CN112307667A - Method and device for estimating state of charge of storage battery, electronic equipment and storage medium - Google Patents
Method and device for estimating state of charge of storage battery, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN112307667A CN112307667A CN202011173783.5A CN202011173783A CN112307667A CN 112307667 A CN112307667 A CN 112307667A CN 202011173783 A CN202011173783 A CN 202011173783A CN 112307667 A CN112307667 A CN 112307667A
- Authority
- CN
- China
- Prior art keywords
- charge
- state
- population
- storage battery
- network
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 83
- 238000013528 artificial neural network Methods 0.000 claims abstract description 65
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 53
- 238000012549 training Methods 0.000 claims abstract description 35
- 239000002245 particle Substances 0.000 claims abstract description 28
- 230000000739 chaotic effect Effects 0.000 claims abstract description 25
- 230000007246 mechanism Effects 0.000 claims abstract description 16
- 238000004364 calculation method Methods 0.000 claims description 12
- 238000013507 mapping Methods 0.000 claims description 8
- 230000035772 mutation Effects 0.000 claims description 5
- 238000004590 computer program Methods 0.000 claims description 4
- 230000002068 genetic effect Effects 0.000 claims description 4
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 7
- 230000006870 function Effects 0.000 description 29
- 238000005457 optimization Methods 0.000 description 10
- 230000003044 adaptive effect Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 238000004891 communication Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 2
- GNFTZDOKVXKIBK-UHFFFAOYSA-N 3-(2-methoxyethoxy)benzohydrazide Chemical compound COCCOC1=CC=CC(C(=O)NN)=C1 GNFTZDOKVXKIBK-UHFFFAOYSA-N 0.000 description 1
- FGUUSXIOTUKUDN-IBGZPJMESA-N C1(=CC=CC=C1)N1C2=C(NC([C@H](C1)NC=1OC(=NN=1)C1=CC=CC=C1)=O)C=CC=C2 Chemical compound C1(=CC=CC=C1)N1C2=C(NC([C@H](C1)NC=1OC(=NN=1)C1=CC=CC=C1)=O)C=CC=C2 FGUUSXIOTUKUDN-IBGZPJMESA-N 0.000 description 1
- 239000002253 acid Substances 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 238000007599 discharging Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 230000008707 rearrangement Effects 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000010845 search algorithm Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/25—Design optimisation, verification or simulation using particle-based methods
-
- 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]
- G01R31/378—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC] specially adapted for the type of battery or accumulator
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- 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
- G06N3/086—Learning methods using evolutionary algorithms, e.g. genetic algorithms or genetic programming
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Geometry (AREA)
- Computer Hardware Design (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Bioinformatics & Computational Biology (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Data Mining & Analysis (AREA)
- Biomedical Technology (AREA)
- Physiology (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Secondary Cells (AREA)
Abstract
The invention discloses a method and a device for estimating the state of charge of a storage battery, electronic equipment and a storage medium. The method comprises the following steps: coding the weight and threshold parameters of the neural network into a coding string, and generating an initial population based on a chaotic mechanism, wherein p is the size of the initial population; setting the encoding information of the population individuals into an initial weight and a threshold of a neural network for network training, and solving a training result by using a fitness function to obtain a fitness value of each individual; sorting population individuals according to fitness values, taking k individuals with the top rank, and performing iterative updating by adopting a particle swarm algorithm; carrying out iterative updating on the remaining p-k individuals by adopting a differential evolution algorithm; setting the coding information of the last generation of population individuals into an initial weight and a threshold of a neural network, and carrying out network training to obtain a network prediction model; and estimating the state of charge of the storage battery according to the network prediction model. The invention realizes the effect of more accurately estimating the state of charge of the storage battery.
Description
Technical Field
The embodiment of the invention relates to a state of charge estimation technology of a storage battery, in particular to a state of charge estimation method and device of the storage battery, electronic equipment and a storage medium.
Background
With the continuous popularization and use of Valve Regulated Lead Acid batteries (VRLA), which have become the most important part of backup dc power supplies for various power systems, the state of charge calculation of the VRLA Battery is one of several main parameters representing the performance and health conditions of the Battery, and has a highly nonlinear relationship with many factors, so how to accurately estimate the state of charge calculation of the Battery is always the key point and difficulty of research in related fields.
The neural network has strong self-adaptability and self-learning capability, adopts a parallel processing structure, and does not need a specific mathematical model, so that the state of charge of the VRLA storage battery is generally calculated by using a neural network algorithm at present.
However, the direct use of the Back Propagation Neural Network (BPNN) algorithm is very likely to fall into local optimality during the training process, which results in an inaccurate estimation result.
Disclosure of Invention
The invention provides a method and a device for estimating the state of charge of a storage battery, electronic equipment and a storage medium, which are used for realizing more accurate calculation of the state of charge of the storage battery.
In a first aspect, an embodiment of the present invention provides a method for estimating a state of charge of a battery, where the method for estimating the state of charge of the battery includes:
step 1, coding weight values and threshold parameters of a neural network into coding strings, and generating an initial population based on a chaotic mechanism, wherein p is the size of the initial population;
step 4, repeatedly executing the steps 2 and 3 on the population obtained by iterative updating until the population is updated to an algebraic threshold value;
and 6, estimating the state of charge of the storage battery according to the network prediction model.
Optionally, the encoding the weight and the threshold parameter of the neural network into a coding string, and generating the initial population based on a chaos mechanism includes:
the network weight and the threshold parameter are coded into a coding string as an individual in the mixed evolution algorithm group, and the length of the individual variable is as follows: θ is mxl + l + l × n + n;
introducing chaotic Logistic mapping to generate an initial population P, wherein the scale is P;
in the formula: mu is a control parameter; i is the serial number of the chaotic variable, i is 1, 2. r denotes the population serial number, r 1, 2.., p; alpha is alphai rExpressing the chaotic variable, and assigning theta initial values alpha with small difference to the formula (1)i 0；
Substituting the obtained theta chaotic variables into the following formula for inverse mapping to obtain an individual variable xr iSequentially taking r ═ 1, 2.., p, and generating an initial population with the scale of p;
in the formula: [ x ] ofi min,xi max]Is a variable xr iThe value range of (2).
Optionally, the fitness function F (y)i,oi) Taking the sum of absolute errors between the estimated value of the state of charge of the network output storage battery and the expected value, and taking an absolute value, wherein the formula is as follows:
in the formula: eta is a coefficient; n is the number of nodes of the network output layer; y isiOutputting a state of charge expected value of a storage battery for the ith node of the network; oiThe state of charge estimate of the output battery for the ith node of the network.
Optionally, the performing iterative update by using a particle swarm algorithm includes:
step a), initializing k individual populations with the top rank;
step b) calculating first fitness values of k individuals at the top of the rank according to the fitness function;
step c) updating the speed and position of the particles;
step d) judging whether the first fitness value is smaller than a fitness value threshold value;
if not, returning to the step b).
Optionally, the performing iterative update by using a differential evolution algorithm includes:
step a) selecting a coding strategy and determining a genetic strategy;
step b) generating an initial population by using the remaining p-k individuals in the population;
step c) calculating a second fitness value of the remaining individuals according to the fitness function;
step d) judging whether the second fitness value is smaller than the fitness value threshold value;
if not, carrying out mutation, intersection and selection, and returning to the step c).
Optionally, after estimating the state of charge of the storage battery according to the network prediction model, the method further includes:
and comparing the relative error value of the state of charge of the storage battery calculated by the state of charge estimation method of the storage battery with the relative error value of the state of charge of the storage battery calculated by only adopting a neural network method.
Optionally, after estimating the state of charge of the storage battery according to the network prediction model, the method further includes:
and comparing the evolution algebra of the state of charge of the storage battery calculated by the method for estimating the state of charge of the storage battery with the evolution algebra of the state of charge of the storage battery estimated by only adopting a neural network method.
In a second aspect, an embodiment of the present invention further provides a state of charge estimation device for a storage battery, where the state of charge estimation device for a storage battery includes:
the initial population generation module is used for coding the weight and the threshold parameter of the neural network into a coding string and generating an initial population based on a chaotic mechanism, wherein p is the size of the initial population;
the fitness value calculation module is used for setting the encoding information of the population individuals into the initial weight and the threshold of the neural network for network training, and solving the training result by using a fitness function to obtain the fitness value of each individual;
the iterative updating module is used for sequencing the population individuals according to the fitness value, taking k individuals which are ranked at the top, and performing iterative updating by adopting a particle swarm algorithm; carrying out iterative updating on the remaining p-k individuals by adopting a differential evolution algorithm;
the algebra judgment module is used for repeatedly executing the fitness value calculation module and the iterative update module on the population obtained by iterative update until the population is updated to an algebra threshold value;
the prediction model establishing module is used for setting the coding information of the last generation of optimal individuals into the initial weight and the threshold value of the network, and performing network training to obtain a network prediction model;
and the state of charge estimation module is used for estimating the state of charge of the storage battery according to the network prediction model.
In a third aspect, an embodiment of the present invention further provides an electronic device for estimating a state of charge of a battery, where the electronic device for estimating a state of charge of a battery includes:
one or more processors;
storage means for storing one or more programs;
when the one or more programs are executed by the one or more processors, the one or more processors implement the method for estimating the state of charge of a battery according to the first aspect.
In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, which stores a computer program, and the computer program, when executed by a processor, implements the method for estimating the state of charge of a storage battery according to the first aspect.
The method comprises the steps of constructing a neural network structure, introducing a chaos mechanism to generate an initial population, obtaining the fitness value of each individual in the population according to a fitness function, conducting goodness and badness sequencing on the fitness values, taking k individuals with the top rank, conducting iterative updating by adopting a particle swarm algorithm, conducting iterative updating on the remaining p-k individuals in the population by adopting a differential evolution method until the population is updated to an algebraic threshold value, setting the coding information of the last generation of population individuals into the initial weight and the threshold value of the neural network, conducting network training to obtain a network prediction model, and estimating the charge state of the storage battery by adopting the network prediction model. The problem that the estimation result is not accurate enough by only adopting a neural network is solved, and the effect of more accurately estimating the charge state of the storage battery is realized.
Drawings
Fig. 1 is a flowchart of a state of charge estimation method for a storage battery according to an embodiment of the present invention;
FIG. 2 is a flow chart of a particle swarm algorithm according to an embodiment of the present invention;
FIG. 3 is a flowchart of a differential evolution algorithm according to an embodiment of the present invention;
fig. 4 is a flowchart of a state of charge estimation method for a storage battery according to a second embodiment of the present invention;
fig. 5 is a graph comparing the relative error value of the state of charge of the storage battery calculated by the state of charge estimation method of the storage battery according to the second embodiment of the present invention with the relative error value of the state of charge of the storage battery estimated by using only the neural network method;
fig. 6 is a comparison graph of an evolution algebra of the state of charge of the storage battery calculated by using the state of charge estimation method of the storage battery according to the second embodiment of the present invention and an evolution algebra of the state of charge of the storage battery estimated by using only the neural network method;
fig. 7 is a schematic structural diagram of a state of charge estimation device for a storage battery according to a third embodiment of the present invention;
fig. 8 is a schematic structural diagram of an electronic device for estimating a state of charge of a battery according to a fourth embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.
Example one
Fig. 1 is a flowchart of a state of charge estimation method for a storage battery according to an embodiment of the present invention, where the present embodiment is applicable to a state of charge calculation of the storage battery, and the method may be executed by a state of charge estimation device for the storage battery, and specifically includes the following steps:
s110, coding the weight and threshold parameters of the neural network into a coding string, and generating an initial population based on a chaos mechanism, wherein p is the size of the initial population.
Neural networks are essentially simplified models of the human brain's way of processing information, and work by simulating a large number of simple processing units connected to each other. The topological structure of the neural network can be determined according to the number of the input and output variables, then the network weight and the threshold parameter are coded into a coding string according to the determined network structure, a chaos mechanism is introduced to generate an initial population, and the initial population comprises p individuals. The chaos mechanism is that the result is used as input again after the input is amplified and some small random changes are added, and the cycle is continued, and the result in such a system cannot be predicted.
The method comprises the following steps of coding a weight value and a threshold parameter of a neural network into a coding string, and generating an initial population based on a chaotic mechanism, wherein the coding string comprises the following steps:
the network weight and the threshold parameter are coded into a coding string as an individual in the mixed evolution algorithm group, and the length of the individual variable is as follows: θ is mxl + l + l × n + n;
introducing chaotic Logistic mapping to generate an initial population P, wherein the scale is P;
in the formula: mu is a control parameter; i is the serial number of the chaotic variable, i is 1, 2. r denotes the population serial number, r 1, 2.., p; alpha is alphai rExpressing the chaotic variable, and assigning theta initial values alpha with small difference to the formula (1)i 0；
Substituting the obtained theta chaotic variables into the following formula for inverse mapping to obtain an individual variable xr iSequentially taking r ═ 1, 2.., p, and generating an initial population with the scale of p;
in the formula: [ x ] ofi min,xi max]Is a variable xr iThe value range of (2).
Thus, an initial population of size p can be generated using a neural network.
S120, setting the encoding information of the population individuals into the initial weight and the threshold of the neural network for network training, and solving the training result by using a fitness function to obtain the fitness value of each individual.
The encoding information of the population individuals is used as the initial weight and the threshold of the neural network, so that the neural network can be continuously trained, the network parameters (the weight and the threshold) corresponding to the minimum error are determined through repeated learning and training, and the training is stopped immediately. And solving the training result by using the fitness function to obtain the fitness value of each individual. The selection of the fitness function directly influences the convergence speed of the algorithm and whether the optimal solution can be found, and the algorithm basically does not utilize external information in evolutionary search, only takes the fitness function as a basis, and utilizes the fitness of each individual in a population to search. Because the complexity of the fitness function is a major component of the algorithm complexity, the fitness function should be designed as simply as possible, minimizing the computational time complexity.
Wherein the fitness function F (y)i,oi) Taking the sum of absolute errors between the estimated value of the state of charge of the network output storage battery and the expected value, and taking an absolute value, wherein the formula is as follows:
in the formula: eta is coefficient(ii) a N is the number of nodes of the network output layer; y isiOutputting a state of charge expected value of a storage battery for the ith node of the network; oiThe state of charge estimate of the output battery for the ith node of the network.
S130, sorting the population individuals according to the fitness value, taking k individuals with the top rank, and performing iterative updating by adopting a particle swarm algorithm; and carrying out iterative updating on the remaining p-k individuals by adopting a differential evolution algorithm.
The particle swarm optimization is a biological heuristic method in the field of computational intelligence, and belongs to a swarm intelligence optimization algorithm. The Differential Evolution Algorithm (DE) is an efficient global optimization Algorithm. It is also a group-based heuristic search algorithm, with each individual in the group corresponding to a solution vector. The population individuals are sorted according to the fitness value, namely, the population individuals are sorted according to the fitness value, k individuals which are ranked at the top are taken, iterative updating is carried out by adopting a particle swarm algorithm, and optimization of the k individuals is realized. And carrying out iterative updating on the remaining p-k individuals by adopting a differential evolution algorithm to realize the optimization of the remaining p-k individuals.
Referring to fig. 2, fig. 2 is a flowchart of a particle swarm algorithm provided in the first embodiment of the present invention, where the iterative updating performed by using the particle swarm algorithm includes S1311 to S1315, and S1311 to S1315 specifically include:
s1311, initializing k individual populations with the top rank in step a);
s1312, step b) calculating first fitness values of k individuals at the top of the rank according to the fitness function;
s1313, updating the speed and the position of the particles in the step c);
s1314, step d) judging whether the first fitness value is smaller than the fitness value threshold value;
s1315 and step e), if not, returning to step b).
Specifically, iteration updating is carried out on k individuals with the top rank by adopting a particle swarm algorithm, a population of the k individuals is initialized, first fitness values of the k individuals are calculated according to a fitness function, the speed and the positions of the particles are updated, the size relation between the first fitness values and a fitness value threshold is judged after the first fitness values are calculated, if the first fitness values are smaller than the fitness value threshold, optimization is completed, iteration is finished, and if the first fitness values are not smaller than the fitness value threshold, the first fitness values of the k individuals are continuously calculated according to the fitness function until the first fitness values are smaller than the fitness value threshold. Therefore, iterative updating of k individuals with the top rank is achieved by using a particle swarm algorithm, and optimization of a network prediction model is achieved.
Referring to fig. 3, fig. 3 is a flowchart of a differential evolution algorithm provided in an embodiment of the present invention, where the iterative updating performed by using the differential evolution algorithm includes S1321 to S1325, and S1321 to S1325 specifically include:
s1321, selecting a coding strategy in the step a), and determining a genetic strategy;
s1322, step b) utilizing the remaining p-k individuals in the population to generate an initial population;
s1323, calculating the fitness value of the remaining individuals according to the fitness function in the step c);
s1324, step d) judges whether the second fitness value is smaller than the fitness value threshold value;
s1325, if not, carrying out mutation, intersection and selection, and returning to the step c).
Specifically, iterative updating is carried out on the remaining p-k individuals by adopting a differential evolution algorithm, a coding strategy is selected firstly, the remaining p-k individuals are utilized to generate an initial population, the p-k individuals are calculated according to a fitness function to obtain a second fitness value of each individual, the size relation between the second fitness value and a fitness value threshold is judged, if the second fitness value is smaller than the fitness value threshold, optimization is completed, and iteration is finished. And if the second fitness value is not smaller than the fitness value threshold, performing mutation, intersection and selection. The variation, crossing and selection means that variation and crossing operations are performed to obtain an intermediate population, individuals are selected from the original population and the intermediate population to obtain a new generation population, and the new generation population is continuously calculated by using a fitness function until the second fitness value is smaller than a fitness value threshold. Therefore, iterative updating of the remaining p-k individuals is achieved by using a differential evolution algorithm, and optimization of a network prediction model is achieved.
And S140, repeating the steps S120 and S130 on the population obtained by iterative updating until the population is updated to an algebraic threshold value.
Specifically, steps S120 and S130 are repeatedly performed on the population obtained by iterative update, and iterative update and optimization are continuously performed until the population is updated to the algebraic threshold, where the algebraic threshold may be input in advance, and the algebraic threshold may be set according to an actual situation.
S150, setting the coding information of the last generation of population individuals into the initial weight and the threshold of the neural network, and carrying out network training to obtain a network prediction model.
Specifically, the encoding information of the iteratively updated last generation population individuals is set as the initial weight and the threshold of the neural network, and network training is continuously performed to optimize the network model and obtain the network prediction model.
And S160, estimating the state of charge of the storage battery according to the network prediction model.
The state of charge (SOC) of the battery refers to a ratio of a remaining capacity of the battery after being used for a certain period of time or left unused for a long time to a capacity of a fully charged state thereof, and is usually expressed as a percentage. The state of charge of the battery ranges from 0 to 1, indicating that the battery is fully discharged when the SOC is 0, and indicating that the battery is fully charged when the SOC is 1. And finally, estimating the state of charge of the storage battery according to the obtained network prediction model, wherein the network prediction model is an optimized model, so that the state of charge of the storage battery can be estimated more accurately.
According to the technical scheme, a neural network structure is built, a chaos mechanism is introduced to generate an initial population, the fitness value of each individual in the population is obtained according to a fitness function, the fitness values are ranked in good or bad mode, k individuals with the top rank are selected, a particle swarm algorithm is adopted to conduct iterative updating, the remaining p-k individuals in the population are conducted iterative updating through a differential evolution method until the population is updated to an algebraic threshold value, therefore, the coding information of the population of the last generation is set to be the initial weight and the threshold value of the neural network, network training is conducted, a network prediction model is obtained, and the charge state of the storage battery can be estimated through the network prediction model. The problem that the estimation result is not accurate enough by only adopting a neural network is solved, and the effect of more accurately estimating the charge state of the storage battery is realized.
Example two
Fig. 4 is a flowchart of a state of charge estimation method for a storage battery according to a second embodiment of the present invention, where the present embodiment is applicable to a state of charge calculation of the storage battery, and the method may be executed by a state of charge estimation device for the storage battery, and referring to fig. 4, the method specifically includes the following steps:
s210, coding the weight and threshold parameters of the neural network into a coding string, and generating an initial population based on a chaos mechanism, wherein p is the size of the initial population.
S220, setting the encoding information of the population individuals into the initial weight and the threshold of the neural network for network training, and solving the training result by using a fitness function to obtain the fitness value of each individual.
S230, sorting the population individuals according to the fitness value, taking k individuals with the top rank, and performing iterative updating by adopting a particle swarm algorithm; and carrying out iterative updating on the remaining p-k individuals by adopting a differential evolution algorithm.
And S240, repeating the steps S220 and S230 on the population obtained by iterative updating until the population is updated to an algebraic threshold value.
And S250, setting the coding information of the population individuals of the last generation into an initial weight and a threshold of the neural network, and carrying out network training to obtain a network prediction model.
And S260, estimating the state of charge of the storage battery according to the network prediction model.
And S270, comparing the relative error value of the state of charge of the storage battery calculated by the state of charge estimation method of the storage battery with the relative error value of the state of charge of the storage battery calculated by only adopting a neural network method.
Referring to fig. 5, fig. 5 is a comparison graph of the relative error value of the state of charge of the storage battery calculated by using the state of charge estimation method of the storage battery according to the second embodiment of the present invention and the relative error value of the state of charge of the storage battery calculated by using only the neural network method, a curve 2 in fig. 5 is the relative error value of the state of charge of the storage battery calculated by using the state of charge estimation method of the storage battery, and a curve 1 in fig. 5 is the relative error value of the state of charge of the storage battery calculated by using only the neural network method. In the whole discharging process, the change of the relative error value of the state of charge of the storage battery calculated by the state of charge estimation method of the storage battery is more stable and is kept within 2%, the maximum change is 1.27%, the change range of the relative error value of the state of charge of the storage battery calculated by only adopting a neural network method is large, and the maximum relative error reaches 5.243%, which shows that the estimation of the state of charge of the storage battery by the state of charge estimation method of the storage battery provided by the embodiment has better accuracy and stability.
S280, comparing the evolution algebra of the charge state of the storage battery calculated by the charge state estimation method of the storage battery with the evolution algebra of the charge state of the storage battery estimated by only adopting a neural network method.
Referring to fig. 6, fig. 6 is a comparison graph of the evolution algebra of the state of charge of the storage battery calculated by using the state of charge estimation method of the storage battery according to the second embodiment of the present invention and the evolution algebra of the state of charge of the storage battery calculated by using only the neural network method, a curve 2 in fig. 6 is a relative error value of the state of charge of the storage battery calculated by using the state of charge estimation method of the storage battery, and a curve 1 in fig. 6 is a relative error value of the state of charge of the storage battery calculated by using only the neural network method. Along with population evolution, when the charge state of the storage battery is calculated by using a charge state estimation method of the storage battery, the average fitness value of the individual is gradually reduced, the fact that the initial weight and the threshold represented by the individual code of each generation of population are closer to the optimal value is reflected, and when the population evolves to the 14 th generation, the average fitness reaches the optimal value. The fitness average value repeatedly trained only by the neural network method fluctuates greatly, and is higher than the fitness average value in the state of charge estimation method using the storage battery provided by the embodiment as a whole. This shows that the method for estimating the state of charge using the storage battery according to the present embodiment has an obvious effect of optimizing the initial values of the threshold and the weight of the neural network, and can avoid falling into local optimum.
According to the technical scheme of the embodiment, a neural network structure is constructed, the fitness value of each individual in a population is obtained according to a fitness function, the fitness values are ranked according to the quality, k individuals with the top rank are selected, iterative updating is carried out by adopting a particle swarm algorithm, the remaining p-k individuals in the population are iteratively updated by adopting a differential evolution method until the population is updated to an algebraic threshold value, therefore, the coding information of the population of the last generation is set to be the initial weight and the threshold value of the neural network, network training is carried out, a network prediction model is obtained, and the charge state of the storage battery can be estimated by adopting the network prediction model. The relative error value of the state of charge of the storage battery calculated by the state of charge estimation method of the storage battery is compared with the relative error value of the state of charge of the storage battery calculated by only adopting a neural network method, and the comparison result shows that the estimation method of the state of charge of the storage battery provided by the embodiment has better accuracy and stability for the estimation of the state of charge of the storage battery. The evolution algebra of the state of charge of the storage battery calculated by the method for estimating the state of charge of the storage battery is compared with the evolution algebra of the state of charge of the storage battery estimated by only adopting a neural network method, and the comparison result shows that the method for estimating the state of charge of the storage battery has obvious effect of optimizing the initial values of the threshold and the weight of the neural network, and can avoid falling into local optimum. The technical scheme of the embodiment solves the problem that the estimation result is not accurate enough only by adopting the neural network, and realizes the effect of more accurately estimating the state of charge of the storage battery.
EXAMPLE III
Fig. 7 is a schematic structural diagram of a state of charge estimation device for a storage battery according to a third embodiment of the present invention, where the third embodiment is applicable to a state of charge calculation of the storage battery, and the state of charge estimation device for the storage battery includes:
the initial population generation module 110 is configured to encode the weight and the threshold parameter of the neural network into a coding string, and generate an initial population based on a chaotic mechanism, where p is the size of the initial population;
the initial population generation module 110 is specifically configured to encode the network weight and the threshold parameter into a coding string as an individual in the hybrid evolution algorithm population, where the individual variable length is: θ is mxl + l + l × n + n;
introducing chaotic Logistic mapping to generate an initial population P, wherein the scale is P;
in the formula: mu is a control parameter; i is the serial number of the chaotic variable, i is 1, 2. r denotes the population serial number, r 1, 2.., p; alpha is alphai rExpressing the chaotic variable, and assigning theta initial values alpha with small difference to the formula (1)i 0；
Substituting the obtained theta chaotic variables into the following formula for inverse mapping to obtain an individual variable xr iSequentially taking r ═ 1, 2.., p, and generating an initial population with the scale of p;
in the formula: [ x ] ofi min,xi max]Is a variable xr iThe value range of (2).
A fitness value calculation module 120, configured to set the encoding information of the population individuals to the initial weight and the threshold of the neural network, perform network training, and solve the training result by using a fitness function to obtain a fitness value of each individual;
wherein the fitness function F (y)i,oi) Taking the sum of absolute errors between the estimated value of the state of charge of the network output storage battery and the expected value, and taking an absolute value, wherein the formula is as follows:
in the formula: eta is a coefficient; n is the number of nodes of the network output layer; y isiOutputting a state of charge expected value of a storage battery for the ith node of the network; oiThe state of charge estimate of the output battery for the ith node of the network.
The iterative updating module 130 is configured to sort the population individuals according to the fitness value, take k individuals ranked at the top, and perform iterative updating by using a particle swarm algorithm; carrying out iterative updating on the remaining p-k individuals by adopting a differential evolution algorithm;
the iterative updating module 130 further includes a particle swarm algorithm executing module 131 and a differential evolution algorithm executing module 132;
the particle swarm algorithm executing module 131 is specifically configured to execute steps a) -e), where steps a) -e) specifically include:
step a), initializing k individual populations with the top rank;
step b) calculating first fitness values of k individuals at the top of the rank according to a fitness function;
step c) updating the speed and position of the particles;
step d) judging whether the first adaptive value is smaller than an adaptive value threshold value;
if not, returning to the step b).
The differential evolution algorithm execution module 132 is specifically configured to execute steps f) -j), where the steps f) -j) specifically include:
step f) selecting a coding strategy and determining a genetic strategy;
step g) generating an initial population by using the remaining p-k individuals in the population;
step h) calculating a second fitness value of the remaining individuals according to the fitness function;
step i) determining whether the second adaptive value is less than the adaptive value threshold;
and j) if not, performing mutation, intersection and selection, and returning to the step h).
An algebra judgment module 140, configured to repeatedly execute the fitness value calculation module and the iterative update module on the population obtained through the iterative update until the population is updated to an algebra threshold;
the prediction model establishing module 150 is used for setting the coding information of the last generation of optimal individuals into the initial weight and the threshold value of the network, and performing network training to obtain a network prediction model;
and the state of charge estimation module 160 is used for estimating the state of charge of the storage battery according to the network prediction model.
Optionally, the device for estimating the state of charge of the storage battery further includes a relative error comparison module 170, configured to compare the relative error value of the state of charge of the storage battery calculated by the device for estimating the state of charge of the storage battery with the relative error value of the state of charge of the storage battery calculated by using only the neural network method.
Optionally, the device for estimating the state of charge of the storage battery further includes an evolution algebra comparison module 180, configured to compare the evolution algebra of the state of charge of the storage battery calculated by the device for estimating the state of charge of the storage battery with the evolution algebra of the state of charge of the storage battery estimated only by using a neural network method.
The state of charge estimation device of the storage battery provided in this embodiment is a method for estimating the state of charge of the storage battery according to the foregoing embodiment, and the implementation principle and technical effect of the state of charge estimation device of the storage battery provided in this embodiment are similar to those of the foregoing embodiment, and are not described here again.
Example four
Fig. 8 is a schematic structural diagram of an electronic device for estimating a state of charge of a battery according to a fourth embodiment of the present invention, as shown in fig. 8, the electronic device includes a processor 40, a memory 41, and a communication interface 42; the number of the processors 40 in the battery state of charge estimation electronic device may be one or more, and one processor 40 is taken as an example in fig. 8; the processor 40, the memory 41, and the communication interface 42 in the state-of-charge estimation electronic device of the storage battery may be connected by a bus or other means, and fig. 8 illustrates the connection by the bus as an example. A bus represents one or more of any of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, and a processor or local bus using any of a variety of bus architectures.
The memory 41, which is a computer-readable storage medium, may be configured to store software programs, computer-executable programs, and modules, such as program instructions/modules corresponding to the methods in the embodiments of the present invention. The processor 40 executes at least one functional application of the battery state of charge estimation electronic device and data processing by executing software programs, instructions and modules stored in the memory 41, so as to implement the method.
The memory 41 may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created from use of the storage battery charge state estimation electronic device, and the like. Further, the memory 41 may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device. In some examples, memory 41 may include memory remotely located from processor 40, which may be connected to the state of charge estimation electronics of the battery via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The communication interface 42 may be configured for the reception and transmission of data.
EXAMPLE five
An embodiment of the present invention further provides a storage medium containing computer-executable instructions, which when executed by a computer processor, perform a method for estimating a state of charge of a battery, the method including:
step 1, coding weight values and threshold parameters of a neural network into coding strings, and generating an initial population based on a chaotic mechanism, wherein p is the size of the initial population;
step 4, repeatedly executing the steps 2 and 3 on the population obtained by iterative updating until the population is updated to an algebraic threshold value;
and 6, estimating the state of charge of the storage battery according to the network prediction model.
Of course, the storage medium provided by the embodiment of the present invention contains computer-executable instructions, and the computer-executable instructions are not limited to the operations of the method described above, and may also perform related operations in the method for estimating the state of charge of the storage battery provided by any embodiment of the present invention.
From the above description of the embodiments, it is obvious for those skilled in the art that the present invention can be implemented by software and necessary general hardware, and certainly, can also be implemented by hardware, but the former is a better embodiment in many cases. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as a floppy disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a FLASH Memory (FLASH), a hard disk or an optical disk of a computer, and includes several instructions for enabling a computer device (which may be a personal computer, a server, or a network device) to execute the methods according to the embodiments of the present invention.
It should be noted that, in the embodiment of the state of charge estimation device for a storage battery, the units and modules included in the embodiment are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be realized; in addition, specific names of the functional units are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present invention.
It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.
Claims (10)
1. A method of estimating a state of charge of a battery, comprising:
step 1, coding weight values and threshold parameters of a neural network into coding strings, and generating an initial population based on a chaotic mechanism, wherein p is the size of the initial population;
step 2, setting the encoding information of the population individuals into the initial weight and the threshold of the neural network for network training, and solving a training result by using a fitness function to obtain the fitness value of each individual;
step 3, sorting the population individuals according to the fitness value, taking k individuals with the top rank, and performing iterative updating by adopting a particle swarm algorithm; carrying out iterative updating on the remaining p-k individuals by adopting a differential evolution algorithm;
step 4, repeatedly executing the steps 2 and 3 on the population obtained by iterative updating until the population is updated to an algebraic threshold value;
step 5, setting the coding information of the last generation of population individuals into an initial weight and a threshold of a neural network, and carrying out network training to obtain a network prediction model;
and 6, estimating the state of charge of the storage battery according to the network prediction model.
2. The method for estimating the state of charge of the storage battery according to claim 1, wherein the encoding the weight and the threshold parameter of the neural network into a code string and generating the initial population based on a chaotic mechanism comprises:
the network weight and the threshold parameter are coded into a coding string as an individual in the mixed evolution algorithm group, and the length of the individual variable is as follows: θ is mxl + l + l × n + n;
introducing chaotic Logistic mapping to generate an initial population P, wherein the scale is P;
in the formula: mu is a control parameter; i is the serial number of the chaotic variable, i is 1, 2. r denotes the population serial number, r 1, 2.., p; alpha is alphai rExpressing the chaotic variable, and assigning theta initial values alpha with small difference to the formula (1)i 0；
Substituting the obtained theta chaotic variables into the following formula for inverse mapping to obtain an individual variable xr iSequentially taking r ═ 1, 2.., p, and generating an initial population with the scale of p;
in the formula: [ x ] ofi min,xi max]Is a variable xr iThe value range of (2).
3. The battery state-of-charge estimation method of claim 1, wherein the fitness function F (y)i,oi) Taking the sum of absolute errors between the estimated value of the state of charge of the network output storage battery and the expected value, and taking an absolute value, wherein the formula is as follows:
in the formula: eta is a coefficient; n is the number of nodes of the network output layer; y isiOutputting a state of charge expected value of a storage battery for the ith node of the network; oiThe state of charge estimate of the output battery for the ith node of the network.
4. The battery state-of-charge estimation method according to claim 1, wherein the iterative update using a particle swarm algorithm comprises:
step a), initializing k individual populations with the top rank;
step b) calculating first fitness values of k individuals at the top of the rank according to the fitness function;
step c) updating the speed and position of the particles;
step d) judging whether the first fitness value is smaller than a fitness value threshold value;
if not, returning to the step b).
5. The battery state-of-charge estimation method according to claim 1, wherein the iterative updating using a differential evolution algorithm comprises:
step a) selecting a coding strategy and determining a genetic strategy;
step b) generating an initial population by using the remaining p-k individuals in the population;
step c) calculating a second fitness value of the remaining individuals according to the fitness function;
step d) judging whether the second fitness value is smaller than the fitness value threshold value;
if not, carrying out mutation, intersection and selection, and returning to the step c).
6. The method of estimating state of charge of a battery according to claim 1, further comprising, after said estimating state of charge of a battery according to said network prediction model:
and comparing the relative error value of the state of charge of the storage battery calculated by the state of charge estimation method of the storage battery with the relative error value of the state of charge of the storage battery calculated by only adopting a neural network method.
7. The method of estimating state of charge of a battery according to claim 1, further comprising, after said estimating state of charge of a battery according to said network prediction model:
and comparing the evolution algebra of the state of charge of the storage battery calculated by the method for estimating the state of charge of the storage battery with the evolution algebra of the state of charge of the storage battery estimated by only adopting a neural network method.
8. A state-of-charge estimation device for a storage battery, comprising:
the initial population generation module is used for coding the weight and the threshold parameter of the neural network into a coding string and generating an initial population based on a chaotic mechanism, wherein p is the size of the initial population;
the fitness value calculation module is used for setting the encoding information of the population individuals into the initial weight and the threshold of the neural network for network training, and solving the training result by using a fitness function to obtain the fitness value of each individual;
the iterative updating module is used for sequencing the population individuals according to the fitness value, taking k individuals which are ranked at the top, and performing iterative updating by adopting a particle swarm algorithm; carrying out iterative updating on the remaining p-k individuals by adopting a differential evolution algorithm;
the algebra judgment module is used for repeatedly executing the fitness value calculation module and the iterative update module on the population obtained by iterative update until the population is updated to an algebra threshold value;
the prediction model establishing module is used for setting the coding information of the last generation of optimal individuals into the initial weight and the threshold value of the network, and performing network training to obtain a network prediction model;
and the state of charge estimation module is used for estimating the state of charge of the storage battery according to the network prediction model.
9. An electronic device for estimating a state of charge of a battery, comprising:
one or more processors;
storage means for storing one or more programs;
when executed by the one or more processors, cause the one or more processors to implement the state of charge estimation method for a battery of any of claims 1-7.
10. A computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements a state of charge estimation method for a battery according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011173783.5A CN112307667A (en) | 2020-10-28 | 2020-10-28 | Method and device for estimating state of charge of storage battery, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011173783.5A CN112307667A (en) | 2020-10-28 | 2020-10-28 | Method and device for estimating state of charge of storage battery, electronic equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112307667A true CN112307667A (en) | 2021-02-02 |
Family
ID=74331599
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011173783.5A Pending CN112307667A (en) | 2020-10-28 | 2020-10-28 | Method and device for estimating state of charge of storage battery, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112307667A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114236401A (en) * | 2021-12-20 | 2022-03-25 | 上海正泰电源***有限公司 | Battery state estimation method based on adaptive particle swarm optimization |
| CN114280490A (en) * | 2021-09-08 | 2022-04-05 | 国网湖北省电力有限公司荆门供电公司 | Lithium ion battery state of charge estimation method and system |
| CN115099163A (en) * | 2022-08-25 | 2022-09-23 | 苏州清研精准汽车科技有限公司 | Charge state determining method and device, electronic equipment and storage medium |
| CN117878360A (en) * | 2024-03-12 | 2024-04-12 | 长春师范大学 | Fuel cell efficiency prediction method and system |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106647262A (en) * | 2016-11-28 | 2017-05-10 | 中国人民解放军国防科学技术大学 | Differential evolution method facing agile satellite multi-object task planning |
-
2020
- 2020-10-28 CN CN202011173783.5A patent/CN112307667A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106647262A (en) * | 2016-11-28 | 2017-05-10 | 中国人民解放军国防科学技术大学 | Differential evolution method facing agile satellite multi-object task planning |
Non-Patent Citations (2)
| Title |
|---|
| 杜瑞等: "启发式BP神经网络估算铅酸电池的SOC", 《电池》, vol. 49, no. 1, 28 February 2019 (2019-02-28), pages 51 - 54 * |
| 陈雷雨等: "基于DEPSO-RBF神经网络的锌银电池SOC估计", 《计算机测量与控制》, 25 September 2015 (2015-09-25), pages 3225 - 3227 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114280490A (en) * | 2021-09-08 | 2022-04-05 | 国网湖北省电力有限公司荆门供电公司 | Lithium ion battery state of charge estimation method and system |
| CN114280490B (en) * | 2021-09-08 | 2024-02-09 | 国网湖北省电力有限公司荆门供电公司 | Lithium ion battery state of charge estimation method and system |
| CN114236401A (en) * | 2021-12-20 | 2022-03-25 | 上海正泰电源***有限公司 | Battery state estimation method based on adaptive particle swarm optimization |
| CN114236401B (en) * | 2021-12-20 | 2023-11-28 | 上海正泰电源***有限公司 | Battery state estimation method based on self-adaptive particle swarm algorithm |
| CN115099163A (en) * | 2022-08-25 | 2022-09-23 | 苏州清研精准汽车科技有限公司 | Charge state determining method and device, electronic equipment and storage medium |
| CN117878360A (en) * | 2024-03-12 | 2024-04-12 | 长春师范大学 | Fuel cell efficiency prediction method and system |
| CN117878360B (en) * | 2024-03-12 | 2024-05-28 | 长春师范大学 | Fuel cell efficiency prediction method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112307667A (en) | Method and device for estimating state of charge of storage battery, electronic equipment and storage medium | |
| CN109815995B (en) | Method for predicting remaining life of lithium battery under condition of missing observed value | |
| CN112700060B (en) | Station terminal load prediction method and prediction device | |
| CN111709186A (en) | Integrated estimation method for health state of retired power lithium battery | |
| CN110007235A (en) | A kind of accumulator of electric car SOC on-line prediction method | |
| CN108052696B (en) | Three-value FPRM circuit area and delay optimization method using particle swarm optimization | |
| CN111832817A (en) | Small world echo state network time sequence prediction method based on MCP penalty function | |
| CN113962362A (en) | Reinforced learning model training method, decision-making method, device, equipment and medium | |
| WO2010048758A1 (en) | Classification of a document according to a weighted search tree created by genetic algorithms | |
| CN113376540A (en) | LSTM battery health state estimation method based on evolution attention mechanism | |
| CN111260015A (en) | Lithium ion battery model parameter identification method based on chaotic cat swarm algorithm | |
| Wei et al. | A hyperheuristic algorithm based on evolutionary strategy for complex mission planning of AUVs in marine environment | |
| CN112131089B (en) | Software defect prediction method, classifier, computer device and storage medium | |
| Phan et al. | Efficiency enhancement of evolutionary neural architecture search via training-free initialization | |
| CN116796821A (en) | Efficient neural network architecture searching method and device for 3D target detection algorithm | |
| CN113743012B (en) | Cloud-edge collaborative mode task unloading optimization method under multi-user scene | |
| CN111352650A (en) | Software modularization multi-objective optimization method and system based on INSGA-II | |
| CN111079902A (en) | Decomposition fuzzy system optimization method and device based on neural network | |
| CN115984025A (en) | Influence propagation estimation method and system based on deep learning graph network model | |
| Li et al. | A surrogate-assisted offspring generation method for expensive multi-objective optimization problems | |
| Shibata et al. | A learning method for dynamic Bayesian network structures using a multi-objective particle swarm optimizer | |
| CN114625886A (en) | Entity query method and system based on knowledge graph small sample relation learning model | |
| CN113641496A (en) | DIDS task scheduling optimization method based on deep reinforcement learning | |
| Zhang et al. | Adaptive truncation technique for constrained multi-objective optimization | |
| Georgiou et al. | Evolutionary Adaptive Schemes of Probabilistic Neural Networks |
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 |