CN115498283A - Lithium ion battery health state prediction method optimized by improved sparrow algorithm - Google Patents

Lithium ion battery health state prediction method optimized by improved sparrow algorithm Download PDF

Info

Publication number
CN115498283A
CN115498283A CN202211036227.2A CN202211036227A CN115498283A CN 115498283 A CN115498283 A CN 115498283A CN 202211036227 A CN202211036227 A CN 202211036227A CN 115498283 A CN115498283 A CN 115498283A
Authority
CN
China
Prior art keywords
sparrow
lithium ion
ion battery
population
algorithm
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
CN202211036227.2A
Other languages
Chinese (zh)
Inventor
李祖欣
侯佳雨
周哲
蔡志端
李会朋
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.)
Huzhou University
Original Assignee
Huzhou 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 Huzhou University filed Critical Huzhou University
Priority to CN202211036227.2A priority Critical patent/CN115498283A/en
Publication of CN115498283A publication Critical patent/CN115498283A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01M—PROCESSES OR MEANS, e.g. BATTERIES, FOR THE DIRECT CONVERSION OF CHEMICAL ENERGY INTO ELECTRICAL ENERGY
    • H01M10/00—Secondary cells; Manufacture thereof
    • H01M10/42—Methods or arrangements for servicing or maintenance of secondary cells or secondary half-cells
    • H01M10/425—Structural combination with electronic components, e.g. electronic circuits integrated to the outside of the casing
    • H01M10/4257—Smart batteries, e.g. electronic circuits inside the housing of the cells or batteries
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
    • 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]
    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01M—PROCESSES OR MEANS, e.g. BATTERIES, FOR THE DIRECT CONVERSION OF CHEMICAL ENERGY INTO ELECTRICAL ENERGY
    • H01M10/00—Secondary cells; Manufacture thereof
    • H01M10/05—Accumulators with non-aqueous electrolyte
    • H01M10/052—Li-accumulators
    • H01M10/0525—Rocking-chair batteries, i.e. batteries with lithium insertion or intercalation in both electrodes; Lithium-ion batteries
    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01M—PROCESSES OR MEANS, e.g. BATTERIES, FOR THE DIRECT CONVERSION OF CHEMICAL ENERGY INTO ELECTRICAL ENERGY
    • H01M10/00—Secondary cells; Manufacture thereof
    • H01M10/42—Methods or arrangements for servicing or maintenance of secondary cells or secondary half-cells
    • H01M10/425—Structural combination with electronic components, e.g. electronic circuits integrated to the outside of the casing
    • H01M2010/4271—Battery management systems including electronic circuits, e.g. control of current or voltage to keep battery in healthy state, cell balancing
    • 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
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E60/00—Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
    • Y02E60/10—Energy storage using batteries

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Software Systems (AREA)
  • Chemical & Material Sciences (AREA)
  • Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Electrochemistry (AREA)
  • Chemical Kinetics & Catalysis (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • General Chemical & Material Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Manufacturing & Machinery (AREA)
  • Materials Engineering (AREA)
  • Medical Informatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Microelectronics & Electronic Packaging (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a lithium ion battery health state prediction method for improving sparrow algorithm optimization, which comprises the following steps of: s1: acquiring training data to construct a training set test set; s2: extracting the characteristics of the historical data, and establishing a support vector machine model; s3: optimizing parameters of the support vector machine by using an improved sparrow algorithm to obtain optimized parameters; s4: constructing a support vector machine lithium ion battery health state prediction model optimized by an improved sparrow algorithm by using the optimized parameters; s5: and predicting the health state of the lithium ion battery by using the obtained battery health state prediction model. The prediction method can overcome the defect that the sparrow algorithm is easy to fall into a local optimal value, improve the accuracy of a support vector machine prediction model, reduce the error of lithium ion battery health state prediction and improve a battery management system.

Description

Lithium ion battery health state prediction method optimized by improved sparrow algorithm
Technical Field
The invention belongs to the field of lithium ion battery health state prediction, and particularly relates to a lithium ion battery health state prediction method for improving sparrow algorithm optimization.
Background
With the development of science and technology, lithium batteries have become mainstream, the demand of new energy automobiles, electric bicycles and energy storage on lithium batteries is continuously enhanced, and the lithium battery industry is in the gold phase of development. The lithium ion power battery is widely applied to the fields of energy supply of electronic products, electric automobiles and industrial equipment, aerospace and the like due to the advantages of high energy density, low self-discharge rate, reusability, no memory effect, long cycle life and the like. However, the performance of the battery is gradually reduced with the increase of the number of cycles, i.e., the aging of the battery is inevitable, which not only affects the normal operation, but also may cause catastrophic accidents. Therefore, it is necessary to predict the state of health of the lithium ion battery. A Battery Management System (BMS) can effectively supervise the power supply of a device, maintain its normal use, and ultimately avoid potential accidents. One of the main roles of the battery management system is to provide accurate knowledge about the internal state of the battery, such as its state of health (SOH). Since the SOH of a battery is an important index in the BMS, it must be clearly and concisely defined. The purpose of SOH is to demonstrate the degree of deterioration of the battery compared to a fresh battery. Thus, the SOH is defined as the ratio of the current maximum available capacity to the initial value and is expressed as:
Figure BDA0003819172480000011
wherein C is t Denotes the capacity of the t-th cycle, C o Indicating the initial capacity of the lithium ion battery. As the battery ages, the SOH curve exhibits a tendency to degrade overall due to continued charging and discharging.
At present, two methods for predicting the health state of a lithium ion battery are mainly used, namely, the method is based on a model and based on data driving, the model-based method mainly utilizes Kalman filtering, particle filtering and the like for modeling, but the model-based prediction method has the defects of large calculation amount and complex parameters and is difficult to be applied to practice, and the prediction method based on the data driving is opposite because the data driving is based on a black box principle, so that the complicated electrochemical behavior in the battery is not required to be solved, and the method can be predicted and widely used only according to historical data, such as a Support Vector Machine (SVM), a Relevance Vector Machine (RVM), a neural network and the like. Because the support vector machine is suitable for the wide study of small sample data, in the prediction process, the support vector machine needs to optimize the kernel parameters, and the commonly used method is a grid search method and a particle swarm optimization, but the grid search method needs longer time and consumes more time, and the particle swarm algorithm falls into a local optimal value, so that the health state prediction of the lithium ion battery by using the support vector machine still has limitations.
Disclosure of Invention
The invention aims to provide a method for predicting the health state of a lithium ion battery of a support vector machine by improving the optimization of a sparrow algorithm, which can overcome the defect that the sparrow algorithm falls into a local extreme value, improve the global search capability of the support vector machine, improve the accuracy of the prediction of the health state of the lithium ion battery and reduce the root mean square error of the prediction of the health state of the lithium ion battery.
In order to achieve the above object, a method for predicting the state of health of a lithium ion battery optimized by an improved sparrow algorithm is provided, which comprises the following steps:
s1: acquiring historical lithium ion data as training data, and constructing a training set and a test set;
s2: extracting the characteristics of the historical data, and establishing a support vector machine model;
s3: optimizing parameters of the support vector machine by using an improved sparrow algorithm to obtain optimized parameters;
the improved sparrow algorithm adopts a quasi-opponent learning strategy, so that the diversity of the initial sparrow population is increased, and the search range is expanded; an inertial weight factor is introduced into the position updating of the finder, and the position formula of the finder is improved, so that the finder has larger global search capability in the early stage of search, more accurate global search can be performed near the optimal position in the later stage, and the algorithm convergence speed is improved; the optimal position is disturbed by Gaussian variation, so that the solving precision of the algorithm is improved, and the local extreme value is avoided;
s4: constructing a lithium ion battery health state prediction model optimized by improving a sparrow algorithm by using the optimized parameters;
s5: and predicting the health state of the lithium ion battery by using the obtained battery health state prediction model. Preferably, the training data includes the time at which the discharge temperature of the B0005, B0006, B0007 batteries in the NASA dataset peaks, the temperature sample entropy of the iso-discharge voltage, the iso-discharge drop time, and the load-averaged voltage of discharge.
Preferably, the parameters of the support vector machine include a support vector machine setting type and a kernel parameter type.
Preferably, the objective function of the improved sparrow algorithm is designed as a root mean square error of a training set of a support vector machine, and the expression is as follows:
Figure BDA0003819172480000021
wherein, y predict Represents the predicted value of SOH of the lithium ion battery, y real Representing the real SOH value of the lithium ion battery, wherein n is the number of samples of the training set. Leading in a training set and a test set sample of lithium ion battery health state prediction before training, determining a support vector machine setting type, and selecting a linear kernel function type; and after the training is finished, comparing the data of the original training set and the original test set with the predicted health state data of the lithium ion battery to obtain the root mean square error of the training set and the root mean square error of the test set.
Preferably, the step S3 specifically includes the following steps:
s31: initializing parameters of an improved sparrow algorithm, including the sparrow population number pop, the finder number PD, the follower number SD and the population maximum iteration number Max iter Group alert person ST；
S32: a quasi-opponent learning strategy is fused, and the initial population diversity is increased;
s33: establishing a fitness function and sequencing;
s34: introducing an inertia weight factor and updating the position of the finder;
s35: updating the follower position;
s36: randomly selecting an alertor and updating the position of the alertor;
s37: calculating and sequencing the updated fitness value;
s38: disturbing the position of the optimal value by using Gaussian difference variation, and reordering;
s39: and (4) whether the stopping condition is met or not, exiting if the stopping condition is met, and outputting a result, otherwise, repeatedly executing the steps S32-S38.
Preferably, in step S32, the quasi-opponent learning strategy is:
for a real number x with a feasible field [ a, b ], the opposite number ox is defined as follows:
ox＝a+b-x(2)
however, the inverse learning only considers the opposites of a certain solution, and in order to increase the diversity of the population and ensure the global optimization capability, the initialized population is optimized by adopting a quasi-opposites learning strategy, and the quasi-opposites qox of x is defined as follows:
qox＝rand((a+b)/2,ox)
then M = [ M ] for D-sparrow individuals 1 ,M 2 ,M 3 ...M D ]Quasi-oppositional individuals
Figure BDA0003819172480000031
The value of the variable in the ith dimension is:
Figure BDA0003819172480000032
preferably, in step S34, the finder position update formula is:
Figure BDA0003819172480000033
Figure BDA0003819172480000034
wherein, w start Is an initial inertial weight, w end And t is the inertia weight when the iteration reaches the maximum number of times, and is the current iteration number. In general, the inertia weight w start ＝0.9，w end The best algorithm performance. Therefore, along with the iteration, the inertia weight is linearly reduced to 0.4 from 0.9, the larger inertia weight at the initial stage of the iteration enables the algorithm to keep stronger global search capability, and the smaller inertia weight at the later stage of the iteration is beneficial to more accurate local search of the algorithm; wherein
Figure BDA0003819172480000035
Is the jth dimension position value, iter, of the ith sparrow individual of the population at the tth iteration max For maximum number of iterations, R 2 ∈[0,1]Is a uniform random number which represents the early warning value of sparrow population, and ST belongs to the field of 0.5,1]Is the safety value of sparrow population, Q is obedience [0,1]A standard normally distributed random number; when R is 2 <ST, meaning that no natural enemy is found near the sparrow population, no danger exists around the sparrow population, and a finder enters an extensive search mode to search for food; when R is 2 >ST, to mean that some sparrows have detected natural enemies, found danger and sent early warnings to other sparrows, then the sparrow population will quickly fly to other safe areas to avoid danger at this time.
Preferably, in step S35, the location update formula of the follower is:
Figure BDA0003819172480000036
wherein X p Is the optimum position occupied by the finder at present, X worst Then the location of the current globally worst sparrow individual is indicated. A represents a 1 × d matrix, each of whichThe individual elements are randomly assigned a value of 1 or-1, and A + ＝A T (AA T ) -1 . When i is>n/2, this shows that the ith follower with lower fitness value does not obtain food, is in a state of full hunger, and needs to fly to other areas to forage to obtain more energy; when i is less than or equal to n/2, the method indicates that the ith follower with higher fitness value obtains food near the finder, and the follower is likely to compete with the finder for food, so that the role of the follower is changed into the finder.
Preferably, in step S36, the position update formula of the alert person is:
Figure BDA0003819172480000041
wherein, X best Is the global optimal position of the current sparrow population. Beta is taken as a step length control parameter and is a random number which follows normal distribution with the mean value of 0 and the variance of 1; f. of i Then is the fitness value of the current sparrow individual, f g And f w Respectively the current global best and worst fitness value. ε is the smallest constant to avoid zero at the denominator.
When f is i >f g This indicates that the sparrows are now at the edge of the population and are most likely to be attacked by the finder.
When f is i ＝f g This indicates that sparrows in the middle of the population were detected as dangerous and need to be approached to other sparrows to minimize the risk of predation. K is [ -1,1]Is a random number representing the direction of movement of the sparrow.
Preferably, in step S38, the gaussian difference variation position update formula is:
X(t+1)＝p 1 *f 1 *(X(t)-X(2,:))+p 2 *f 2 *(X rand -X(t)) (9)
wherein p1 and p2 are weight coefficients, values are both 0.5, f1 and f2 are Gaussian distribution random numbers with mean value of 0 and variance of 1, X (t) is the position of the current optimal individual, and X (2:) is the current timeThe position of the optimal sparrow body, X (t + 1) is the position of the original optimal sparrow body after disturbance, X rand Is a randomly selected sparrow location in the population.
The invention at least comprises the following beneficial effects:
firstly, the invention adopts the improved sparrow algorithm to carry out kernel parameter optimization on the support vector machine, thereby improving the data processing capability of the support vector machine.
Secondly, aiming at the problem that the sparrow algorithm is easy to fall into the local extreme value, the quasi-opponent learning strategy is adopted, the diversity of the initial sparrow population is increased, the search range is expanded, the problem that the sparrow algorithm is easy to fall into the local extreme value is effectively improved, and the optimization effect of the algorithm is improved.
And thirdly, when the sparrow algorithm carries out predation, an inertia weight factor is introduced to improve a position formula of a finder so that the finder has larger global search capability in the early stage of search and can carry out more accurate global search near the optimal position in the later stage, and the convergence speed of the algorithm is improved.
And fourthly, after the optimal position is searched by the sparrow algorithm, the optimal position is disturbed to avoid falling into a local extreme value, so that the optimal position is dynamically updated by utilizing a Gaussian difference variation strategy, and the solving precision of the algorithm is improved.
Fifthly, the health state of the lithium ion battery is predicted by adopting a support vector machine, the model training speed is high, and the prediction error can be reduced to a certain extent.
Drawings
Fig. 1 is a flow chart of a lithium ion battery health state prediction method optimized by an improved sparrow algorithm.
Fig. 2 is a flow chart of the sparrow algorithm.
Fig. 3 is a flow chart of the improved sparrow algorithm of the present invention.
FIG. 4 is a block diagram of a support vector machine according to the present invention.
Fig. 5 is a graph showing a comparison of the results when the predicted starting point of the B0005-pack battery data used in the present invention is 100.
Fig. 6 is a graph comparing the results when the predicted starting point of the B0006 battery data used in the present invention is 100.
Fig. 7 is a graph comparing the results when the predicted starting point of the B0007 battery data used in the present invention is 100.
Detailed Description
The present invention is further described in detail below with reference to the attached drawings so that those skilled in the art can implement the invention by referring to the description text.
Examples
As shown in fig. 1, a method for predicting the state of health of a lithium ion battery optimized by an improved sparrow algorithm includes the following steps:
s1: acquiring historical lithium ion data as training data, and constructing a training set and a test set;
s2: extracting the characteristics of the historical data, and establishing a support vector machine model;
s3: optimizing parameters of the support vector machine by using an improved sparrow algorithm to obtain optimized parameters;
the improved sparrow algorithm adopts a quasi-opponent learning strategy, so that the diversity of the initial sparrow population is increased, and the search range is expanded;
an inertial weight factor is introduced into the position updating of the finder, and the position formula of the finder is improved, so that the finder has larger global search capability in the early stage of search, more accurate global search can be performed near the optimal position in the later stage, and the algorithm convergence speed is improved; the optimal position is disturbed by Gaussian variation, so that the solving precision of the algorithm is improved, and the local extreme value is avoided;
s4: constructing a lithium ion battery health state prediction model optimized by improving a sparrow algorithm by using the optimized parameters;
s5: and predicting the health state of the lithium ion battery by using the obtained battery health state prediction model. Further, the training data includes the time at which the discharge temperature of the B0005, B0006, B0007 batteries in the NASA data set reaches the maximum, the temperature sample entropy of the equal discharge voltage, the equal discharge voltage drop time, and the load average voltage of the discharge.
Further, the parameters of the support vector machine include a support vector machine setting type and a kernel parameter type.
Further, the objective function of the improved sparrow algorithm is designed as a root mean square error of a training set of a support vector machine, and an expression of the objective function is as follows:
Figure BDA0003819172480000051
wherein, y predict Represents the predicted value of SOH of the lithium ion battery, y real Representing the real SOH value of the lithium ion battery, wherein n is the number of samples of the training set. Leading in a training set and a test set sample of lithium ion battery health state prediction before training, determining a support vector machine setting type, and selecting a linear kernel function type; and after the training is finished, comparing the data of the original training set and the original test set with the predicted health state data of the lithium ion battery to obtain the root mean square error of the training set and the root mean square error of the test set.
Further, step S3 specifically includes the following steps:
s31: initializing parameters of an improved sparrow algorithm, including the sparrow population number pop, the finder number PD, the follower number SD and the population maximum iteration number Max iter The population alert ST;
s32: a quasi-opponent learning strategy is fused, and the initial population diversity is increased;
s33: establishing a fitness function and sequencing;
s34: introducing an inertia weight factor and updating the position of the finder;
s35: updating the follower position;
s36: randomly selecting an alertor and updating the position of the alertor;
s37: calculating and sequencing the updated fitness value;
s38: disturbing the position of the optimal value by using Gaussian difference variation, and reordering;
s39: and (4) whether the stopping condition is met or not, exiting if the stopping condition is met, and outputting a result, otherwise, repeatedly executing the steps S32-S38.
The health condition of the lithium ion battery has a close relation with the discharge load voltage, the time of the discharge voltage reaching the highest point, the equal discharge voltage drop time and the temperature sample entropy of the equal discharge voltage, and generally speaking, as the aging of the lithium ion battery is increased, except that the temperature sample entropy of the equal discharge voltage is increased, the other characteristics are decreased along with the aging of the lithium ion battery.
The Sparrow Algorithm (SSA) mainly simulates a process of searching for food by a Sparrow population, divides the Sparrow population into a finder, a follower and a vigilant according to different proportions, and promotes the Sparrow population to gradually Search for an optimal solution, namely the position of the optimal food through iteration, and a flow chart of the Algorithm is shown in fig. 2.
The sparrow algorithm is originally proposed by Xue Jiankai, and has the advantages of high convergence speed, high global search capability and the like, so that the sparrow algorithm is widely applied, but still has local extreme values, and the problem that the local optimal solution cannot be found out is solved. The invention improves the algorithm as shown in fig. 3, and mainly comprises the following three points:
1. by adopting a quasi-opponent learning strategy, the diversity of the initial sparrow population is increased, the search range is expanded, the problem that the sparrow population is easy to fall into a local extreme value is effectively solved, and the algorithm optimization effect is improved.
2. An inertial weight factor is introduced into the position updating of the finder, and the position formula of the finder is improved, so that the finder has larger global search capability in the early stage of search, more accurate global search can be performed near the optimal position in the later stage, and the algorithm convergence speed is improved.
3. And the optimal position is disturbed by Gaussian variation, so that the solving precision of the algorithm is improved, and the local extreme value is avoided.
The quasi-opponent learning strategy is based on the inverse learning, and on the basis of the inverse learning, the opponent solution of the whole population is considered, so that the initialized population is optimized, and the quasi-opponent learning strategy is as follows:
ox＝a+b-x(2)
however, the inverse learning only considers the opposites of a certain solution, and in order to increase the diversity of the population and ensure the global optimization capability, the initialized population is optimized by adopting a quasi-opposites learning strategy, and the quasi-opposites qox of x is defined as follows:
qox＝rand((a+b)/2,ox)
then M = [ M ] for D-sparrow individuals 1 ,M 2 ,M 3 ...M D ]Quasi-oppositional individuals
Figure BDA0003819172480000061
The value of the variable in the ith dimension is as follows:
Figure BDA0003819172480000062
the inertia weight w embodies the ability of the particle to inherit the previous speed, shi.Y firstly introduces the inertia weight w into the PSO algorithm, and analyzes that a larger inertia weight is beneficial to the global search ability, and a smaller inertia weight is beneficial to the local search.
Specifically, the finder location update formula is:
Figure BDA0003819172480000063
Figure BDA0003819172480000064
wherein, w start Is an initial inertial weight, w end And t is the inertia weight when the iteration reaches the maximum number, and is the current iteration number. In general, the inertia weight w start ＝0.9，w end Algorithm performance is best when = 0.4. Therefore, along with the iteration, the inertia weight is linearly reduced to 0.4 from 0.9, the larger inertia weight at the initial stage of the iteration enables the algorithm to keep stronger global search capability, and the smaller inertia weight at the later stage of the iteration is beneficial to more accurate local search of the algorithm; wherein
Figure BDA0003819172480000071
Is the jth dimension position value, iter, of the ith sparrow individual of the population at the tth iteration max For maximum number of iterations, R 2 ∈[0,1]Is a uniform random number which represents the early warning value of sparrow population, and ST belongs to the field of 0.5,1]Is the safety value of sparrow population, Q is obedience [0,1]A standard normally distributed random number; when R is 2 <ST, meaning that no natural enemy is found near the sparrow population, no danger exists around the sparrow population, and a finder enters an extensive search mode to search for food; when R is 2 >ST, to mean that some sparrows have detected natural enemies, found danger and sent early warnings to other sparrows, then the sparrow population will quickly fly to other safe areas to avoid danger at this time.
Specifically, the location update formula of the follower is:
Figure BDA0003819172480000072
wherein, X p Is the optimum position occupied by the finder at present, X worst Then the location of the current globally worst sparrow individual is indicated. A represents a 1 × d matrix in which each element is randomly assigned a value of 1 or-1, and A + ＝A T (AA T ) -1 . When i is>n/2, this shows that the ith follower with lower fitness value does not obtain food, is in a state of full hunger, and needs to fly to other areas to forage to obtain more energy; when i is less than or equal to n/2, the method indicates that the ith follower with higher fitness value obtains food near the finder, and the follower is likely to compete with the finder for food, so that the role of the follower is changed into the finder.
Specifically, the position update formula of the alert person is as follows:
Figure BDA0003819172480000073
wherein, X best Is the current sparrowGlobal optimal position of the population. Beta is taken as a step length control parameter and is a random number which follows normal distribution with the mean value of 0 and the variance of 1; f. of i It is the fitness value of the current sparrow individual. f. of g And f w Respectively the current global best and worst fitness value. ε is the smallest constant to avoid zero at the denominator.
When f is i >f g This indicates that the sparrow at this point is at the edge of the population and is highly likely to be attacked by the finder.
When f is i ＝f g This indicates that the sparrow in the middle of the population was detected as dangerous and needs to be approached to other sparrows to minimize the risk of predation. K is [ -1,1]Is a random number which represents the direction of the movement of the sparrows and is a step size control coefficient.
Specifically, the gaussian difference variation position updating formula is:
X(t+1)＝p 1 *f 1 *(X(t)-X(2,:))+p 2 *f 2 *(X rand -X(t)) (9)
wherein p1 and p2 are weight coefficients, values are both 0.5, f1 and f2 are Gaussian distribution random numbers with a mean value of 0 and a variance of 1, X (t) is the position of the current optimal individual, X (2) is the position of the current suboptimal individual, X (t + 1) is the position of the original optimal sparrow individual after being disturbed, X +1 is the position of the original optimal sparrow individual rand Is a randomly selected sparrow location in the population.
The SOH prediction of the lithium ion battery is performed by using the improved sparrow algorithm, where the table is a prediction accuracy index table when the prediction starting point of the three groups of battery data (B0005, B0006, B0007) of the present invention is 100, fig. 5 is a result map when the prediction starting point of the B0005 group of battery data is 100, fig. 6 is a result map when the prediction starting point of the B0006 group of battery data is 100, and fig. 7 is a result map when the prediction starting point of the B0007 group of battery data is 100.
Table-three groups of batteries prediction accuracy index table by different algorithms
Figure BDA0003819172480000081
A Support Vector Machine (SVM) is a machine learning method based on the statistical learning theory (STL) created by Vapnik. The statistical learning theory adopts a structure risk minimization criterion, minimizes the structure risk while minimizing the error of the sample point, improves the generalization capability of the model, and has no limit of data dimension. When the SVM is applied to the regression fitting analysis, the basic idea is to find an optimal classification surface so that the error of all training samples from the optimal classification surface is minimized, and the structure of the SVM is shown in fig. 4. For the support vector machine, a kernel function needs to be determined to be a linear kernel function in a training stage, C is a penalty factor, the larger C is, the larger the penalty is to a sample with large training error is, overfitting is easy, and the smaller C is, underfitting is easy.
Although the embodiments of the present invention have been described with reference to the accompanying drawings, it is not intended to limit the scope of the present invention, and it should be understood by those skilled in the art that various modifications and variations can be made without inventive efforts by those skilled in the art based on the technical solution of the present invention.

Claims (10)

1. A lithium ion battery health state prediction method optimized by an improved sparrow algorithm is characterized by comprising the following steps:
s1: acquiring historical lithium ion data as training data, and constructing a training set and a test set;
s2: extracting the characteristics of the historical data, and establishing a support vector machine model;
s3: optimizing parameters of the support vector machine by using an improved sparrow algorithm to obtain optimized parameters;
the improved sparrow algorithm adopts a quasi-opponent learning strategy, so that the diversity of the initial sparrow population is increased, and the search range is expanded;
an inertia weight factor is introduced into the position updating of the finder, and a position formula of the finder is improved, so that the finder has larger global search capability in the early stage of search, more accurate global search can be performed near the optimal position in the later stage, and the convergence speed of the algorithm is improved; the optimal position is disturbed by Gaussian variation, so that the solving precision of the algorithm is improved, and the local extreme value is avoided;
s4: constructing a lithium ion battery health state prediction model optimized by improving a sparrow algorithm by using the optimized parameters;
s5: and predicting the health state of the lithium ion battery by using the obtained battery health state prediction model.
2. The improved sparrow algorithm optimized lithium ion battery state of health prediction method of claim 1, wherein the training data comprises time to peak discharge temperature of B0005, B0006, B0007 batteries in NASA dataset, temperature sample entropy of equal discharge voltage, equal discharge voltage drop time, and load average voltage of discharge.
3. The improved sparrow algorithm optimized lithium ion battery state of health prediction method of claim 1, wherein the parameters of the support vector machine comprise a support vector machine setting type and a kernel parameter type.
4. The improved sparrow algorithm optimized lithium ion battery state of health prediction method according to claim 1, wherein the objective function of the improved sparrow algorithm is designed as a root mean square error of a training set of a support vector machine, and the expression is as follows:
Figure FDA0003819172470000011
wherein, y predict Represents the predicted value of SOH of the lithium ion battery, y real Representing the real SOH value of the lithium ion battery, wherein n is the number of samples of the training set; leading in a training set and a test set sample of lithium ion battery health state prediction before training, determining a support vector machine setting type, and selecting a linear kernel function type; after the training is completed, willAnd comparing the data of the original training set and the test set with the health state data of the lithium ion battery obtained through prediction to obtain the root mean square error of the training set and the root mean square error of the test set.
5. The improved sparrow algorithm optimized lithium ion battery state of health prediction method according to claim 1, wherein step S3 comprises the following steps:
s31: initializing parameters of an improved sparrow algorithm, including the sparrow population number pop, the finder number PD, the follower number SD and the population maximum iteration number Max iter A population early warning value ST;
s32: a quasi-opponent learning strategy is fused, and the initial population diversity is increased;
s33: establishing a fitness function and sequencing;
s34: introducing an inertia weight factor and updating the position of the finder;
s35: updating the position of the follower;
s36: randomly selecting an alertor and updating the position of the alertor;
s37: calculating and sequencing the updated fitness value;
s38: disturbing the position of the optimal value by using Gaussian difference variation, and reordering;
s39: and (4) whether the stopping condition is met or not, exiting if the stopping condition is met, and outputting a result, otherwise, repeatedly executing the steps S32-S38.
6. The improved sparrow algorithm optimized lithium ion battery state of health prediction method of claim 5, wherein in step S32, the quasi-opponent learning strategy is:
for a real number x with a feasible field [ a, b ], the opposite number ox is defined as follows:
ox＝a+b-x (2)
however, the inverse learning only considers the opposites of a certain solution, and in order to increase the diversity of the population and ensure the global optimization capability, the initialized population is optimized by adopting a quasi-opposites learning strategy, and the quasi-opposites qox of x is defined as follows:
qox＝rand((a+b)/2,ox)
then M = [ M ] for D-sparrow individuals 1 ,M 2 ,M 3 ...M D ]Quasi-oppositional individuals
Figure FDA0003819172470000021
The value of the variable in the ith dimension is as follows:
Figure FDA0003819172470000022
7. the improved sparrow algorithm optimized lithium ion battery state of health prediction method of claim 6, wherein in step S34, the discoverer location update formula is:
Figure FDA0003819172470000023
Figure FDA0003819172470000024
wherein w start Is an initial inertial weight, w end The inertia weight when the iteration reaches the maximum number of times, and t is the current iteration number;
Figure FDA0003819172470000025
is the j-dimension position value, iter, of the ith sparrow individual of the population at the t-th iteration max For maximum number of iterations, R 2 ∈[0,1]Is a uniform random number which represents the early warning value of sparrow population, and ST belongs to the field of 0.5,1]Is the safety value of sparrow population, Q is obedience [0,1]A standard normally distributed random number; when R is 2 <ST, meaning that no natural enemy is found near the sparrow population, no danger exists around the sparrow population, and a finder enters an extensive search mode to search for food; when R is 2 >ST time, meanSmelling that some sparrows have detected natural enemies, found danger and signaled other sparrows to remind groups of sparrows to fly to other safe areas to avoid the danger.
8. The improved sparrow algorithm optimized lithium ion battery state of health prediction method of claim 7, wherein in step S35, the position update formula of the follower is:
Figure FDA0003819172470000026
wherein, X p Is the optimal position occupied by the finder at present, X worst Then the position of the current global worst sparrow individual is represented; a represents a 1 × d matrix in which each element is randomly assigned a value of 1 or-1, and A + ＝A T (AA T ) -1 (ii) a When i is>n/2, this shows that the ith follower with lower fitness value does not obtain food, is in a state of full hunger, and needs to fly to other areas to forage to obtain more energy; when i is less than or equal to n/2, the method indicates that the ith follower with higher fitness value obtains food near the finder, and the follower is likely to compete with the finder for food, so that the role of the follower is changed into the finder.
9. The improved sparrow algorithm optimized li-ion battery state of health prediction method of claim 8, wherein in step S36, the position update formula of the alerter is:
Figure FDA0003819172470000031
wherein, X best Is the global optimal position of the current sparrow population; beta is taken as a step length control parameter and is a random number which follows normal distribution with the mean value of 0 and the variance of 1; f. of i The fitness value of the current sparrow individual is obtained; f. of g And f w The current global best and worst fitness values, respectively; ε is the smallest constant to avoid zero denominator;
when f is i >f g This indicates that the sparrows are at the edge of the population and are most likely to be attacked by the discoverer;
when f is i ＝f g In time, this indicates that sparrows in the middle of the population have detected danger and need to approach other sparrows to minimize the risk of being preyed; k E [ -1,1]Is a random number representing the orientation of the movement of the sparrow.
10. The improved sparrow algorithm optimized lithium ion battery state of health prediction method of claim 9, wherein in step S38, the gaussian difference variation position update formula:
X(t+1)＝p 1 *f 1 *(X(t)-X(2,:))+p 2 *f 2 *(X rand -X(t)) (9)
wherein p1 and p2 are weight coefficients, values are both 0.5, f1 and f2 are Gaussian distribution random numbers with a mean value of 0 and a variance of 1, X (t) is the position of the current optimal individual, X (2) is the position of the current suboptimal individual, X (t + 1) is the position of the original optimal sparrow individual after being disturbed, X +1 is the position of the original optimal sparrow individual rand Is a randomly selected sparrow location in the population.
CN202211036227.2A 2022-08-27 2022-08-27 Lithium ion battery health state prediction method optimized by improved sparrow algorithm Pending CN115498283A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211036227.2A CN115498283A (en) 2022-08-27 2022-08-27 Lithium ion battery health state prediction method optimized by improved sparrow algorithm

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211036227.2A CN115498283A (en) 2022-08-27 2022-08-27 Lithium ion battery health state prediction method optimized by improved sparrow algorithm

Publications (1)

Publication Number Publication Date
CN115498283A true CN115498283A (en) 2022-12-20

Family

ID=84466030

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211036227.2A Pending CN115498283A (en) 2022-08-27 2022-08-27 Lithium ion battery health state prediction method optimized by improved sparrow algorithm

Country Status (1)

Country Link
CN (1) CN115498283A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115983134A (en) * 2023-01-17 2023-04-18 淮阴工学院 Battery power state prediction method and system based on neural network
CN116449218A (en) * 2023-04-26 2023-07-18 淮阴工学院 Lithium battery health state estimation method
CN117278338A (en) * 2023-11-23 2023-12-22 江苏君立华域信息安全技术股份有限公司 Network intrusion detection method and system based on deep learning optimization
CN117423067A (en) * 2023-12-18 2024-01-19 成都华芯智云科技有限公司 Passenger flow statistics terminal based on TOF technology

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115983134A (en) * 2023-01-17 2023-04-18 淮阴工学院 Battery power state prediction method and system based on neural network
CN115983134B (en) * 2023-01-17 2024-02-23 淮阴工学院 Battery power state prediction method and system based on neural network
CN116449218A (en) * 2023-04-26 2023-07-18 淮阴工学院 Lithium battery health state estimation method
CN116449218B (en) * 2023-04-26 2024-03-12 淮阴工学院 Lithium battery health state estimation method
CN117278338A (en) * 2023-11-23 2023-12-22 江苏君立华域信息安全技术股份有限公司 Network intrusion detection method and system based on deep learning optimization
CN117423067A (en) * 2023-12-18 2024-01-19 成都华芯智云科技有限公司 Passenger flow statistics terminal based on TOF technology
CN117423067B (en) * 2023-12-18 2024-03-12 成都华芯智云科技有限公司 Passenger flow statistics terminal based on TOF technology

Similar Documents

Publication Publication Date Title
CN115498283A (en) Lithium ion battery health state prediction method optimized by improved sparrow algorithm
Huang et al. Convolutional gated recurrent unit–recurrent neural network for state-of-charge estimation of lithium-ion batteries
Ren et al. A review of machine learning state-of-charge and state-of-health estimation algorithms for lithium-ion batteries
CN106600059B (en) Intelligent power grid short-term load prediction method based on improved RBF neural network
Li et al. A merged fuzzy neural network and its applications in battery state-of-charge estimation
CN113156325B (en) Method for estimating health state of battery
CN108537337A (en) Lithium ion battery SOC prediction techniques based on optimization depth belief network
CN113064093A (en) Energy storage battery state of charge and state of health joint estimation method and system
Ahmad et al. Feature and instance selection via cooperative PSO
CN109917292A (en) A kind of lithium ion battery life-span prediction method based on DAUPF
Peng et al. An improved PSO algorithm for battery parameters identification optimization based on Thevenin battery model
CN110705794A (en) Method for predicting window state based on support vector machine algorithm
CN115219906A (en) Multi-model fusion battery state of charge prediction method and system based on GA-PSO optimization
CN115808633A (en) POA-SVR-based method for monitoring state of energy storage battery in island environment
Zou et al. Advancements in Artificial Neural Networks for health management of energy storage lithium-ion batteries: A comprehensive review
Gong et al. Early prediction of cycle life for lithium-ion batteries based on evolutionary computation and machine learning
Savargaonkar et al. Uncorrelated sparse autoencoder with long short-term memory for state-of-charge estimations in lithium-ion battery cells
CN114487890A (en) Lithium battery health state estimation method for improving long-term and short-term memory neural network
Wu et al. Remaining useful life prediction of Lithium-ion batteries based on PSO-RF algorithm
Ren et al. A novel multiple training-scale dynamic adaptive cuckoo search optimized long short-term memory neural network and multi-dimensional health indicators acquisition strategy for whole life cycle health evaluation of lithium-ion batteries
CN116029183A (en) Power battery temperature prediction method based on iPSO-LSTM model
Singhal et al. An efficient approach for availability analysis through fuzzy differential equations and particle swarm optimization
Zhou et al. Combined sparse Bayesian learning strategy for remaining useful life forecasting of lithium-ion battery
Li et al. Hybrid Data-Driven Approach for Predicting the Remaining Useful Life of Lithium-Ion Batteries
CN116796624A (en) Self-adaptive incremental battery RUL prediction method combining multi-source degradation characteristics

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