to obtain the processed target point cloud data. Here, the 3D point cloud data is subjected to noise filtering to improve the quality of the 3D point cloud data.

[0061] B: Align the standard point cloud data of the battery in the initial state with the target point cloud data, and compare the aligned standard point cloud data and target point cloud data point by point to determine the offset of each point.

[0062] Here, the initial state battery refers to an unused battery whose appearance meets the inspection standards. Standard point cloud data can also be obtained through structured light sensors or lidar.

[0063] Regarding step B above, during implementation, the standard point cloud data of the initial battery state and the target point cloud data are first registered. Here, when performing point cloud data registration, a coarse registration is first performed, such as using feature matching or global registration methods, followed by a fine registration, such as using ICP. The registered standard point cloud data and target point cloud data are then compared point by point to determine the offset of each point. Here, when performing the point-by-point comparison, the nearest point to each point is first found in the registered standard point cloud data, and the distance between the two points is then used as the offset.

[0064] C: When the offset of each point is less than or equal to the preset threshold, it is determined that the appearance of the retired battery meets the inspection standard.

[0065] Regarding step C above, when the offset of each point is less than or equal to the preset threshold, the retired battery's appearance is considered to meet the inspection standard. This ensures comprehensive and accurate inspection of physical properties through point cloud registration and comparison, providing a reliable foundation for subsequent analysis.

[0066] As an optional implementation, with respect to step S101 above, performing an electrochemical impedance spectroscopy test on the retired battery to determine multiple characteristic parameters of the retired battery includes:

[0067] Step 1011: Apply voltage excitation signals of different frequencies to the retired battery within a frequency measurement interval of the retired battery's impedance spectrum to obtain current response signals of the retired battery at the different frequencies, and calculate resistance values of the retired battery at the different frequencies based on the current response signals of the retired battery at the different frequencies to obtain impedance spectrum data of the retired battery.

[0068] Here, impedance spectrum data of retired batteries is first collected. Regarding step 1011 above, in a specific implementation, the impedance spectrum frequency measurement range of the retired batteries is first determined. Within this impedance spectrum frequency measurement range, voltage excitation signals of different frequencies are applied to the retired batteries to obtain current response signals of the retired batteries at different frequencies. Based on the current response signals of the retired batteries at different frequencies, the resistance values of the retired batteries at different frequencies are calculated to obtain the impedance spectrum data of the retired batteries.

[0069] Step 1012: Using a DRT algorithm, the frequency domain information of the impedance spectrum data is converted into time domain features to obtain a relaxation time distribution curve.

[0070] The DRT algorithm, or Distribution of Relaxation Times (DRT), is an analytical method for extracting time-domain information from electrochemical impedance spectroscopy (EIS) data. It converts impedance frequency-domain data into a time-domain distribution, revealing the timescale characteristics of electrochemical processes within batteries.

[0071] In the specific implementation of step 1012, the DRT algorithm is used to convert the frequency domain information of the impedance spectrum data into time domain characteristics to obtain a relaxation time distribution curve. This is used to explore the dynamic response characteristics of the retired battery at different frequencies. The relaxation time distribution curve shows the intensity of the electrochemical process corresponding to different relaxation times.

[0072] Step 1013 , determining multiple characteristic peaks from the relaxation time distribution curve, and using characteristic information of each characteristic peak as multiple characteristic parameters of the retired battery.

[0073] For the above step 1013, during the specific implementation, each peak in the curve represents a specific electrochemical process (such as ion diffusion, charge transfer, etc.), and its position and intensity reflect the battery performance and aging characteristics. Multiple characteristic peaks are determined from the relaxation time distribution curve, which represent different electrochemical processes and their corresponding relaxation times. The characteristic information of each characteristic peak is used as multiple characteristic parameters of the retired battery to characterize the battery performance and aging characteristics. Specifically, the multiple characteristic parameters include resistance R, time constant τ and relaxation time distribution γ(τ). Here, resistance is the resistance value corresponding to each peak in the curve, reflecting the difficulty of the electrochemical process. The time constant is the time constant corresponding to each peak in the curve, indicating the time scale of the electrochemical process. The relaxation time distribution is the width and shape of each peak in the curve, reflecting the complexity of the electrochemical process. Here, as an example, when there are four characteristic peaks in the relaxation time distribution curve, the multiple characteristic parameters extracted include τ1, τ2, τ3, τ4, γ(τ1), γ(τ2), γ(τ3), γ(τ4), R1, R2, R3, and R4.

[0074] Thus, according to the above steps 1011 to 1013, DRT technology can provide more intuitive time domain information. The DRT curve directly shows the time scale characteristics of different electrochemical processes, which facilitates understanding of complex battery aging mechanisms. Moreover, by extracting information from multiple peaks, different aging mechanisms can be distinguished.

[0075] S102: Perform dimensionality reduction processing on the terminal voltage and the plurality of characteristic parameters to obtain a target characteristic data set of the retired battery.

[0076] In the specific implementation of step S102, dimensionality reduction is performed on the terminal voltage and multiple characteristic parameters obtained in step S101 to obtain a target feature dataset for retired batteries. Specifically, the PCA algorithm is used to reduce the dimensionality of the characteristic parameters, remove redundant features, extract key feature data, and ultimately obtain the feature dataset. This dimensionality reduction process reduces the computational complexity of subsequent models, significantly reducing computation time and memory consumption.

[0077] As an optional implementation, with respect to step S102, performing dimensionality reduction processing on the terminal voltage and the plurality of characteristic parameters to obtain a target characteristic data set of the retired battery includes:

[0078] Step 1021 : constructing an original feature data set using the terminal voltage and the plurality of feature parameters, and preprocessing the original feature data set to obtain a feature data set to be reduced in dimensionality.

[0079] In the specific implementation of step 1021, an original feature dataset is constructed using the terminal voltage and multiple feature parameters. Continuing with the above example, the constructed original feature dataset is represented as X = (U, τ1, τ2, τ3, τ4, γ(τ1), γ(τ2), γ(τ3), γ(τ4), R1, R2, R3, R4). The original feature dataset is then preprocessed and normalized to ensure comparability between different features, thereby obtaining the feature dataset to be reduced in dimension.

[0080] Step 1022: construct a kernel matrix using a preset kernel function and the feature data set to be reduced in dimension, and perform centralization processing on the kernel matrix to obtain a centralized kernel matrix.

[0081] Here, the role of the kernel function is to map the original data to a high-dimensional feature space so that the nonlinear relationship can be linearized. In the embodiment provided in this application, a Gaussian radial basis function (RBF) is used as the preset kernel function, and the preset kernel function is expressed as:

[0082]

[0083] Among them, σ is the kernel width parameter, which controls the smoothness of the kernel function.

[0084] Regarding step 1022, in a specific implementation, a kernel matrix is first constructed using a preset kernel function and the feature dataset to be reduced in dimension, and the kernel matrix is then centered to obtain a centralized kernel matrix. Specifically, the preset kernel function is K(x, y), and the preset kernel function K(x, y) is used to capture nonlinear relationships in the data. Then, the kernel function values between all sample pairs are calculated using the selected kernel function to construct the kernel matrix K. The kernel matrix K is expressed by the following formula:

[0085] K ij =K(x i ,x j )

[0086] Finally, the kernel matrix K is centered to remove the influence of the mean in the data. The kernel matrix K is centered using the following formula:

[0087]

[0088] in, is the centralized kernel matrix, I is the identity matrix, 1 is the all-one column vector, and n is the number of samples.

[0089] Step 1023 : Perform eigendecomposition on the centralized kernel matrix to obtain multiple eigenvalues of the centralized kernel matrix and an eigenvector corresponding to each eigenvalue.

[0090] Regarding the above step 1023, in the specific implementation, after obtaining the centralized kernel matrix, the centralized kernel matrix is subjected to eigendecomposition to obtain multiple eigenvalues of the centralized kernel matrix and the eigenvectors corresponding to each eigenvalue. Specifically, the centralized kernel matrix Perform eigendecomposition: turn up Multiple eigenvalues λ1≥λ2≥…≥λ n The eigenvector v1,v2,…,v corresponding to each eigenvalue n .

[0091] Step 1024 : Select eigenvectors corresponding to a preset number of eigenvalues as principal components according to the size of each eigenvalue, so as to obtain a low-dimensional eigenvector matrix.

[0092] Regarding step 1024, in the specific implementation, the eigenvectors corresponding to the preset number of eigenvalues are selected as principal components according to the size of each eigenvalue to obtain a low-dimensional eigenvector matrix. Here, the eigenvectors corresponding to the first d eigenvalues are selected as principal components according to the size of the eigenvalue, where d is the dimension of the target after dimensionality reduction. These eigenvectors are denoted as V = [v1, v2, ..., v d], where V is an n×d matrix.

[0093] Step 1025 : Project the original feature data set into the low-dimensional feature vector matrix to obtain the target feature data set.

[0094] Regarding step 1025, in a specific implementation, the original feature dataset is projected into a low-dimensional feature vector matrix to obtain a target feature dataset Z = X V after dimensionality reduction. Here, continuing with the above example, the input original feature dataset is represented as X = (U, τ1, τ2, τ3, τ4, γ(τ1), γ(τ2), γ(τ3), γ(τ4), R1, R2, R3, R4) 13 dimensions. After dimensionality reduction using the PCA algorithm, the resulting target feature dataset is represented as Z = (U, τ2, γ(τ2), γ(τ3), R2, R3).

[0095] Thus, according to steps 1021 through 1025, the PCA algorithm can capture nonlinear relationships and map data into a high-dimensional space using a kernel function, enabling better processing of complex nonlinear data distributions. It also reduces redundant features and extracts the most important principal components through feature decomposition, effectively reducing the data dimensionality. This dimensionality reduction allows the subsequent battery residual value score prediction model to run more quickly, enabling the determination of the residual value assessment score for retired batteries to be more quickly determined.

[0096] S103: Input the target feature data set into a pre-trained battery residual value score prediction model to determine the residual value assessment score of the retired battery.

[0097] Regarding the above step S103, during specific implementation, the target feature data set after dimensionality reduction is input into a pre-trained battery residual value score prediction model to determine the residual value assessment score of the retired battery.

[0098] As an optional implementation, the battery residual value score prediction model is trained through the following steps:

[0099] I: Get sample data.

[0100] II: Input the sample data into the original prediction model of the battery residual value score, and train the original prediction model of the battery residual value score using the least squares regression tree combined with the gradient boosting algorithm to obtain the battery residual value score prediction model.

[0101] During the implementation of Steps I and II above, sample data is obtained. Specifically, the sample data includes a sample feature dataset and sample values of the battery residual value scores for multiple sample batteries. Here, the sample feature dataset also undergoes dimensionality reduction according to the above steps. The sample data is then input into the original battery residual value score prediction model. This original battery residual value score prediction model is trained using a least squares regression tree combined with a gradient boosting algorithm to obtain a battery residual value score prediction model.

[0102] Specifically, when the least squares regression tree is combined with the gradient boosting algorithm to train the original prediction model of the battery residual value score, the model is first initialized and the initial prediction value is set to a constant Usually taken as the training set target value y i The mean or median of , that is:

[0103]

[0104] Where n is the number of training samples.

[0105] Next, multiple rounds of iterations are performed. For m = 1, 2, ..., M (M is the number of iterations), the negative gradient (pseudo residual) of the original prediction model of the battery residual value score is first calculated, and the negative gradient of the least squares loss function is used as the pseudo residual. The calculation formula is:

[0106]

[0107] in, is the least squares loss function.

[0108] Then fit the pseudo residuals using the regression tree f m (x) fitting pseudo residual r im , find the optimal tree structure by optimizing the following objective function:

[0109]

[0110] Where F represents the set of all possible regression trees.

[0111] Finally, update the model and add the newly trained tree to the model. However, in order to prevent overfitting, the newly trained tree f m (x) After weighting with the learning rate λ, update the prediction value of the ensemble model:

[0112]

[0113] Repeat the above steps until M rounds of iterations have passed or the residual is small enough to obtain the final integrated model, whose prediction function is:

[0114]

[0115] in, Estimate the score for the residual value.

[0116] According to the residual value assessment method provided in this application, as an optional embodiment, after determining the residual value assessment score of the retired battery, the retired batteries can be sorted according to the residual value assessment result. After determining the residual value assessment score of the retired battery, the residual value assessment method further includes:

[0117] (1) When the application field of the retired battery is the energy storage field, if it is determined that the residual value assessment score of the retired battery is greater than or equal to a first preset score, it is considered that the retired battery meets the sorting criteria.

[0118] (2) When the application field of the retired battery is low-speed electric vehicles, if it is determined that the residual value assessment score of the retired battery is greater than or equal to a second preset score, the retired battery is considered to meet the sorting criteria.

[0119] Here, the second preset score is greater than or equal to the first preset score.

[0120] Regarding the above steps (1) to (2), in specific implementation, when the application field of the retired battery is the energy storage field, if the residual value assessment score of the retired battery is determined to be greater than or equal to the first preset score, the retired battery is considered to meet the sorting criteria and can be reused. When the application field of the retired battery is low-speed electric vehicles, if the residual value assessment score of the retired battery is determined to be greater than or equal to the second preset score, the retired battery is considered to meet the sorting criteria and can be reused.

[0121] The residual value assessment method for retired batteries provided in an embodiment of the present application includes: first, when the appearance of the retired battery is judged to meet the inspection standard using the three-dimensional point cloud data of the retired battery, the terminal voltage of the retired battery is detected to determine the terminal voltage of the retired battery, and the electrochemical impedance spectroscopy test is performed on the retired battery to determine multiple characteristic parameters of the retired battery; then, dimensionality reduction processing is performed on the terminal voltage and the multiple characteristic parameters to obtain a target feature data set of the retired battery; finally, the target feature data set is input into a pre-trained battery residual value score prediction model to determine the residual value assessment score of the retired battery.

[0122] This application obtains characteristic parameters by performing terminal voltage detection and electrochemical impedance spectroscopy testing on retired batteries, performs dimensionality reduction on the characteristic parameters, and inputs the reduced-dimensionality data set into a battery residual value score prediction model to determine the residual value assessment score of retired batteries. This application does not require long-term capacity testing, nor does it require parameter identification through complex equivalent circuit models, which greatly improves the efficiency of battery residual value assessment. In addition, electrochemical impedance spectroscopy testing can monitor the battery status in situ and in real time. The process does not involve deep discharge or charging of the battery and will not damage the battery. It analyzes and explains the battery aging status from a mechanistic perspective, thereby improving the consistency of battery sorting. By combining terminal voltage detection technology with electrochemical impedance spectroscopy testing and using a battery residual value score prediction model, the battery residual value assessment score can be quickly and accurately determined to evaluate the health status of retired batteries.

[0123] See also Figure 2 、 Figure 3 , Figure 2 This is one of the structural diagrams of a residual value assessment device for retired batteries provided in an embodiment of the present application. Figure 3 This is a second structural diagram of a residual value assessment device for retired batteries provided in an embodiment of the present application. Figure 2 As shown in , the residual value evaluation device 200 includes:

[0124] a characteristic parameter acquisition module 201 configured to, when determining that the appearance of a retired battery meets a test standard using the three-dimensional point cloud data of the retired battery, perform a terminal voltage test on the retired battery to determine the terminal voltage of the retired battery, and perform an electrochemical impedance spectroscopy test on the retired battery to determine a plurality of characteristic parameters of the retired battery;

[0125] a characteristic parameter dimension reduction module 202, configured to perform dimension reduction processing on the terminal voltage and a plurality of the characteristic parameters to obtain a target characteristic data set of the retired battery;

[0126] The residual value score determination module 203 is configured to input the target feature data set into a pre-trained battery residual value score prediction model to determine the residual value assessment score of the retired battery.

[0127] Furthermore, the characteristic parameter acquisition module 201 is further configured to determine whether the appearance of the retired battery meets the inspection standard through the following steps:

[0128] Preprocessing the three-dimensional point cloud data to obtain processed target point cloud data; wherein the preprocessing includes removing large-scale noise and smoothing small-scale noise;

[0129] Registering the standard point cloud data of the battery in the initial state with the target point cloud data, and comparing the registered standard point cloud data with the target point cloud data point by point to determine the offset of each point;

[0130] When the offset of each point is less than or equal to a preset threshold, it is determined that the appearance of the retired battery meets the inspection standard.

[0131] Furthermore, when the characteristic parameter acquisition module 201 is used to perform an electrochemical impedance spectroscopy test on the retired battery to determine a plurality of characteristic parameters of the retired battery, the characteristic parameter acquisition module 201 is further used to:

[0132] applying voltage excitation signals of different frequencies to the retired battery within a frequency measurement interval of the retired battery's impedance spectrum to obtain current response signals of the retired battery at the different frequencies, and calculating resistance values of the retired battery at the different frequencies based on the current response signals of the retired battery at the different frequencies to obtain impedance spectrum data of the retired battery;

[0133] The DRT algorithm is used to convert the frequency domain information of the impedance spectrum data into time domain characteristics to obtain a relaxation time distribution curve;

[0134] A plurality of characteristic peaks are determined from the relaxation time distribution curve, and characteristic information of each characteristic peak is used as a plurality of characteristic parameters of the retired battery; wherein the plurality of characteristic parameters include resistance, time constant and relaxation time distribution.

[0135] Furthermore, when the characteristic parameter dimension reduction module 202 is used to perform dimension reduction processing on the terminal voltage and the plurality of characteristic parameters to obtain the target characteristic data set of the retired battery, the characteristic parameter dimension reduction module 202 is further used to:

[0136] constructing an original feature data set using the terminal voltage and the plurality of feature parameters, and preprocessing the original feature data set to obtain a feature data set to be reduced in dimension;

[0137] Constructing a kernel matrix using a preset kernel function and the feature data set to be reduced in dimension, and performing centralization processing on the kernel matrix to obtain a centralized kernel matrix;

[0138] Performing eigendecomposition on the centralized kernel matrix to obtain multiple eigenvalues of the centralized kernel matrix and an eigenvector corresponding to each eigenvalue;

[0139] According to the size of each eigenvalue, a preset number of eigenvalues corresponding to the eigenvectors are selected as principal components to obtain a low-dimensional eigenvector matrix;

[0140] The original feature data set is projected into the low-dimensional feature vector matrix to obtain the target feature data set.

[0141] For further information, see Figure 3 The residual value evaluation device 200 further includes a model training module 204, which is configured to train the battery residual value score prediction model through the following steps:

[0142] Acquire sample data; wherein the sample data includes a sample feature data set and a battery residual value score sample value of a plurality of sample batteries;

[0143] The sample data is input into the original prediction model of the battery residual value score, and the original prediction model of the battery residual value score is trained using a least squares regression tree combined with a gradient boosting algorithm to obtain the battery residual value score prediction model.

[0144] For further information, see Figure 3 The residual value evaluation device 200 further includes a battery sorting module 205. After determining the residual value evaluation score of the retired battery, the battery sorting module 205 is configured to:

[0145] When the application field of the retired battery is the energy storage field, if it is determined that the residual value assessment score of the retired battery is greater than or equal to a first preset score, then the retired battery is considered to meet the sorting criteria;

[0146] When the application field of the retired battery is low-speed electric vehicles, if it is determined that the residual value assessment score of the retired battery is greater than or equal to a second preset score, the retired battery is considered to meet the sorting criteria; wherein, the second preset score is greater than or equal to the first preset score.

[0147] See also Figure 4 , Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 4 As shown in FIG, the electronic device 400 includes a processor 410 , a memory 420 and a bus 430 .

[0148] The memory 420 stores machine-readable instructions executable by the processor 410. When the electronic device 400 is running, the processor 410 communicates with the memory 420 via the bus 430. When the machine-readable instructions are executed by the processor 410, the above-mentioned Figure 1 The steps of the residual value assessment method for retired batteries in the method embodiment shown are specifically implemented in accordance with the method embodiment and will not be described in detail here.

[0149] The embodiment of the present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the computer program can execute the above-mentioned Figure 1 The steps of the residual value assessment method for retired batteries in the method embodiment shown are specifically implemented in accordance with the method embodiment and will not be described in detail here.

[0150] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0151] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. There may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed may be through some communication interface, indirect coupling or communication connection of devices or units, which may be electrical, mechanical or other forms.

[0152] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0153] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0154] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0155] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The scope of protection of the present application is not limited thereto. Although the present application has been described in detail with reference to the above-mentioned embodiments, those skilled in the art should understand that any person skilled in the art can modify or easily conceive of changes to the technical solutions described in the above-mentioned embodiments within the technical scope disclosed in the present application, or perform equivalent replacements for some of the technical features thereof. These modifications, changes, or replacements do not deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A method for evaluating the residual value of retired batteries, characterized in that: The residual value assessment method includes: When it is determined that the appearance of the retired battery meets the inspection standard using the three-dimensional point cloud data of the retired battery, a terminal voltage test is performed on the retired battery to determine the terminal voltage of the retired battery, and an electrochemical impedance spectroscopy test is performed on the retired battery to determine multiple characteristic parameters of the retired battery; performing dimensionality reduction processing on the terminal voltage and the plurality of characteristic parameters to obtain a target characteristic data set of the retired battery; The target feature data set is input into a pre-trained battery residual value score prediction model to determine the residual value assessment score of the retired battery.

2. The residual value assessment method according to claim 1, characterized in that: The following steps are used to determine whether the appearance of the retired battery meets the inspection standards, including: Preprocessing the three-dimensional point cloud data to obtain processed target point cloud data; wherein the preprocessing includes removing large-scale noise and smoothing small-scale noise; Registering the standard point cloud data of the battery in the initial state with the target point cloud data, and comparing the registered standard point cloud data with the target point cloud data point by point to determine the offset of each point; When the offset of each point is less than or equal to a preset threshold, it is determined that the appearance of the retired battery meets the inspection standard.

3. The residual value assessment method according to claim 1, characterized in that: The electrochemical impedance spectroscopy test is performed on the retired battery to determine multiple characteristic parameters of the retired battery, including: applying voltage excitation signals of different frequencies to the retired battery within a frequency measurement interval of the retired battery's impedance spectrum to obtain current response signals of the retired battery at the different frequencies, and calculating resistance values of the retired battery at the different frequencies based on the current response signals of the retired battery at the different frequencies to obtain impedance spectrum data of the retired battery; The DRT algorithm is used to convert the frequency domain information of the impedance spectrum data into time domain characteristics to obtain a relaxation time distribution curve; A plurality of characteristic peaks are determined from the relaxation time distribution curve, and characteristic information of each characteristic peak is used as a plurality of characteristic parameters of the retired battery; wherein the plurality of characteristic parameters include resistance, time constant and relaxation time distribution.

4. The residual value assessment method according to claim 1, characterized in that: The performing dimensionality reduction processing on the terminal voltage and the plurality of characteristic parameters to obtain a target characteristic data set of the retired battery includes: constructing an original feature data set using the terminal voltage and the plurality of feature parameters, and preprocessing the original feature data set to obtain a feature data set to be reduced in dimension; Constructing a kernel matrix using a preset kernel function and the feature data set to be reduced in dimension, and performing centralization processing on the kernel matrix to obtain a centralized kernel matrix; Performing eigendecomposition on the centralized kernel matrix to obtain multiple eigenvalues of the centralized kernel matrix and an eigenvector corresponding to each eigenvalue; According to the size of each eigenvalue, a preset number of eigenvalues corresponding to the eigenvectors are selected as principal components to obtain a low-dimensional eigenvector matrix; The original feature data set is projected into the low-dimensional feature vector matrix to obtain the target feature data set.

5. The residual value assessment method according to claim 1, characterized in that: The battery residual value score prediction model is trained by the following steps: Acquire sample data; wherein the sample data includes a sample feature data set and a battery residual value score sample value of a plurality of sample batteries; The sample data is input into the original prediction model of the battery residual value score, and the original prediction model of the battery residual value score is trained using a least squares regression tree combined with a gradient boosting algorithm to obtain the battery residual value score prediction model.

6. The residual value assessment method according to claim 1, characterized in that: After determining the residual value assessment score of the retired battery, the residual value assessment method further includes: When the application field of the retired battery is the energy storage field, if it is determined that the residual value assessment score of the retired battery is greater than or equal to a first preset score, then the retired battery is considered to meet the sorting criteria; When the application field of the retired battery is low-speed electric vehicles, if it is determined that the residual value assessment score of the retired battery is greater than or equal to a second preset score, the retired battery is considered to meet the sorting criteria; wherein, the second preset score is greater than or equal to the first preset score.

7. A residual value assessment device for retired batteries, characterized in that: The residual value assessment device comprises: a characteristic parameter acquisition module, configured to, when judging by using the three-dimensional point cloud data of the retired battery that the appearance of the retired battery meets the inspection standard, perform a terminal voltage test on the retired battery to determine the terminal voltage of the retired battery, and perform an electrochemical impedance spectroscopy test on the retired battery to determine a plurality of characteristic parameters of the retired battery; a characteristic parameter dimensionality reduction module, configured to perform dimensionality reduction processing on the terminal voltage and a plurality of the characteristic parameters to obtain a target characteristic data set of the retired battery; The residual value score determination module is used to input the target feature data set into a pre-trained battery residual value score prediction model to determine the residual value assessment score of the retired battery.

8. The residual value evaluation device according to claim 7, characterized in that: The characteristic parameter acquisition module is further configured to determine whether the appearance of the retired battery meets the inspection standard through the following steps: Preprocessing the three-dimensional point cloud data to obtain processed target point cloud data; wherein the preprocessing includes removing large-scale noise and smoothing small-scale noise; Registering the standard point cloud data of the battery in the initial state with the target point cloud data, and comparing the registered standard point cloud data with the target point cloud data point by point to determine the offset of each point; When the offset of each point is less than or equal to a preset threshold, it is determined that the appearance of the retired battery meets the inspection standard.

9. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor and the memory communicate via the bus. When the processor runs the machine-readable instructions, the steps of the method for residual value assessment of retired batteries as described in any one of claims 1 to 6 are executed.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method for residual value assessment of retired batteries according to any one of claims 1 to 6 are executed.

Images