Lithium battery aging monitoring method based on bayesian optimization of equal voltage difference charging time
By optimizing the equal voltage difference charging time of lithium batteries using a Bayesian optimization algorithm and combining it with a generalized linear regression model, the accuracy and real-time issues of existing lithium battery aging monitoring are solved, achieving efficient and accurate aging status assessment.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- 武汉华海通用电气有限公司
- Filing Date
- 2026-01-12
- Publication Date
- 2026-06-09
Smart Images
Figure CN122172039A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of battery monitoring technology, and more specifically, to a method for monitoring lithium battery aging based on Bayesian optimization of equal voltage difference charging time. Background Technology
[0002] Lithium-ion batteries undergo irreversible chemical and physical changes during charge-discharge cycles, including loss of active lithium, damage to electrode material structure, growth of the solid electrolyte membrane, lithium dendrite deposition, and electrolyte decomposition and consumption. These changes manifest electrically as reduced capacity and increased internal resistance, leading to performance degradation. Simultaneously, lithium dendrites may puncture the separator, causing short circuits, and gases produced by electrolyte decomposition (such as carbon monoxide and methane) can cause battery bulging, posing safety risks. Therefore, real-time monitoring of the aging status of lithium-ion batteries is crucial. This can alert maintenance personnel to replace batteries or implement health management strategies to slow down aging, thereby improving user experience and mitigating potential risks.
[0003] Currently, lithium battery aging monitoring methods mainly rely on capacity and internal resistance as health parameters. Capacity calculation typically uses the coulomb method, which requires accumulating the current throughout the entire charge-discharge cycle. However, accumulated errors in current detection can lead to inaccurate capacity calculations. Internal resistance detection employs DC, AC, or electrochemical impedance spectroscopy methods. To avoid the influence of polarization resistance, the battery needs to be allowed to rest sufficiently before testing, which cannot meet the real-time requirements of online monitoring in practical applications. These methods are either susceptible to hardware errors or cumbersome to operate, limiting their applicability in real-world scenarios.
[0004] For the charging process of ternary lithium batteries (typically with a terminal voltage range of 2.7V to 4.2V), the equal voltage difference charging time is related to the aging state and is an ideal aging health parameter because battery aging shortens the charging time. However, existing methods using equal voltage difference charging time often do not optimize the voltage boundaries in the calculation, resulting in the inability to establish a significant relationship with capacity decay, which in turn leads to biases in the assessment of aging state.
[0005] Therefore, it is necessary to study a better scheme for accurately monitoring the aging status of lithium batteries. Summary of the Invention
[0006] This invention addresses the technical problems existing in the prior art by providing a lithium battery aging monitoring method based on Bayesian optimization of equal voltage difference charging time. By optimizing voltage selection, the correlation between equal voltage difference charging time and capacity is enhanced, thus solving the problems of insufficient accuracy and applicability of existing methods.
[0007] According to a first aspect of the present invention, a method for monitoring the aging of lithium batteries based on Bayesian optimization of equal voltage difference charging time is provided, comprising: S1, Obtain the aging capacity sequence obtained from the nth charge-discharge cycle of the lithium battery. and equal voltage difference charging time Among them, the equal voltage difference charging time It is a function related to the low charging voltage VL and the high charging voltage VH; S2, aging capacity sequence and equal voltage difference charging time Using the Pearson correlation coefficient between the two voltage levels as the optimization objective, a Bayesian optimization process was performed to obtain the optimized equal-voltage differential charging time. ; S3, based on the aging capacity sequence And optimized equal voltage difference charging time Fit a generalized linear regression model based on the aging capacity threshold. The generalized linear regression model is used to calculate the aging threshold of the equal voltage difference charging time. ; S4, based on the equal voltage difference charging time aging threshold To assess the aging status of the battery under test.
[0008] Based on the above technical solution, the present invention can also be improved as follows.
[0009] Optionally, the equal voltage difference charging time , represented as:
[0010] Where m is the index of the sampling point during the nth charging process. This represents the charging voltage sampled at the m-th time during the n-th charging process. This represents the time value corresponding to when the charging voltage reaches the high charging voltage VH. This represents the time value corresponding to when the charging voltage reaches the low charging voltage VL.
[0011] Optionally, step S2 includes: S201, Define the Bayesian optimization objective function as equal voltage difference charging time. and aging capacity sequence The Pearson correlation coefficient between them; Low charging voltage Value range and charging high voltage The initial random sample points are obtained from the range of values to obtain the sample set. The probability distribution of the Pearson correlation coefficient is obtained by modeling the sample set using a Gaussian process. S202, define the Bayesian optimization acquisition function as the expected improvement function:
[0012] in, and These are the low charging voltage and high charging voltage values that maximize the Pearson correlation coefficient within the current sample set; The new charging low voltage is obtained by maximizing the expected improvement function. and charging high voltage Samples are used to update the sample set and the Gaussian process model; S203, when the preset iteration stop condition is met, outputs the Bayesian-optimized low-voltage charging value. and charging high voltage value And thus obtain the first Optimized equal voltage difference charging time during the second charging process , represented as:
[0013] in, To achieve the optimized high charging voltage value The corresponding time value, To achieve the optimized low charging voltage value when the charging voltage reaches the target value The corresponding time value.
[0014] Optionally, in step S2, the aging capacity sequence and equal voltage difference charging time The Pearson correlation coefficient between them is expressed as:
[0015] in, The average isovoltage charging time for N charge-discharge cycles under given VL and VH. This represents the average aging capacity over N charge-discharge cycles at given VL and VH, where n is the index of the number of charge cycles. .
[0016] Optionally, in step S3, the step based on the aging capacity sequence... And optimized equal voltage difference charging time Fitting a generalized linear regression model includes: The generalized linear regression model for aging capacity to equal voltage difference charging time is defined as follows:
[0017] in, Here are the predicted values for equal voltage difference charging time, where A is the coefficient of the linear term, B is the coefficient of the nonlinear term, and C is the constant term. According to the aging capacity sequence And optimized equal voltage difference charging time The values of A, B, and C are estimated using the least squares method as follows: , and , represented as:
[0018] in, It is an N×1 aging capacity matrix. It is an N-fold optimized equal voltage difference charging time matrix of size N×1.
[0019] Optionally, in step S3, the step of determining the aging capacity threshold... The generalized linear regression model is used to calculate the aging threshold of the equal voltage difference charging time. ,include: The aging capacity threshold Substituting into the generalized linear regression model, the output is the aging threshold for equal voltage difference charging time. .
[0020] Optionally, step S4 includes: During the k-th charging process of the lithium battery under test, the optimized equal voltage difference charging time obtained by performing a Bayesian optimization process is acquired. and the aging threshold of the equal voltage difference charging time. In comparison: like ＜ If the test result is positive, the lithium battery under test is determined to be in an aging and failure state; otherwise, the lithium battery under test is determined to be in a healthy state.
[0021] According to a second aspect of the present invention, a lithium battery aging monitoring system based on Bayesian optimization of equal voltage difference charging time is provided, comprising: The data acquisition module is configured to acquire the aging capacity sequence obtained from the nth charge-discharge cycle of the lithium battery. and equal voltage difference charging time Among them, the equal voltage difference charging time It is a function related to the low charging voltage VL and the high charging voltage VH; The optimization calculation module is configured to optimize the aging capacity sequence. and equal voltage difference charging time Using the Pearson correlation coefficient between the two voltage levels as the optimization objective, a Bayesian optimization process was performed to obtain the optimized equal-voltage differential charging time. ; The threshold calculation module is configured to be based on the aging capacity sequence. And optimized equal voltage difference charging time Fit a generalized linear regression model and base it on the aging capacity threshold. The generalized linear regression model is used to calculate the aging threshold of the equal voltage difference charging time. ; The condition assessment module is configured to perform aging based on the equal voltage difference charging time threshold. To assess the aging status of the battery under test.
[0022] According to a third aspect of the present invention, an electronic device is provided, including a memory and a processor, wherein the processor is configured to execute a computer management program stored in the memory to implement the steps of the above-described lithium battery aging monitoring method based on Bayesian optimization of equal voltage difference charging time.
[0023] According to a fourth aspect of the present invention, a computer-readable storage medium is provided, on which a computer management program is stored, wherein when the computer management program is executed by a processor, the steps of the above-described lithium battery aging monitoring method based on Bayesian optimization of equal voltage difference charging time are implemented.
[0024] This invention provides a lithium battery aging monitoring method, system, electronic device, and storage medium based on Bayesian optimization of equal voltage difference charging time. Firstly, it uses a Bayesian optimization algorithm to optimize the equal voltage difference charging time... With aging capacity sequence Using the Pearson correlation coefficient between them as the objective function, the system automatically searches for and determines a pair of optimal charging voltage boundaries (VL and VH), thereby... Enhanced to a health characteristic highly sensitive to battery capacity degradation (e.g., differential charging time) Furthermore, using the optimized features and aging capacity data, a generalized linear regression model is fitted to determine the known capacity failure threshold ( This is converted into a directly comparable equal voltage difference charging time aging threshold. Ultimately, during the monitoring phase, only the optimal charging time of the battery under test needs to be determined. Is it below the threshold? This invention enables rapid and accurate assessment of aging conditions. It transforms the complex problem of aging monitoring into a simple threshold comparison, overcoming the dependence of traditional methods on complete charge-discharge curves or battery resting conditions. It achieves efficient and reliable online monitoring using only fragmented charging data, making it particularly suitable for battery management systems with high real-time requirements. Attached Figure Description
[0025] Figure 1 A flowchart of a lithium battery aging monitoring method based on Bayesian optimization of equal voltage difference charging time provided by the present invention; Figure 2 is a schematic diagram of the Bayesian optimization results provided in a certain embodiment, wherein 2(a) is a schematic diagram of the objective function model and 2(b) is a schematic diagram of the optimization convergence history; Figure 3 A schematic diagram of the fitting results of a generalized linear regression model provided for a certain embodiment; Figure 4 is a schematic diagram of the aging state detection results provided in a certain embodiment, wherein 4(a) is a schematic diagram of aging capacity and 4(b) is a schematic diagram of equal voltage difference charging time. Figure 5 A block diagram of a lithium battery aging monitoring system based on Bayesian optimization of equal voltage difference charging time is provided for this invention. Figure 6 A schematic diagram of the hardware structure of a possible electronic device provided by the present invention; Figure 7 This is a schematic diagram of the hardware structure of a possible computer-readable storage medium provided by the present invention. Detailed Implementation
[0026] The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.
[0027] Figure 1 A flowchart of a lithium battery aging monitoring method based on Bayesian optimization of equal voltage difference charging time is provided for an embodiment of the present invention, as shown below. Figure 1 As shown, the method includes steps S1 to S4: S1 defines the range of values for the low charging voltage VL of the lithium battery. and the range of values for the high charging voltage VH Obtain the aging capacity sequence obtained from the nth charge-discharge cycle in the lithium battery aging experiment. and equal voltage difference charging time ,in, Equal voltage difference charging time It is a function related to the low charging voltage VL and the high charging voltage VH.
[0028] This step involves obtaining the aging capacity sequence during lithium battery charge-discharge cycles. and equal voltage difference charging time ,in Defined as the time difference between the battery voltage increasing from a preset low threshold (low charging voltage VL) to a high threshold (high charging voltage VH). This step utilizes the physical characteristic that the increase in internal resistance during battery aging leads to a change in charging rate, using the charging time difference as a quantifiable feature characterizing the aging state. This step transforms the complex electrochemical aging phenomenon into an easily monitorable time parameter, providing fundamental data support for subsequent optimization and avoiding the dependence on complete charge-discharge cycles in traditional capacity testing.
[0029] S2, aging capacity sequence and equal voltage difference charging time Using the Pearson correlation coefficient between the two voltage levels as the optimization objective, a Bayesian optimization process was performed to obtain the optimized equal-voltage differential charging time. .
[0030] Understandably, the Pearson correlation coefficient is used to quantify the strength and direction of the linear relationship between two variables. For example, this step uses equal voltage difference charging time... With aging capacity sequence Using the Pearson correlation coefficient as the objective function, the optimal voltage boundaries for VL and VH are dynamically searched through Bayesian optimization. This step guides sampling through Gaussian process modeling and the expectation boosting function (EI), maximizing the statistical correlation between time and capacity in the selection of voltage boundaries. By automatically selecting the voltage range most sensitive to battery aging, the original charging time is... Enhanced as an optimized feature This improves the distinguishability of aging characteristics and the reliability of monitoring.
[0031] S3, based on the aging capacity sequence And optimized equal voltage difference charging time Fit a generalized linear regression model based on the aging capacity threshold. The generalized linear regression model is used to calculate the aging threshold of the equal voltage difference charging time. .
[0032] This step utilizes a statistical model to establish a nonlinear mapping relationship between capacity and time, transforming abstract capacity decay into specific time-based criteria, specifically by setting an aging capacity threshold. Mapped to time threshold A lightweight threshold determination tool was built, which only requires inputting the aging capacity threshold. This allows us to obtain an operable equal voltage difference charging time aging threshold. This provides direct evidence for real-time monitoring.
[0033] S4, based on the equal voltage difference charging time aging threshold To assess the aging status of the battery under test.
[0034] This step compares the optimized equal voltage difference charging time of the battery under test (e.g., ) and aging threshold of equal voltage difference charging time Determine the aging status (e.g.) ＜ (Time-based failure). By directly comparing the optimized time characteristics with preset thresholds, binary state decisions are achieved, simplifying complex aging assessments into a single threshold comparison. There is no need for real-time calculation of capacity or internal resistance, supporting online and rapid diagnostics, and it is especially suitable for embedded real-time applications of battery management systems (BMS).
[0035] Understandably, given the shortcomings in the background technology, this invention proposes a lithium battery aging monitoring method based on Bayesian optimization of equal voltage difference charging time. This method automatically searches for and determines the optimal charging voltage boundaries (VL and VH) through Bayesian optimization, maximizing the correlation between the measured equal voltage difference charging time and the battery's aged capacity within this range. This strengthens the charging time as a highly sensitive feature to aging. Furthermore, this optimized feature is used to establish a generalized linear regression model, transforming the known capacity failure threshold into a directly comparable charging time threshold. Ultimately, this method simplifies the complex aging state assessment to a single charging time threshold comparison, enabling online and rapid lithium battery aging failure judgment using only fragmented charging data. This effectively overcomes the dependence of traditional methods on complete charge-discharge curves, current accumulation, or battery static operation, significantly improving the applicability, efficiency, and reliability of monitoring.
[0036] Based on the above technical solutions, the embodiments of the present invention can be further improved as follows.
[0037] In one possible embodiment, in step S1, the equal voltage difference charging time , represented as:
[0038] Where m is the index of the sampling point during the nth charging process. This represents the charging voltage sampled at the m-th time during the n-th charging process. This represents the time value corresponding to when the charging voltage reaches the high charging voltage VH. This represents the time value corresponding to when the charging voltage reaches the low charging voltage VL. Therefore, the difference between the two can be used to calculate the equal voltage difference charging time.
[0039] Understandably, this embodiment transforms the continuous charging voltage-time curve into a discretized time feature. By locking the charging timestamps corresponding to two specific voltage points, the difference between them is calculated to characterize the battery's charging rate within this voltage range. This embodiment utilizes the electrochemical characteristic that the increased internal resistance during battery aging leads to a longer charging time per unit voltage range. It transforms the abstract differences in charging curve morphology into precisely measurable time parameters, providing quantifiable feature inputs for subsequent Bayesian optimization, and achieving effective extraction from complex waveforms to single feature values.
[0040] In one possible embodiment, step S2 includes sub-steps S201-S203: S201, Define the Bayesian optimization objective function as equal voltage difference charging time. and aging capacity sequence The Pearson correlation coefficient between the two is used to quantify the strength of their linear association. The objective function is expressed as:
[0041] in, The average isovoltage charging time for N charge-discharge cycles under given VL and VH. This represents the average aging capacity over N charge-discharge cycles at given VL and VH, where n is the index of the number of charge cycles. .
[0042] Initial random sample points are obtained within the range of low charging voltage VL and high charging voltage VH to obtain an initial sample set, thereby exploring the parameter space. Then, a Gaussian process is used to model the initial sample set to obtain the probability distribution of the Pearson correlation coefficient, including the mean and variance, to characterize the uncertainty of the objective function.
[0043] This step actively learns the global behavior of the objective function through a probabilistic model, providing a foundation for the iterative optimization of subsequent acquisition functions (such as the expected improvement function), ensuring that the search for voltage boundaries can converge efficiently and avoid local optima, ultimately improving the robustness and efficiency of aging feature selection.
[0044] S202, define the Bayesian optimization acquisition function as the Expected Improvement (EI) function:
[0045] in, and These are the low charging voltage and high charging voltage values that maximize the Pearson correlation coefficient within the current sample set; New charging low voltage (VL) and charging high voltage (VH) samples are obtained by maximizing the expected boost function, in order to update the sample set and the Gaussian process model.
[0046] Understandably, this step is the core of the Bayesian optimization process (the iterative step), which uses the expected improvement function as the acquisition function to intelligently guide the search direction. Specifically, EI The function is defined as the Pearson correlation coefficient R(VL,VH) of the current candidate point (VL, VH) and the historical best value. R( , ) The expected positive difference essentially represents a trade-off between the potentially high returns of exploring unknown regions and the expected probability of local refinement using known, better regions. By continuously maximizing the EI function, the system can dynamically select the next most promising voltage boundary sample point (VL, VH) and use this sample to update the probability distribution (mean and variance) of the Gaussian process model, thereby gradually approaching the global optimum. This embodiment effectively avoids the shortcomings of traditional optimization methods that are prone to getting trapped in local optima, improving the efficiency and robustness of voltage boundary search.
[0047] S203, when the preset iteration stop condition is met, outputs the Bayesian-optimized low-voltage charging value. and charging high voltage value And thus obtain the first Optimized equal voltage difference charging time during the second charging process , represented as:
[0048] in, To achieve the optimized high charging voltage value The corresponding time value, To achieve the optimized low charging voltage value when the charging voltage reaches the target value The corresponding time value.
[0049] This step is the convergence and output phase of the Bayesian optimization process. When the preset iteration stopping condition is reached (such as the maximum number of iterations or the objective function convergence threshold), the search terminates and the optimal charging voltage boundary value is output. and The optimized voltage value was determined through iterative optimization using a Gaussian process model and a data acquisition function (such as the desired boost function), maximizing the Pearson correlation coefficient between the equal voltage difference charging time and the aging capacity. Furthermore, the optimized equal voltage difference charging time for the nth charging process was calculated using the obtained optimized voltage value. That is, the voltage from Rise to The time difference. This step ensures the aging characteristics ( The sensitivity and reliability of the data provide high-quality input for the subsequent fitting of the generalized linear regression model and the calculation of the aging threshold, thereby improving the overall accuracy and efficiency of lithium battery aging monitoring.
[0050] In one possible embodiment, in step S3, the aging capacity sequence is... And optimized equal voltage difference charging time Fitting the generalized linear regression model, including sub-step S301: S301 defines the generalized linear regression model from aging capacity to equal voltage difference charging time as follows:
[0051] in, Here are the predicted values for equal voltage difference charging time, where A is the coefficient of the linear term, B is the coefficient of the nonlinear term, and C is the constant term. According to the aging capacity sequence And optimized equal voltage difference charging time The values of A, B, and C are estimated using the least squares method as follows: , and , represented as:
[0052] in, It is an N×1 aging capacity matrix. It is an N-fold optimized equal voltage difference charging time matrix of size N×1.
[0053] It is understandable that this embodiment defines a generalized linear regression model to measure the battery aging capacity. Compared with the optimized equal voltage difference charging time The nonlinear relationship between the parameters is modeled as a mathematical expression containing a linear term A, a square root term B, and a constant term C, to more accurately capture the complex correlation between capacity decay and charging time changes during battery aging. Subsequently, the model parameters are estimated using the least squares method, and the optimal parameters are solved through matrix operations to ensure the predicted values are accurate. Compared with actual value To minimize the error. This embodiment is based on the high-sensitivity features optimized by Bayes. Establishing a quantitative relationship can improve the accuracy of the mapping between aging capacity and charging time, thereby enhancing the robustness and practicality of the entire monitoring system in judging the aging status of lithium batteries.
[0054] In one possible embodiment, in step S3, the step of determining the aging capacity threshold... The generalized linear regression model is used to calculate the aging threshold of the equal voltage difference charging time. This includes sub-step S302: S302, the aging capacity threshold Substituting into the generalized linear regression model, the output is the aging threshold for equal voltage difference charging time. .
[0055] It is understandable that a preset aging capacity threshold based on the degree of battery capacity degradation would be used. Substituting (e.g., 80% of the initial capacity) into the generalized linear regression model established through the aforementioned embodiments, the corresponding equal voltage difference charging time aging threshold is directly output through mathematical calculation. This embodiment transforms the capacity failure standard, which is difficult to measure directly online and in real time, into a charging time threshold that can be directly and quickly measured in actual monitoring. This simplifies the complex aging state judgment into a simple time value comparison, greatly reducing the computational complexity of the online monitoring system.
[0056] In one possible embodiment, step S4 includes: In the kth (k)th lithium battery under test During the second charging process, the optimized equal voltage difference charging time obtained by performing the Bayesian optimization process is acquired. and the aging threshold of the equal voltage difference charging time. In comparison: like ＜ If the test result is positive, the lithium battery under test is determined to be in an aging and failure state; otherwise, the lithium battery under test is determined to be in a healthy state.
[0057] Understandably, in the actual monitoring of the battery under test, this step directly compares the optimized equal voltage difference charging time of its k-th charging process. Compared with the preset time aging threshold This allows for intuitive assessment of aging failure or health status. In this embodiment, the complex model (Bayesian optimization, regression analysis) established in the preceding steps is ultimately transformed into an extremely simple, real-time executable logical comparison operation. This enables the method of this invention to be integrated into resource-constrained battery management systems (BMS), achieving online automatic diagnosis without manual intervention and with low computational costs. It fundamentally solves the engineering challenges of traditional aging monitoring methods, such as computational complexity, high latency, and difficulty in embedded deployment.
[0058] The method of the present invention will now be verified in a specific implementation scenario.
[0059] This implementation scenario uses charging voltage and aging capacity data from the B0005 lithium battery aging dataset provided by NASA. The maximum number of optimization iterations is 100. The charging low voltage VL ranges from [3.4V to 3.8V], and the charging high voltage VH ranges from [3.9V to 4.2V]. Figure 2 shows the results obtained through Bayesian optimization. Figure 2(a) shows the initial sample and the sample space of charging low voltage VL and charging high voltage VH obtained during Bayesian optimization, as well as the Pearson correlation coefficient values of the objective function for each sample, i.e., a schematic diagram of the objective function model. Figure 2(b) shows the changing trend of the Pearson correlation coefficient values during iterative optimization, i.e., a schematic diagram of the optimization convergence history. Figure 2(b) shows that the optimization finally converged, and the optimized charging low voltage value at the voltage boundary was obtained. and optimized charging high voltage value The voltages are 3.784V and 4.135V respectively.
[0060] Figure 3 Based on the optimized voltage boundaries obtained from Figure 2, namely a low charging voltage of 3.784V and a high charging voltage of 4.135V, the fitting results were obtained by performing a generalized linear regression model on the equal voltage charging time and aging capacity of the B0005 dataset. The fitted parameters are as follows: , and . Figure 3 middle This represents a true isovoltage charging time series. The fact that the equal voltage difference charging time predicted by the aging capacity and the generalized linear regression model are basically consistent with each other demonstrates the effectiveness of the Bayesian optimization and generalized linear regression model proposed in this invention.
[0061] Figure 4 is based on Figure 2 and Figure 3 Based on the optimized voltage boundary and generalized linear regression model obtained from the B0005 dataset, the results of aging state detection on the NASA-provided battery aging dataset B0006 are presented. Figure 4(a) shows the trend of aging capacity with charge-discharge cycles, where the value continuously decreases. 80% of the initial capacity is used as the threshold for aging failure judgment (i.e., the aging threshold of equal voltage difference charging time). As shown in Figure 4(a), the battery reaches an aging failure state after approximately 60 charge-discharge cycles. Figure 4(b) shows the trend of equal voltage differential charging time with charge-discharge cycles, and its value continuously decreases. Based on the generalized linear regression model, 80% of the initial capacity value is converted into the aging threshold of equal voltage differential charging time. After approximately 55 charge-discharge cycles, the battery reached an aging failure state, with a difference of only 5 cycles between the two. The high accuracy of aging monitoring indicates that the method proposed in this invention can replace capacity with equal voltage charging time to detect the aging state of lithium batteries. It has the characteristics of strong usability, few constraints, high computational efficiency, and accurate calculation.
[0062] Figure 5 A structural diagram of a lithium battery aging monitoring system based on Bayesian optimization of equal voltage difference charging time is provided for an embodiment of the present invention, as shown below. Figure 5 As shown, a lithium battery aging monitoring system based on Bayesian optimization of equal voltage difference charging time includes a data acquisition module, an optimization calculation module, a threshold calculation module, and a state assessment module, wherein: The data acquisition module is configured to acquire the aging capacity sequence obtained from the nth charge-discharge cycle of the lithium battery. and equal voltage difference charging time Among them, the equal voltage difference charging time It is a function related to the low charging voltage VL and the high charging voltage VH; The optimization calculation module is configured to optimize the aging capacity sequence. and equal voltage difference charging time Using the Pearson correlation coefficient between the two voltage levels as the optimization objective, a Bayesian optimization process was performed to obtain the optimized equal-voltage differential charging time. ; The threshold calculation module is configured to be based on the aging capacity sequence. And optimized equal voltage difference charging time Fit a generalized linear regression model and base it on the aging capacity threshold. The generalized linear regression model is used to calculate the aging threshold of the equal voltage difference charging time. ; The condition assessment module is configured to perform aging based on the equal voltage difference charging time threshold. To assess the aging status of the battery under test.
[0063] It is understood that the lithium battery aging monitoring system based on Bayesian optimization of equal voltage difference charging time provided by the present invention corresponds to the lithium battery aging monitoring method based on Bayesian optimization of equal voltage difference charging time provided in the foregoing embodiments. The relevant technical features of the lithium battery aging monitoring system based on Bayesian optimization of equal voltage difference charging time can be referred to the relevant technical features of the lithium battery aging monitoring method based on Bayesian optimization of equal voltage difference charging time, and will not be repeated here.
[0064] Please see Figure 6 , Figure 6 This is a schematic diagram illustrating an embodiment of the electronic device provided in this invention. For example... Figure 6 As shown, this embodiment of the invention provides an electronic device 600, including a memory 610, a processor 620, and a computer program 611 stored in the memory 610 and executable on the processor 620. When the processor 620 executes the computer program 611, it performs the following steps: S1, Obtain the aging capacity sequence obtained from the nth charge-discharge cycle of the lithium battery. and equal voltage difference charging time Among them, the equal voltage difference charging time It is a function related to the low charging voltage VL and the high charging voltage VH; S2, aging capacity sequence and equal voltage difference charging time Using the Pearson correlation coefficient between the two voltage levels as the optimization objective, a Bayesian optimization process was performed to obtain the optimized equal-voltage differential charging time. ; S3, based on the aging capacity sequence And optimized equal voltage difference charging time Fit a generalized linear regression model based on the aging capacity threshold. The generalized linear regression model is used to calculate the aging threshold of the equal voltage difference charging time. ; S4, based on the equal voltage difference charging time aging threshold To assess the aging status of the battery under test.
[0065] Please see Figure 7 , Figure 7 This is a schematic diagram illustrating an embodiment of a computer-readable storage medium provided by the present invention. (See diagram below.) Figure 7 As shown, this embodiment provides a computer-readable storage medium 700, on which a computer program 611 is stored. When the computer program 611 is executed by a processor, it performs the following steps: S1, Obtain the aging capacity sequence obtained from the nth charge-discharge cycle of the lithium battery. and equal voltage difference charging time Among them, the equal voltage difference charging time It is a function related to the low charging voltage VL and the high charging voltage VH; S2, aging capacity sequence and equal voltage difference charging time Using the Pearson correlation coefficient between the two voltage levels as the optimization objective, a Bayesian optimization process was performed to obtain the optimized equal-voltage differential charging time. ; S3, based on the aging capacity sequence And optimized equal voltage difference charging time Fit a generalized linear regression model based on the aging capacity threshold. The generalized linear regression model is used to calculate the aging threshold of the equal voltage difference charging time. ; S4, based on the equal voltage difference charging time aging threshold To assess the aging status of the battery under test.
[0066] This invention provides a lithium battery aging monitoring method, system, and storage medium based on Bayesian optimization of equal voltage difference charging time. First, it calculates the charging time difference for any voltage range (VL to VH) during lithium battery charging. This is used as a fundamental characteristic to characterize the aging state; subsequently, this time series is compared with the actual aging capacity series. The Pearson correlation coefficient between the two is used as the optimization objective. A Bayesian optimization algorithm is used to automatically search for and lock the optimal voltage boundary that maximizes the correlation. and Thus, the original features Enhanced to be highly sensitive to capacity decay. Next, based on this optimized feature and aging capacity data, a generalized linear regression model is fitted, thereby accurately mapping the known capacity failure threshold (e.g., 80%) to a charging time threshold that can be directly used for comparison. Ultimately, in actual monitoring, it is only necessary to calculate the charging time of the battery under test within the optimized voltage range and compare it with the time threshold. A simple comparison can quickly diagnose the aging state.
[0067] This invention transforms the complex and time-consuming battery health assessment into a lightweight charging time measurement and comparison, eliminating the reliance on complete charge-discharge curves, current integration, or long periods of inactivity. It enables online, real-time, and low-computation-cost lithium battery aging and failure assessment, significantly improving the engineering practicality and system reliability of the monitoring.
[0068] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.
[0069] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0070] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0071] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0072] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0073] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.
[0074] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.
Claims
1. A lithium battery aging monitoring method based on Bayesian optimization of equal voltage difference charging time, characterized in that, include: S1, Obtain the aging capacity sequence obtained from the nth charge-discharge cycle of the lithium battery. and equal voltage difference charging time Among them, the equal voltage difference charging time It is a function related to the low charging voltage VL and the high charging voltage VH; S2, aging capacity sequence and equal voltage difference charging time Using the Pearson correlation coefficient between the two voltage levels as the optimization objective, a Bayesian optimization process was performed to obtain the optimized equal-voltage differential charging time. ; S3, based on the aging capacity sequence And optimized equal voltage difference charging time Fit a generalized linear regression model based on the aging capacity threshold. The generalized linear regression model is used to calculate the aging threshold of the equal voltage difference charging time. ; S4, based on the equal voltage difference charging time aging threshold To assess the aging status of the battery under test.
2. The lithium battery aging monitoring method based on Bayesian optimization of equal voltage difference charging time according to claim 1, characterized in that, The equal voltage difference charging time , is represented as: Where m is the index of the sampling point during the nth charging process. This represents the charging voltage sampled at the m-th time during the n-th charging process. This represents the time value corresponding to when the charging voltage reaches the high charging voltage VH. This represents the time value corresponding to when the charging voltage reaches the low charging voltage VL.
3. The lithium battery aging monitoring method based on Bayesian optimization of equal voltage difference charging time according to claim 1, characterized in that, Step S2 includes: S201, Define the Bayesian optimization objective function as equal voltage difference charging time. and aging capacity sequence The Pearson correlation coefficient between them; Initial random sample points are obtained within the range of low charging voltage VL and high charging voltage VH to form a sample set. A Gaussian process is used to model the sample set to obtain the probability distribution of the Pearson correlation coefficient. S202, define the Bayesian optimization acquisition function as the expected improvement function: in, and These are the low charging voltage and high charging voltage values that maximize the Pearson correlation coefficient within the current sample set; New charging low voltage VL and charging high voltage VH samples are obtained by maximizing the expected boost function, so as to update the sample set and the Gaussian process model. S203, when the preset iteration stop condition is met, outputs the Bayesian-optimized low-voltage charging value. and charging high voltage value This leads to the optimized equal voltage difference charging time for the nth charging process. , is represented as: 。 in, To achieve the optimized high charging voltage value The corresponding time value, To achieve the optimized low charging voltage value The corresponding time value.
4. The lithium battery aging monitoring method based on Bayesian optimization of equal voltage difference charging time according to claim 3, characterized in that, In step S2, the aging capacity sequence and equal voltage difference charging time The Pearson correlation coefficient between them is expressed as: in, The average isovoltage charging time for N charge-discharge cycles under given VL and VH. This represents the average aging capacity over N charge-discharge cycles at given VL and VH, where n is the index of the number of charge cycles. .
5. The lithium battery aging monitoring method based on Bayesian optimization of equal voltage difference charging time according to claim 1, characterized in that, In step S3, the aging capacity sequence is used as a basis for... And optimized equal voltage difference charging time Fitting a generalized linear regression model includes: The generalized linear regression model for aging capacity to equal voltage difference charging time is defined as follows: in, Here are the predicted values for equal voltage difference charging time, where A is the coefficient of the linear term, B is the coefficient of the nonlinear term, and C is the constant term. According to the aging capacity sequence And optimized equal voltage difference charging time The values of A, B, and C are estimated using the least squares method as follows: , and , is represented as: in, It is an N×1 aging capacity matrix. It is an N-fold optimized equal voltage difference charging time matrix of size N×1.
6. The lithium battery aging monitoring method based on Bayesian optimization of equal voltage difference charging time according to claim 5, characterized in that, In step S3, the step of determining the aging capacity threshold... The generalized linear regression model is used to calculate the aging threshold of the equal voltage difference charging time. ,include: The aging capacity threshold Substituting into the generalized linear regression model, the output is the aging threshold for equal voltage difference charging time. .
7. The lithium battery aging monitoring method based on Bayesian optimization of equal voltage difference charging time according to claim 1, characterized in that, Step S4 includes: During the k-th charging process of the lithium battery under test, the optimized equal voltage difference charging time obtained by performing a Bayesian optimization process is acquired. and the aging threshold of the equal voltage difference charging time. In comparison: like < If the test result is positive, the lithium battery under test is determined to be in an aging and failure state; otherwise, the lithium battery under test is determined to be in a healthy state.
8. A lithium battery aging monitoring system based on Bayesian optimization of equal voltage difference charging time, characterized in that, include: The data acquisition module is configured to acquire the aging capacity sequence obtained from the nth charge-discharge cycle of the lithium battery. and equal voltage difference charging time Among them, the equal voltage difference charging time It is a function related to the low charging voltage VL and the high charging voltage VH; The optimization calculation module is configured to optimize the aging capacity sequence. and equal voltage difference charging time Using the Pearson correlation coefficient between the two voltage levels as the optimization objective, a Bayesian optimization process was performed to obtain the optimized equal-voltage differential charging time. ; The threshold calculation module is configured to be based on the aging capacity sequence. And optimized equal voltage difference charging time Fit a generalized linear regression model and base it on the aging capacity threshold. The generalized linear regression model is used to calculate the aging threshold of the equal voltage difference charging time. ; The condition assessment module is configured to perform aging based on the equal voltage difference charging time threshold. To assess the aging status of the battery under test.
9. An electronic device, characterized in that, The system includes a memory and a processor, wherein the processor is used to execute computer management programs stored in the memory to implement the steps of the lithium battery aging monitoring method based on Bayesian optimization of equal voltage difference charging time as described in any one of claims 1-7.
10. A computer-readable storage medium, characterized in that, It stores a computer management program, which, when executed by a processor, implements the steps of a lithium battery aging monitoring method based on Bayesian optimization of equal voltage difference charging time as described in any one of claims 1-7.