CN111856282B - Vehicle-mounted lithium battery state estimation method based on improved genetic unscented Kalman filtering - Google Patents

Vehicle-mounted lithium battery state estimation method based on improved genetic unscented Kalman filtering Download PDF

Info

Publication number
CN111856282B
CN111856282B CN201910317198.9A CN201910317198A CN111856282B CN 111856282 B CN111856282 B CN 111856282B CN 201910317198 A CN201910317198 A CN 201910317198A CN 111856282 B CN111856282 B CN 111856282B
Authority
CN
China
Prior art keywords
lithium battery
open
circuit voltage
current
value
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.)
Active
Application number
CN201910317198.9A
Other languages
Chinese (zh)
Other versions
CN111856282A (en
Inventor
谢长君
房伟
麦立强
陈伟
曾春年
黄亮
蔡振华
熊斌宇
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.)
Wuhan University of Technology WUT
Original Assignee
Wuhan University of Technology WUT
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 Wuhan University of Technology WUT filed Critical Wuhan University of Technology WUT
Priority to CN201910317198.9A priority Critical patent/CN111856282B/en
Publication of CN111856282A publication Critical patent/CN111856282A/en
Application granted granted Critical
Publication of CN111856282B publication Critical patent/CN111856282B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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]
    • G01R31/374—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC] with means for correcting the measurement for temperature or ageing
    • 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]
    • G01R31/385—Arrangements for measuring battery or accumulator variables
    • G01R31/387—Determining ampere-hour charge capacity or SoC
    • G01R31/388—Determining ampere-hour charge capacity or SoC involving voltage measurements
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/11—Complex mathematical operations for solving equations, e.g. nonlinear equations, general mathematical optimization problems
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/15—Correlation function computation including computation of convolution operations
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/12—Computing arrangements based on biological models using genetic models
    • G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/60—Other road transportation technologies with climate change mitigation effect
    • Y02T10/70—Energy storage systems for electromobility, e.g. batteries

Abstract

The invention provides a vehicle-mounted lithium battery state estimation method based on improved genetic unscented Kalman filtering, which is characterized in that open-circuit voltage and current of a lithium battery are sampled under the condition of constant-current discharge of the lithium battery; establishing a state space equation of the lithium battery according to a second-order RC equivalent circuit model, and calculating an estimated value of open-circuit voltage; estimating and updating the second-order RC equivalent circuit model in real time by using a Kalman filtering algorithm; searching for an optimal noise covariance matrix by using an improved genetic algorithm; estimating the battery state by using an unscented Kalman filtering algorithm through an open-circuit voltage estimation value to obtain a current open-circuit voltage estimation value; and then outputting an SOC estimated value corresponding to the current open-circuit voltage estimated value according to the relation between the open-circuit voltage of the lithium battery and the SOC. The method solves the problem of serious system state variable nonlinearity caused by electrochemical reaction in the lithium battery, and improves the real-time performance and the accuracy of estimation.

Description

Vehicle-mounted lithium battery state estimation method based on improved genetic unscented Kalman filtering
Technical Field
The invention belongs to the field of vehicle-mounted lithium batteries, and particularly relates to a vehicle-mounted lithium battery state estimation method based on improved genetic unscented Kalman filtering.
Background
The lithium ion battery has the advantages of long cycle life, small self-discharge rate, high energy, no memory effect and the like, and is widely applied to the field of electric automobiles. In order to ensure safe, reliable and efficient operation of a lithium ion battery, it is necessary to accurately estimate the operating state of the battery and accurately establish a battery management system. The state of charge (SOC) of the lithium ion battery directly reflects the amount of the residual electric quantity, and the overcharge and overdischarge behaviors of the lithium ion battery can be avoided only by accurately estimating the SOC of the battery, so that the battery is kept in a good working state.
At present, the estimation method for the SOC of the lithium ion battery at home and abroad mainly comprises the following steps: 1) the method comprises the following steps that an open-circuit voltage method (OCV) is used, the SOC value is obtained by measuring the open-circuit voltage by utilizing the nonlinear relation between the open-circuit voltage and the SOC, and the method is not suitable for online estimation of the SOC because a battery needs to be kept still for a long time in the measuring process; 2) the ampere-hour integration method (CC) is used for integrating the current under a known initial value, has extremely high requirements on the initial value of the SOC, and ignores the influence caused by the capacity attenuation caused by the accumulation error generated in the current detection and the battery aging; 3) the method is characterized by comprising the following steps of (1) searching the relation between the ohmic-polarization internal resistance and the SOC of the lithium ion battery by an Electrochemical Impedance Spectroscopy (EIS), wherein the method is poor in stability, complex in detection, long in running time and less in application in real-time monitoring; 4) kalman Filtering (KF) corrects the state estimation by using observation data on the basis of an equivalent model, and the algorithm has higher requirement on model precision and is not suitable for a nonlinear system; 5) the Extended Kalman Filter (EKF) is used for linearly processing a nonlinear function and then finishing the estimation of a target by Kalman filtering, and the estimation value is diverged due to the non-orthostatic property of a variance matrix in the calculation process; 6) unscented Kalman (UKF), abandon the nonlinear function to process the linearization, utilize the unscented transformation to process the nonlinear transfer problem of mean value and covariance, the method has higher precision to the statistic of nonlinear distribution; 7) particle Filter (PF), which uses discrete particle set to describe the probability density of system random variable approximately, to express the posterior probability distribution problem of observation measurement and control quantity accurately; 8) the neural network method (NN) has strong processing capacity on a nonlinear system, but needs to train a large amount of data, and meanwhile, the estimation error is greatly influenced by training data and a training method; 9) the method is a Sliding Mode Observation (SMO) method, and can effectively solve the influence of a nonlinear model on state estimation, but frequent switching of control states can cause the system to generate buffeting.
In the actual working process of the lithium ion battery, factors such as environmental temperature, cycle frequency, detection precision and the like have important influence on the state estimation of the lithium ion battery. In practical application, the problem of estimation value divergence caused by system nonlinearity often occurs, and the phenomenon of particle shortage occurs because the number of particles is small by simply using a particle filter algorithm.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: the vehicle-mounted lithium battery state estimation method based on the improved genetic unscented Kalman filtering is provided, and estimation precision is improved.
The technical scheme adopted by the invention for solving the technical problems is as follows: a vehicle-mounted lithium battery state estimation method based on improved genetic unscented Kalman filtering is characterized by comprising the following steps: it comprises the following steps:
s1, sampling the open-circuit voltage and current of the lithium battery under the condition of constant-current discharge of the lithium battery;
s2, establishing a state space equation of the lithium battery according to the second-order RC equivalent circuit model, and calculating an estimated value of open-circuit voltage by using the state space equation and the current sampled by the S1; comparing the sampled open circuit voltage with the calculated open circuit voltage estimate;
s3, estimating and updating the second-order RC equivalent circuit model in real time by using a Kalman filtering algorithm and through the open-circuit voltage estimated value and the measured value calculated in S2;
s4, through constant current charging and discharging of lithium batteries with different SOC values, open-circuit voltage is measured after standing for a period of time, and the relation between the open-circuit voltage and the SOC of the lithium batteries is obtained through fitting;
s5, searching an optimal noise covariance matrix by using an improved genetic algorithm;
s6, estimating the state of the battery by adopting the optimal noise covariance matrix obtained in S5 and an unscented Kalman filtering algorithm through the open-circuit voltage estimated value obtained in S2 to obtain a current open-circuit voltage estimated value; and then outputting an SOC estimated value corresponding to the current open-circuit voltage estimated value according to the relation between the open-circuit voltage of the lithium battery and the SOC obtained in S4.
According to the above method, the S1 specifically includes:
selecting vehicle-mounted lithium battery packs with the same consistency to perform a charge-discharge experiment at normal temperature, and selecting the average voltage value of the lithium batteries as effective data;
preventing the lithium battery from being over-discharged, and setting a discharge cut-off voltage;
and setting the constant current discharge rate of the lithium battery, and sampling the open-circuit voltage and the current of the lithium battery in a fixed sampling period.
According to the method, the S3 specifically comprises the following steps:
3-1, obtaining a system transfer function by a kirchhoff law according to a second-order RC equivalent circuit model of the lithium battery;
3-2, carrying out bilinear transformation on the transfer function, and selecting a state equation and an observation equation under the condition of a fixed sampling period;
and 3-3, performing online estimation on the state variable through a Kalman filtering algorithm, and obtaining the identification value of each parameter in the second-order RC equivalent circuit model by utilizing the inverse transformation of z.
According to the method, the S4 specifically comprises the following steps:
4-1, fully discharging/charging the lithium battery and standing for a period of time;
4-2, performing an equidistant constant-current pulse charging/discharging experiment on the lithium battery, standing after each pulse experiment, and measuring the terminal voltage of the lithium battery;
4-3, obtaining the average voltage value of the lithium battery pack at each stage through multiple experiments;
and 4, obtaining the average voltage value of the lithium battery packs with the same consistency as the open-circuit voltage of a single lithium battery through experiments, and fitting a function.
According to the method, the step S5 specifically includes:
5-1, selecting a chromosome coding mode, and randomly generating an initial population according to the selected coding mode;
5-2, judging whether a convergence condition is met, generally judging whether the iteration times are reached, and if the execution is met, executing the step 5-6, and not executing the step 5-3;
5-3, selecting crossed individuals by using an inverse dichotomy;
and 5-4, performing cross operation on the selected cross individuals by using an improved cross probability algorithm.
5-5, selecting variation, and returning to 5-2;
5-6, outputting the optimal solution.
According to the method, the step S6 specifically includes:
6-1, analyzing the electrochemical process of the battery, and listing a state equation and a measurement equation of the system;
6-2, calculating initial values of all state quantities;
6-3, building a Sigma point;
6-4, updating a state equation;
6-5, updating a measurement equation;
6-6, repeating the steps 6-2 to 6-5.
The invention has the beneficial effects that:
1. by employing S5 and S6, the system noise and observed noise covariance matrix Q w And R v Better correction can be obtained in each iteration process to obtain a stable estimation result, and the unscented Kalman filtering algorithm better solves the problem of serious nonlinearity of system state variables caused by electrochemical reaction inside a lithium ion battery, so that the SOC estimation precision is greatly improved compared with the traditional method.
2. Because the step S5 is adopted, the cross individuals are selected by applying the inverse dichotomy, and the improved genetic algorithm of the better individuals and the worse individuals is distinguished by using a brand-new probability algorithm, the blindness of the genetic algorithm operation is reduced, and the preservation of the excellent genes and the elimination of the worse genes are facilitated.
Drawings
FIG. 1 is a flow chart of the present invention.
Fig. 2 is a schematic diagram of a second-order RC equivalent circuit model.
FIG. 3 is a parameter identification graph.
FIG. 4 is a graph of open circuit voltage versus SOC.
FIG. 5 is a flow chart of an improved genetic algorithm.
FIG. 6 is a graph of the relationship between the charging and discharging current and time under the test condition of pulse charging and discharging current.
FIG. 7 is a graph of SOC estimation for different initial settings.
Fig. 8 is a SOC estimation error graph for different initial values.
FIG. 9 is a UDDS cyclic operating condition current plot.
FIG. 10 is a graph of SOC estimation for different algorithms.
FIG. 11 is a graph of SOC estimation error for different algorithms.
Detailed Description
The present invention is further illustrated by the following specific examples.
The invention provides a vehicle-mounted lithium battery state estimation method based on improved genetic unscented Kalman filtering, which comprises the following steps as shown in figure 1:
and S1, sampling the open-circuit voltage and current of the lithium battery under the condition of constant-current discharge of the lithium battery.
In this embodiment, S1 specifically includes the following sub-steps:
(1-1) selecting a lithium iron phosphate battery pack with the same consistency and the rated capacity of a battery monomer of 40Ah at normal temperature, carrying out a charge-discharge experiment, and selecting the average voltage value of a lithium battery as effective data;
(1-2) preventing the over-discharge of the lithium battery, setting a discharge cutoff voltage by software;
(1-3) programming and setting an electronic load to realize the constant-current discharge rate of the battery 1/5C, and sampling the open-circuit voltage and the current of the lithium battery by taking 1s as a sampling period.
S2, establishing a state space equation of the lithium battery according to the second-order RC equivalent circuit model, and calculating an estimated value of the open-circuit voltage by using the state space equation by using the current sampled in S1; the sampled open circuit voltage is compared to the calculated open circuit voltage estimate.
The method specifically comprises the following substeps:
(2-1) first, a second-order RC equivalent circuit model based on current and charge-discharge direction is established as shown in FIG. 2, and the model comprises a voltage control voltage source U oc Characterizing a nonlinear relationship of state of charge and open circuit voltage; r 0 Is the ohmic internal resistance of the cell; r is s ，C s Characterizing the polarization response of the circuit in a short time for the electrochemical polarization resistance and capacitance of the cell; r is l ，C l Polarizing resistance and capacitance for battery concentration, and representing long-time circuit polarization response; i is c Is the working current; v c Is the battery terminal voltage;
(2-2) establishing a state space equation of the lithium ion battery according to the second-order RC equivalent circuit model:
Figure BDA0002033502600000041
V c ＝V oc (SOC)-V s -V l -R 0 I c (k)；②
wherein: t is the sampling time, C N Is the battery capacity, k is the discrete time variable; the same applies below;
(2-3) obtaining an estimated value of the open-circuit voltage by combining a state space equation (i) and a state space equation (ii) according to the sampled current value;
(2-4) comparing the estimated value with the measured value.
And S3, estimating and updating the second-order RC equivalent circuit model in real time by using a Kalman filtering algorithm and through the open-circuit voltage estimated value and the measured value calculated in S2.
The method specifically comprises the following substeps:
(3-1): according to a second-order RC equivalent circuit model of the lithium ion battery, a system transfer function can be obtained through kirchhoff's law:
Figure BDA0002033502600000051
the time constants are set as follows: t is s ＝R s C,T l ＝R l C l ,a＝T s T l ,b＝R s T l +R l T s +R 0 (T s +T l ),c＝R 0 +R s +R l ,d＝T s +T l Then there are:
Figure BDA0002033502600000052
(3-2): order to
Figure BDA0002033502600000053
Derived from the bilinear transformation:
Figure BDA0002033502600000054
wherein:
Figure BDA0002033502600000055
Figure BDA0002033502600000056
selecting a state equation and an observation equation under the condition that the fixed sampling period is T:
θ＝[k 1 ,k 2 ,k 3 ,k 4 ,k 5 ] T
y(k)＝-k 1 ·y(k-1)-k 2 ·y(k-1)+k 3 ·I(k)+k 4 ·I(k-1)+k 5 ·I(k-2)
(3-3): and performing online estimation on the state variable through a Kalman filtering algorithm, and obtaining the identification value of each parameter in the equivalent circuit model by utilizing the inverse transformation of z. Through carrying out a discharge experiment on lithium iron phosphate battery packs with the same consistency, selecting the average voltage value of the lithium ion batteries as effective data, and identifying parameters of 100 groups of voltages by adopting a KF algorithm. Its parameter identification curveAs shown in fig. 3. It can be seen that the ohmic internal resistance R of the battery is obtained throughout the 100 successive discharge stages 0 The variation is small; polarization resistance R s ，R l And a polarization capacitor C s ，C l The converse trend is reversed, with polarization capacitance decreasing as polarization resistance increases.
And S4, charging and discharging lithium batteries with different SOC values at constant current, measuring open-circuit voltage after standing for a period of time, and fitting to obtain the relation between the open-circuit voltage and the SOC of the lithium batteries.
The method specifically comprises the following substeps:
(4-1): fully discharging/charging the lithium ion battery and standing for a period of time;
(4-2): then 10 constant current pulse charging/discharging experiments with equal interval are carried out on the battery, the battery is kept stand for 3 hours after each pulse experiment, the terminal voltage of the battery is measured, and the terminal voltage at the moment can be approximate to open-circuit voltage;
(4-3): obtaining the average value of the lithium ion battery packs in each stage through multiple experiments;
(4-4): the average voltage of the lithium ion battery packs with the same consistency is obtained through experiments and is used as the open-circuit voltage value of the single battery, and a 5-order fitting function is utilized. The relationship curve for identifying the open circuit voltage and the state of charge is shown in fig. 4. The open circuit voltage and SOC of the high order fit are as follows:
V oc ＝15.4647SOC 5 -44.9905SOC 4 +49.5110SOC 3 -25.1893SOC 2 +5.8724SOC+2.7669；
and S5, searching an optimal noise covariance matrix by using an improved genetic algorithm.
As shown in fig. 5, the method specifically includes the following sub-steps:
(5-1): selecting a chromosome coding mode, and randomly generating an initial population according to the selected coding mode;
(5-2): judging whether a convergence condition is met, generally judging whether the iteration times are reached, and if the execution step (5-6) is met, not meeting the execution step (5-3);
(5-3): cross individuals were selected using the inverse dichotomy: the n individuals in the chromosome set are randomly equally divided into n/2 sets. Crossover operations are performed on chromosomes within the set to produce offspring chromosomes. Two sets of offspring chromosomes are randomly selected for combination, generating n/4 combinations, and the set from which each chromosome comes is labeled. Pairwise crossing operations are performed on chromosomes from different sets within a set. Two sets of offspring chromosomes were randomly selected for combination, generating n/8 combinations, and labeling which set each chromosome came from. And continuously performing the operations until the operation is finally combined into a set. (ii) a
(5-4): and performing cross operation on the selected cross individuals by using an improved cross probability algorithm:
Figure BDA0002033502600000061
wherein, P e0 The value can be taken from 0.85 to 0.95 according to the actual situation as the reference cross probability; f best The fitness value of the optimal individual in the current population is obtained;
Figure BDA0002033502600000062
the current population average fitness value is obtained; f is the fitness value of the individual entering the cross-pairing operation.
(5-5): selecting variation and returning to the step (5-2);
(5-6): and outputting the optimal solution.
S6, estimating the battery state by adopting the optimal noise covariance matrix obtained in S5 and an unscented Kalman filtering algorithm through the open-circuit voltage estimation value obtained in S2 to obtain the current open-circuit voltage estimation value; and then outputting an SOC estimated value corresponding to the current open-circuit voltage estimated value according to the relation between the open-circuit voltage of the lithium battery and the SOC obtained in S4.
The method specifically comprises the following substeps:
(6-1): the state equation and the measurement equation of the system are listed:
X k ＝f(X k-1 ,U k )+W k
Y k ＝g(X k-1 )+V k
k is the current time f (X) k-1 ,U k ) For the nonlinear system state transition equation, g (X) k-1 ) For non-linear measurement equations, X k Is a state variable, U k For known input, Y k Is a measurement signal; w k Is process noise, V k To measure noise. We assume W k And V k Is white Gaussian noise with uncorrelated mean values of zero and covariance values of Q w And R v 。
The battery equivalent model is subjected to electrochemical process analysis such as battery self-discharge, electrochemical polarization, concentration difference polarization and the like, and can be divided into the following two parts: based on a run-time model and on a voltage-current characteristic model. The voltage-current model-based part can be obtained by analyzing the external discharge characteristics of the lithium ion battery, and the charge state value of the battery is between 0 and 1.
Based on the battery equivalent model, the state equation and the measurement equation of the lithium ion battery system are listed as follows:
Figure BDA0002033502600000071
V c ＝V OC +I c R 0 +V s +V l
wherein the current I c Taking positive values during charging and negative values during discharging, C q Is the nominal capacity of a lithium ion battery.
(6-2): calculating initial values of various state quantities:
Figure BDA0002033502600000072
Figure BDA0002033502600000073
where k | k-1 is an estimate based on time k-1 versus time k.
(6-3): build Sigma points:
Figure BDA0002033502600000074
Figure BDA0002033502600000081
(6-4): updating the state equation:
Figure BDA0002033502600000082
Figure BDA0002033502600000083
(6-5): updating a measurement equation:
Figure BDA0002033502600000084
Figure BDA0002033502600000085
Figure BDA0002033502600000086
(6-6): the four steps are repeated, namely the optimal state estimation value X at the k moment can be estimated according to the state value at the k-1 moment and the observed value obtained at the k moment k 。
The lithium ion battery was tested with pulse charge and discharge currents under the conditions of setting the initial SOC values to 0.2, 0.4, and 0.6 as described in step (1-1), and the pulse charge and discharge experiment is shown in fig. 6. Fig. 7 and 8 are respectively an SOC estimation value and an estimation error curve based on an unscented particle filter algorithm under a pulse charge-discharge current experimental condition. It can be known that the closer the initial value is to the true value, the faster the convergence rate is; even if the initial set value and the true value have large errors, the correction and the iteration can quickly converge to the vicinity of a theoretical value after a period of time; under the working condition of pulse current, when the initial charge state is large, the theoretical value can be stably tracked through about 200s of adjustment; when the estimated value is stable, the estimation error is within 1.5%.
The method is used for further verifying the trackability of the unscented particle filter algorithm under the complex working condition. The invention takes the power demand value of the American small electric car under the UDDS working condition as the driving working condition, and then reduces the power demand value to the working condition of the single battery according to a certain proportion. And taking 2 times of UDDS cycle working condition as the charge-discharge test condition of the lithium ion battery pack, wherein the cycle period is 2792s, and comparing the unscented particle filter algorithm with EKF, UKF and EPF under the working condition for simulation analysis. The UDDS-based current under the cyclic condition is shown in fig. 9, and the sampling period is set to 1s in the experimental process.
Fig. 10 and 11 are respectively a lithium ion battery SOC estimation value and an estimation error curve obtained based on different filtering algorithms under the UDDS cycle condition. Fig. 10 shows that under the condition of unknown initial values, the estimation of the state of charge of the lithium ion battery by using the unscented particle filter algorithm is obviously superior to other estimation methods in the aspects of convergence speed and tracking accuracy. Under the conditions of complex UDDS cycle working conditions and large initial estimation error, the convergence time is about 250 s; from fig. 11, it can be known that the tracking accuracy of the algorithm after the estimated value is stable is less than 2.0% under the UDDS cycle condition.
The invention discloses a vehicle-mounted lithium ion battery state estimation method based on an improved genetic unscented Kalman filtering algorithm, which aims to accurately estimate the working state of a lithium ion battery, and has important significance in accurately establishing a battery management system and ensuring the safe, reliable and efficient operation of the lithium ion battery. According to the method, through a lithium ion battery charging and discharging experiment, an equivalent model is identified by using a Kalman filtering algorithm, a second-order RC equivalent circuit model is established, further, the State of Charge estimation (SOC estimation for short) of a lithium ion battery is carried out by using an unscented Kalman algorithm, and meanwhile, the system noise and observation noise covariance in the iteration process of the unscented Kalman filtering algorithm are optimized by using an improved genetic algorithm. Compared with the traditional method, the method provided by the invention solves the problem of serious non-linearization of system state variables caused by electrochemical reaction inside the lithium ion battery, and obviously improves the real-time performance and accuracy of estimation.
The above embodiments are only used for illustrating the design idea and features of the present invention, and the purpose of the present invention is to enable those skilled in the art to understand the content of the present invention and implement the present invention accordingly, and the protection scope of the present invention is not limited to the above embodiments. Therefore, all equivalent changes or modifications based on the principles and design concepts disclosed herein are intended to be included within the scope of the present invention.

Claims (5)

1. A vehicle-mounted lithium battery state estimation method based on improved genetic unscented Kalman filtering is characterized by comprising the following steps: it comprises the following steps:
s1, sampling the open-circuit voltage and current of the lithium battery under the condition of constant-current discharge of the lithium battery;
s2, establishing a state space equation of the lithium battery according to the second-order RC equivalent circuit model, and calculating an estimated value of the open-circuit voltage by using the state space equation by using the current sampled in S1; comparing the sampled open circuit voltage with the calculated open circuit voltage estimate;
s3, estimating and updating the second-order RC equivalent circuit model in real time by using a Kalman filtering algorithm and through the open-circuit voltage estimated value and the measured value calculated in S2;
s4, through constant current charging and discharging of lithium batteries with different SOC values, open-circuit voltage is measured after standing for a period of time, and the relation between the open-circuit voltage and the SOC of the lithium batteries is obtained through fitting;
s5, searching an optimal noise covariance matrix by using an improved genetic algorithm;
s6, estimating the state of the battery by adopting the optimal noise covariance matrix obtained in S5 and an unscented Kalman filtering algorithm through the open-circuit voltage estimated value obtained in S2 to obtain a current open-circuit voltage estimated value; then, outputting an SOC estimated value corresponding to the current open-circuit voltage estimated value according to the relation between the open-circuit voltage of the lithium battery and the SOC obtained in S4;
the S5 specifically includes:
5-1, selecting a chromosome coding mode, and randomly generating an initial population according to the selected coding mode;
5-2, judging whether a convergence condition is met, judging whether the iteration times are reached, and if the convergence condition is met, executing the step 5-6, and not executing the step 5-3;
5-3, selecting crossed individuals by using an inverse dichotomy:
dividing n individuals in the chromosome set into n/2 sets at random; performing crossover operation on chromosomes in the set to generate offspring chromosomes; randomly selecting two offspring chromosome sets for combination to generate n/4 combinations, and marking the set from which each chromosome comes; performing pairing crossing operation on chromosomes from different sets in the sets; randomly selecting two offspring chromosome sets for combination to generate n/8 combinations, and marking which set each chromosome comes from; continuously performing the operations according to the operation until the operations are finally combined into a set;
5-4, performing cross operation on the selected cross individuals by using an improved cross probability algorithm:
Figure FDA0003739381380000011
wherein, P e0 Taking the value of the reference cross probability as a reference, wherein the value is between 0.85 and 0.95 according to the actual condition; f best The fitness value of the optimal individual in the current population is obtained;
Figure FDA0003739381380000012
the current population average fitness value is obtained; f is the fitness value of the individual entering the cross-pairing operation;
5-5, selecting variation, and returning to 5-2;
5-6, outputting the optimal solution.
2. The vehicle-mounted lithium battery state estimation method based on the improved genetic unscented kalman filter according to claim 1, characterized in that: the S1 specifically includes:
selecting vehicle-mounted lithium battery packs with the same consistency to perform a charge-discharge experiment at normal temperature, and selecting the average voltage value of the lithium batteries as effective data;
preventing the lithium battery from being over-discharged, and setting a discharge cut-off voltage;
and setting the constant current discharge rate of the lithium battery, and sampling the open-circuit voltage and current of the lithium battery at a fixed sampling period.
3. The vehicle-mounted lithium battery state estimation method based on the improved genetic unscented Kalman filter as claimed in claim 1, characterized in that: the S3 specifically includes:
3-1, obtaining a system transfer function according to a second-order RC equivalent circuit model of the lithium battery by using kirchhoff's law;
3-2, carrying out bilinear transformation on the transfer function, and selecting a state equation and an observation equation under the condition of a fixed sampling period;
and 3-3, carrying out online estimation on the state variable through a Kalman filtering algorithm, and obtaining the identification value of each parameter in the second-order RC equivalent circuit model by utilizing the inverse transformation of z.
4. The vehicle-mounted lithium battery state estimation method based on the improved genetic unscented Kalman filter as claimed in claim 1, characterized in that: the S4 specifically includes:
4-1, fully discharging/charging the lithium battery and standing for a period of time;
4-2, performing an equidistant constant-current pulse charging/discharging experiment on the lithium battery, standing after each pulse experiment, and measuring the terminal voltage of the lithium battery;
4-3, obtaining the average voltage value of the lithium battery pack at each stage through multiple experiments;
and 4, obtaining the average voltage value of the lithium battery packs with the same consistency as the open-circuit voltage of a single lithium battery through experiments, and fitting a function.
5. The vehicle-mounted lithium battery state estimation method based on the improved genetic unscented Kalman filter as claimed in claim 1, characterized in that: the S6 specifically includes:
6-1, analyzing the electrochemical process of the battery, and listing a state equation and a measurement equation of the system;
6-2, calculating initial values of all state quantities;
6-3, establishing a Sigma point;
6-4, updating a state equation;
6-5, updating a measurement equation;
6-6, repeating the steps 6-2 to 6-5.
CN201910317198.9A 2019-04-19 2019-04-19 Vehicle-mounted lithium battery state estimation method based on improved genetic unscented Kalman filtering Active CN111856282B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910317198.9A CN111856282B (en) 2019-04-19 2019-04-19 Vehicle-mounted lithium battery state estimation method based on improved genetic unscented Kalman filtering

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910317198.9A CN111856282B (en) 2019-04-19 2019-04-19 Vehicle-mounted lithium battery state estimation method based on improved genetic unscented Kalman filtering

Publications (2)

Publication Number Publication Date
CN111856282A CN111856282A (en) 2020-10-30
CN111856282B true CN111856282B (en) 2022-08-26

Family

ID=72951916

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910317198.9A Active CN111856282B (en) 2019-04-19 2019-04-19 Vehicle-mounted lithium battery state estimation method based on improved genetic unscented Kalman filtering

Country Status (1)

Country Link
CN (1) CN111856282B (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112710939A (en) * 2020-12-18 2021-04-27 电子科技大学 Dynamic stability performance evaluation method of power device
CN113138344B (en) * 2021-04-23 2024-03-05 安徽理工大学 SOC estimation method based on fractional order extended Kalman filtering algorithm
CN113381700B (en) * 2021-06-29 2024-01-30 哈尔滨工业大学 RC parameter design method of high-frequency passive RC integral amplifier
CN116203432B (en) * 2023-03-23 2023-10-20 广东工业大学 CSO optimization-based unscented Kalman filtering method for predicting battery state of charge
CN116699415A (en) * 2023-05-26 2023-09-05 云储新能源科技有限公司 Method and system for estimating electric quantity of dynamic reconfigurable battery system and electronic equipment

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR3050278B1 (en) * 2016-04-15 2018-03-30 Saft METHOD FOR DETERMINING THE VALUE OF PARAMETERS RELATING TO THE STATUS OF AN BATTERY BATTERY BATTERY BATTERY AND ELECTRONIC BATTERY MANAGEMENT SYSTEM
CN106019164A (en) * 2016-07-07 2016-10-12 武汉理工大学 Lithium battery SOC estimation algorithm based on dual adaptive unscented Kalman filter
KR20180084358A (en) * 2017-01-17 2018-07-25 삼성전자주식회사 Method and apparatus for estimating state of battery
CN107390127A (en) * 2017-07-11 2017-11-24 欣旺达电动汽车电池有限公司 A kind of SOC estimation method
CN108872873A (en) * 2018-06-29 2018-11-23 长沙理工大学 A kind of lithium iron phosphate dynamic battery state-of-charge joint estimate method based on GA-AUKF

Also Published As

Publication number Publication date
CN111856282A (en) 2020-10-30

Similar Documents

Publication Publication Date Title
CN111856282B (en) Vehicle-mounted lithium battery state estimation method based on improved genetic unscented Kalman filtering
Xiong et al. A novel practical state of charge estimation method: an adaptive improved ampere‐hour method based on composite correction factor
CN112526348B (en) Battery model parameter identification method based on multi-innovation recursive Bayesian algorithm
CN112615075B (en) Battery quick charging method and computer equipment
CN105425153B (en) A kind of method of the state-of-charge for the electrokinetic cell for estimating electric vehicle
CN108732508B (en) Real-time estimation method for lithium ion battery capacity
Sidhu et al. State of charge estimation of lithium-ion batteries using hybrid machine learning technique
CN109839599B (en) Lithium ion battery SOC estimation method based on second-order EKF algorithm
CN108445422B (en) Battery state of charge estimation method based on polarization voltage recovery characteristics
CN111781504A (en) Method for identifying aging state and reconstructing open-circuit voltage of lithium ion power battery
Wang et al. Adaptive state-of-charge estimation method for an aeronautical lithium-ion battery pack based on a reduced particle-unscented kalman filter
CN111208438B (en) Method for cooperatively estimating residual capacity of lithium-ion battery and sensor deviation based on neural network and unscented Kalman filter
Tan et al. Joint estimation of ternary lithium-ion battery state of charge and state of power based on dual polarization model
CN111428433A (en) Lithium ion battery state calculation method based on hybrid filtering
CN115389936A (en) Online prediction method for continuous peak power capability of digital-analog hybrid driven lithium battery
CN113484771A (en) Method for estimating wide-temperature full-life SOC and capacity of lithium ion battery
CN115327415A (en) Lithium battery SOC estimation method based on limited memory recursive least square algorithm
CN115219918A (en) Lithium ion battery life prediction method based on capacity decline combined model
CN112698217B (en) Battery monomer capacity estimation method based on particle swarm optimization algorithm
CN113420444A (en) Lithium ion battery SOC estimation method based on parameter online identification
Li et al. State-of-charge estimation of lithium-ion battery using multi-state estimate technic for electric vehicle applications
CN115494400B (en) Lithium battery lithium separation state online monitoring method based on ensemble learning
CN112580289A (en) Hybrid capacitor power state online estimation method and system
CN114720881A (en) Lithium battery parameter identification method based on improved initial value forgetting factor recursive least square method
Liu et al. RBF network-aided adaptive unscented kalman filter for lithium-ion battery SOC estimation in electric vehicles

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
GR01 Patent grant
GR01 Patent grant