CN109143108B

CN109143108B - Lithium ion battery SOH estimation method based on electrochemical impedance spectroscopy

Info

Publication number: CN109143108B
Application number: CN201810826685.3A
Authority: CN
Inventors: 刘征宇; 杨昆; 姚利阳; 武银行
Original assignee: Hefei University of Technology
Current assignee: Hefei University of Technology
Priority date: 2018-07-25
Filing date: 2018-07-25
Publication date: 2021-08-20
Anticipated expiration: 2038-07-25
Also published as: CN109143108A

Abstract

The invention discloses a lithium ion battery SOH estimation method based on electrochemical impedance spectroscopy, which comprises the following steps: measuring an electrochemical impedance spectrum of the lithium ion battery; establishing an equivalent circuit model; measuring electrochemical impedance spectra at different SOCs and different cycle times; performing parameter identification by using the impedance spectrum; the neural network is trained for evaluation of SOH. The method utilizes the electrochemical impedance spectrum to identify the parameters, and can complete the estimation of the SOH of the battery under the condition of not damaging the internal structure of the battery.

Description

Lithium ion battery SOH estimation method based on electrochemical impedance spectroscopy

Technical Field

The invention relates to the technical field of health state estimation of lithium ion batteries, in particular to a lithium ion battery SOH estimation method based on electrochemical impedance spectroscopy.

Background

Lithium ion batteries are gaining increasing attention in automotive, stationary and hybrid energy production systems. High energy density, high power density and cycle number are the main reasons for their development. However, as the number of charge and discharge cycles of the battery increases, the battery gradually ages, the internal resistance gradually increases and the capacity gradually decreases, and the state of health (soh) of the battery represents the degree of aging of the battery. In EVs (electric vehicles), the use of lithium ion batteries requires special attention to the evaluation of the battery state, in particular SOH, which indicates a battery failure. However, SOH is a difficult parameter to determine. In fact, degradation of the battery occurs at the interface between the electrolyte and the electrodes due to the growth of an SEI (solid electrolyte interface) film on the anode side of the battery, degradation of the battery resulting in capacity fade due to increased SEI thickness leading to increased impedance.

Conventional estimation methods for a battery are a definition method, a capacity fade method, a chemical analysis method, a partial discharge method, and the like. The definition method needs repeated charge and discharge experiments on the battery and is difficult to realize in practical application; the capacity attenuation method is easy to be interfered by the outside world, and the measurement precision is too low; chemical analysis methods require disassembly of the cell for analysis, which can render the cell unusable; the partial discharge method has long test time and high test difficulty. In view of the foregoing, it is desirable to provide a measurement method with high estimation accuracy and fast estimation speed without damaging the battery result.

Disclosure of Invention

The invention aims to make up for the defects of the prior art and provides a lithium ion battery SOH estimation method based on electrochemical impedance spectroscopy.

The invention is realized by the following technical scheme:

a lithium ion battery SOH estimation method based on electrochemical impedance spectroscopy comprises the following steps:

(1) measuring the electrochemical impedance spectrum of the lithium ion battery at the temperature of 25 ℃ and the SOC of 60%;

(2) establishing an equivalent circuit model based on the electrochemical impedance spectrum measured in the step (1);

(3) measuring electrochemical impedance spectrums of the lithium ion battery under different charge-discharge cycle times;

(4) measuring electrochemical impedance spectrums of the lithium ion battery under different SOC conditions;

(5) performing parameter identification on the equivalent circuit model according to the electrochemical impedance spectrum measured in the step (3) and the step (4);

(6) and (5) training the three layers of neural networks according to the parameters identified in the step (5) to obtain the SOH of the lithium ion battery.

The specific method for identifying the parameters in the step (5) comprises the following steps:

1) r can be achieved according to the first intersection point of the electrochemical impedance spectrum and the x axis measured in the step (3) and the step (4)_ohmIdentifying the parameters;

2) r in equivalent circuit model_seiAnd C_seiIn parallelThen with R_ohmThe relationship between Re (Z) and im (Z) can be obtained by simplifying the series circuit:

where Re (Z) represents the real part of the impedance, im (Z) represents the imaginary part of the impedance, R_ohmIs the ohmic resistance, C, in the equivalent circuit model_seiIs a capacitance value in an equivalent circuit model, R_seiIs the resistance representing the SEI film in the equivalent circuit model;

3) the expression can be obtained from the formula (1) that one circle center is

Radius of

According to the circle center or radius of the first semicircle of the electrochemical impedance spectrum measured in the step (3) and the step (4), R is obtained_seiOf (d) completes the pair R_seiThe parameter identification.

In the step (6), the SOH of the lithium ion battery is obtained by training the three-layer neural network according to the parameters identified in the step (5), and the specific formula is as follows:

in the formula, R_EOLRepresenting the internal resistance of the battery at the end of the service life of the battery, R is the internal resistance of the battery in the current state of the battery, R_NEWThe internal resistance of the battery when leaving factory; training a three-layer neural network by using the obtained data, wherein the input is SOC and R_ohmAnd R_seiThe output is the SOH of the battery.

The invention has the advantages that: the method disclosed by the invention uses the electrochemical impedance spectrum to identify the parameters, can estimate the SOH of the battery under the condition of not damaging the internal structure of the battery, and has the advantages of higher precision, lower requirement on hardware and higher practical value.

Drawings

Fig. 1 is an electrochemical impedance spectrum of a lithium ion battery at a temperature T of 25 ℃ and an SOC of 60%.

Fig. 2 is an equivalent circuit model of a lithium ion battery.

Fig. 3 is an electrochemical impedance spectrum of a lithium ion battery at different SOCs (temperature T25 ℃, cycle of battery 1).

Fig. 4 shows electrochemical impedance spectra of lithium ion batteries at different cycle numbers (temperature T25 ℃, SOH 60%).

FIG. 5 is the ohmic internal resistance R of the cell at different cycle numbers_ohmRelation to SOC.

FIG. 6 shows the R of the cell at different cycle numbers_seiRelation to SOC.

Detailed Description

(1) the electrochemical impedance spectrum of the lithium ion battery with the measurement temperature of 25 ℃ and the SOC of 60% is shown in figure 1;

(2) the electrochemical impedance spectrum was analyzed and a corresponding equivalent circuit model was established as shown in fig. 2. The proposed model can reproduce the impedance spectrum of the battery in different states of charge and different states of health, and the impedance of each circuit element can be expressed as:

ohmic internal resistance: r_ohm＝R_Ω

b:R_seiAnd C_seiThe parallel circuit of (1):

c:R_ct、Z_win series with C_dParallel circuits:

as the battery ages, some impedances will change necessarily, and for some impedances with obvious changes, we can benefitPerforming parameter identification by using an electrochemical impedance spectrum and determining the SOH;

(3) keeping the experiment temperature at 25 ℃ and the battery charge-discharge cycle number unchanged for one time, and measuring the electrochemical impedance spectrum of the battery under different SOC conditions, as shown in FIG. 3;

(4) keeping the experiment temperature at 25 ℃ and the battery SOC at 60%, and measuring the electrochemical impedance spectrum of the battery under different charge-discharge cycle times, as shown in FIG. 4;

(5) analysis of the electrochemical impedance spectrum of fig. 3 reveals that the internal resistance of the battery increases and then decreases at different SOC levels as the SOC increases. Analysis of FIG. 4 shows that as the number of battery cycles increases, the impedance of the battery increases significantly, and the R of the battery increases_ohm、R_seiIs more pronounced, especially for R_seiThe variation is very obvious, so we choose to use R_ohm、R_seiAs a parameter for estimating the SOH of the lithium ion battery;

(6) the specific method for identifying the parameters of the selected parameters comprises the following steps:

1) observing the first intersection point of the electrochemical impedance spectrum and the x-axis in FIGS. 3 and 4, the intersection point corresponds to R in the equivalent circuit model_ohmThe values of these intersections are extracted and are shown in fig. 5;

2) r in equivalent circuit model_seiAnd C_seiAre connected in parallel with R_ohmThe relationship between Re (Z) and im (Z) can be obtained by simplifying the series circuit:

in the formula, R_e(Z) represents the real part of the impedance, im (Z) represents the imaginary part of the impedance, R_ohmRepresenting ohmic resistance, R, in an equivalent circuit model_seiRepresenting the SEI resistance in the equivalent circuit model;

Radius of

Observing the first semicircle of the electrochemical impedance spectroscopy shown in fig. 3 and 4, finding the value of the center or radius of the circle, and completing the comparison of R_seiIdentification of the parameter R_seiThe results of (a) are shown in FIG. 6;

(7) obtaining a series of impedance values of the battery under different SOC and aging states, and representing the SOH of the battery by using the battery impedance obtained by parameter identification, wherein the specific formula is as follows:

in the formula, R_EOLRepresenting the internal resistance of the battery at the end of the service life of the battery, R is the internal resistance of the battery in the current state of the battery, R_NEWThe internal resistance of the battery when the battery leaves the factory.

(8) Training a three-layer neural network by using the obtained data, wherein the input is SOC and R_ohmAnd R_seiThe SOH of the battery can be estimated after a sufficient training amount, and the estimation accuracy is higher when the training amount is larger.

Claims

1. A lithium ion battery SOH estimation method based on electrochemical impedance spectroscopy is characterized in that: comprises the following steps:

(5) performing parameter identification on the equivalent circuit model in the step (2) according to the electrochemical impedance spectrums measured in the steps (3) and (4);

(6) training the three-layer neural network according to the parameters identified in the step (5) to obtain the SOH of the lithium ion battery;

where Re (Z) represents the real part of the impedance, im (Z) represents the imaginary part of the impedance, R_ohmIs the ohmic resistance in the equivalent circuit model, c_seiIs a capacitance value in an equivalent circuit model, R_seiIs the resistance representing the SEI film in the equivalent circuit model;

Radius of

According to the circle center or radius of the first semicircle of the electrochemical impedance spectrum measured in the step (3) and the step (4), R is obtained_seiOf (d) completes the pair R_seiIdentifying the parameters;

in the formula, R_EOLRepresenting the internal resistance of the battery at the end of its life, R beingInternal resistance of the battery, R, at the current state of the battery_NEWThe internal resistance of the battery when leaving factory; training a three-layer neural network by using the obtained data, wherein the input is SOC and R_ohmAnd R_seiThe output is the SOH of the battery.