A lithium battery self-discharge rate testing method and device
Patent Information
- Application Number
- CN202610798216.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-06-04
- Publication Date
- 2026-10-23
AI Technical Summary
但现有技术仍存在缺陷：经过加速后仍然需要数天时间，测试周期仍然较长，无法满足产线快速筛选的需求；现有快速测试方法（如直流内阻法、交流阻抗法）往往只能定性判断异常电芯，难以准确量化自放电率，无法兼顾速度与精度；微安级的自放电电流容易受温度波动、极化效应等环境因素干扰，现有技术在加速测试时对干扰的抑制不足
[0021]本发明首先在标准温度下将待测锂电池的荷电状态调整至预设值，向待测锂电池施加脉冲充放电以消除极化，并采集待测锂电池的初始动态内阻和初始开路电压，然后将待测锂电池在比标准温度更低的预设低温下静置预设时长，静置过程中采集待测锂电池的实时自放电电流、实时动态内阻、实时荷电状态、以及待测锂电池所处环境的实时温度，静置结束后，在预设低温下，向待测锂电池施加脉冲充放电的反向脉冲充放电以消除残余极化，采集待测锂电池的终止开路电压和终止动态内阻，最后根据所采集的数据，利用修正的阿伦尼乌斯模型对实时自放电电流进行修正，利用卡尔曼滤波器对修正后的实时自放电电流、实时动态内阻、开路电压漂移量进行降噪，基于降噪后的数据、实时温度以及实时荷电状态（SOC）作为BP神经网络的输入，得到动态补偿系数，并基于动态补偿系数计算待测锂电池的自放电率。与现有技术相比，能够大幅缩短测试时间，有效提升测试效率，满足锂电池产线100%在线测试的需求；利用多维数据，结合阿伦尼乌斯修正、卡尔曼滤波和BP神经网络构建动态补偿体系，抑制温度、电磁、内阻漂移等干扰，有效提高测试精度；可适用于多种体系电池、适配不同老化程度的电芯，适用范围更广、适配性更强；整个过程无需人工干预，避免了人为误差，适合规模化生产测试场景。
Smart Images
Figure CN122962415A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of lithium battery performance testing technology, and in particular to a method and apparatus for testing the self-discharge rate of lithium batteries. Background Technology
[0002] The self-discharge rate of lithium batteries is one of the key indicators for measuring battery quality. It refers to the rate at which the battery spontaneously loses its charge when it is left in an open-circuit, static state. An excessively high self-discharge rate will lead to poor consistency in the battery pack, affecting the overall lifespan and safety of the battery pack.
[0003] Traditional self-discharge rate testing methods typically employ the "static voltage drop measurement method": after fully charging the battery, it is left to stand in an open circuit at room temperature (around 25°C) for 7-28 days, and the voltage difference or capacity loss before and after the standing period is measured to calculate the self-discharge rate. This method is time-consuming and severely restricts the production efficiency and R&D iteration speed of lithium batteries.
[0004] Currently, there have been some attempts in the industry to accelerate testing, such as shortening the current stabilization time by adjusting the power control strategy or using series capacitors for rapid testing. However, existing technologies still have shortcomings: even after acceleration, it still takes several days, and the testing cycle is still long, which cannot meet the needs of rapid screening on production lines; existing rapid testing methods (such as DC internal resistance method and AC impedance method) can often only qualitatively identify abnormal cells, and it is difficult to accurately quantify the self-discharge rate, thus failing to balance speed and accuracy; the microampere-level self-discharge current is easily affected by environmental factors such as temperature fluctuations and polarization effects, and existing technologies are insufficient in suppressing interference during accelerated testing. Summary of the Invention
[0005] The main objective of this invention is to propose a method and apparatus for testing the self-discharge rate of lithium batteries, which can effectively balance testing speed and accuracy.
[0006] This invention is achieved through the following technical solution:
[0007] A method for testing the self-discharge rate of lithium batteries includes the following steps:
[0008] Step S1: Adjust the state of charge of the lithium battery under test to a preset value at a standard temperature, apply pulse charge and discharge to the lithium battery under test to eliminate polarization, and collect the initial dynamic internal resistance and initial open circuit voltage of the lithium battery under test.
[0009] Step S2: Place the lithium battery under test at a preset low temperature for a preset time. During the placement process, collect the real-time self-discharge current, real-time dynamic internal resistance, real-time terminal voltage, real-time state of charge, and real-time temperature of the environment where the lithium battery is located. The preset low temperature is lower than the standard temperature. The preset time is 2h to 8h.
[0010] Step S3: After the resting period, at a preset low temperature, apply a reverse pulse charge / discharge to the lithium battery under test to eliminate residual polarization, and collect the termination open circuit voltage and termination dynamic internal resistance of the lithium battery under test.
[0011] Step S4: Based on the data collected in steps S1 to S3, the real-time self-discharge current is corrected using the modified Arrhenius model. The real-time self-discharge current, real-time dynamic internal resistance, and open-circuit voltage drift are denoised using a Kalman filter. Based on the denoised data, real-time temperature, and real-time state of charge as inputs to the BP neural network, dynamic compensation coefficients are obtained, and the self-discharge rate of the lithium battery under test is calculated based on the dynamic compensation coefficients.
[0012] Furthermore, the standard temperature is 25±2℃, and the preset low temperature is -10℃ to 0℃.
[0013] Furthermore, in step S1, the charge / discharge rate of the pulse and the reverse pulse is 0.5C~2C, and the charging or discharging time is 5s~30s.
[0014] Furthermore, in step S4, according to the formula Calculate the corrected real-time self-discharge current ,in, For real-time self-discharge current, Let r be the known activation energy of the self-discharge reaction, and r be the gas constant. Standard temperature This is the real-time temperature.
[0015] Furthermore, in step S4, the open-circuit voltage drift before noise reduction by the Kalman filter is: ,in, This is the initial open-circuit voltage. To terminate the open circuit voltage.
[0016] Furthermore, in step S4, standard battery samples with different aging degrees, different states of charge, and different temperatures under different systems are selected. Through long-term static testing, the accurate self-discharge rate of each standard battery sample is obtained. This accurate self-discharge rate is then compared with the average temperature of the environment in which each standard battery sample is located during the static testing process. Average state of charge of each standard battery sample and the average dynamic internal resistance of each standard battery sample A training set is constructed to train the BP neural network.
[0017] Furthermore, in step S4, according to the formula Calculate the self-discharge rate of the lithium battery under test, where, To calculate the self-discharge rate at standard temperature, For preset duration, This represents the open-circuit voltage drift after noise reduction using a Kalman filter. This is a polarization voltage correction value obtained based on the real-time terminal voltage and the termination open-circuit voltage. These are the dynamic compensation coefficients output by the BP neural network.
[0018] This invention is also achieved through the following technical solutions:
[0019] An apparatus for implementing the lithium battery self-discharge rate testing method as described above includes a temperature-controlled chamber for housing the lithium battery under test, a charge / discharge module for applying pulse charge / discharge and measuring open-circuit voltage, a micro-current detection module for measuring self-discharge current, an internal resistance detection module for measuring dynamic internal resistance, and a control module. The control module corrects the real-time self-discharge current using a modified Arrhenius model based on the collected data, reduces noise in the corrected real-time self-discharge current, real-time dynamic internal resistance, and open-circuit voltage drift using a Kalman filter, obtains dynamic compensation coefficients based on the noise-reduced data, real-time temperature, and real-time state of charge as inputs to a BP neural network, and calculates the self-discharge rate of the lithium battery under test based on the dynamic compensation coefficients.
[0020] As can be seen from the above description of the present invention, compared with the prior art, the present invention has the following beneficial effects:
[0021] This invention first adjusts the state of charge (SOC) of the lithium battery under test to a preset value at a standard temperature. Then, pulsed charge-discharge is applied to the battery to eliminate polarization, and the initial dynamic internal resistance and initial open-circuit voltage of the battery are collected. Next, the battery is placed at a preset low temperature (lower than the standard temperature) for a preset time. During this time, the real-time self-discharge current, real-time dynamic internal resistance, real-time SOC, and the real-time temperature of the environment surrounding the battery are collected. After the placement period, reverse pulsed charge-discharge is applied to the battery at the preset low temperature to eliminate residual polarization, and the termination open-circuit voltage and termination dynamic internal resistance are collected. Finally, based on the collected data, a modified Arrhenius model is used to correct the real-time self-discharge current. A Kalman filter is used to reduce noise in the corrected real-time self-discharge current, real-time dynamic internal resistance, and open-circuit voltage drift. Based on the denoised data, real-time temperature, and real-time SOC as input to a BP neural network, dynamic compensation coefficients are obtained, and the self-discharge rate of the battery is calculated based on these coefficients. Compared with existing technologies, it can significantly shorten testing time, effectively improve testing efficiency, and meet the needs of 100% online testing in lithium battery production lines. By utilizing multi-dimensional data and combining Arrhenius correction, Kalman filtering, and BP neural network to construct a dynamic compensation system, it can suppress interference from temperature, electromagnetic fields, and internal resistance drift, effectively improving testing accuracy. It is applicable to various battery systems and can be adapted to cells with different aging levels, making it more widely applicable and adaptable. The entire process requires no manual intervention, avoiding human error and making it suitable for large-scale production testing scenarios. Attached Figure Description
[0022] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.
[0023] Figure 1 This is a flowchart of the present invention. Detailed Implementation
[0024] The present invention will be further described below through specific embodiments.
[0025] like Figure 1 As shown, the method for testing the self-discharge rate of lithium batteries includes the following steps:
[0026] Step S1: Adjust the state of charge of the lithium battery under test to a preset value at a standard temperature, apply pulse charge and discharge to the lithium battery under test to eliminate polarization, and collect the initial dynamic internal resistance and initial open circuit voltage of the lithium battery under test.
[0027] The lithium battery under test is placed in a temperature-controlled chamber, with the chamber temperature set to a standard temperature of 25±2℃. The charge / discharge module charges and discharges the lithium battery under test to adjust its state of charge (SOC) to a preset value (e.g., 50% SOC or 100% SOC). Subsequently, the charge / discharge module applies a pulse charge / discharge cycle to the lithium battery under test to eliminate polarization. Specifically, the pulse charge / discharge process involves charging at a constant current of 1C for 10 seconds, resting for 5 seconds, and then discharging at a constant current of 1C for 10 seconds. After the charge / discharge cycle is complete, the internal resistance detection module measures the battery's dynamic internal resistance R0, and the charge / discharge module measures and records the open-circuit voltage OCV0. The OCV0, R0, preset SOC value, and ambient temperature T0 are then uploaded to the control module.
[0028] Step S2: Place the lithium battery under test at a preset low temperature for a preset time. During the placement process, collect the real-time self-discharge current, real-time dynamic internal resistance, real-time state of charge, and real-time temperature of the environment where the lithium battery is located. The preset low temperature is lower than the standard temperature. The preset time is 2h to 8h.
[0029] Specifically, the temperature control chamber is controlled to reduce the temperature from the standard temperature to the preset temperature at a set cooling rate (e.g., 1℃ / min), allowing the lithium battery under test to stand at the preset temperature for a preset time of 4 hours. During the standing period, the real-time self-discharge current I(t) of the lithium battery under test is continuously monitored by a micro-current detection module, the real-time dynamic internal resistance R(t) of the lithium battery under test is detected by an internal resistance detection module, the real-time temperature T(t) is collected by the temperature control chamber, and the real-time terminal voltage V(t) change of the lithium battery under test is recorded in real time by a charge / discharge module. The real-time terminal voltage V(t) is uploaded to the control module at a frequency of 100ms / time, and the remaining real-time data is uploaded to the control module at a frequency of 1s / time. The preset low temperature can be -10℃ to 0℃.
[0030] Step S3: After the settling period, at a preset low temperature, apply a reverse pulse charge / discharge to the lithium battery under test to eliminate residual polarization, and collect the termination open circuit voltage and termination dynamic internal resistance of the lithium battery under test.
[0031] The reverse pulse charge and discharge in this step refers to using a pulse with the opposite pulse direction and the same amplitude as the pulse in step S1 to charge and discharge, in order to eliminate the residual polarization that may be caused by charge accumulation at low temperature. Then, the termination open circuit voltage OCV1 at this time is measured and recorded by the charge and discharge module, and the termination dynamic internal resistance R1 at this time is measured by the internal resistance check module. OCV1 and R1 are then uploaded to the control module.
[0032] Step S4: Based on the data collected in steps S1 to S3, the real-time self-discharge current is corrected using the modified Arrhenius model. The real-time self-discharge current, real-time dynamic internal resistance, and open-circuit voltage drift are denoised using a Kalman filter. Based on the denoised data, real-time temperature, and real-time state of charge as inputs to the BP neural network, dynamic compensation coefficients are obtained, and the self-discharge rate of the lithium battery under test is calculated based on the dynamic compensation coefficients.
[0033] Specifically, based on the modified Arrhenius equation, the real-time self-discharge current I(t) during the low-temperature static process is initially corrected to eliminate the fundamental influence of temperature on the self-discharge reaction rate, thus obtaining the initially corrected self-discharge current. ,in, Let r be the known activation energy of the self-discharge reaction, and r be the gas constant. Standard temperature The temperature is real-time. The activation energy of the self-discharge reaction is obtained through a pre-calibrated standard battery sample. The acquisition process is based on existing technology, and different values correspond to different battery systems. The energy is pre-stored in a data storage module connected to the control module.
[0034] Since the self-discharge current is in the microampere range, it is easily affected by factors such as temperature fluctuations, electromagnetic interference, and internal resistance drift. Therefore, a Kalman filter is introduced to reduce noise in the pre-corrected self-discharge current, real-time dynamic internal resistance, and open-circuit drift, in order to eliminate random interference and ensure the accuracy of the data input to the BP neural network. The open-circuit voltage drift before noise reduction is... ,in, This is the initial open-circuit voltage. To terminate the open circuit voltage.
[0035] The core computational steps of Kalman filtering include:
[0036] Prediction: Based on the filtering results of the previous moment, predict the self-discharge current, dynamic resistance and open-circuit voltage drift at the current moment, and establish prediction equations;
[0037] Update: Compare the measured data at the current moment with the predicted value, calculate the error covariance, and correct the predicted value through adaptive weight adjustment to obtain stable data after noise reduction;
[0038] Iteration: A prediction-update iteration is completed every 1 second, following the data changes during the resting process and dynamically suppressing interference. This process is existing technology.
[0039] A backpropagation (BP) neural network consists of an input layer, hidden layers, and an output layer. The denoised data, real-time temperature, and real-time state of charge are used as input layer parameters, and the output layer outputs the final compensation coefficients. This approach aims to overcome the limitations of traditional single-temperature-SOC compensation and improve accuracy. Specifically, the real-time state of charge (SOC) is calculated in real time using the known mapping relationship between OCV and SOC.
[0040] For training the BP neural network, standard battery samples with different systems (lithium iron phosphate, ternary, lithium cobalt oxide) and varying degrees of aging (new battery, 500 cycles, 1000 cycles), different states of charge (30%, 50%, 80%, 100%), and different temperatures (-10℃, -5℃, 0℃, 5℃, 15℃, 25℃) were selected. Through long-term (7-28 days) static testing, the accurate self-discharge rate of each standard battery sample was obtained. This accurate self-discharge rate was then compared with the average temperature of the environment in which each standard battery sample was placed during the static testing. Average state of charge of each standard battery sample and the average dynamic internal resistance of each standard battery sample A training set is constructed to train the BP neural network. After training until the error is less than 0.5%, the network parameters are stored in the data storage module.
[0041] During the test, the control module compares the calculated self-discharge rate with the preset standard value. If the error exceeds 1%, the weight parameters of the neural network are automatically adjusted to achieve adaptive optimization of the algorithm and improve the adaptability to different types of battery cells.
[0042] Finally, the self-discharge rate of the tested lithium battery at standard temperature was... ,in, To calculate the self-discharge rate at standard temperature, For preset duration, This represents the open-circuit voltage drift after noise reduction using a Kalman filter. These are the dynamic compensation coefficients output by the BP neural network. This is a polarization voltage correction value obtained based on the real-time terminal voltage and the termination open-circuit voltage. Specifically, it is based on the collected real-time terminal voltage V(t) and termination open-circuit voltage of the lithium battery under test. The polarization rebound curve was fitted using an exponential model. ,but Where A is the initial polarization voltage amplitude, representing the maximum voltage offset caused by the polarization effect immediately after the reverse pulse ends. This is the polarization relaxation time constant, which reflects the rate at which the polarization voltage decays.
[0043] The initial calibration of the compensation function is combined with algorithm iteration. The pre-calibrated temperature-self-discharge rate acceleration coefficient table serves as the initial input reference for the BP neural network, rather than the sole basis. Dynamic optimization is achieved through algorithm iteration, which solves the problems of poor adaptability and insufficient accuracy of traditional calibration methods.
[0044] In this invention, the terms "first," "second," and "third," etc., are used only to distinguish similar objects and are not necessarily used to describe a specific order or sequence, nor should they be construed as indicating or implying relative importance. The use of terms such as "upper," "lower," "left," "right," "front," and "rear" to indicate orientation or positional relationships is based on the orientation or positional relationships shown in the accompanying drawings and is only for the convenience of describing the invention, not to indicate or imply that the device referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation on the scope of protection of this invention. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.
[0045] Furthermore, in the description of this application, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the preceding and following related objects have an "or" relationship.
[0046] The above are merely specific embodiments of the present invention, but the design concept of the present invention is not limited thereto. Any non-substantial modifications made to the present invention using this concept shall be considered as infringing upon the protection scope of the present invention.
Claims
1. A method for testing the self-discharge rate of a lithium battery, characterized in that: Includes the following steps: Step S1: Adjust the state of charge of the lithium battery under test to a preset value at a standard temperature, apply pulse charge and discharge to the lithium battery under test to eliminate polarization, and collect the initial dynamic internal resistance and initial open circuit voltage of the lithium battery under test. Step S2: Place the lithium battery under test at a preset low temperature for a preset time. During the placement process, collect the real-time self-discharge current, real-time dynamic internal resistance, real-time terminal voltage, real-time state of charge, and real-time temperature of the environment where the lithium battery is located. The preset low temperature is lower than the standard temperature. The preset time is 2h to 8h. Step S3: After the resting period, at a preset low temperature, apply a reverse pulse charge / discharge to the lithium battery under test to eliminate residual polarization, and collect the termination open circuit voltage and termination dynamic internal resistance of the lithium battery under test. Step S4: Based on the data collected in steps S1 to S3, the real-time self-discharge current is corrected using the modified Arrhenius model. The real-time self-discharge current, real-time dynamic internal resistance, and open-circuit voltage drift are denoised using a Kalman filter. Based on the denoised data, real-time temperature, and real-time state of charge as inputs to the BP neural network, dynamic compensation coefficients are obtained, and the self-discharge rate of the lithium battery under test is calculated based on the dynamic compensation coefficients.
2. The method for testing the self-discharge rate of a lithium battery according to claim 1, characterized in that: The standard temperature is 25±2℃, and the preset low temperature is -10℃ to 0℃.
3. The method for testing the self-discharge rate of a lithium battery according to claim 1, characterized in that: In step S1, the charge / discharge rate of the pulse and the reverse pulse is 0.5C to 2C, and the charging or discharging time is 5s to 30s.
4. A method for testing the self-discharge rate of a lithium battery according to claim 1, 2, or 3, characterized in that: In step S4, according to the formula Calculate the corrected real-time self-discharge current ,in, For real-time self-discharge current, Let r be the known activation energy of the self-discharge reaction, and r be the gas constant. Standard temperature This is the real-time temperature.
5. The method for testing the self-discharge rate of a lithium battery according to claim 4, characterized in that: In step S4, the open-circuit voltage drift before noise reduction by the Kalman filter is: ,in, This is the initial open-circuit voltage. To terminate the open circuit voltage.
6. The method for testing the self-discharge rate of a lithium battery according to claim 5, characterized in that: In step S4, standard battery samples with different aging degrees, different states of charge, and different temperatures under different systems are selected. Through long-term static testing, the accurate self-discharge rate of each standard battery sample is obtained. This accurate self-discharge rate is then compared with the average temperature of the environment in which each standard battery sample was placed during the static testing process. Average state of charge of each standard battery sample and the average dynamic internal resistance of each standard battery sample A training set is constructed to train the BP neural network.
7. The method for testing the self-discharge rate of a lithium battery according to claim 6, characterized in that: In step S4, according to the formula Calculate the self-discharge rate of the lithium battery under test, where, To calculate the self-discharge rate at standard temperature, For preset duration, This represents the open-circuit voltage drift after noise reduction using a Kalman filter. This is a polarization voltage correction value obtained based on the real-time terminal voltage and the termination open-circuit voltage. These are the dynamic compensation coefficients output by the BP neural network.
8. An apparatus for implementing the lithium battery self-discharge rate testing method as described in any one of claims 1 to 7, characterized in that: The system includes a temperature control chamber for housing the lithium battery under test, a charge / discharge module for applying pulse charge / discharge and measuring open-circuit voltage, a micro-current detection module for measuring self-discharge current, an internal resistance detection module for measuring dynamic internal resistance, and a control module. The control module corrects the real-time self-discharge current using a modified Arrhenius model based on the collected data, and uses a Kalman filter to reduce noise in the corrected real-time self-discharge current, real-time dynamic internal resistance, and open-circuit voltage drift. Based on the noise-reduced data, real-time temperature, and real-time state of charge as inputs to a BP neural network, dynamic compensation coefficients are obtained, and the self-discharge rate of the lithium battery under test is calculated based on the dynamic compensation coefficients.