CN111563576A - Lithium battery capacity estimation method based on bat detection-extreme learning machine - Google Patents

Lithium battery capacity estimation method based on bat detection-extreme learning machine Download PDF

Info

Publication number
CN111563576A
CN111563576A CN202010299233.1A CN202010299233A CN111563576A CN 111563576 A CN111563576 A CN 111563576A CN 202010299233 A CN202010299233 A CN 202010299233A CN 111563576 A CN111563576 A CN 111563576A
Authority
CN
China
Prior art keywords
lithium battery
bat
fitness
learning machine
output
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
CN202010299233.1A
Other languages
Chinese (zh)
Inventor
葛东东
吴壮文
万志平
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.)
Zhejiang Industry Polytechnic College
Original Assignee
Zhejiang Industry Polytechnic College
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 Zhejiang Industry Polytechnic College filed Critical Zhejiang Industry Polytechnic College
Priority to CN202010299233.1A priority Critical patent/CN111563576A/en
Publication of CN111563576A publication Critical patent/CN111563576A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • 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]
    • G01R31/385—Arrangements for measuring battery or accumulator variables
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Computing Systems (AREA)
  • Artificial Intelligence (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Tests Of Electric Status Of Batteries (AREA)

Abstract

A lithium battery capacity estimation method based on bat detection-extreme learning machine belongs to the technical field of batteries. The present invention includes the following steps, S1: carrying out charge-discharge cycle working condition test on the lithium battery, recording test data, and determining an input variable and an output variable through sensitivity analysis; s2: forming a training set and a test set by using the input variable and the output variable; designing a bat detection algorithm and importing a training set for iterative optimization to obtain an optimal output weight; s3: calculating an input connection weight and a hidden layer neuron threshold value, and constructing a feedforward neural network structure extreme learning machine; s4: and importing the test set into the extreme learning machine constructed in the step S3 to estimate the capacity of the lithium battery, and evaluating the performance of the lithium battery capacity estimation. The method has good generalization capability, can effectively reduce the estimation error of the lithium battery capacity, and improves the estimation precision of the lithium battery capacity.

Description

Lithium battery capacity estimation method based on bat detection-extreme learning machine
Technical Field
The invention belongs to the technical field of batteries, and particularly relates to a lithium battery capacity estimation method based on a bat detection-extreme learning machine.
Background
The electric automobile industry is growing rapidly and has become a strategic emerging industry in China. The lithium battery has the advantages of light weight, large energy storage, large power and the like, and is the first choice of an ideal power source of the electric automobile. The capacity of the battery can be expressed as the capacity discharged from the battery in a full charge state to a cut-off voltage under a certain condition, namely a discharge current time integral value, which is also an important index for measuring the performance of the battery.
Lithium batteries generally suffer from capacity fade with increasing charge and discharge cycle times. The capacity of lithium batteries also varies with the use environment, aging, and the like. Therefore, inaccurate estimation of the capacity of the lithium battery can lead to inaccurate prediction of the remaining service life of the battery and also influence accurate estimation of the driving range of the electric vehicle.
At present, the battery capacity estimation mainly comprises methods of empirical estimation, estimation based on a parameter model, data driving and the like. The experience estimation generally carries out a life aging test firstly, a certain aging rule is searched, and a capacity attenuation curve is established to carry out the actual capacity estimation of the battery. The empirical estimation method is greatly influenced by the type of the lithium battery and the service environment of the battery. The estimation algorithm based on the parameter model has higher requirements on the accuracy of the battery model, and the inaccuracy of the model parameter identification directly influences the estimation accuracy of the actual capacity of the battery. At present, a method with good generalization performance, strong noise fault tolerance and high estimation precision is needed to estimate the capacity of the lithium battery.
Disclosure of Invention
The invention mainly solves the technical problems in the prior art and provides a lithium battery capacity estimation method based on a bat detection-extreme learning machine, which has good generalization capability and strong noise fault tolerance capability, can effectively reduce the battery capacity estimation error and improve the battery capacity estimation precision.
The technical problem of the invention is mainly solved by the following technical scheme: a lithium battery capacity estimation method based on a bat detection-extreme learning machine comprises the following steps:
s1: carrying out charge-discharge cycle working condition test on the lithium battery, recording test data, and determining an input variable and an output variable through sensitivity analysis;
s2: forming a training set and a test set by using the input variable and the output variable; designing a bat detection algorithm and importing a training set for iterative optimization to obtain an optimal output weight;
s3: calculating an input connection weight and a hidden layer neuron threshold value, and constructing a feedforward neural network structure extreme learning machine;
s4: and importing the test set into the extreme learning machine constructed in the step S3 to estimate the capacity of the lithium battery, and evaluating the performance of the lithium battery capacity estimation.
Preferably, in step S1, the test data includes a constant current charging time TIConstant voltage charging time TVTotal charging time Ttotal,charge△ V battery terminal voltage change value at charging instantchargeTotal discharge time Ttotal,dischargeAnd the battery temperature change value △ Temp in the discharging processdischarge△ V battery terminal voltage variation value at discharging momentdischargeAnd the actual remaining battery capacity Caged。
Preferably, in step S2, the input variables are normalized and then selected to form a training set of input data with a certain proportion, and the rest is a test set of input data; after normalization processing is carried out on the output variables, selecting the output variables to form an output data training set with a certain proportion, and taking the rest as an output data test set; the selection proportion of the input variables is the same as that of the output variables.
Preferably, step S2 includes the steps of:
s2-1: carrying out initialization parameter setting on the bat detection algorithm, wherein the parameters comprise a population size n and an initial pulse loudness Ai 0Initial pulse emissivity ri 0Maximum frequency range QmaxMinimum frequency range QminIteration end condition, implicitThe method comprises the following steps of (1) counting the number of layer neurons S1, the maximum iteration number N, the minimum Fitness threshold value and a Fitness evaluation function Fitness;
s2-2: calculating the frequency Q of each batiPosition SiVelocity ViFitness value FitnessiOutput weight βikSearching a bat with the minimum fitness and a group of corresponding output weights according to the following formula, namely, the current population optimal solution;
fmin＝min(Fitness)
Qi＝Qmin+(Qmin-Qmax)×rand
Figure BDA0002453352600000021
Figure BDA0002453352600000022
in the formula, rand ∈ [0,1 ]]Is a random vector, fminThe bat with the minimum fitness is BestS, namely the position corresponding to the bat with the minimum fitness is the current best position, t is the iteration number, and i is 1,2,3 … n;
s2-3: under random variation, calculating new frequency Q of each batiPosition SiAnd velocity Vi；
S2-4: judging whether the pulse emissivity of each bat meets the requirement, if so, turning to the next step, otherwise, turning to the step S2-3;
s2-5: calculating the Fitness value Fitness of each bat at a new positioniAnd output weights βik；
S2-6: judging the new Fitness value Fitness of each batiAnd loudness AiIf the requirements are met, turning to the next step if the requirements are met, or turning to the step S2-3;
s2-7: according to the new position S accepted by each batiCalculating the corresponding Fitness FitnessiAnd output weights βikAnd make loudness AiAnd pulse emissivity riUpdating of (1);
s2-8, finding the bat with the minimum fitness, moving all the bats to the position of the bat with the minimum fitness, updating the current optimal position BestS, the current optimal output weight Best β and the current minimum fitness fminI.e. the update of the current optimal solution;
s2-9: judging whether a termination condition is met, if so, saving the optimal output weight beta, otherwise, turning to the step S2-3, and performing iterative computation again; wherein, the termination condition is that the maximum iteration number N is reached, or the minimum fitness reaches a threshold value.
Preferably, in step S3, the input connection weights and hidden layer neuron thresholds are calculated by reshape function, where the function variables are the current best position BestS, the number of hidden layer neurons S1, the number of input layer neurons R, and the number of test set samples.
Preferably, in step S4, the performance evaluation index of the lithium battery capacity estimation includes an absolute error, a root mean square error, a mean absolute error percentage, a maximum absolute error, and a minimum absolute error between the lithium battery capacity estimation value and the actual capacity value of the lithium battery.
The invention has the following beneficial effects: according to the bat detection bionic principle, the iterative optimization process is designed to obtain the optimized output weight. And (4) constructing an extreme learning machine by calculating the input connection weight and the hidden layer neuron threshold value, thereby estimating the capacity of the lithium battery. The invention can process a complex nonlinear battery system, improves the generalization capability and the noise fault-tolerant capability, can effectively reduce the estimation error of the battery capacity and improve the estimation precision of the battery capacity.
Drawings
FIG. 1 is a schematic flow diagram of the present invention;
FIG. 2 is a graph showing the relationship between the actual capacity and the number of charge and discharge cycles of the lithium battery according to the present invention;
FIG. 3 is a comparison graph of an estimated lithium battery capacity value, an estimated lithium battery capacity value of a conventional Extreme Learning Machine (ELM), and an actual lithium battery capacity value according to the present invention;
FIG. 4 is a graph comparing the error of the present invention in estimating the capacity of a lithium battery with a conventional Extreme Learning Machine (ELM);
fig. 5 is a graph comparing the average absolute error and the root mean square error of the present invention with the conventional Extreme Learning Machine (ELM) in estimating the capacity of a lithium battery.
Detailed Description
The technical scheme of the invention is further specifically described by the following embodiments and the accompanying drawings.
Example (b): a lithium battery capacity estimation method based on a bat detection-extreme learning machine is shown in figure 1 and comprises the following steps:
s1: carrying out charge-discharge cycle working condition test on the lithium battery, recording test data, and determining an input variable and an output variable through sensitivity analysis;
s2: forming a training set and a test set by using the input variable and the output variable; designing a bat detection algorithm and importing a training set for iterative optimization to obtain an optimal output weight;
s3: calculating an input connection weight and a hidden layer neuron threshold value, and constructing a feedforward neural network structure extreme learning machine;
s4: and importing the test set into the extreme learning machine constructed in the step S3 to estimate the capacity of the lithium battery, and evaluating the performance of the lithium battery capacity estimation.
In step S1, the charge-discharge cycle condition test refers to a full-charge-discharge cycle test performed on a brand new lithium battery. The rated capacity of the lithium battery is 2AH, the charge cut-off voltage is 4.2V, the discharge cut-off voltage is 2.5V, and the ambient temperature is 24 ℃. The operation of the charge-discharge cycle working condition test is as follows: the lithium battery is charged with a constant current of 1.5A to a cut-off voltage of 4.2V, and then charged with a constant voltage of 4.2V until the current is reduced to 20 mA; after the lithium battery is fully charged, constant current discharge is carried out at a current of 2A until the cut-off voltage is 2.5V. Simultaneously recording test data, wherein the test data comprises input data and output data, and the input data comprises constant current charging time TIConstant voltage charging time TVTotal charging time Ttotal,charge△ V battery terminal voltage change value at charging instantchargeTotal time of dischargeTtotal,dischargeAnd the battery temperature change value △ Temp in the discharging processdischargeAnd △ V battery terminal voltage change value at discharging momentdischarge(ii) a The output data includes the actual remaining capacity C of the batteryaged。
In step S1, the sensitivity analysis means that a certain input variable is fixed and the other input variables are changed, and the degree of influence of the fixed input variable on the output result is calculated. All input variables are sorted by degree of influence. Through sensitivity analysis, five input variables (constant current charging time T) with large influence are obtainedIConstant voltage charging time TV△ V battery terminal voltage change value at charging instantchargeTotal discharge time Ttotal,dischargeAnd a battery temperature change value △ Temp during dischargingdischarge) As a final input variable, the actual remaining capacity C of the batteryagedAs an output variable, thereby reducing the amount of training data.
In step S2, the input variables are normalized by a normalization function, and then selected to form a training set of input data in a certain proportion, and the rest is a test set of input data; the output variables are firstly subjected to normalization processing through a normalization function and then selected to form an output data training set in a certain proportion, and the rest is an output data test set; the selection proportion of the input variables is the same as that of the output variables. The selection mode comprises sorting and then selecting in sequence, disordering and then randomly selecting and the like. Here, an input data training set P, an output data training set T, an input data test set tv.p, and an output data test set tv.t are formed by random selection after scrambling, and the selection ratio is 7:3, i.e., the ratio of the input data training set P to the input data test set tv.p, and the ratio of the output data training set T to the output data test set tv.t are all 7: 3.
In step S2, the method includes the steps of:
s2-1: setting initialization parameters of a bat detection algorithm, wherein the population size n is 10, and the initial pulse loudness A isi 00.8, initial pulse emissivity ri 0Is 0.1Maximum frequency range QmaxIs 1, minimum frequency range QminThe Fitness evaluation function Fitness is Fitness ═ sqrt (mse (T-Y));
s2-2: calculating the frequency Q of each batiPosition SiVelocity ViFitness value FitnessiOutput weight βikSearching a bat with the minimum fitness and a group of corresponding output weights according to the following formula, namely, the current population optimal solution;
fmin＝min(Fitness)
Qi＝Qmin+(Qmin-Qmax)×rand
Figure BDA0002453352600000041
Figure BDA0002453352600000042
in the formula, rand ∈ [0,1 ]]Is a random vector, fminThe bat with the minimum fitness is BestS, namely the position corresponding to the bat with the minimum fitness is the current best position, t is the iteration number, and i is 1,2,3 … n;
s2-3: under random variation, calculating new frequency Q of each batiPosition SiAnd velocity Vi；
S2-4: judging whether the pulse emissivity of each bat meets the requirement, if so, turning to the next step, otherwise, turning to the step S2-3; the requirement is satisfied that the pulse emissivity of the time is larger than the pulse emissivity of the last time, namely rand (r)>ri；
S2-5: calculating the Fitness value Fitness of each bat at a new positioniAnd output weights βik；
S2-6: judging the new Fitness value Fitness of each batiAnd loudness AiIf the requirements are met, turning to the next step if the requirements are met, or turning to the step S2-3; meet the requirementsThe value of the adaptability is smaller than the value of the adaptability last time, and the loudness of the time is smaller than the loudness of the last time, namely Fitnessinew<Fitnessiold，rand(A)<Ai；
S2-7: according to the new position S accepted by each batiCalculating the corresponding Fitness FitnessiAnd output weights βikAnd make loudness AiAnd pulse emissivity riUpdating of (1);
s2-8, finding the bat with the minimum fitness, moving all the bats to the position of the bat with the minimum fitness, updating the current optimal position BestS, the current optimal output weight Best β and the current minimum fitness fminI.e. the update of the current optimal solution;
s2-9: judging whether a termination condition is met, if so, saving the optimal output weight beta, otherwise, turning to the step S2-3, and performing iterative computation again; the termination condition may be that the maximum number of iterations N is reached, or that the minimum fitness reaches a threshold. Here, the termination condition is that the maximum number of iterations is reached. When the iteration is finished, the optimal output weight beta is as follows:
β＝[-0.0079 0.7251 -0.3232 0.6134 -0.1418 1.3757 0.3096 0.1380 -0.1523 0.9382 -1.1780 0.0491]T
in step S3, the input connection weight w and the hidden layer neuron threshold b are calculated by reshape function, where the function variables are the best current position BestS, the number of hidden layer neurons S1, the number of input layer neurons R, the number of test set samples, and so on.
Figure BDA0002453352600000051
In step S4, the performance evaluation index of the lithium battery capacity estimation includes an absolute error, a root mean square error, an average absolute error, a percentage of average absolute error, a maximum absolute error, and a minimum absolute error between the lithium battery capacity estimation value and the actual capacity value of the lithium battery.
The capacity of the lithium battery is estimated and compared with a traditional Extreme Learning Machine (ELM), wherein the number of hidden layer neuron nodes, a training set, a test set, an activation function of hidden layer neurons and other settings of the traditional extreme learning machine are consistent with those of the invention. The smaller the error, the higher the estimation accuracy of the lithium battery capacity.
Fig. 2 is a graph showing the relationship between the actual capacity and the number of charge and discharge cycles of the lithium battery according to the present invention.
Fig. 3 is a comparison graph of the estimated value of the capacity of the lithium battery according to the present invention, the estimated value of the capacity of the lithium battery of the conventional Extreme Learning Machine (ELM), and the actual capacity value of the lithium battery.
Fig. 4 is a comparison graph of the error of the lithium battery capacity estimation performed by the conventional Extreme Learning Machine (ELM) according to the present invention, and it can be seen that the error curve of the lithium battery capacity estimation of the present invention is closer to 0.
Fig. 5 is a graph comparing the average absolute error and the root mean square error of the present invention with the conventional Extreme Learning Machine (ELM) in estimating the capacity of a lithium battery. Through calculation, the performance evaluation index value of the lithium battery capacity estimation is as follows:
the average absolute error of the extreme learning machine is 0.2122%, the average absolute error of the traditional extreme learning machine is 1.6169%, and the average absolute error of the extreme learning machine is reduced by 86.88%;
the root mean square error of the extreme learning machine is 0.3159%, the root mean square error of the traditional extreme learning machine is 2.3109%, and the root mean square error is reduced by 86.33%;
the maximum absolute error of the extreme learning machine is 0.9346%, the maximum absolute error of the traditional extreme learning machine is 7.40%, and the maximum absolute error of the extreme learning machine is reduced by 87.37%.
In conclusion, the invention designs an iterative optimization flow according to the bat detection bionic principle to obtain an optimized output weight. And (4) constructing an extreme learning machine by calculating the input connection weight and the hidden layer neuron threshold value, thereby estimating the capacity of the lithium battery. The invention can process a complex nonlinear battery system, improves the generalization capability and the noise fault-tolerant capability, can effectively reduce the estimation error of the battery capacity and improve the estimation precision of the battery capacity.
It will be understood by those skilled in the art that all or part of the steps in the method for implementing the above embodiments may be implemented by relevant hardware instructed by a program, and the program may be stored in a computer-readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc.
Finally, it should be noted that the above embodiments are merely representative examples of the present invention. It is obvious that the invention is not limited to the above-described embodiments, but that many variations are possible. Any simple modification, equivalent change and modification made to the above embodiments in accordance with the technical spirit of the present invention should be considered to be within the scope of the present invention.

Claims (6)

1. A lithium battery capacity estimation method based on a bat detection-extreme learning machine is characterized by comprising the following steps:
s1: carrying out charge-discharge cycle working condition test on the lithium battery, recording test data, and determining an input variable and an output variable through sensitivity analysis;
s2: forming a training set and a test set by using the input variable and the output variable; designing a bat detection algorithm and importing a training set for iterative optimization to obtain an optimal output weight;
s3: calculating an input connection weight and a hidden layer neuron threshold value, and constructing a feedforward neural network structure extreme learning machine;
s4: and importing the test set into the extreme learning machine constructed in the step S3 to estimate the capacity of the lithium battery, and evaluating the performance of the lithium battery capacity estimation.
2. The method for estimating the capacity of a lithium battery based on a bat detection-extreme learning machine as claimed in claim 1, wherein in step S1, the test data comprises a constant current charging time TIConstant voltage charging time TVTotal charging time Ttotal,charge△ V battery terminal voltage change value at charging instantchargeTotal discharge time Ttotal,dischargeAnd the battery temperature change value △ Temp in the discharging processdischargeBattery terminal at the moment of dischargeVoltage variation value △ VdischargeAnd the actual remaining battery capacity Caged。
3. The method for estimating the capacity of a lithium battery based on a bat detection-extreme learning machine as claimed in claim 1, wherein in step S2, the input variables are normalized and then selected to form a training set of input data with a certain proportion, and the rest is a test set of input data; after normalization processing is carried out on the output variables, selecting the output variables to form an output data training set with a certain proportion, and taking the rest as an output data test set; the selection proportion of the input variables is the same as that of the output variables.
4. The method for estimating the capacity of the lithium battery based on the bat detection-limit learning machine as claimed in claim 1, wherein the step S2 comprises the steps of:
s2-1: carrying out initialization parameter setting on the bat detection algorithm, wherein the parameters comprise a population size n and an initial pulse loudness Ai 0Initial pulse emissivity ri 0Maximum frequency range QmaxMinimum frequency range QminIteration termination conditions, the number of hidden layer neurons S1, the maximum iteration number N, the minimum Fitness threshold value and a Fitness evaluation function Fitness;
s2-2: calculating the frequency Q of each batiPosition SiVelocity ViFitness value FitnessiOutput weight βikSearching a bat with the minimum fitness and a group of corresponding output weights according to the following formula, namely, the current population optimal solution;
fmin＝min(Fitness)
Qi＝Qmin+(Qmin-Qmax)×rand
Vi t＝Vi t-1+(Si t-1-BestS)×Qi
Si＝Vi t+Si t-1
in the formula, rand ∈ [0,1 ]]Is a random vector, fminThe bat with the minimum fitness is BestS, namely the position corresponding to the bat with the minimum fitness is the current best position, t is the iteration number, and i is 1,2,3 … n;
s2-3: under random variation, calculating new frequency Q of each batiPosition SiAnd velocity Vi；
S2-4: judging whether the pulse emissivity of each bat meets the requirement, if so, turning to the next step, otherwise, turning to the step S2-3;
s2-5: calculating the Fitness value Fitness of each bat at a new positioniAnd output weights βik；
S2-6: judging the new Fitness value Fitness of each batiAnd loudness AiIf the requirements are met, turning to the next step if the requirements are met, or turning to the step S2-3;
s2-7: according to the new position S accepted by each batiCalculating the corresponding Fitness FitnessiAnd output weights βikAnd make loudness AiAnd pulse emissivity riUpdating of (1);
s2-8, finding the bat with the minimum fitness, moving all the bats to the position of the bat with the minimum fitness, updating the current optimal position BestS, the current optimal output weight Best β and the current minimum fitness fminI.e. the update of the current optimal solution;
s2-9: judging whether a termination condition is met, if so, saving the optimal output weight beta, otherwise, turning to the step S2-3, and performing iterative computation again; wherein, the termination condition is that the maximum iteration number N is reached, or the minimum fitness reaches a threshold value.
5. The bat detection-extreme learning machine-based lithium battery capacity estimation method of claim 1, wherein in step S3, the input connection weights and hidden layer neuron thresholds are calculated by reshape function, wherein the function variables are current best position BestS, number of hidden layer neurons S1, number of input layer neurons R, and number of test set samples.
6. The method as claimed in claim 1, wherein in step S4, the performance evaluation indexes of the lithium battery capacity estimation include absolute error, root mean square error, average absolute error percentage, maximum absolute error, and minimum absolute error between the estimated lithium battery capacity value and the actual lithium battery capacity value.
CN202010299233.1A 2020-04-16 2020-04-16 Lithium battery capacity estimation method based on bat detection-extreme learning machine Pending CN111563576A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010299233.1A CN111563576A (en) 2020-04-16 2020-04-16 Lithium battery capacity estimation method based on bat detection-extreme learning machine

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010299233.1A CN111563576A (en) 2020-04-16 2020-04-16 Lithium battery capacity estimation method based on bat detection-extreme learning machine

Publications (1)

Publication Number Publication Date
CN111563576A true CN111563576A (en) 2020-08-21

Family

ID=72071611

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010299233.1A Pending CN111563576A (en) 2020-04-16 2020-04-16 Lithium battery capacity estimation method based on bat detection-extreme learning machine

Country Status (1)

Country Link
CN (1) CN111563576A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111476355A (en) * 2020-04-16 2020-07-31 浙江工业职业技术学院 Lithium battery SOC estimation method based on bat detection-extreme learning machine
CN113850016A (en) * 2021-08-16 2021-12-28 国网江苏省电力有限公司技能培训中心 Simulation transformer substation storage battery residual life prediction method in intermittent working mode

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103399276A (en) * 2013-07-25 2013-11-20 哈尔滨工业大学 Lithium-ion battery capacity estimation and residual cycling life prediction method
CN109163911A (en) * 2018-09-21 2019-01-08 昆明理工大学 A kind of fault of engine fuel system diagnostic method based on improved bat algorithm optimization ELM
CN110470477A (en) * 2019-09-19 2019-11-19 福州大学 A kind of Fault Diagnosis of Roller Bearings based on SSAE and BA-ELM

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103399276A (en) * 2013-07-25 2013-11-20 哈尔滨工业大学 Lithium-ion battery capacity estimation and residual cycling life prediction method
CN109163911A (en) * 2018-09-21 2019-01-08 昆明理工大学 A kind of fault of engine fuel system diagnostic method based on improved bat algorithm optimization ELM
CN110470477A (en) * 2019-09-19 2019-11-19 福州大学 A kind of Fault Diagnosis of Roller Bearings based on SSAE and BA-ELM

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
刘凯文;刘聪聪;李凯;王桂丽;张持健;: "基于 GA-ELM模型的锂电池 SOH预测", 无线电通信技术, vol. 45, no. 3, pages 248 - 252 *
耿银凤,: "基于极限学习机的脑卒中TCD数据分类研究", 《中国优秀硕士学位论文全文数据库信息科技辑》, no. 8, pages 18 - 21 *
耿银凤: "基于极限学习机的脑卒中 TCD数据分类研究", 中国优秀硕士学位论文全文数据库 信息科技辑, no. 8, pages 140 - 170 *
耿银凤: "基于极限学习机的脑卒中 TCD数据分类研究", 中国优秀硕士学位论文全文数据库信息科技辑, no. 8, pages 18 - 21 *
陈婕;金馨;: "基于极限学习机的锂电池健康状态预测", 软件, vol. 39, no. 2, pages 191 - 196 *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111476355A (en) * 2020-04-16 2020-07-31 浙江工业职业技术学院 Lithium battery SOC estimation method based on bat detection-extreme learning machine
CN113850016A (en) * 2021-08-16 2021-12-28 国网江苏省电力有限公司技能培训中心 Simulation transformer substation storage battery residual life prediction method in intermittent working mode
CN113850016B (en) * 2021-08-16 2024-04-05 国网江苏省电力有限公司技能培训中心 Method for predicting residual life of storage battery of simulated transformer substation in intermittent working mode

Similar Documents

Publication Publication Date Title
CN110632528B (en) Lithium battery SOH estimation method based on internal resistance detection
CN110888059B (en) Charge state estimation algorithm based on improved random forest combined volume Kalman
CN112018465B (en) Multi-physical-field-constrained intelligent quick charging method for lithium ion battery
CN110596593A (en) Lithium ion battery SOC estimation method based on intelligent adaptive extended Kalman filtering
CN111366848A (en) Battery health state prediction method based on PSO-ELM algorithm
CN110133507B (en) Battery remaining capacity estimation method based on NARX-UKF algorithm
CN107741568B (en) Lithium battery SOC estimation method based on state transition optimization RBF neural network
CN113702843B (en) Lithium battery parameter identification and SOC estimation method based on suburb optimization algorithm
CN113406521B (en) Lithium battery health state online estimation method based on feature analysis
CN114861527A (en) Lithium battery life prediction method based on time series characteristics
CN111458646A (en) Lithium battery SOC estimation method based on PSO-RBF neural network
CN113189490B (en) Lithium battery health state estimation method based on feature screening and Gaussian process regression
CN111856287A (en) Lithium battery health state detection method based on stacked residual causal convolutional neural network
CN111563576A (en) Lithium battery capacity estimation method based on bat detection-extreme learning machine
CN115308606B (en) Lithium ion battery health state estimation method based on adjacent characteristics
CN115389946A (en) Lithium battery health state estimation method based on isobaric rise energy and improved GRU
CN114740375A (en) Online lithium battery state of charge estimation method based on state estimation
CN110232432B (en) Lithium battery pack SOC prediction method based on artificial life model
CN117022048A (en) Electric automobile battery state of charge evaluation method
CN116736133A (en) Early prediction method for capacity degradation track of lithium ion battery in full life cycle
CN116029183A (en) Power battery temperature prediction method based on iPSO-LSTM model
CN114895206A (en) Lithium ion battery SOH estimation method based on RBF neural network of improved wolf optimization algorithm
CN114325393A (en) Lithium ion battery pack SOH self-adaptive estimation method based on PF and GPR
CN114896865A (en) Digital twin-oriented self-adaptive evolutionary neural network health state online prediction method
CN110515001B (en) Two-stage battery performance prediction method based on charging and discharging

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