Lithium ion battery capacity evolution prediction method based on damage-impedance relationship model

Abstract

The application discloses a lithium ion battery capacity evolution prediction method based on a damage-impedance relation model, comprising the following steps: obtaining evolution data of electrode capacity and average charge transfer impedance; constructing a damage-impedance relation model describing the growth rate evolution of average charge transfer impedance before and after discharging with the increase of charge and discharge cycles; determining the damage-impedance relation; predicting the evolution law of average charge transfer impedance; constructing a phenomenological relation model between average charge transfer impedance and capacity; determining the phenomenological relation between average charge transfer impedance and capacity; and predicting the capacity evolution law in the cycle process. The model is established from the mechanical mechanism of electrode material degradation, and capacity evolution prediction results superior to traditional empirical models can be obtained, thereby providing a new scheme for life prediction. The model is simple and clear, easy to understand, and convenient and simple to predict, and can be popularized to capacity and life prediction of nano-ion batteries and potassium ion batteries.

Description

Technical Field

[0001] This invention belongs to the field of lithium-ion battery new energy, specifically involving a method for predicting the capacity evolution of lithium-ion batteries based on a damage-impedance relationship model. Background Technology

[0002] The mechanochemical behavior of electrode materials during charge and discharge processes affects the performance of lithium-ion batteries, primarily including capacity, stability, and cycle life. Predicting the capacity and lifespan of lithium-ion batteries is crucial for ensuring the safe and efficient operation of power systems and the reuse of batteries.

[0003] Currently, traditional empirical models are frequently used in practical prediction processes. These models employ exponential or multinomial models to fit parameters to a sample of circulating capacity to obtain the prediction results. These predictions have two limitations:

[0004] First, it does not consider the internal mechanism of capacity decay, especially the mechanical damage caused by electrochemistry.

[0005] Secondly, the prediction results depend on the selected mathematical model.

[0006] Therefore, there is an urgent need for a method to predict the capacity evolution of lithium-ion batteries that takes into account the internal damage of electrode materials. Summary of the Invention

[0007] This invention addresses the problems existing in the prior art and aims to provide a method for predicting the capacity evolution of lithium-ion batteries based on a damage-impedance relationship model.

[0008] The technical solution of this invention is: a method for predicting the capacity evolution of lithium-ion batteries based on a damage-impedance relationship model, comprising the following steps:

[0009] i. Obtain evolution data on electrode capacity and average charge transfer impedance;

[0010] ii. Construct a damage-impedance relationship model to describe the evolution of the average charge transfer impedance before and after discharge with the growth rate of charge-discharge cycles;

[0011] iii. Determine the damage-impedance relationship;

[0012] iv. Predict the evolution of average charge transfer impedance;

[0013] v. Construct a phenomenological model of the relationship between average charge transfer impedance and capacitance;

[0014] vi. Determine the phenomenological relationship between average charge transfer impedance and capacitance;

[0015] vii. Predict the capacity evolution law of the cyclic process.

[0016] Furthermore, step i obtains the evolution data of electrode capacity and average charge transfer impedance, the specific process of which is as follows:

[0017] First, impedance and capacity evolution data of lithium-ion batteries during cycling were obtained using an electrochemical workstation and Blue Battery charging and discharging equipment.

[0018] Then, the average charge transfer impedance evolution data is obtained based on the impedance evolution data.

[0019] Furthermore, step ii involves constructing a damage-impedance relationship model describing the evolution of the average charge transfer impedance before and after discharge with the rate of increase of charge-discharge cycles. The specific process is as follows:

[0020] First, we obtain the expression of the fatigue propagation theory of materials under cyclic loading in mechanics;

[0021] Then, a general expression describing the rate of increase of the average charge transfer impedance before and after discharge with charge-discharge cycles is constructed;

[0022] Finally, the above general expression is denoted as the damage-impedance relationship model.

[0023] Furthermore, step iv predicts the evolution of the average charge transfer impedance, and the specific process is as follows:

[0024] First, the average charge transfer impedance sample is obtained based on step i;

[0025] Then, the average charge transfer impedance samples are divided to obtain the training average charge transfer impedance sample set;

[0026] Next, the training average charge transfer impedance sample set is used to train the damage-impedance relationship model to determine the model feature parameters and the damage-impedance relationship.

[0027] Finally, the damage-impedance relationship is used for prediction to obtain prediction data A based on the training average charge transfer impedance sample set, and the prediction data A is saved.

[0028] Furthermore, step v constructs a phenomenological model of the relationship between average charge transfer impedance and capacitance, as detailed below:

[0029] An exponential model is used to construct a phenomenological relationship between average charge transfer impedance and capacity, which is denoted as the phenomenological relationship model between average charge transfer impedance and capacity.

[0030] Furthermore, the method for obtaining the average charge transfer impedance is as follows:

[0031] First, the impedance is selected as the charge transfer impedance;

[0032] Then, the impedance evolution data is taken as the average of the values ​​before and after discharge;

[0033] Finally, the above average value is the average charge transfer impedance.

[0034] Furthermore, step vi determines the phenomenological relationship between the average charge transfer impedance and the capacitance, as follows:

[0035] First, the capacity samples are divided into a training capacity sample set and an analysis capacity set;

[0036] Then, the training average charge transfer impedance sample set is used;

[0037] Next, the training average charge transfer impedance sample set and capacity sample set are substituted into the phenomenological relation model for training;

[0038] Finally, an exponential model relationship between average charge transfer impedance and capacitance was determined.

[0039] Furthermore, step vii predicts the capacity evolution law of the cyclic process, and the specific process is as follows:

[0040] First, the predicted data A is substituted into the exponential model relationship between the average charge transfer impedance and the capacity after training to obtain the predicted data for the capacity sample set.

[0041] Then, error analysis is performed between the predicted data and the measured data of the analysis capacity sample set.

[0042] The beneficial effects of this invention are as follows:

[0043] This invention extends the theory of mechanical fatigue damage to the prediction of battery capacity in the field of electrochemistry. Based on the inherent causal relationship between "fatigue damage → increased impedance → capacity decay" of electrode materials under charge-discharge cycle conditions, a damage-impedance relationship model is established using impedance, which reflects the damage of electrode materials and can be directly measured, as the "key". This model can predict the evolution of electrode capacity and is applicable to all electrodes.

[0044] The model established by this invention based on the mechanical mechanism of electrode material degradation can obtain capacity evolution prediction results that are superior to those of traditional empirical models, providing a new solution for lifetime prediction.

[0045] The model of this invention is simple and clear, easy to understand, and convenient and simple to predict. It can be extended to the prediction of capacity and lifespan of nano-ion batteries and potassium-ion batteries. Attached Figure Description

[0046] Figure 1 This invention relates to the electrochemical impedance spectroscopy and its equivalent circuit.

[0047] Figure 2This is a graph showing the evolution of the specific capacity of the electrode at different rates with the number of cycles in this invention.

[0048] Figure 3 This is the evolution of charge transfer impedance before and after discharge under different cycle numbers in this invention;

[0049] Figure 4 These are the prediction results and their errors obtained from two different prediction methods in this invention.

[0050] Among them, (a) and (b) are the prediction results at high magnification;

[0051] (c) and (d) Prediction results at low magnification. Detailed Implementation

[0052] The present invention will now be described in detail with reference to the accompanying drawings and embodiments:

[0053] like Figures 1 to 4 As shown, a method for predicting the capacity evolution of lithium-ion batteries based on a damage-impedance relationship model includes the following steps:

[0054] i. Obtain evolution data on electrode capacity and average charge transfer impedance;

[0055] ii. Construct a damage-impedance relationship model to describe the evolution of the average charge transfer impedance before and after discharge with the growth rate of charge-discharge cycles;

[0056] iii. Determine the damage-impedance relationship;

[0057] iv. Predict the evolution of average charge transfer impedance;

[0058] v. Construct a phenomenological model of the relationship between average charge transfer impedance and capacitance;

[0059] vi. Determine the phenomenological relationship between average charge transfer impedance and capacitance;

[0060] vii. Predict the capacity evolution law of the cyclic process.

[0061] Step i: Obtain the evolution data of electrode capacity and average charge transfer impedance. The specific process is as follows:

[0062] First, impedance and capacity evolution data of lithium-ion batteries during cycling were obtained using an electrochemical workstation and Blue Battery charging and discharging equipment.

[0063] Then, the average charge transfer impedance evolution data is obtained based on the impedance evolution data.

[0064] Step ii: Construct a damage-impedance relationship model describing the evolution of the average charge transfer impedance before and after discharge with the rate of increase of charge-discharge cycles. The specific process is as follows:

[0065] First, we obtain the expression of the fatigue propagation theory of materials under cyclic loading in mechanics;

[0066] Then, a general expression describing the rate of increase of the average charge transfer impedance before and after discharge with charge-discharge cycles is constructed;

[0067] Finally, the above general expression is denoted as the damage-impedance relationship model.

[0068] Step iv predicts the evolution of the average charge transfer impedance, and the specific process is as follows:

[0069] First, the average charge transfer impedance sample is obtained based on step i;

[0070] Then, the average charge transfer impedance samples are divided to obtain the training average charge transfer impedance sample set;

[0071] Next, the training average charge transfer impedance sample set is used to train the damage-impedance relationship model to determine the model feature parameters and the damage-impedance relationship.

[0072] Finally, the damage-impedance relationship is used for prediction to obtain prediction data A based on the training average charge transfer impedance sample set, and the prediction data A is saved.

[0073] Step v: Construct a phenomenological model of the relationship between average charge transfer impedance and capacitance. The specific process is as follows:

[0074] An exponential model is used to construct a phenomenological relationship between average charge transfer impedance and capacity, which is denoted as the phenomenological relationship model between average charge transfer impedance and capacity.

[0075] The method for obtaining the average charge transfer impedance is as follows:

[0076] First, the impedance is selected as the charge transfer impedance;

[0077] Then, the impedance evolution data is taken as the average of the values ​​before and after discharge;

[0078] Finally, the above average value is the average charge transfer impedance.

[0079] Step ⅵ determines the phenomenological relationship between average charge transfer impedance and capacitance. The specific process is as follows:

[0080] First, the capacity samples are divided into a training capacity sample set and an analysis capacity set;

[0081] Then, the training average charge transfer impedance sample set is used;

[0082] Next, the training average charge transfer impedance sample set and capacity sample set are substituted into the phenomenological relation model for training;

[0083] Finally, an exponential model relationship between average charge transfer impedance and capacitance was determined.

[0084] Step vii: Predict the capacity evolution of the cyclic process. The specific process is as follows:

[0085] First, the predicted data A is substituted into the exponential model relationship between the average charge transfer impedance and the capacity after training to obtain the predicted data for the capacity sample set.

[0086] Then, error analysis is performed between the predicted data and the measured data of the analysis capacity sample set.

[0087] Specifically, the damage-impedance relationship model draws on the fatigue propagation theory of materials under cyclic loading (Paris theory) in mechanics, and constructs a description of the average charge transfer impedance before and after discharge using Paris theory. An evolution model of the rate of increase of charge-discharge cycles.

[0088] The damage-impedance relationship model is as follows:

[0089]

[0090] Where N is the number of iterations, and C and m are the model feature parameters.

[0091] Specifically, an exponential model is used to construct the average charge transfer impedance. Phenomenological relation with capacity Q:

[0092]

[0093] Where a and b are model feature parameters.

[0094] More specifically, C, m, a, and b are obtained using a particle filtering algorithm.

[0095] Specifically, the average charge transfer impedance samples are divided into a training average charge transfer impedance sample set and an analysis average charge transfer impedance sample set. The top n% of the average charge transfer impedance samples are taken as the training average charge transfer impedance sample set, and the remaining part is the analysis average charge transfer impedance sample set.

[0096] Similarly, the capacity samples are divided into a training capacity sample set and an analysis capacity sample set. The first n% of the capacity samples are taken as the training capacity sample set, and the remaining part is the analysis capacity sample set.

[0097] In one embodiment of the present invention, n is taken as 50 in the following examples.

[0098] Specifically, the process of establishing the damage-impedance relationship model in this invention is as follows:

[0099] During the cyclic charging and discharging process of lithium-ion batteries, the electrode material typically undergoes reciprocating deformation, which is mechanically equivalent to applying a generalized cyclic load to the electrode material. Similar to fatigue damage under cyclic loading, electrode materials also experience fatigue damage under generalized cyclic loading. This damage increases the impedance of the electrode material, and the increased impedance leads to capacity decay. Therefore, there is an inherent causal relationship between "fatigue damage → increased impedance → capacity decay" in the electrochemical process.

[0100] In fatigue damage theory, crack length is usually used as the key variable, and the service life is predicted by constructing a Paris model that describes the crack propagation rate with cycles.

[0101] This invention uses average charge transfer impedance as the key variable and predicts the evolution of electrode capacity by constructing an evolution model that describes the growth rate of average charge transfer impedance with charge-discharge cycles.

[0102] Specifically, the training method in the damage-impedance relationship model is implemented using the particle filter algorithm.

[0103] Specifically, the training method in the phenomenological relation model is implemented using the particle filter algorithm.

[0104] Another embodiment

[0105] First, obtain experimental data samples of impedance and capacitance.

[0106] As a case study, graphene-coated silicon electrodes were selected for the experiment. The required electrodes were fabricated, and CR2032 lithium-ion batteries were assembled. A Blue Battery charge-discharge system was used to perform charge-discharge cycles at two rates (C / 2 and C / 4). Every four cycles, an electrochemical impedance spectroscopy test was performed using an electrochemical workstation. Each impedance measurement was conducted before and after discharge. Figure 1 The equivalent circuit shown is used to fit the electrochemical impedance spectroscopy to obtain charge transfer impedance data before and after discharge. After removing the initial activation and late failure stages corresponding to the charge transfer impedance, only the stable decay stage of interest in the actual prediction is retained. The resulting evolution data of capacity and charge transfer impedance are shown below. Figure 2 and 3 As shown.

[0107] right Figure 3 The average charge transfer impedance at different cycle numbers is obtained by averaging the impedance data in (a) and (b). Data sample.

[0108] Then, the average charge transfer impedance is predicted.

[0109] Pick The first 50% of the evolution data samples (corresponding to the training average charge transfer impedance sample set) are substituted into formula (1) for training. A particle filter algorithm is developed using Matlab to obtain the parameters C and m in the model. The initial values ​​of C and m can be referenced to 0.001 and 1, and selected according to the actual situation. After determining the parameters C and m, the damage-impedance relationship can be determined, and thus the damage-impedance relationship can be predicted. The evolution pattern in the latter 50% of the cycle, i.e., the predicted data A.

[0110] Next, the phenomenological relationship between average charge transfer impedance and capacitance was determined.

[0111] Will The first 50% of the data samples (corresponding to the training average charge transfer impedance sample set) and the first 50% of the Q data samples (corresponding to the training capacity sample set) are substituted into formula (2) for training. Similarly, a particle filter algorithm is developed using Matlab to obtain the parameters a and b in the model, where the initial values ​​of a and b are obtained using least squares fitting. After determining the parameters a and b through the particle filter algorithm, the model is obtained. The phenomenological relationship with Q.

[0112] Finally, a comparison of the prediction accuracy of capacity evolution patterns was conducted.

[0113] Will The evolutionary pattern in the latter 50% of the cycle, i.e., the regression of predicted data A. The phenomenological relationship between Q and formula (2) can predict the evolution of capacity in the last 50% of the cycle.

[0114] To evaluate the predictive capability of the damage-impedance relationship model, the error between the predicted capacity and the actual capacity at two different capacities was calculated, and the results were compared with those of a traditional empirical model. The results are as follows: Figure 4 As shown, the traditional empirical model uses an exponential model to directly construct the phenomenological relationship between capacity Q and the number of cycles N, i.e.

[0115] Q = a²exp(b²N) (3)

[0116] Similarly, using the first 50% of the data samples and combining the particle filter algorithm for experimental prediction, the initial values ​​of the model parameters a2 and b2 were also obtained by least squares fitting.

[0117] from Figure 4The comparison shows that the damage-impedance relationship model of the present invention provides prediction results that are closer to the measured results at both magnification rates. A comparison of the prediction errors of the two models reveals that the prediction error of the damage-impedance relationship model is below 3% at all magnification rates, while the maximum prediction error of the traditional empirical model exceeds 6%. Therefore, the damage-impedance relationship model of the present invention demonstrates superior predictive ability.

[0118] The prediction process of this invention takes into account the damage mechanism inside the electrode material, and the model is simple and clear, achieving higher accuracy prediction. The operation process is simple and can be extended to predict the actual remaining service life of lithium-ion batteries.

Claims

1. A method for predicting the capacity evolution of lithium-ion batteries based on a damage-impedance relationship model, characterized in that: Includes the following steps: (i) Obtain evolution data of electrode capacity and average charge transfer impedance; (ii) Construct a damage-impedance relationship model to describe the evolution of the average charge transfer impedance before and after discharge with the growth rate of charge-discharge cycles; (iii) Determine the damage-impedance relationship; (iv) Predict the evolution of average charge transfer impedance; (v) Construct a phenomenological model of the relationship between average charge transfer impedance and capacitance; (vi) Determine the phenomenological relationship between average charge transfer impedance and capacitance; (vii) Predict the capacity evolution law of the cyclic process; Step vii: Predict the capacity evolution of the cyclic process. The specific process is as follows: First, the predicted data A is substituted into the exponential model relationship between the average charge transfer impedance and the capacity after training to obtain the predicted data for the capacity sample set. Then, error analysis is performed between the predicted data and the measured data of the analysis capacity sample set; The average charge transfer impedance before and after discharge is constructed using Paris theory. An evolution model of the rate of increase with the charge-discharge cycle; The damage-impedance relationship model is as follows: (1)； in, N The number of loops. C and m These are the model feature parameters; The average charge transfer impedance is constructed using an exponential model. With capacity Q The phenomenological relationship between them: (2)； in, a and b These are the model feature parameters; The C, m, a and b Obtained using a particle filter algorithm; The average charge transfer impedance samples are divided into a training average charge transfer impedance sample set and an analysis average charge transfer impedance sample set. The first n% of the average charge transfer impedance samples are taken as the training average charge transfer impedance sample set, and the remaining part is the analysis average charge transfer impedance sample set.

2. The lithium-ion battery capacity evolution prediction method based on a damage-impedance relationship model according to claim 1, characterized in that: Step (i) involves obtaining the evolution data of electrode capacity and average charge transfer impedance. The specific process is as follows: First, impedance and capacity evolution data of lithium-ion batteries during cycling were obtained using an electrochemical workstation and Blue Battery charging and discharging equipment. Then, the average charge transfer impedance evolution data is obtained based on the impedance evolution data.

3. The lithium-ion battery capacity evolution prediction method based on a damage-impedance relationship model according to claim 1, characterized in that: Step (ii) constructs a damage-impedance relationship model describing the evolution of the average charge transfer impedance before and after discharge with the rate of increase of charge-discharge cycles. The specific process is as follows: First, we obtain the expression of the fatigue propagation theory of materials under cyclic loading in mechanics; Then, a general expression describing the rate of increase of the average charge transfer impedance before and after discharge with charge-discharge cycles is constructed; Finally, the above general expression is denoted as the damage-impedance relationship model.

4. The lithium-ion battery capacity evolution prediction method based on a damage-impedance relationship model according to claim 1, characterized in that: Step (iv) predicts the evolution of the average charge transfer impedance, and the specific process is as follows: First, the average charge transfer impedance sample is obtained based on step (i); Then, the average charge transfer impedance samples are divided to obtain the training average charge transfer impedance sample set; Next, the training average charge transfer impedance sample set is used to train the damage-impedance relationship model to determine the model feature parameters and the damage-impedance relationship. Finally, the damage-impedance relationship is used for prediction to obtain prediction data A based on the training average charge transfer impedance sample set, and the prediction data A is saved.

5. The lithium-ion battery capacity evolution prediction method based on the damage-impedance relationship model according to claim 1, characterized in that: Step (v) constructs a phenomenological model of the relationship between average charge transfer impedance and capacitance. The specific process is as follows: An exponential model is used to construct a phenomenological relationship between average charge transfer impedance and capacity, which is denoted as the phenomenological relationship model between average charge transfer impedance and capacity.

6. The lithium-ion battery capacity evolution prediction method based on the damage-impedance relationship model according to claim 2, characterized in that: The method for obtaining the average charge transfer impedance is as follows: First, the impedance is selected as the charge transfer impedance; Then, the impedance evolution data is taken as the average of the values ​​before and after discharge; Finally, the above average value is the average charge transfer impedance.

7. The method for predicting the capacity evolution of lithium-ion batteries based on a damage-impedance relationship model according to claim 1, characterized in that: Step (vi) determines the phenomenological relationship between average charge transfer impedance and capacitance. The specific process is as follows: First, the capacity samples are divided into a training capacity sample set and an analysis capacity set; Then, the training average charge transfer impedance sample set is used; Next, the training average charge transfer impedance sample set and capacity sample set are substituted into the phenomenological relation model for training; Finally, an exponential model relationship between average charge transfer impedance and capacitance was determined.

8. The method for predicting the capacity evolution of lithium-ion batteries based on a damage-impedance relationship model according to claim 1, characterized in that: Step (vii) predicts the capacity evolution law of the cyclic process, and the specific process is as follows: First, the predicted data A is substituted into the exponential model relationship between the average charge transfer impedance and the capacity after training to obtain the predicted data for the capacity sample set. Then, error analysis is performed between the predicted data and the measured data of the analysis capacity sample set.

Images