CN112986831A - Lithium ion battery life prediction method based on correlation coefficient particle filtering - Google Patents

Lithium ion battery life prediction method based on correlation coefficient particle filtering Download PDF

Info

Publication number
CN112986831A
CN112986831A CN202110479300.2A CN202110479300A CN112986831A CN 112986831 A CN112986831 A CN 112986831A CN 202110479300 A CN202110479300 A CN 202110479300A CN 112986831 A CN112986831 A CN 112986831A
Authority
CN
China
Prior art keywords
capacity
lithium ion
ion battery
battery
correlation coefficient
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202110479300.2A
Other languages
Chinese (zh)
Inventor
周勇
高迪驹
张松勇
王硕丰
顾伟
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shanghai Maritime University
Original Assignee
Shanghai Maritime University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shanghai Maritime University filed Critical Shanghai Maritime University
Priority to CN202110479300.2A priority Critical patent/CN112986831A/en
Publication of CN112986831A publication Critical patent/CN112986831A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
    • G01R31/367—Software therefor, e.g. for battery testing using modelling or look-up tables
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
    • G01R31/392—Determining battery ageing or deterioration, e.g. state of health

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Secondary Cells (AREA)
  • Tests Of Electric Status Of Batteries (AREA)

Abstract

The invention discloses a lithium ion battery service life prediction method based on correlation coefficient particle filtering, which comprises the following steps: s1, setting a prediction starting point and a battery life threshold; s2, obtaining data of lithium ion battery to be predicted and capacity estimation value thereof
Figure DDA0003048574440000011
S3, establishing a state space of a lithium ion battery capacity index decay model, and performing parameter estimation; s4, setting the number N of sampling particles and the process noise variance sigmawAnd measure the noise variance σvAnd a resampling threshold
Figure DDA0003048574440000012
S5, estimating the capacity of the lithium ion battery to be predicted
Figure DDA0003048574440000013
Substituting the measured value into a lithium ion battery capacity exponential decay model, continuously updating the weight of particles based on a correlation coefficient particle filter algorithm, and obtaining state posterior estimation when the starting point is predicted; and S6, iterating the state posterior estimation to the life threshold value based on the capacity attenuation model, and obtaining the residual life prediction result. The advantages are that: according to the method, a correlation coefficient particle filter algorithm is introduced into the prediction of the RUL of the battery, so that the prediction precision of the RUL of the lithium battery can be effectively improved.

Description

Lithium ion battery life prediction method based on correlation coefficient particle filtering
Technical Field
The invention relates to the technical field of health prediction and diagnosis of a battery management system, in particular to a lithium ion battery service life prediction method based on correlation coefficient particle filtering.
Background
Lithium ion batteries have characteristics of high energy density, high open circuit voltage, wide temperature range, rapid charge and discharge, and high output power, and have been widely used in almost all industrial fields with energy supply. In many fields, lithium ion batteries have gradually become their key devices. However, unlike other rechargeable batteries, the performance of lithium ion batteries slowly degrades during use, which is manifested by a decrease in the capacity and an increase in the internal resistance of the lithium ion battery. The continuous use of the battery after the battery reaches the service life threshold value may bring a series of safety problems, and the accurate prediction of the remaining service life (RUL) of the battery is very important for ensuring the reliable and safe operation of the lithium ion battery, but the problem that the remaining capacity cannot be measured in real time exists in the RUL prediction of the lithium ion battery at present.
Disclosure of Invention
The invention aims to provide a lithium ion battery life prediction method based on correlation coefficient particle filtering, which aims at the problem that the capacity is difficult to directly measure in the RUL prediction of a lithium ion battery, and can estimate the residual capacity of the battery in real time by extracting measurable indirect parameters in a discharge period when the battery runs; in addition, the method introduces a relative particle filter algorithm into the battery RUL prediction aiming at the problems of particle degradation and sample shortage in the standard particle filter algorithm, and can effectively improve the RUL prediction precision of the lithium battery.
In order to achieve the purpose, the invention is realized by the following technical scheme:
a lithium ion battery life prediction method based on correlation coefficient particle filtering is characterized by comprising the following steps:
s1, setting a prediction starting point and a battery life threshold;
s2, obtaining the data of the lithium ion battery to be predicted, and obtaining the capacity estimation value of the lithium ion battery to be predicted based on the lithium ion battery capacity estimation method
Figure BDA0003048574420000021
S3, establishing a state space of a lithium ion battery capacity index decay model, and performing parameter estimation;
s4, setting the number N of sampling particles and the process noise variance sigmawAnd measure the noise variance σvAnd a resampling threshold
Figure BDA0003048574420000022
S5, estimating the capacity of the lithium ion battery to be predicted
Figure BDA0003048574420000023
Substituting the measured value into a lithium ion battery capacity exponential decay model, continuously updating the weight of particles based on a correlation coefficient particle filter algorithm, and obtaining state posterior estimation when the starting point is predicted;
and S6, iterating the state posterior estimation to the life threshold value based on the capacity attenuation model, and obtaining the residual life prediction result.
Optionally, the step S2 includes:
s21, acquiring and analyzing data of the lithium ion battery during operation;
s22, extracting physical parameters capable of representing battery performance degradation in the data obtained in the step S21;
s23, reducing the dimensionality of the physical parameters extracted in the step S22 based on a principal component analysis method, and acquiring fusion health factors capable of representing battery performance degradation;
s24, taking the fusion health factor as an input parameter of the NARX neural network, and taking actually measured capacity data as an output parameter to obtain a relation model of the fusion health factor and the battery residual capacity;
s25, obtaining characteristic parameters of the lithium ion battery to be predicted before the prediction starting point, reducing the dimension, and using the characteristic parameters as the input of a relation model fusing the health factor and the battery residual capacity, thereby obtaining the capacity estimation value of the lithium ion battery to be predicted
Figure BDA0003048574420000024
Optionally, the physical parameters capable of characterizing the battery performance degradation in step S22 include:
the time when the voltage is reduced to the minimum peak value, the constant voltage drop discharge time, the time when the current at the load end and the output current are reduced to the minimum peak value, and the time when the temperature is increased to the maximum peak value.
Optionally, the state space for establishing the lithium ion battery capacity exponential decay model in step S3 includes:
the lithium ion battery capacity exponential decay model is as follows:
Qk＝a·exp(b·k)+c·exp(d·k) (1)
wherein a, b, c, d are parameters of the capacity exponential decay model, a is a first initial value of the capacity of the battery, c is a second initial value of the capacity of the battery, b is a first capacity decay rate, d is a second capacity decay rate, k is the number of charge-discharge cycles, QkThe residual capacity of the battery at the moment k is the residual capacity of the battery at the kth charge-discharge cycle;
the exponential decay model of the capacity of the lithium ion battery in the formula (1) can be converted into the following by polynomial operation:
Qk＝Qk-1·exp(b)+c·exp[d·(k-1)]·[1-exp(b-d)] (2)
the state space equation of the lithium ion battery capacity exponential decay model is as follows:
Figure BDA0003048574420000031
Figure BDA0003048574420000032
wherein wkIs the process noise at time k, vkMeasurement noise at time k, wk～N(0,σw) Denotes wkObedience is expected to be 0 and variance is σwNormal distribution of (v)k～N(0,σv) Denotes vkObedience is expected to be 0 and variance is σvNormal distribution of (2), Qk-1The residual capacity at the k-1 st charge-discharge cycle.
Optionally, the performing parameter estimation in step S3 includes:
taking the parameters in the lithium ion battery capacity exponential decay model as the system state to obtain a state space model x at the moment kk：
xk＝[ak,bk,ck,dk]
ak＝ak-1+wa wa～N(0,σa)
bk＝bk-1+wb wb～N(0,σb)
ck＝ck-1+wc wc～N(0,σc)
dk＝dk-1+wd wd～N(0,σd) (6)
Qk＝ak·exp(bk·k)+ck·exp(dk·k)+vk vk～N(0,σn)
Wherein, ak,bk,ck,dkA first battery capacity initial value, a second battery capacity initial value, a first capacity fade rate and a second capacity fade rate at time k, respectivelyk-1,bk-1,ck-1,dk-1Respectively corresponding parameter at time k-1, wa,wb,wc,wdAre respectively a constant;
fitting the full-period capacity data of the training lithium ion battery according to the lithium ion battery capacity exponential decay model, and taking the fitted parameters as initial parameters of the lithium ion battery to be predicted;
estimating the capacity of the lithium ion battery to be predicted
Figure BDA0003048574420000033
State space model x as time kkThe parameters are updated iteratively based on a particle filter algorithm, and the values of b, c and d after iteration are parameters in a lithium ion battery capacity exponential decay model.
Optionally, the step S5 includes:
s51, initialization: for i 1,2, N, from the distribution
Figure BDA0003048574420000041
Extracting N particles to form a particle set
Figure BDA0003048574420000042
S52, calculating the sample weight of the particle:
Figure BDA0003048574420000043
wherein z iskAs individual measurements;
s53, weight normalization:
Figure BDA0003048574420000044
s54, calculating the number N of effective particleseffSetting a resampling threshold NthIf the number of effective particles Neff＜NthThen resampling is carried out;
s55, actual measurement of construction systemValue sequence and sample estimation measurement matrix: for the actual measured value sequence, a length L sequence is taken
Figure BDA0003048574420000045
For a sequence of sample estimation measurements, take
Figure BDA0003048574420000046
Figure BDA0003048574420000047
S56, calculating a correlation coefficient: for the sequence ZkAnd
Figure BDA0003048574420000048
calculating the correlation coefficient cc;
s57, recalculating the weight of the particle sample: the range of the correlation coefficient cc is [ -1,1], and in order to transform it into the positive range, an exponential function with a parameter α is taken to process the correlation coefficient:
Figure BDA0003048574420000049
wherein, the parameter alpha is more than 0, the function is to adjust the discrete degree of the sample weight, and beta is the correlation coefficient after conversion;
the recalculated particle sample weights are:
Figure BDA00030485744200000410
and (4) normalizing the recalculated sample particle weight value according to the formula (8).
Compared with the prior art, the invention has the following advantages:
aiming at the problem that the capacity is difficult to directly measure in the RUL prediction of the lithium ion battery, the method can be used for establishing a relation model fusing a health factor and the residual capacity of the battery by analyzing measurable battery performance parameters in a discharge period during the operation of the battery so as to obtain the estimated value of the residual capacity of the lithium ion battery in real time; in addition, the method introduces a relative particle filter algorithm into the battery RUL prediction aiming at the problems of particle degradation and sample shortage in the standard particle filter algorithm, and can effectively improve the RUL prediction precision of the lithium battery.
Further, the method of the present invention can predict the RUL of the battery and give an uncertainty expression of the prediction result by substituting the obtained capacity estimation value as a measurement value into a correlation coefficient particle filter algorithm.
Drawings
FIG. 1 is a schematic diagram of a lithium ion battery life prediction method based on correlation coefficient particle filtering according to the present invention;
FIG. 2 is a graph of B5 cell and B6 cell capacity versus cycle number for an example of the present invention;
FIG. 3 is a diagram illustrating the capacity estimation result of a B6 battery according to an embodiment of the present invention;
FIG. 4 is a schematic diagram illustrating the RUL prediction result of a lithium ion battery based on a standard particle filtering algorithm according to an embodiment of the present invention;
fig. 5 is a schematic diagram of a RUL prediction result of a lithium ion battery based on a correlation coefficient particle filter algorithm according to an embodiment of the present invention.
Detailed Description
The present invention will now be further described by way of the following detailed description of a preferred embodiment thereof, taken in conjunction with the accompanying drawings.
As shown in fig. 1, a schematic diagram of a lithium ion battery life prediction method based on correlation coefficient particle filtering according to the present invention is shown, and the method includes:
and S1, setting a prediction starting point and a battery life threshold.
S2, obtaining the data of the lithium ion battery to be predicted, and obtaining the capacity estimation value of the lithium ion battery to be predicted based on the lithium ion battery capacity estimation method
Figure BDA0003048574420000051
The step S2 includes:
and S21, acquiring and analyzing the data of the training lithium ion battery during operation.
And S22, extracting physical parameters capable of representing the battery performance degradation from the data obtained in the step S21.
Wherein the physical parameters capable of characterizing the battery performance degradation in step S22 include: the time when the voltage is reduced to the minimum peak value, the constant voltage drop discharge time, the time when the current at the load end and the output current are reduced to the minimum peak value, the time when the temperature is increased to the maximum peak value and the like.
And S23, reducing the dimensionality of the physical parameters extracted in the step S22 based on the principal component analysis method, and acquiring fusion health factors capable of representing battery performance degradation.
And S24, taking the fusion health factor as an input parameter of the NARX neural network, and taking actually measured capacity data as an output parameter to obtain a relation model of the fusion health factor and the battery residual capacity.
S25, obtaining characteristic parameters of the lithium ion battery to be predicted before the prediction starting point, reducing the dimension, and using the characteristic parameters as the input of a relation model fusing the health factor and the battery residual capacity, thereby obtaining the capacity estimation value of the lithium ion battery to be predicted
Figure BDA0003048574420000061
And S3, establishing a state space of the lithium ion battery capacity exponential decay model, and performing parameter estimation.
The state space for establishing the lithium ion battery capacity exponential decay model in the step S3 includes:
the lithium ion battery capacity exponential decay model is as follows:
Qk＝a·exp(b·k)+c·exp(d·k) (1)
wherein a, b, c, d are parameters of the capacity exponential decay model, a is a first initial value of the capacity of the battery, c is a second initial value of the capacity of the battery, b is a first capacity decay rate, d is a second capacity decay rate, k is the number of charge-discharge cycles, QkBattery at time kThe residual capacity is the residual capacity in the k-th charge-discharge cycle;
the exponential decay model of the capacity of the lithium ion battery in the formula (1) can be converted into the following by polynomial operation:
Qk＝Qk-1·exp(b)+c·exp[d·(k-1)]·[1-exp(b-d)] (2)
the state space equation of the lithium ion battery capacity exponential decay model is as follows:
Figure BDA0003048574420000062
Figure BDA0003048574420000063
wherein wkIs the process noise at time k, vkMeasurement noise at time k, wk～N(0,σw) Denotes wkObedience is expected to be 0 and variance is σwNormal distribution of (v)k～N(0,σv) Denotes vkObedience is expected to be 0 and variance is σvNormal distribution of (2), Qk-1The residual capacity at the k-1 st charge-discharge cycle.
Further, the performing of parameter estimation in step S3 includes:
taking the parameters in the lithium ion battery capacity exponential decay model as the system state to obtain a state space model x at the moment kk：
xk＝[ak,bk,ck,dk]
ak＝ak-1+wa wa～N(0,σa)
bk＝bk-1+wb wb～N(0,σb)
ck＝ck-1+wc wc～N(0,σc)
dk＝dk-1+wd wd～N(0,σd) (6)
Qk＝ak·exp(bk·k)+ck·exp(dk·k)+vk vk～N(0,σn)
Wherein, ak,bk,ck,dkA first battery capacity initial value, a second battery capacity initial value, a first capacity fade rate and a second capacity fade rate at time k, respectivelyk-1,bk-1,ck-1,dk-1Respectively corresponding parameter at time k-1, wa,wb,wc,wdAre constants (the constants are not necessarily equal).
And fitting the full-period capacity data of the training lithium ion battery according to the lithium ion battery capacity exponential decay model, and taking the fitted parameters as initial parameters of the lithium ion battery to be predicted.
Estimating the capacity of the lithium ion battery to be predicted
Figure BDA0003048574420000071
State space model x as time kkThe parameters are updated iteratively based on a particle filter algorithm, and the values of b, c and d after iteration are parameters in a lithium ion battery capacity exponential decay model.
S4, setting the number N of sampling particles and the process noise variance sigmawAnd measure the noise variance σvAnd a resampling threshold
Figure BDA0003048574420000072
S5, estimating the capacity of the lithium ion battery to be predicted
Figure BDA0003048574420000073
Substituting the measured value into a lithium ion battery capacity exponential decay model, continuously updating the particle weight based on a correlation coefficient particle filter algorithm, and obtaining the state posterior estimation at the starting point of prediction.
Specifically, the step S5 includes:
s51, initialization: for i 1,2, N, from the distribution
Figure BDA0003048574420000074
Extracting N particles to form a particle set
Figure BDA0003048574420000075
S52, calculating the sample weight of the particle:
Figure BDA0003048574420000076
wherein z iskAre individual measurements.
S53, weight normalization:
Figure BDA0003048574420000081
s54, calculating the number N of effective particleseffSetting a resampling threshold NthIf the number of effective particles Neff＜NthResampling is performed.
S55, constructing a system actual measurement value sequence and a sample estimation measurement value matrix: for the actual measured value sequence, a length L sequence is taken
Figure BDA0003048574420000082
For a sequence of sample estimation measurements, take
Figure BDA0003048574420000083
Figure BDA0003048574420000084
S56, calculating a correlation coefficient: for the sequence ZkAnd
Figure BDA0003048574420000085
the correlation coefficient cc is calculated.
S57, recalculating the weight of the particle sample: the range of the correlation coefficient cc is [ -1,1], and in order to transform it into the positive range, an exponential function with a parameter α is taken to process the correlation coefficient:
Figure BDA0003048574420000086
wherein, the parameter alpha is more than 0, the function is to adjust the discrete degree of the sample weight, and beta is the correlation coefficient after conversion;
the recalculated particle sample weights are:
Figure BDA0003048574420000087
and (4) normalizing the recalculated sample particle weight value according to the formula (8).
And S6, iterating the state posterior estimation to the life threshold value based on the capacity attenuation model, and obtaining the residual life prediction result.
In this embodiment, the validity of the method of the present invention is demonstrated by combining with an example, where the test set is test data obtained by performing an accelerated life test on a lithium ion battery by National Aeronautics and astronautics (NASA), and the data set includes test data of temperature, current, voltage, and the like during charging and discharging processes of four lithium ion batteries with numbers B5, B6, B7, and B18. The test samples had selected B5 and B6 cells. The B5 battery data are used as a training set to establish a relation model fused with the capacity of the health factor, so that the B6 battery capacity is estimated, and the B6 battery data are used for verification and RUL prediction.
As shown in fig. 2, the capacity of the B5 cell and the B6 cell were plotted against the number of cycles. The battery capacity generally shows a tendency to gradually decay, locally accompanied by a capacity recovery effect.
As shown in fig. 3, the result is the capacity estimation of the B6 battery. And establishing a relation model fusing the health factor and the residual capacity of the battery by taking the full life cycle data of the B5 battery as training data. Indirect parameters in the circulation process of the B6 battery are extracted, a main component analysis method is used for obtaining health factors, the health factors are used as input, the full life cycle capacity of the B6 battery is estimated, and the root mean square error of the estimation result is 0.0247. From the above results, it can be seen that the proposed lithium ion capacity estimation method has high accuracy and strong adaptability.
Fig. 4 shows the result of RUL prediction of a lithium ion battery using a standard particle filter algorithm. The prediction starting point is the 80 th cycle, the service life threshold value for judging whether the battery is failed is Q < 1.38, the particle number is selected to be 200, the actual failure time of the battery is the 113 th cycle, the predicted failure time is the 105 th cycle, and the prediction error is 8 cycle cycles.
Fig. 5 shows the result of RUL prediction of a lithium ion battery using the correlation coefficient particle filter algorithm according to the present invention. The prediction starting point is also selected as the 80 th cycle, the life threshold for judging whether the battery is failed is Q < 1.38, the particle number is selected to be 200, the actual failure time of the battery is the 113 th cycle, the predicted failure time is the 109 th cycle, and the prediction error is 4 cycle cycles. The simulation result can show that the method provided by the invention can effectively predict the service life of the lithium ion battery, the prediction error is within an acceptable range, and the particle filter algorithm based on the correlation coefficient can effectively improve the accuracy of the long-term prediction of the RUL.
In summary, according to the lithium ion battery life prediction method based on the correlation coefficient particle filter, aiming at the problem that the capacity is difficult to directly measure in the RUL prediction of the lithium ion battery, a relation model fusing a health factor and the battery residual capacity is established by analyzing measurable battery performance parameters in a discharge cycle during the operation of the battery, so as to obtain an estimated value of the lithium ion battery residual capacity in real time; in addition, the method introduces a relative particle filter algorithm into the battery RUL prediction aiming at the problems of particle degradation and sample shortage in the standard particle filter algorithm, and can effectively improve the RUL prediction precision of the lithium battery.
Further, the method of the present invention can predict the RUL of the battery and give an uncertainty expression of the prediction result by substituting the obtained capacity estimation value as a measurement value into a correlation coefficient particle filter algorithm.
While the present invention has been described in detail with reference to the preferred embodiments, it should be understood that the above description should not be taken as limiting the invention. Various modifications and alterations to this invention will become apparent to those skilled in the art upon reading the foregoing description. Accordingly, the scope of the invention should be determined from the following claims.

Claims (6)

1. A lithium ion battery life prediction method based on correlation coefficient particle filtering is characterized by comprising the following steps:
s1, setting a prediction starting point and a battery life threshold;
s2, obtaining the data of the lithium ion battery to be predicted, and obtaining the capacity estimation value of the lithium ion battery to be predicted based on the lithium ion battery capacity estimation method
Figure FDA0003048574410000011
S3, establishing a state space of a lithium ion battery capacity index decay model, and performing parameter estimation;
s4, setting the number N of sampling particles and the process noise variance sigmawAnd measure the noise variance σvAnd a resampling threshold
Figure FDA0003048574410000012
S5, estimating the capacity of the lithium ion battery to be predicted
Figure FDA0003048574410000013
Substituting the measured value into a lithium ion battery capacity exponential decay model, continuously updating the weight of particles based on a correlation coefficient particle filter algorithm, and obtaining state posterior estimation when the starting point is predicted;
and S6, iterating the state posterior estimation to the life threshold value based on the capacity attenuation model, and obtaining the residual life prediction result.
2. The method for predicting the life of a lithium ion battery based on the correlation coefficient particle filter according to claim 1, wherein the step S2 comprises:
s21, acquiring and analyzing data of the lithium ion battery during operation;
s22, extracting physical parameters capable of representing battery performance degradation in the data obtained in the step S21;
s23, reducing the dimensionality of the physical parameters extracted in the step S22 based on a principal component analysis method, and acquiring fusion health factors capable of representing battery performance degradation;
s24, taking the fusion health factor as an input parameter of the NARX neural network, and taking actually measured capacity data as an output parameter to obtain a relation model of the fusion health factor and the battery residual capacity;
s25, obtaining characteristic parameters of the lithium ion battery to be predicted before the prediction starting point, reducing the dimension, and using the characteristic parameters as the input of a relation model fusing the health factor and the battery residual capacity, thereby obtaining the capacity estimation value of the lithium ion battery to be predicted
Figure FDA0003048574410000014
3. The correlation coefficient particle filter-based lithium ion battery life prediction method of claim 2, wherein the physical parameters capable of characterizing battery performance degradation in step S22 comprise:
the time when the voltage is reduced to the minimum peak value, the constant voltage drop discharge time, the time when the current at the load end and the output current are reduced to the minimum peak value, and the time when the temperature is increased to the maximum peak value.
4. The method for predicting the life of a lithium ion battery based on the correlation coefficient particle filter of claim 1, wherein the establishing the state space of the exponential decay model of the capacity of the lithium ion battery in the step S3 comprises:
the lithium ion battery capacity exponential decay model is as follows:
Qk＝a·exp(b·k)+c·exp(d·k) (1)
wherein a, b, c, d are parameters of the capacity exponential decay model, a is a first initial value of the capacity of the battery, c is a second initial value of the capacity of the battery, b is a first capacity decay rate, d is a second capacity decay rate, k is the number of charge-discharge cycles, QkThe residual capacity of the battery at the moment k is the residual capacity of the battery at the kth charge-discharge cycle;
the exponential decay model of the capacity of the lithium ion battery in the formula (1) can be converted into the following by polynomial operation:
Qk＝Qk-1·exp(b)+c·exp[d·(k-1)]·[1-exp(b-d)] (2)
the state space equation of the lithium ion battery capacity exponential decay model is as follows:
Figure FDA0003048574410000021
Figure FDA0003048574410000022
wherein wkIs the process noise at time k, vkMeasurement noise at time k, wk～N(0,σw) Denotes wkObedience is expected to be 0 and variance is σwNormal distribution of (v)k～N(0,σv) Denotes vkObedience is expected to be 0 and variance is σvNormal distribution of (2), Qk-1The residual capacity at the k-1 st charge-discharge cycle.
5. The method for predicting the life of a lithium ion battery based on the correlation coefficient particle filter according to claim 4, wherein the performing parameter estimation in step S3 comprises:
taking the parameters in the lithium ion battery capacity exponential decay model as the system state to obtain a state space model x at the moment kk：
xk＝[ak,bk,ck,dk]
ak＝ak-1+wa wa～N(0,σa)
bk＝bk-1+wb wb～N(0,σb)
ck＝ck-1+wc wc～N(0,σc)
dk＝dk-1+wd wd～N(0,σd) (6)
Qk＝ak·exp(bk·k)+ck·exp(dk·k)+vk vk～N(0,σn)
Wherein, ak,bk,ck,dkA first battery capacity initial value, a second battery capacity initial value, a first capacity fade rate and a second capacity fade rate at time k, respectivelyk-1,bk-1,ck-1,dk-1Respectively corresponding parameter at time k-1, wa,wb,wc,wdAre respectively a constant;
fitting the full-period capacity data of the training lithium ion battery according to the lithium ion battery capacity exponential decay model, and taking the fitted parameters as initial parameters of the lithium ion battery to be predicted;
estimating the capacity of the lithium ion battery to be predicted
Figure FDA0003048574410000031
State space model x as time kkThe parameters are updated iteratively based on a particle filter algorithm, and the values of b, c and d after iteration are parameters in a lithium ion battery capacity exponential decay model.
6. The method for predicting the life of a lithium ion battery based on the correlation coefficient particle filter according to claim 5, wherein the step S5 comprises:
s51, initialization: for i 1,2, N, from the distribution
Figure FDA0003048574410000032
Extracting N particles to form a particle set
Figure FDA0003048574410000033
S52, calculating the sample weight of the particle:
Figure FDA0003048574410000034
wherein z iskAs individual measurements;
s53, weight normalization:
Figure FDA0003048574410000035
s54, calculating the number N of effective particleseffSetting a resampling threshold NthIf the number of effective particles Neff＜NthThen resampling is carried out;
s55, constructing a system actual measurement value sequence and a sample estimation measurement value matrix: for the actual measured value sequence, a length L sequence is taken
Figure FDA0003048574410000036
For a sequence of sample estimation measurements, take
Figure FDA0003048574410000037
Figure FDA0003048574410000041
S56, calculating a correlation coefficient: for the sequence ZkAnd
Figure FDA0003048574410000042
calculating the correlation coefficient cc;
s57, recalculating the weight of the particle sample: the range of the correlation coefficient cc is [ -1,1], and in order to transform it into the positive range, an exponential function with a parameter α is taken to process the correlation coefficient:
Figure FDA0003048574410000043
wherein, the parameter alpha is more than 0, the function is to adjust the discrete degree of the sample weight, and beta is the correlation coefficient after conversion;
the recalculated particle sample weights are:
Figure FDA0003048574410000044
and (4) normalizing the recalculated sample particle weight value according to the formula (8).
CN202110479300.2A 2021-04-30 2021-04-30 Lithium ion battery life prediction method based on correlation coefficient particle filtering Pending CN112986831A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110479300.2A CN112986831A (en) 2021-04-30 2021-04-30 Lithium ion battery life prediction method based on correlation coefficient particle filtering

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110479300.2A CN112986831A (en) 2021-04-30 2021-04-30 Lithium ion battery life prediction method based on correlation coefficient particle filtering

Publications (1)

Publication Number Publication Date
CN112986831A true CN112986831A (en) 2021-06-18

Family

ID=76336858

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110479300.2A Pending CN112986831A (en) 2021-04-30 2021-04-30 Lithium ion battery life prediction method based on correlation coefficient particle filtering

Country Status (1)

Country Link
CN (1) CN112986831A (en)

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113657012A (en) * 2021-07-21 2021-11-16 西安理工大学 TCN and particle filter-based method for predicting residual life of key equipment
CN113721158A (en) * 2021-07-19 2021-11-30 恒大海拉电子(扬州)有限公司 Lithium ion battery life state estimation method based on temperature coupling model
CN113839107A (en) * 2021-09-22 2021-12-24 北京航空航天大学 Early warning method for soft package lithium ion battery diving degradation mode
CN114167284A (en) * 2021-11-02 2022-03-11 江苏博强新能源科技股份有限公司 Lithium battery RUL prediction method and device based on BMS big data and ensemble learning
CN114236414A (en) * 2021-12-16 2022-03-25 上海交通大学 Satellite lithium battery life prediction method based on KCC-PF algorithm
CN114252797A (en) * 2021-12-17 2022-03-29 华中科技大学 Uncertainty estimation-based lithium battery remaining service life prediction method
CN114325393A (en) * 2021-12-28 2022-04-12 江苏大学 Lithium ion battery pack SOH self-adaptive estimation method based on PF and GPR
CN115575843A (en) * 2022-10-25 2023-01-06 楚能新能源股份有限公司 Lithium ion battery life prediction method
CN115701851A (en) * 2022-09-19 2023-02-14 楚能新能源股份有限公司 Soft package lithium ion battery thickness prediction method
CN117074965A (en) * 2023-10-17 2023-11-17 深圳市神通天下科技有限公司 Lithium ion battery remaining life prediction method and system

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105445671A (en) * 2015-12-29 2016-03-30 北京航天测控技术有限公司 Lithium ion battery service life prediction method based on traceless particle filtering
CN107238800A (en) * 2017-06-12 2017-10-10 北京理工大学 A kind of remaining probable life Forecasting Methodology of the electrokinetic cell system based on Method Using Relevance Vector Machine and particle filter
CN110457789A (en) * 2019-07-25 2019-11-15 桂林电子科技大学 A kind of lithium ion battery residual life prediction technique merged based on improvement particle filter with double exponential decay experience physical models
CN110703120A (en) * 2019-09-29 2020-01-17 上海海事大学 Lithium ion battery service life prediction method based on particle filtering and long-and-short time memory network
CN111931415A (en) * 2019-07-18 2020-11-13 东莞理工学院 Global optimal particle filter-based life prediction method for lithium ion battery
CN112001114A (en) * 2020-07-02 2020-11-27 浙江大学 Battery life prediction method based on LSTM improved regularized particle filter algorithm

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105445671A (en) * 2015-12-29 2016-03-30 北京航天测控技术有限公司 Lithium ion battery service life prediction method based on traceless particle filtering
CN107238800A (en) * 2017-06-12 2017-10-10 北京理工大学 A kind of remaining probable life Forecasting Methodology of the electrokinetic cell system based on Method Using Relevance Vector Machine and particle filter
CN111931415A (en) * 2019-07-18 2020-11-13 东莞理工学院 Global optimal particle filter-based life prediction method for lithium ion battery
CN110457789A (en) * 2019-07-25 2019-11-15 桂林电子科技大学 A kind of lithium ion battery residual life prediction technique merged based on improvement particle filter with double exponential decay experience physical models
CN110703120A (en) * 2019-09-29 2020-01-17 上海海事大学 Lithium ion battery service life prediction method based on particle filtering and long-and-short time memory network
CN112001114A (en) * 2020-07-02 2020-11-27 浙江大学 Battery life prediction method based on LSTM improved regularized particle filter algorithm

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
DIJU GAO 等: "A Method for Predicting the Remaining Useful Life of Lithium-Ion Batteries Based on Particle Filter Using Kendall Rank Correlation Coefficient", 《ENERGIES》, 13 August 2020 (2020-08-13), pages 1 - 13 *
崔显 等: "基于KCC-PF 的锂离子电池剩余使用寿命预测", 《装备环境工程》, vol. 19, no. 4, 30 April 2022 (2022-04-30), pages 86 - 94 *

Cited By (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113721158A (en) * 2021-07-19 2021-11-30 恒大海拉电子(扬州)有限公司 Lithium ion battery life state estimation method based on temperature coupling model
CN113657012A (en) * 2021-07-21 2021-11-16 西安理工大学 TCN and particle filter-based method for predicting residual life of key equipment
CN113839107A (en) * 2021-09-22 2021-12-24 北京航空航天大学 Early warning method for soft package lithium ion battery diving degradation mode
CN113839107B (en) * 2021-09-22 2023-03-21 北京航空航天大学 Early warning method for soft package lithium ion battery diving degradation mode
CN114167284A (en) * 2021-11-02 2022-03-11 江苏博强新能源科技股份有限公司 Lithium battery RUL prediction method and device based on BMS big data and ensemble learning
CN114167284B (en) * 2021-11-02 2023-12-22 江苏博强新能源科技股份有限公司 Lithium battery RUL prediction method and equipment based on BMS big data and integrated learning
CN114236414A (en) * 2021-12-16 2022-03-25 上海交通大学 Satellite lithium battery life prediction method based on KCC-PF algorithm
CN114252797B (en) * 2021-12-17 2023-03-10 华中科技大学 Uncertainty estimation-based lithium battery remaining service life prediction method
CN114252797A (en) * 2021-12-17 2022-03-29 华中科技大学 Uncertainty estimation-based lithium battery remaining service life prediction method
CN114325393A (en) * 2021-12-28 2022-04-12 江苏大学 Lithium ion battery pack SOH self-adaptive estimation method based on PF and GPR
CN115701851A (en) * 2022-09-19 2023-02-14 楚能新能源股份有限公司 Soft package lithium ion battery thickness prediction method
CN115701851B (en) * 2022-09-19 2023-09-29 楚能新能源股份有限公司 Soft package lithium ion battery thickness prediction method
CN115575843A (en) * 2022-10-25 2023-01-06 楚能新能源股份有限公司 Lithium ion battery life prediction method
CN117074965A (en) * 2023-10-17 2023-11-17 深圳市神通天下科技有限公司 Lithium ion battery remaining life prediction method and system
CN117074965B (en) * 2023-10-17 2023-12-12 深圳市神通天下科技有限公司 Lithium ion battery remaining life prediction method and system

Similar Documents

Publication Publication Date Title
CN112986831A (en) Lithium ion battery life prediction method based on correlation coefficient particle filtering
CN111103544B (en) Lithium ion battery remaining service life prediction method based on long-time and short-time memory LSTM and particle filter PF
CN113009349B (en) Lithium ion battery health state diagnosis method based on deep learning model
CN107957562B (en) Online prediction method for residual life of lithium ion battery
CN108254696B (en) Battery health state evaluation method and system
CN111239630A (en) Energy storage battery service life prediction method and management system
CN108037462B (en) Method and system for quantifying health condition of storage battery
CN109828220B (en) Linear evaluation method for health state of lithium ion battery
CN111965559B (en) On-line estimation method for SOH of lithium ion battery
CN112989690B (en) Multi-time-scale state-of-charge estimation method for lithium battery of hybrid electric vehicle
CN111220920B (en) Retired lithium ion battery state of charge calculation method based on H-infinity unscented Kalman filtering algorithm
CN112198434B (en) Identification method, system, equipment, device and medium for battery capacity attenuation model parameters
CN115902647B (en) Intelligent battery state monitoring method
CN112684363A (en) Lithium ion battery health state estimation method based on discharge process
CN110703120A (en) Lithium ion battery service life prediction method based on particle filtering and long-and-short time memory network
CN115267555A (en) Battery SOH (State of health) evaluation system of energy storage system based on battery multipoint temperature measurement
CN114839538A (en) Method for extracting degradation characteristics of lithium ion battery for estimating residual life
CN112305426B (en) Lithium ion battery power state estimation system under multi-constraint condition
CN113392507A (en) Method for predicting residual life of lithium ion power battery
CN114114038A (en) Lithium battery SOC and available capacity joint estimation method under full-life and full-temperature conditions
CN113359048A (en) Indirect prediction method for remaining service life of lithium ion battery
CN112765794A (en) Reliability analysis method of lithium iron phosphate battery based on accelerated life test
CN115219918A (en) Lithium ion battery life prediction method based on capacity decline combined model
CN114280490B (en) Lithium ion battery state of charge estimation method and system
CN113466728B (en) Method and system for online identification of two-stage battery model parameters

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination