Methods, apparatus, equipment and dielectrics for predicting the remaining lifespan of lithium-ion batteries

Abstract

This invention relates to the field of mathematical modeling technology, and more particularly to a method, apparatus, device, and computer-readable storage medium for predicting the remaining lifespan of lithium-ion batteries. The method includes: establishing a segmented empirical degradation model based on inflection point data of the lithium-ion battery; obtaining an initial prediction result of the remaining lifespan of the lithium-ion battery according to a preset particle filter algorithm and the segmented empirical degradation model, and obtaining an original error sequence based on the initial prediction result; determining a reconstruction error sequence corresponding to the capacity of the lithium-ion battery according to the original error sequence and a preset discrete wavelet transform algorithm; constructing a prediction model using a support vector regression algorithm based on the reconstruction error sequence, and correcting the initial prediction result based on the prediction model to determine the final prediction result of the remaining lifespan. This invention improves the prediction efficiency and accuracy of the remaining lifespan of lithium-ion batteries.

Description

Technical Field

[0001] This invention relates to the field of mathematical modeling technology, and in particular to a method, apparatus, terminal device, and computer-readable storage medium for predicting the remaining lifespan of lithium-ion batteries. Background Technology

[0002] With the continuous development of science and technology and the increasing demand for lithium batteries in people's production and life, lithium batteries have a very broad market prospect due to their wide application. At the same time, users have put forward higher requirements for the prediction of the remaining service life of lithium batteries through battery management.

[0003] Existing methods for predicting the remaining lifespan of lithium-ion batteries typically employ electrochemical and equivalent circuit models or data-driven approaches. On the one hand, using electrochemical and equivalent circuit models requires extensive and complex calculations, and the modeling process is also very complicated. Furthermore, measurement difficulties can easily lead to quadrants with large prediction errors, making it unsuitable for predicting the remaining lifespan of batteries with long prediction cycles. On the other hand, using data-driven approaches requires a large amount of data recording and computational costs to predict the degradation trend of lithium batteries.

[0004] In summary, existing methods for predicting the remaining lifespan of lithium-ion batteries suffer from technical problems such as low accuracy of prediction data and cumbersome prediction processes. Summary of the Invention

[0005] The main objective of this invention is to provide a method, apparatus, device, and computer-readable storage medium for predicting the remaining lifespan of lithium-ion batteries, aiming to optimize the modeling process for predicting the remaining lifespan of lithium-ion batteries and improve the accuracy and efficiency of the prediction data.

[0006] To achieve the above objectives, the present invention provides a method for predicting the remaining service life of a lithium-ion battery, the method comprising:

[0007] A segmented empirical degradation model was established based on the inflection point data of lithium-ion batteries;

[0008] The initial prediction result of the remaining service life of the lithium-ion battery is obtained based on the preset particle filter algorithm and the segmented empirical degradation model, and the original error sequence is obtained based on the initial prediction result.

[0009] The reconstructed error sequence corresponding to the capacity of the lithium-ion battery is determined based on the original error sequence and the preset discrete wavelet transform algorithm.

[0010] A prediction model for a support vector regression algorithm is constructed based on the reconstructed error sequence, and the initial prediction results are corrected based on the prediction model to determine the final prediction result for the remaining useful life.

[0011] Optionally, the step of establishing a segmented empirical degradation model based on the inflection point data of lithium-ion batteries includes:

[0012] Acquire the capacity degradation data of the lithium-ion battery, and smooth the capacity degradation data according to a preset filter;

[0013] After the filter completes the smoothing process of the capacity degradation data, the difference in the capacity decay curve of the lithium-ion battery is determined according to a preset first-order differential equation.

[0014] Based on a preset parameter interval, the inflection point data of the lithium-ion battery corresponding to the difference is identified, and the segmented empirical degradation model is established based on the inflection point data.

[0015] Optionally, the step of obtaining the initial prediction result of the remaining service life of the lithium-ion battery based on the preset particle filter algorithm and the segmented empirical degradation model includes:

[0016] Determine whether the current cycle data of the lithium-ion battery is less than the inflection point data;

[0017] If so, then determine the double exponential model corresponding to the segmented empirical degradation model, and obtain the initial prediction result based on the double exponential model and the preset particle filter algorithm;

[0018] If not, then determine the autoregressive integrated moving average model corresponding to the segmented empirical degradation model, and obtain the initial prediction result based on the autoregressive integrated moving average model and the preset particle filter algorithm.

[0019] Optionally, before the step of obtaining the initial prediction result based on the autoregressive integrated moving average model and a preset particle filter algorithm, the method includes:

[0020] The order difference equation is obtained based on the stationarity of the lithium-ion battery, and the first parameter of the autoregressive integrated moving average model is determined based on the order difference equation.

[0021] The second and third parameters of the autoregressive integrated moving average model are determined based on the preset Akaike information criterion and the preset Bayesian information criterion.

[0022] Optionally, the step of determining the reconstructed error sequence corresponding to the capacity of the lithium-ion battery based on the original error sequence and a preset discrete wavelet transform algorithm includes:

[0023] The decomposed signal is obtained based on the original error sequence and the preset discrete wavelet transform algorithm;

[0024] The reconstruction error sequence corresponding to the capacity of the lithium-ion battery is obtained by reconstructing the decomposed signal.

[0025] Optionally, the step of reconstructing the reconstruction error sequence corresponding to the capacity of the lithium-ion battery based on the decomposed signal includes:

[0026] The high-frequency component signal of the decomposed signal is removed based on the preset discrete wavelet transform algorithm to obtain the low-frequency component signal of the decomposed signal.

[0027] The reconstruction error sequence is obtained by reconstructing the low-frequency component signal.

[0028] Optionally, the step of correcting the initial prediction result based on the prediction model to determine the final prediction result of the remaining useful life includes:

[0029] Generate a prediction error sequence based on the prediction model;

[0030] The initial prediction result is corrected based on the prediction error sequence to obtain the final prediction result of the remaining useful life.

[0031] Furthermore, to achieve the above objectives, the present invention also provides a device for predicting the remaining service life of a lithium-ion battery. The device for predicting the remaining service life of a lithium-ion battery includes:

[0032] The modeling module is used to build segmented empirical degradation models based on the inflection point data of lithium-ion batteries.

[0033] The filtering calculation module is used to obtain the initial prediction result of the remaining service life of the lithium-ion battery according to the preset particle filtering algorithm and the segmented empirical degradation model, and to obtain the original error sequence based on the initial prediction result.

[0034] The reconstructed data determination module is used to determine the reconstructed error sequence corresponding to the capacity of the lithium-ion battery based on the original error sequence and a preset discrete wavelet transform algorithm.

[0035] The prediction module is used to construct a prediction model of the support vector regression algorithm based on the reconstructed error sequence, and to correct the initial prediction result based on the prediction model to determine the final prediction result of the remaining useful life.

[0036] Each functional module of the lithium-ion battery remaining life prediction device of the present invention implements the steps of the lithium-ion battery remaining life prediction method of the present invention as described above during operation.

[0037] In addition, to achieve the above objectives, the present invention also provides a terminal device, the terminal device including a memory, a processor, and a lithium-ion battery remaining life prediction program stored in the memory and executable on the processor, wherein when the lithium-ion battery remaining life prediction program is executed by the processor, the steps of the above-described lithium-ion battery remaining life prediction method are implemented.

[0038] Furthermore, to achieve the above objectives, the present invention also provides a computer-readable storage medium storing a program for predicting the remaining lifespan of a lithium-ion battery, wherein when the program is executed by a processor, it implements the steps of the above-described method for predicting the remaining lifespan of a lithium-ion battery.

[0039] This invention first determines the inflection point data of lithium-ion batteries to establish a segmented empirical degradation model, thereby improving the universality and accuracy of the algorithm. Then, based on the empirical degradation model, a pre-defined particle filter (PF) algorithm is implemented to determine the initial prediction results, and an original error sequence is generated based on these results. Next, a pre-defined discrete wavelet transform (DWT) algorithm is used to decompose and reconstruct the original error sequence to determine the reconstructed error sequence, improving data validity by reducing local noise distribution information. Finally, a support vector regression (SVR) algorithm is constructed based on the reconstructed error sequence to determine the prediction error sequence, and the initial prediction results based on the PF algorithm are corrected based on the prediction error sequence, thus determining the final prediction result of the remaining lifespan of the lithium-ion battery.

[0040] Unlike existing methods for predicting the remaining lifespan of lithium-ion batteries, this invention implements the Power Factor (PF) algorithm based on a segmented empirical degradation model to determine the initial prediction result for the lithium-ion battery. The original error sequence is then passed as a byproduct to a pre-defined Decomposition and Reconstruction (DWT) algorithm. The approximate signal with high coefficients is used as the reconstruction error sequence, which is then used as training data to a pre-defined Suspension-Variation (SVR) algorithm. The prediction error sequence obtained from the SVR algorithm is used to correct the initial prediction result based on the PF algorithm to determine the final prediction result for the remaining lifespan of the lithium-ion battery. This effectively avoids the complex modeling and calculation required for each prediction of the remaining lifespan of lithium-ion batteries, as well as the low measurement accuracy. By implementing the PF algorithm based on a segmented empirical degradation model, and then sequentially using the DWT and SVR algorithms to determine the final prediction result, this invention effectively optimizes the modeling and calculation processes for predicting the remaining lifespan of lithium-ion batteries, thereby improving prediction efficiency and the accuracy of the prediction results. Attached Figure Description

[0041] Figure 1This is a flowchart illustrating the first embodiment of the method for predicting the remaining lifespan of a lithium-ion battery according to the present invention.

[0042] Figure 2 This is a schematic diagram illustrating the specific application process of an embodiment of the lithium-ion battery remaining life prediction method of the present invention.

[0043] Figures 3(a), 3(b), 3(c), and 3(d) are schematic diagrams of capacity degradation data involved in an embodiment of the lithium-ion battery remaining service life prediction method of the present invention.

[0044] Figures 4(a), 4(b), 4(c), and 4(d) are iterative schematic diagrams based on the PF algorithm in an embodiment of the lithium-ion battery remaining life prediction method of the present invention.

[0045] Figure 5 This is a schematic diagram of an embodiment of the lithium-ion battery remaining life prediction method of the present invention, based on the DWT algorithm.

[0046] Figures 6(A01), 6(A02), 6(A03), and 6(A04) are schematic diagrams of an embodiment of the lithium-ion battery remaining life prediction method of the present invention, based on the SVR algorithm.

[0047] Figure 7 This is a schematic diagram of the remaining lifespan prediction device module for lithium-ion batteries of the present invention.

[0048] Figure 8 This is a schematic diagram of the structure of the terminal device involved in the embodiment of the present invention;

[0049] Figure 9 This is a schematic diagram of the structure of a computer-readable storage medium involved in an embodiment of the present invention.

[0050] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0051] This invention provides a method for predicting the remaining lifespan of a lithium-ion battery, referring to... Figure 1 As shown, Figure 1 This is a flowchart illustrating the first embodiment of the method for predicting the remaining lifespan of a lithium-ion battery according to the present invention.

[0052] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application.

[0053] In this embodiment, the remaining lifespan prediction method for lithium-ion batteries of the present invention is applied to a terminal device for predicting the remaining lifespan of lithium-ion batteries. The remaining lifespan prediction method for lithium-ion batteries of the present invention includes:

[0054] Step S10: Establish a segmented empirical degradation model based on the inflection point data of lithium-ion batteries;

[0055] The terminal device first smooths the capacity degradation data of lithium-ion batteries (LIBs) according to a preset filter. After the smoothing process, it determines the difference in the capacity decay curve of the lithium-ion battery according to a preset first-order differential equation. Then, it identifies the inflection point data corresponding to the difference based on a preset parameter interval and establishes the segmented empirical degradation model based on the inflection point data.

[0056] It should be noted that the segmented empirical degradation model refers to the double exponential model and the ARIMA model. The double exponential model can be expressed as:

[0057] Q(k) = ae b·k +ce d·k (1)

[0058] Where k is the number of charge / discharge cycles, Q(k) is the battery capacity at the kth cycle, and a, b, c, and d are constants that change with time (related to the battery's internal impedance and are aging parameters).

[0059] The improved double-exponential model, by subtracting the non-positive value expression, significantly mitigates the sharp degradation in the later stages and can be expressed as:

[0060] Q(k) = ae a+b·k -ce c+d / k (2)

[0061] Among them, e a+b·k and -e c+d / k The scale represents the initial ability decline and the accelerated ability decline process.

[0062] The performance fitting results of models (1) and (2) are shown in Table 2. It is clear that throughout the degradation process, model (2) outperforms model (1) in both SSE and RMSE, and R2 is also closer to 1. However, when the field of view is placed in the early stage of degradation, model (1) performs even better. Table 2 shows the fitting performance comparison of models [i.e., models (1) and (2)] in the two degradation stages.

[0063]

[0064] Table 2

[0065] ARIMA models can observe the current moment based on historical moments. ARIMA models are suitable for real-time state modeling and can be used as a measurement model for PF models. An ARIMA model consists of three parts: ① an autoregressive (AR) process; ② establishing a stationary time series through difference ordering; and ③ a moving average (MA) process. The general form of ARIMA(p,d,q) is as follows:

[0066]

[0067] Wherein, p is the backward Q(kj) of the measured value and the modeling error ε k Used to obtain the current measurement value Q(k); r j and θ i q represents the parameters of the observation at time j and the parameters of the error term at time i, respectively; q represents the number of shift error terms.

[0068] p represents the lags of the lithium-ion battery charge / discharge cycle data used in the prediction model, also known as the AR / Auto-Regressive term; d represents the number of differential steps required for the lithium-ion battery charge / discharge cycle data to be stable, also known as the Integrated term; q represents the lags of the prediction error used in the prediction model, also known as the MA / Moving Average term.

[0069] Inflection point data refers to the charge / discharge cycle data before the actual capacity drop point of a lithium-ion battery.

[0070] In this embodiment, different degradation stages are accurately described by inflection points, namely, a dual exponential model (1) and an ARIMA model (3) are established as early and late model schemes, respectively. This improves the problem of insufficient real-time performance in the existing method of predicting the remaining lifespan of lithium-ion batteries, reduces unnecessary parameter estimation, and greatly reduces the amount of computation.

[0071] Step S20: Obtain the initial prediction result of the remaining service life of the lithium-ion battery according to the preset particle filter algorithm and the segmented empirical degradation model, and obtain the original error sequence based on the initial prediction result;

[0072] In this embodiment, the terminal device adjusts the segmented empirical degradation model using a preset particle filter algorithm to obtain an initial prediction result of the remaining lifespan of the lithium-ion battery, and then obtains the original error sequence based on the preset particle filter algorithm according to the initial prediction result.

[0073] It should be noted that remaining useful life can be expressed as Remaining Useful Life, or RUL. The remaining useful life of a lithium-ion battery refers to the number of charge / discharge cycles required for the battery's maximum usable capacity to decay to a specified failure threshold after a certain charge / discharge process, i.e., the number of charge / discharge cycles.

[0074] The particle filter algorithm, also known as the PF algorithm, is a Bayesian filtering algorithm based on the Monte Carlo method. It can handle any nonlinear, non-Gaussian problem. The PF algorithm is expressed as follows:

[0075]

[0076] Where, x k u represents the state of the system at time k. k-1 The process noise at time k-1 is represented by f(·), which is a linear or nonlinear function that establishes the relationship between the current state and the state at the last time step. k v represents the observation value at time k. k It measures noise, h(·), which is a linear or nonlinear function that establishes the relationship between the state value and the simultaneous measurement value. p(x k |x k-1 ) is the prior probability of the state equation, p(z) k |x k ) represents the likelihood function of the observed distribution.

[0077] In this embodiment, the empirical degradation model is used as the measurement equation in the PF algorithm, and the particle filter algorithm flow is as follows:

[0078] Process 1: Initialization

[0079] N particles p(x0) is generated by sampling from the prior probability.

[0080] Process 2: Importance Sampling

[0081] Samples i = 1, ..., N

[0082]

[0083] Step 3: Weight Calculation

[0084]

[0085] Step 4: Resampling

[0086] Calculate the number of valid samples

[0087]

[0088] if Generate based on importance weights, By using residual resampling method.

[0089] Process 4: State Estimation

[0090]

[0091] Referring to Figure 4, which is an iterative schematic diagram of the PF algorithm based on an embodiment of the lithium-ion battery remaining life prediction method of the present invention, Figure 4(c) shows a partial iterative process of particle filtering algorithm measurement based on the segmented empirical degradation model (double exponential model and ARIMA model).

[0092] Step S30: Determine the reconstruction error sequence corresponding to the capacity of the lithium-ion battery based on the original error sequence and the preset discrete wavelet transform algorithm;

[0093] In this embodiment, the terminal device inputs the received original error sequence into a preset Discrete Wavelet Transform (DWT) algorithm, and then analyzes and decomposes the signals of different scales of the original error sequence according to the preset DWT algorithm. Through the finite decomposition and reconstruction process, it extracts the original error sequence into approximate components, thereby obtaining a smooth residual sequence, i.e., the reconstructed error sequence.

[0094] In this embodiment, the reconstructed error sequence obtained by the preset Discrete Wavelet Transform (DWT) algorithm is used as the training dataset for prediction, which lays a solid foundation for the successful prediction of the preset Support Vector Regression algorithm.

[0095] Step S40: Construct a prediction model for the support vector regression algorithm based on the reconstructed error sequence, and correct the initial prediction result based on the prediction model to determine the final prediction result of the remaining useful life.

[0096] In this embodiment, the terminal device first constructs a prediction model of the support vector regression (SVR) algorithm based on the reconstruction error sequence, generates a prediction error sequence of the lithium-ion battery based on the prediction model of the SVR algorithm, and then determines the final prediction result of the remaining service life of the lithium-ion battery based on the prediction error sequence.

[0097] Reference Figure 2 , Figure 2 This is a schematic diagram illustrating the specific application process of an embodiment of the lithium-ion battery remaining lifespan prediction method of the present invention. Initialization represents preprocessing the capacity degradation data of the lithium-ion battery; k represents the current cycle data of the lithium-ion battery, i.e., the current charge / discharge cycle number of the lithium-ion battery; k ipk represents the number of periods of the inflection point data. end This refers to the number of cycles corresponding to the initial capacity percentage of a lithium-ion battery, i.e., the cycle data for the failure threshold.

[0098] Reference Figure 2 As shown, the terminal device first initializes its settings, and then checks whether k is less than k. ip If the current number of charging / discharging cycles is less than the number of cycles of the inflection point data, it can also be understood as performing iterative measurement of the particle filter algorithm in the double exponential model, i.e., Eq.(1), before the inflection point data; if the current number of charging / discharging cycles is greater than the number of cycles of the inflection point data, it can also be understood as performing iterative measurement of the particle filter algorithm in the ARIMA model, i.e., Eq.(3), after the inflection point data, and obtaining the initial prediction result, i.e., PF predicted; then obtain the original error series based on PF predicted; then transmit the original error series to the preset DWT algorithm model, analyze and decompose the signal of the original error series in multiple layers to obtain the reconstructed error series, i.e., Reconstructed error series; then construct the prediction model of the SVR algorithm based on the Reconstructed error series, and obtain the prediction error series based on the prediction model of the SVR algorithm, i.e., Predicted error; finally correct PF predicted based on Predicted error to obtain the prediction result of the remaining service life of the lithium-ion battery, i.e., FinalRULprediction.

[0099] Reference Figure 2 As shown, the iterative measurement of the particle filtering algorithm based on the segmented empirical degradation model first initializes the number of particles of the current charge / discharge cycle of the lithium-ion battery, then performs importance sampling (Eq. (5)); then performs weight calculation (Eq. (6)); next, after calculating the number of effective samples through resampling (Eq. (7), performs state estimation (Eq. (8)); and finally determines whether k is less than k. end If it is less than, output PF predicted; if it is greater than, re-execute the judgment to see if k is less than k. ip The instructions.

[0100] This invention first determines the inflection point data of lithium-ion batteries to establish a segmented empirical degradation model, thereby improving the universality and accuracy of the algorithm. Then, based on the empirical degradation model, a pre-defined particle filter (PF) algorithm is implemented to determine the initial prediction results, and an original error sequence is generated based on these results. Next, a pre-defined discrete wavelet transform (DWT) algorithm is used to decompose and reconstruct the original error sequence to determine the reconstructed error sequence, improving data validity by reducing local noise distribution information. Finally, a support vector regression (SVR) algorithm is constructed based on the reconstructed error sequence to determine the prediction error sequence, and the initial prediction results based on the PF algorithm are corrected based on the prediction error sequence, thus determining the final prediction result of the remaining lifespan of the lithium-ion battery.

[0101] Unlike existing methods for predicting the remaining lifespan of lithium-ion batteries, this invention implements the Power Factor (PF) algorithm based on a segmented empirical degradation model to determine the initial prediction result for the lithium-ion battery. The original error sequence is then passed as a byproduct to a pre-defined Decomposition and Reconstruction (DWT) algorithm. The approximate signal with high coefficients is used as the reconstruction error sequence, which is then used as training data to a pre-defined Suspension-Variation (SVR) algorithm. The prediction error sequence obtained from the SVR algorithm is used to correct the initial prediction result based on the PF algorithm to determine the final prediction result for the remaining lifespan of the lithium-ion battery. This effectively avoids the complex modeling and calculation required for each prediction of the remaining lifespan of lithium-ion batteries, as well as the low measurement accuracy. By implementing the PF algorithm based on a segmented empirical degradation model, and then sequentially using the DWT and SVR algorithms to determine the final prediction result, this invention effectively optimizes the modeling and calculation processes for predicting the remaining lifespan of lithium-ion batteries, thereby improving prediction efficiency and the accuracy of the prediction results.

[0102] Furthermore, based on the first embodiment of the remaining lifespan prediction of the lithium-ion battery of the present invention, a second embodiment of the remaining lifespan prediction of the lithium-ion battery of the present invention is proposed.

[0103] In this embodiment, step S10 above: establishing a segmented empirical degradation model based on the inflection point data of lithium-ion batteries, may specifically include:

[0104] Step S101: Obtain the capacity degradation data of the lithium-ion battery, and smooth the capacity degradation data according to a preset filter;

[0105] In this embodiment, the terminal device first acquires the capacity degradation data of the lithium-ion battery, and then smooths the capacity degradation data according to a preset filter.

[0106] For example, four batteries (lithium-ion phosphate (LFP) / graphite, nominal capacity 1.1Ah, nominal voltage 3.3V) were selected and labeled A01-A04. All batteries used a one- or two-step fast charging protocol [C1(Q1)-C2, where C1 and C2 are the first and second constant current steps respectively, Q1 is the state of charge (SOC) at 80% SOC when charging in 1CCC-CV mode, with a cutoff voltage of 3.6V]. All batteries were discharged at a constant current of 4C with a cutoff voltage of 2V (see Table 1). Table 1 provides a detailed description of the charge and discharge protocols for the four lithium-ion batteries A01-A04. The four lithium batteries were tested at the same temperature (30°C).

[0107] Battery Charging rules barcode aisle cycle room temperature A01 6C (60%) - 3C EL150800460640 29 731 30℃ A02 6C (60%) - 3C EL150800460436 30 757 30℃ A03 7C (40%) - 3C EL150800460601 38 648 30℃ A04 7C (30%) - 3.6C EL150800460622 40 703 30℃

[0108] Table 1

[0109] Referring to Figure 3, which is a schematic diagram of capacity degradation data in an embodiment of the lithium-ion battery remaining life prediction method of the present invention, the capacity of LIBs decreases slowly in the early stage. After about 400-500 cycles, the power begins to drop sharply, as shown in the capacity degradation data of Figure 3(a). After some cycles, the capacity of LIBs does not just decrease unilaterally, but shows a rebound trend.

[0110] Figure 3(b) shows the smoothing results of the Savitzky-Golay filter on the capacity of four LIBs. The Savitzky-Golay filter is the best at smoothing the capacity of four LIBs, but it includes but is not limited to the Savitzky-Golay filter.

[0111] Additionally, it should be noted that, referring to Figure 3, Figure 3(c) shows the true capacity drop of A01. By manually setting a linkage line parallel to the initial capacity of the battery and tangent to the A01 battery curve, the end of life (EOL) point is 80%. This tangent point is determined as the true capacity drop point, i.e., the 518th cycle.

[0112] Step S102: After determining that the filter has completed the smoothing process of the capacity degradation data, the difference between the capacity decay curves of the lithium-ion battery is determined according to the preset first-order differential equation.

[0113] In this embodiment, after the terminal device determines that the filter has completed the smoothing process of the capacity degradation data of the lithium battery, it determines the difference in the capacity decay curve of the lithium-ion battery according to a preset first-order differential equation.

[0114] Step S103: Identify the inflection point data of the lithium-ion battery corresponding to the difference based on the preset parameter interval range, and establish the segmented empirical degradation model based on the inflection point data.

[0115] In this embodiment, the terminal device identifies the inflection point data corresponding to the difference based on a preset parameter range, and establishes a segmented empirical degradation model based on the inflection point data of the lithium-ion battery.

[0116] Referring to Figure 3, Figure 3(d) shows the identification of the difference in capacity decay curves and the 3σ-interval inflection point for cell A01. Point 470, determined using the u±3σ interval criterion of the differential capacity decay curve, indicates that the current capacity of cell A01 is less than 90% of its initial capacity. Early results from the segmented empirical degradation model suggest that... b·k +ce d·k The (double exponential model) shows a slow downward trend, which later indicates... (Autoregressive Integrated Moving Average Model, ARIMA Model) is the plunge phase.

[0117] In this embodiment, a segmented degradation model is established by determining inflection point data to predict diving trends.

[0118] Furthermore, in some feasible embodiments, step S20 above: obtaining the initial prediction result of the remaining service life of the lithium-ion battery based on the preset particle filter algorithm and the segmented empirical degradation model, may further include:

[0119] Step S201: Determine whether the current cycle data of the lithium-ion battery is less than the inflection point data;

[0120] In this embodiment, the terminal device needs to obtain the current cycle data of the lithium-ion battery, and then determine whether the current cycle data of the lithium-ion battery is less than the inflection point data.

[0121] Step S202: If yes, then determine the double exponential model corresponding to the segmented empirical degradation model, and obtain the initial prediction result based on the double exponential model and the preset particle filter algorithm;

[0122] In this embodiment, if the terminal device determines that the current cycle data of the lithium-ion battery is less than the inflection point data, it performs a measurement using a preset particle filter algorithm on the double exponential model and obtains the initial prediction result of the remaining lifespan of the lithium battery.

[0123] Referring to Figure 4, Figure 4 is an iterative schematic diagram of the PF algorithm based on an embodiment of the lithium-ion battery remaining life prediction method of the present invention. Figure 4(a) shows the complete iterative process of particle filtering measurement based on the double exponential model (1), and Figure 4(b) shows a partial iterative process of particle filtering measurement based on the double exponential model (1).

[0124] Step S203: If not, determine the autoregressive integrated moving average model corresponding to the segmented empirical degradation model, and obtain the initial prediction result based on the autoregressive integrated moving average model and the preset particle filter algorithm.

[0125] In this embodiment, if the terminal device determines that the current cycle data of the lithium-ion battery is greater than the inflection point data, it performs a preset particle filter algorithm measurement on the autoregressive integrated moving average model and obtains the initial prediction result of the remaining service life of the lithium battery.

[0126] Referring to Figure 4, Figure 4 is an iterative schematic diagram based on the PF algorithm involved in an embodiment of the lithium-ion battery remaining life prediction method of the present invention. In Figure 4(d), part of the iterative process of particle filtering measurement based on the autoregressive integrated moving average model (3) is shown.

[0127] Furthermore, in some other feasible embodiments, before step S203 above: obtaining the initial prediction result based on the autoregressive integrated moving average model and the preset particle filter algorithm, the method for predicting the remaining service life of lithium-ion batteries may further include:

[0128] Step A10: Obtain the order difference equation based on the stationarity of the lithium-ion battery, and determine the first parameter of the autoregressive integrated moving average model based on the order difference equation.

[0129] In this embodiment, the terminal device obtains the order difference equation based on the stability of the lithium-ion battery, and then determines the first parameter of the autoregressive integrated moving average model based on the order difference equation, which is the parameter d in ARIMA(p,d,q).

[0130] It should be noted that the order of the difference equation corresponding to the current lithium-ion battery is determined based on the stability of the lithium-ion battery, that is, the parameter d in ARIMA(p,d,q) corresponding to the order in the order difference equation.

[0131] Step A20: Determine the second and third parameters of the autoregressive integrated moving average model based on the preset Akaike information criterion and the preset Bayesian information criterion.

[0132] In this embodiment, the terminal device determines the second and third parameters of the autoregressive integrated moving average model based on the preset Akaike information criterion (AIC) and the preset Bayesian information criterion (BIC), which are parameters p and q in ARIMA(p,d,q).

[0133] The AIC and BIC values ​​can be represented as follows:

[0134] AIC=2m-2ln(L) (9)

[0135] BIC=ln(n)*m-2ln(L) (10)

[0136] Where m, n, and L represent the number of model parameters, the number of samples, and the likelihood function, respectively.

[0137] Taking battery A01 as an example, in order to prevent overfitting, the values ​​of AIC and BIC in Table 3 are 2 and 4 respectively to determine the qualitative value. Table 2 shows the AIC and BIC of the four battery models (A01-A04).

[0138] Battery Best Model AIC BIC A01 ARIMA(4,2,2) -5539 -5514 A02 ARIMA(4,2,3) -5520 -5511 A03 ARIMA(4,2,3) -5548 -5532 A04 ARIMA(4,2,2) -5529 -5518

[0139] Table 3

[0140] Furthermore, in some feasible embodiments, step S30 above: determining the reconstruction error sequence corresponding to the capacity of the lithium-ion battery based on the original error sequence and a preset discrete wavelet transform algorithm, may include:

[0141] Step S301: Obtain the decomposed signal based on the original error sequence and the preset discrete wavelet transform algorithm;

[0142] In this embodiment, the terminal device analyzes and decomposes the original error sequence into multiple layers according to a preset discrete wavelet transform algorithm to obtain the decomposed signal.

[0143] Step S302: Reconstruct the reconstruction error sequence corresponding to the capacity of the lithium-ion battery based on the decomposed signal.

[0144] In this embodiment, the terminal device extracts approximate components from the decomposed signal of the original error sequence through a finite decomposition and reconstruction process, thereby obtaining a smooth residual sequence as training data for prediction, i.e., reconstructing the error sequence.

[0145] It's important to note that the Discrete Wavelet Transform denoising process, also known as the Mallat algorithm, involves the original signal passing through two complementary filters in the first decomposition layer to generate approximate and detailed components. The second decomposition layer further decomposes these approximate and detailed components based on the approximate components from the first layer. This process continues; the more decomposition layers, the fewer approximate components are obtained, resulting in a smaller variance in the reconstructed signal. (Refer to...) Figure 5 , Figure 5 This is a schematic diagram of an embodiment of the lithium-ion battery remaining life prediction method of the present invention, based on the DWT algorithm. The low-frequency and high-frequency signals of the original capacity error sequence of battery A01 are obtained by DWT decomposition.

[0146] Furthermore, in some other feasible embodiments, step S302 above: reconstructing the reconstruction error sequence corresponding to the capacity of the lithium-ion battery based on the decomposed signal, may include:

[0147] Step S3021: Based on a preset discrete wavelet transform algorithm, delete the high-frequency component signal of the decomposed signal to obtain the low-frequency component signal of the decomposed signal.

[0148] In this embodiment, the terminal device removes the high-frequency component signal corresponding to the decomposed signal based on a preset discrete wavelet transform algorithm in order to obtain the low-frequency component signal corresponding to the decomposed signal.

[0149] It should be noted that signal correlation is evaluated using the Pearson correlation coefficient ρ (value between -1 and 1) and variance, expressed as eq.(11). When evaluating the correlation between the decomposed signal and the original signal, low correlation is represented by a value close to 0, i.e., high-frequency component signals. eq.(11) can be expressed as:

[0150]

[0151] Where, σ X and σ Y The standard deviation of the signal is represented by covariance (cov(X,Y)), which represents the overall error between the two signals.

[0152] Step S3022: Reconstruct the reconstruction error sequence based on the low-frequency component signal.

[0153] In this embodiment, the terminal device reconstructs the reconstruction error sequence based on the low-frequency component signal.

[0154] It should be noted that, since high-frequency components are considered unimportant to the evolution trend of the error sequence, the reconstructed error sequence can be composed of low-frequency components, and the dominant information of its error evolution is as follows: Figure 5 As shown in the figure. Therefore, the reconstructed low-frequency component signal of the sixth layer (the original signal minus the high-frequency component signals of the first six layers) is selected as the long-term trend of error evolution. The correlation between low-frequency and high-frequency signals is shown in Table 4, the result after discrete wavelet transform of battery A01.

[0155]

[0156] Table 4

[0157] By removing some low-correlation high-frequency signals from the residual components (i.e., the decomposed signal), the data quality is improved. It can be seen that the global fluctuations of the residual components and the trend characteristics of the reconstructed low-frequency component signals are contained in a smoother, less noisy reconstruction error sequence. This reconstruction error sequence, used as a training dataset, is used to build a prediction model based on the SVR algorithm, laying a solid foundation for successful predictions by support vector regression.

[0158] Furthermore, in some feasible embodiments, step 40 above: revising the initial prediction result based on the prediction model to determine the final prediction result of the remaining useful life may include:

[0159] Step 401: Generate a prediction error sequence based on the prediction model;

[0160] In this embodiment, the nonlinear relationship of the prediction model based on the SVR algorithm in the terminal device maps the training dataset (i.e., the reconstructed error sequence, also known as the training sample set) to a high-dimensional space. The input and output relationship of the training sample set can be expressed as:

[0161] E(x)=ωφ(x)+b (12)

[0162] Here, E(x) represents the capacity of the error, i.e., the corresponding outputs x, ω, and b represent the input data (feature vector of the sample), weights, and intercept, respectively. φ(x) is the high-dimensional feature space. Slack variables are introduced. The penalty factor C is solved by transforming it into an optimization problem, as shown in Formula (12).

[0163]

[0164] Here, ε (ε>0) is the insensitive loss coefficient. The Lagrange multiplier algorithm and the Karush-Kuhn-Tucker condition are introduced and transformed into a dual form to solve constrained optimization problems.

[0165]

[0166]

[0167] Where, α k and It is the Lagrange multiplier, and finally ω and b are obtained by solving a convex optimization problem.

[0168]

[0169]

[0170] Where N nK(x) represents the number of support vectors. k ,x j )=φ(x k )φ(x j Let be the kernel function. Finally, the hyperplane can be represented as:

[0171] E(x)=ωφ(x)+b * (19)

[0172] One of the most popular kernel functions in machine learning is the radial basis function (RBF) kernel, which can be represented as:

[0173] K RBF (x,x i )=exp(-γ||xx i || 2 ),γ>0 (20)

[0174] The hyperparameter γ is used to achieve an optimal balance between training the model (a prediction model based on the SVR algorithm) and inductive ability. Another hyperparameter C balances the complexity of the support vectors and the misclassification rate. These two hyperparameters are determined by the particle swarm optimization (PSO) algorithm, which uses a swarm of particles moving in the search space. The optimal particle is obtained through information interaction between the particles.

[0175] The reconstructed error sequence is used as a training dataset to establish an SVR prediction model. The prediction curve is shown in Figure 6. Figure 6 is a schematic diagram of the DWT algorithm based on an embodiment of the lithium-ion battery remaining service life prediction method of the present invention, showing the SVR prediction error results of four batteries A01-A04.

[0176] Step 402: Correct the initial prediction result according to the prediction error sequence to obtain the final prediction result of the remaining useful life.

[0177] In this embodiment, the terminal device corrects the initial prediction result based on the PF algorithm according to the prediction error sequence to obtain the final prediction result of the remaining service life.

[0178] In this embodiment, a prediction error sequence is generated based on the SVR algorithm, which realizes the effective and comprehensive utilization of training information (i.e., reconstructed error sequence), including periodic capability analysis, establishing a segmented model of capability dive, model-based prediction error, and data-driven prediction result correction.

[0179] In summary, this invention proposes a hybrid method for predicting the relative upswing (RUL) of lithium-ion batteries exhibiting capacity drop phenomena, based on an adaptive piecewise empirical degradation model and incorporating a particle filtering algorithm as the model component. The discrete wavelet transform error sequence is then decomposed into a data-driven component based on a support vector regression algorithm, successfully predicting the spread of the error sequence. On one hand, to overcome the poor real-time performance of traditional empirical degradation models, a piecewise model is proposed to ensure better performance in both early and late degradation stages. On the other hand, components of the DWT and SVR algorithms are introduced to further correct prediction errors. This invention enables comprehensive and effective utilization of training information, including cyclic capacity degradation data, the establishment of a piecewise empirical capacity degradation model with capacity drop, and model-based particle filtering to predict errors. The RUL prediction framework proposed in this invention guarantees accurate RUL prediction results, significantly optimizes the modeling and computation process, and improves prediction efficiency and accuracy.

[0180] Furthermore, the present invention also provides a device for predicting the remaining lifespan of a lithium-ion battery. (Refer to...) Figure 7 , Figure 7 This is a schematic diagram of the remaining lifespan prediction device module for lithium-ion batteries of the present invention.

[0181] The remaining lifespan prediction device for lithium-ion batteries of the present invention includes:

[0182] Modeling module H01 is used to build a segmented empirical degradation model based on the inflection point data of lithium-ion batteries;

[0183] The filtering calculation module H02 is used to obtain the initial prediction result of the remaining service life of the lithium-ion battery according to the preset particle filtering algorithm and the segmented empirical degradation model, and to obtain the original error sequence based on the initial prediction result.

[0184] The reconstructed data determination module H03 is used to determine the reconstructed error sequence corresponding to the capacity of the lithium-ion battery based on the original error sequence and a preset discrete wavelet transform algorithm.

[0185] The prediction module H04 is used to construct a prediction model of the support vector regression algorithm based on the reconstructed error sequence, and to correct the initial prediction result based on the prediction model to determine the final prediction result of the remaining useful life.

[0186] Optionally, the modeling module H01 may include:

[0187] The acquisition unit is used to acquire the capacity degradation data of the lithium-ion battery and smooth the capacity degradation data according to a preset filter.

[0188] The difference determination unit is used to identify the inflection point data of the lithium-ion battery corresponding to the difference based on a preset parameter interval range, and to establish the segmented empirical degradation model based on the inflection point data.

[0189] Optionally, the filter calculation module H02 may include:

[0190] The judgment unit is used to determine whether the current cycle data of the lithium-ion battery is less than the inflection point data;

[0191] The first obtaining unit is used to determine the double exponential model corresponding to the segmented empirical degradation model if the condition is met, and to obtain the initial prediction result based on the double exponential model and the preset particle filter algorithm.

[0192] The second obtaining unit is used to determine the autoregressive integrated moving average model corresponding to the segmented empirical degradation model if no, and to obtain the initial prediction result based on the autoregressive integrated moving average model and the preset particle filter algorithm.

[0193] Optionally, the filter calculation module H02 may further include:

[0194] The equation determination unit is used to obtain the order difference equation based on the stationarity of the lithium-ion battery, and to determine the first parameter of the autoregressive integrated moving average model based on the order difference equation.

[0195] The parameter determination unit is used to determine the second and third parameters of the autoregressive integrated moving average model based on the preset Akaike information criterion and the preset Bayesian information criterion.

[0196] Optionally, the data reconstruction determination module H03 may include:

[0197] The decomposition unit is used to obtain the decomposed signal based on the original error sequence and a preset discrete wavelet transform algorithm;

[0198] The reconstruction unit is used to reconstruct the reconstruction error sequence corresponding to the capacity of the lithium-ion battery based on the decomposed signal.

[0199] Optionally, the data reconstruction determination module H03 may further include:

[0200] The signal acquisition unit is used to remove the high-frequency component signal of the decomposed signal based on a preset discrete wavelet transform algorithm in order to obtain the low-frequency component signal of the decomposed signal.

[0201] An error reconstruction unit is used to reconstruct the reconstruction error sequence based on the low-frequency component signal.

[0202] Optionally, the prediction module H04 includes:

[0203] A generation unit is used to generate a prediction error sequence based on the prediction model;

[0204] The final prediction unit is used to correct the initial prediction result based on the prediction error sequence to obtain the final prediction result of the remaining useful life.

[0205] Each functional module of the lithium-ion battery remaining life prediction device of the present invention implements the steps of the lithium-ion battery remaining life prediction method of the present invention as described above during operation.

[0206] Furthermore, the present invention also provides a terminal device. Please refer to... Figure 8 , Figure 8 This is a schematic diagram of the terminal device involved in an embodiment of the present invention. Specifically, the terminal device in this embodiment can be a device for predicting the remaining lifespan of a locally operating lithium-ion battery.

[0207] like Figure 8 As shown, the terminal device in this embodiment of the invention may include: a processor 1001, such as a CPU; a communication bus 1002; a user interface 1003; a network interface 1004; a memory 1005; and a sensing unit 1006. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen and an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface).

[0208] The memory 1005 is disposed on the main body of the terminal device. The memory 1005 stores a program that performs corresponding operations when executed by the processor 1001. The memory 1005 is also used to store parameters for use by the terminal device. The memory 1005 can be a high-speed RAM or a stable, non-volatile memory, such as a disk drive. Optionally, the memory 1005 can also be a storage device independent of the aforementioned processor 1001.

[0209] Those skilled in the art will understand that Figure 8 The terminal device structure shown does not constitute a limitation on the terminal device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0210] like Figure 8 As shown, the memory 1005, which serves as a storage medium, may include an operating system, a network communication module, a user interface module, and a smart connection program for terminal devices.

[0211] exist Figure 8 In the terminal device shown, the processor 1001 can be used to call the smart connection program of the terminal device stored in the memory 1005 and execute the steps of the various embodiments of the lithium-ion battery remaining life prediction method of the present invention.

[0212] Furthermore, the present invention also provides a computer-readable storage medium. Please refer to... Figure 9 , Figure 9 This is a schematic diagram of the structure of a computer-readable storage medium involved in an embodiment of the present invention.

[0213] The present invention also provides a computer-readable storage medium storing a program for predicting the remaining lifespan of a lithium-ion battery. When the program for predicting the remaining lifespan of a lithium-ion battery is executed by a processor, it implements the steps of the method for predicting the remaining lifespan of a lithium-ion battery as described above.

[0214] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0215] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0216] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a computer-readable storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0217] The above are merely preferred embodiments of the present invention and do not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for predicting the remaining service life of a lithium-ion battery, characterized in that, The method for predicting the remaining lifespan of the lithium-ion battery includes: A segmented empirical degradation model was established based on the inflection point data of lithium-ion batteries; The initial prediction result of the remaining service life of the lithium-ion battery is obtained based on the preset particle filter algorithm and the segmented empirical degradation model, and the original error sequence is obtained based on the initial prediction result. The reconstructed error sequence corresponding to the capacity of the lithium-ion battery is determined based on the original error sequence and the preset discrete wavelet transform algorithm. A prediction model for a support vector regression algorithm is constructed based on the reconstructed error sequence, and the initial prediction result is corrected based on the prediction model to determine the final prediction result for the remaining useful life. The step of establishing a segmented empirical degradation model based on the inflection point data of lithium-ion batteries includes: Acquire capacity degradation data of lithium-ion batteries and smooth the capacity degradation data according to a preset filter; After the filter completes the smoothing process of the capacity degradation data, the difference in the capacity decay curve of the lithium-ion battery is determined according to a preset first-order differential equation. Based on a preset parameter interval range, identify the inflection point data of the lithium-ion battery corresponding to the difference, and determine whether the current cycle data is less than the inflection point data: If the current period data is less than the inflection point data, then the double exponential model is used as the segmented empirical degradation model; If the current period data is greater than or equal to the inflection point data, then an autoregressive integrated moving average model is used as the segmented empirical degradation model, wherein the parameters of the autoregressive integrated moving average model are optimized using the Akaike information criterion and the Bayesian information criterion.

2. The method for predicting the remaining service life of a lithium-ion battery as described in claim 1, characterized in that, The step of obtaining the initial prediction result of the remaining service life of the lithium-ion battery based on the preset particle filter algorithm and the segmented empirical degradation model includes: If the segmented empirical degradation model is the double exponential model, then the initial prediction result is obtained according to the double exponential model and the preset particle filter algorithm; If the segmented empirical degradation model is an autoregressive integrated moving average model, then the initial prediction result is obtained based on the autoregressive integrated moving average model and the preset particle filter algorithm.

3. The method for predicting the remaining service life of a lithium-ion battery as described in claim 2, characterized in that, Before the step of obtaining the initial prediction result based on the autoregressive integrated moving average model and the preset particle filter algorithm, the method includes: The order difference equation is obtained based on the stationarity of the lithium-ion battery, and the first parameter of the autoregressive integrated moving average model is determined based on the order difference equation. The second and third parameters of the autoregressive integrated moving average model are determined based on the preset Akaike information criterion and the preset Bayesian information criterion.

4. The method for predicting the remaining service life of a lithium-ion battery as described in claim 1, characterized in that, The step of determining the reconstructed error sequence corresponding to the capacity of the lithium-ion battery based on the original error sequence and a preset discrete wavelet transform algorithm includes: The decomposed signal is obtained based on the original error sequence and the preset discrete wavelet transform algorithm; The reconstruction error sequence corresponding to the capacity of the lithium-ion battery is obtained by reconstructing the decomposed signal.

5. The method for predicting the remaining service life of a lithium-ion battery as described in claim 4, characterized in that, The step of reconstructing the reconstruction error sequence corresponding to the capacity of the lithium-ion battery based on the decomposed signal includes: The high-frequency component signal of the decomposed signal is removed based on the preset discrete wavelet transform algorithm to obtain the low-frequency component signal of the decomposed signal. The reconstruction error sequence is obtained by reconstructing the low-frequency component signal.

6. The method for predicting the remaining service life of a lithium-ion battery as described in claim 1, characterized in that, The step of revising the initial prediction result based on the prediction model to determine the final prediction result of the remaining useful life includes: Generate a prediction error sequence based on the prediction model; The initial prediction result is corrected based on the prediction error sequence to obtain the final prediction result of the remaining useful life.

7. A device for predicting the remaining lifespan of a lithium-ion battery, characterized in that, The remaining lifespan prediction device for the lithium-ion battery includes: The modeling module is used to build segmented empirical degradation models based on the inflection point data of lithium-ion batteries. The filtering calculation module is used to obtain the initial prediction result of the remaining service life of the lithium-ion battery according to the preset particle filtering algorithm and the segmented empirical degradation model, and to obtain the original error sequence based on the initial prediction result. The reconstructed data determination module is used to determine the reconstructed error sequence corresponding to the capacity of the lithium-ion battery based on the original error sequence and a preset discrete wavelet transform algorithm. The prediction module is used to construct a prediction model of the support vector regression algorithm based on the reconstructed error sequence, and to correct the initial prediction result based on the prediction model to determine the final prediction result of the remaining useful life; The modeling module is further configured to: acquire capacity degradation data of lithium-ion batteries; smooth the capacity degradation data according to a preset filter; after determining that the filter has completed the smoothing of the capacity degradation data, determine the difference in the capacity decay curves of the lithium-ion batteries according to a preset first-order differential equation; identify the inflection point data of the lithium-ion batteries corresponding to the difference based on a preset parameter interval, and determine whether the current period data is less than the inflection point data: if the current period data is less than the inflection point data, a double exponential model is used as the piecewise empirical degradation model; if the current period data is greater than or equal to the inflection point data, an autoregressive integrated moving average model is used as the piecewise empirical degradation model, wherein the parameters of the autoregressive integrated moving average model are optimized using the Akaike information criterion and the Bayesian information criterion.

8. A terminal device, characterized in that, The terminal device includes a memory, a processor, and a lithium-ion battery remaining life prediction program stored in the memory and executable on the processor. When the processor executes the lithium-ion battery remaining life prediction program, it implements the steps of the lithium-ion battery remaining life prediction method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program for predicting the remaining lifespan of a lithium-ion battery, which, when executed by a processor, implements the steps of the method for predicting the remaining lifespan of a lithium-ion battery as described in any one of claims 1 to 6.

Images