A battery detection method and application based on magnetic field relaxation multi-exponential fitting

Abstract

This invention relates to a battery testing method and application based on multi-exponential fitting of magnetic field relaxation, belonging to the field of battery testing technology. By continuously monitoring the external magnetic field distribution during the resting phase after charging and discharging, magnetic field time-series data is obtained, and the magnetic field change is calculated by referring to the magnetic field distribution under stable resting conditions. The global magnetic field change is extracted using spatial summation, and then multi-exponential models are applied to fit the global and local magnetic field changes separately. The fitting errors are compared to determine the optimal model, and time constants and amplitude parameters are extracted to characterize the kinetic behavior and spatial distribution of different electrochemical processes. Spatial mapping based on local parameters can identify local anomalies, non-uniform reactions, or potential failure regions, enabling battery performance evaluation, consistency screening, and early fault diagnosis. Compared to traditional voltage relaxation methods, this invention provides higher spatial resolution and reaction sensitivity, and has broad industrial application prospects.

Description

Technical Field

[0001] This invention belongs to the field of battery testing technology, specifically relating to a battery testing method and application based on magnetic field relaxation multi-exponential fitting. Background Technology

[0002] Lithium-ion batteries exhibit concentration gradients, potential differences, and polarization during charging and discharging. When the current is interrupted, the system tends to spontaneously restore electrochemical equilibrium; this process is called "relaxation." During relaxation, ions redistribute within the particles or between components, and the reactions on the electrode surface gradually slow down, manifesting as slow changes in voltage, current, or external fields. This process reflects core performance parameters such as the battery's internal transport behavior, polarization recovery characteristics, and structural stability, and has become an important tool for studying battery kinetics, lifetime evolution, and anomalous behavior.

[0003] Current mainstream relaxation detection methods are mainly based on changes in relaxation voltage. By fitting the voltage recovery curve of the battery after charging / discharging with multiple exponential functions, the decay process at different time scales is extracted to infer internal mechanisms such as ion concentration relaxation, interfacial polarization, or side reactions. The relaxation voltage method has advantages such as ease of operation, stable measurement, and solid research foundation, and is widely used in electrochemical behavior analysis and aging characterization. However, this method has two key limitations: First, voltage, as a global response signal, reflects the average behavior of the entire cell and cannot provide spatially resolved information, making it difficult to identify local anomalies or non-uniform reaction regions; second, in some unsteady-state or side reaction-dominated stages, the voltage response may mask local activity differences or delayed reaction signals, limiting sensitivity and interpretability. Summary of the Invention

[0004] To overcome the shortcomings of the prior art, this invention provides a battery detection method and application based on multi-exponential fitting of magnetic field relaxation. By mapping the magnetic field signal to the natural decay characteristics of the current density during battery relaxation, and extracting reaction time scale and response amplitude parameters from the global and local multi-exponential fitting of magnetic field changes, this method identifies the kinetic characteristics and spatial distribution differences of multiple reaction processes within the battery. This method exhibits higher sensitivity and adaptability in scenarios such as battery consistency screening, aging identification, and local fault diagnosis.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A battery detection method based on magnetic field relaxation multi-exponential fitting includes the following steps:

[0007] Step 1: After the battery has undergone charging or discharging, continuously monitor the magnetic field distribution during its relaxation process to obtain time series data of the magnetic field distribution. Using the magnetic field distribution when the battery reaches a stable static state as a reference, the change in magnetic field distribution at each moment is calculated. ;

[0008] Step 2: Analyze the changes in magnetic field distribution. Spatial summation yields the global magnetic field variation representing the overall relaxation response of the battery. ;

[0009] Step 3, for Multiple exponential decay models were fitted separately, and the fitting errors were compared to determine the optimal fitting model. The overall parameter amplitudes were then extracted. and time constant , used to characterize the global electrochemical reaction behavior of the battery;

[0010] Step 4: For each spatial location of The time-series data were fitted with a multi-exponential decay model, and the fitting errors were compared to determine the optimal fitting model for each spatial location. The local parameter amplitudes were then extracted. and time constant Based on the spatial distribution of local parameters, a battery internal reaction state map is constructed to identify local abnormal reactions, non-uniformity, degradation, or potential fault areas.

[0011] The multi-exponential decay model takes the following form:

[0012]

[0013] Where n is the number of exponent terms, such as single exponent (n=1), double exponent (n=2), triple exponent (n=3), etc. It can also be extended to four exponents or higher-order terms according to the actual reaction complexity, so as to more accurately characterize the electrochemical relaxation process of multiple time scales inside the battery.

[0014] Furthermore, in step one, the magnetic field distribution test plane is the long and wide surface of the battery, and the test coordinates are... These are points with a fixed spacing on the plane;

[0015] Furthermore, in step two, the global magnetic field changes. The calculation method is to perform calculations on all spatial locations. Summation

[0016] .

[0017] Furthermore, in step four, the parameter fitting method for the multi-exponential model can employ least squares, nonlinear regression, the Levenberg-Marquardt algorithm, or a gradient descent-based numerical optimization method. These methods should possess good convergence and parameter estimation capabilities to achieve a robust fit between the exponential coefficients and the time constant.

[0018] Furthermore, in step four, the criterion for the optimal fitting model is that the fitting residual is minimized or the information criterion is satisfied.

[0019] Furthermore, in step four, the time constants in the global and local parameters... Characterizing the kinetic timescale and amplitude of each electrochemical process Characterize the intensity of the contribution of each electrochemical process to the overall magnetic field change;

[0020] The battery testing methods described above are applicable to battery performance evaluation, consistency screening, and detection of localized degradation / faults.

[0021] Compared with the prior art, the beneficial effects of the present invention are:

[0022] 1. This invention breaks through the limitation of the traditional relaxation voltage method lacking spatial resolution. By physically correlating the magnetic field change signal with the current density, it realizes non-destructive monitoring of different electrochemical processes at different sites of the battery, which can reveal local slow decay processes or abnormal reaction areas, and improves the sensitivity and interpretability of the detection.

[0023] 2. The multi-exponential fitting modeling framework proposed in this invention can be applied to both the overall curve and the magnetic field response data of local points. By extracting the temporal characteristics of various dynamic behaviors through time constants and amplitude parameters, it provides a new quantitative, high-throughput, and spatially visualized means for battery performance evaluation, health status identification, and consistency screening. Attached Figure Description

[0024] Figure 1 This is a flowchart of the detection method described in this invention;

[0025] Figure 2 Global magnetic field change for overall battery relaxation response Schematic diagram;

[0026] Figure 3 for Schematic diagram of fitting the single-exponential decay model, double-exponential decay model and triple-exponential decay model;

[0027] Figure 4 for A schematic diagram illustrating the fitting errors of the single-exponential decay model, the double-exponential decay model, and the triple-exponential decay model;

[0028] Figure 5 for The optimal fitting model and fitting error diagram for each spatial location;

[0029] Figure 6 for Local parameters and A schematic diagram of the spatial distribution. Detailed Implementation

[0030] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings and embodiments. Obviously, the described embodiments are only some embodiments of the invention, not all embodiments. 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.

[0031] Example 1:

[0032] This embodiment provides a battery detection method and application based on magnetic field relaxation multi-exponential fitting, such as... Figure 1 As shown, the method includes the following specific steps:

[0033] Step 1: After the battery has undergone charging or discharging, continuously monitor the magnetic field distribution during its relaxation process to obtain time series data of the magnetic field distribution. Using the magnetic field distribution when the battery reaches a stable static state as a reference, the change in magnetic field distribution at each moment is calculated. ;

[0034] In this embodiment, the battery under test is a 5Ah wound soft-pack lithium-ion battery. According to the test method in step one, after the sample battery undergoes a 0.5C constant current discharge, the planar magnetic field distribution at a distance of 5mm from the battery's length and width surfaces is measured. Continuous testing was conducted, with the battery magnetic field test area measuring 65mm × 70mm. The area was scanned every 5mm interval. Based on the stable state 6 hours after the end of discharge, the change in magnetic field distribution at each time point was calculated. ;

[0035] Step 2: Analyze the changes in magnetic field distribution. Spatial summation yields the global magnetic field variation representing the overall relaxation response of the battery. ;

[0036] In this embodiment, the total relaxation test duration is 6 hours, with each test interval being 7 minutes. The change in magnetic field distribution at each time point is measured. By performing spatial summation, the global magnetic field change during the battery relaxation process can be obtained. ,like Figure 2 As shown.

[0037] Step 3, for Multiple exponential decay models were fitted separately, and the fitting errors were compared to determine the optimal fitting model. The overall parameter amplitudes were then extracted. and time constant This is used to characterize the global electrochemical reaction behavior of the battery; the multi-exponential decay model takes the following form:

[0038] , where n is the number of exponent terms.

[0039] In this embodiment, for The single exponential decay model (n=1), double exponential decay model (n=2), and triple exponential decay model (n=3) were fitted respectively, and the fitting results are as follows: Figure 3 As shown, the fitting curve of the single exponential decay model differs somewhat from the experimental data. The fitting curve of the double exponential decay model fits the experimental data well. The fitting curve of the triple exponential decay model is similar to that of the double exponential decay model, but the fitting parameters of the triple exponential decay model are redundant. By comparing the root mean square (RMSE) fitting errors of the single, double, and triple exponential models, the double exponential decay model shows a better improvement than the single exponential decay model, but the triple exponential decay model does not show an improvement and exhibits overfitting due to redundant model parameters. Figure 4 As shown, the double exponential fitting model is therefore the optimal fitting model.

[0040] Step 4: For each spatial location of The time-series data were fitted with a multi-exponential decay model, and the fitting errors were compared to determine the optimal fitting model for each spatial location. The local parameter amplitudes were then extracted. and time constant ;

[0041] In this embodiment, for each spatial location of Time-series data were fitted using a multi-exponential decay model. By comparing the root mean square fitting error results, the optimal fitting model selected for each spatial location of the battery was determined as follows: Figure 5 As shown in (a), the root mean square error of the fit at each spatial location is as follows: Figure 5 As shown in (b), by selecting the corresponding fitting model, the attenuation trend of the relaxation magnetic field at different spatial locations can be well fitted; such as Figure 6 As shown in (a) and (d), these are the extracted local parameters of the single-exponential fitting model. and ,like Figure 6 (b) and (c) are the magnitude parameters in the double exponential fitting model. , Spatial distribution, Figure 6(e) and (f) are the time constants in the double exponential fitting model. and Spatial distribution.

[0042] A battery internal reaction state map is constructed based on the spatial distribution of local parameters to identify local abnormal reactions, non-uniformity, degradation or potential fault areas.

[0043] In this embodiment, according to Figure 6 The spatial distribution of local parameters of the optimal model shows that, due to the non-uniformity of the current density distribution, the high current density region near the tab needs more exponential fitting, indicating that there are two electrochemical processes in this region during relaxation. The low current density region near the edge of the battery only needs a single exponential model to fit, indicating that only one electrochemical process is captured in this region during relaxation.

[0044] This invention detects changes in the magnetic field distribution of a battery during relaxation and extracts global and local time constants and amplitude parameters by fitting a multi-exponential decay model. This enables non-destructive identification of reaction kinetics at different locations within the battery, thereby constructing a spatially resolved reaction state map. This method not only overcomes the limitation of traditional relaxation voltage methods in providing spatial information but also possesses the ability to identify locally degraded or abnormal regions. It is applicable to various scenarios such as battery factory quality inspection, battery pack consistency assessment, aging monitoring, and potential fault early warning. It offers significant advantages such as high throughput, non-destructive testing, and spatial visualization, and has broad engineering application value and promising prospects for widespread adoption.

[0045] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A battery detection method based on magnetic field relaxation multi-exponential fitting, characterized in that, Includes the following steps: Step 1: After the battery has undergone charging or discharging, continuously monitor the magnetic field distribution during its relaxation process to obtain time series data of the magnetic field distribution. Using the magnetic field distribution when the battery reaches a stable static state as a reference, the change in magnetic field distribution at each moment is calculated. ; Step 2: Analyze the changes in magnetic field distribution. Spatial summation yields the global magnetic field variation representing the overall relaxation response of the battery. ; Step 3, for Multiple exponential decay models were fitted separately, and the fitting errors were compared to determine the optimal fitting model. The overall parameter amplitudes were then extracted. and time constant , used to characterize the global electrochemical reaction behavior of the battery; Step 4: For each spatial location of The time-series data were fitted with multiple exponential decay models, and the fitting errors were compared to determine the optimal fitting model for each spatial location. The local parameter amplitudes were then extracted. and time constant Based on the spatial distribution of local parameters, a battery internal reaction state map is constructed to identify local abnormal reactions, non-uniformity, degradation, or potential fault areas. The multi-exponential decay model takes the following form: Where n is the number of exponent terms.

2. The method according to claim 1, characterized in that: The magnetic field distribution test plane is the long and wide surface of the battery, and the test coordinates are... These are the points with a fixed spacing on the plane.

3. The method according to claim 1, characterized in that: Global magnetic field changes The calculation method is to perform calculations on all spatial locations. Summation 。 4. The method according to claim 1, characterized in that: The parameter fitting methods for multi-exponential models include least squares, nonlinear regression, Levenberg-Marquardt algorithm, or numerical optimization methods based on gradient descent.

5. The method according to claim 1, characterized in that: The criterion for the optimal fitting model is either the minimum fitting residual or the optimal information criterion.

6. The method according to claim 1, characterized in that: Time constants in global and local parameters Characterizing the kinetic timescale and amplitude of each electrochemical process Characterize the contribution intensity of each electrochemical process to the overall magnetic field change.

7. An application of the detection method according to any one of claims 1-6, characterized in that: The detection method is applied to battery performance evaluation, consistency screening, and local degradation / fault detection.

Images