CN114062944B - Method and system for autoregressive estimation of state of charge of storage battery based on longhorn beetle whisker identification - Google Patents
Method and system for autoregressive estimation of state of charge of storage battery based on longhorn beetle whisker identification Download PDFInfo
- Publication number
- CN114062944B CN114062944B CN202111308153.9A CN202111308153A CN114062944B CN 114062944 B CN114062944 B CN 114062944B CN 202111308153 A CN202111308153 A CN 202111308153A CN 114062944 B CN114062944 B CN 114062944B
- Authority
- CN
- China
- Prior art keywords
- charge
- state
- storage battery
- model
- estimation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 64
- 241000254173 Coleoptera Species 0.000 title claims abstract description 52
- 210000002569 neuron Anatomy 0.000 claims abstract description 28
- 238000013528 artificial neural network Methods 0.000 claims abstract description 23
- 230000006870 function Effects 0.000 claims description 23
- 238000011156 evaluation Methods 0.000 claims description 21
- 238000004364 calculation method Methods 0.000 claims description 12
- 238000010845 search algorithm Methods 0.000 claims description 12
- 238000012360 testing method Methods 0.000 claims description 8
- 238000010219 correlation analysis Methods 0.000 claims description 6
- 238000012795 verification Methods 0.000 claims description 4
- 230000006872 improvement Effects 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 2
- 238000012545 processing Methods 0.000 claims description 2
- 239000002253 acid Substances 0.000 abstract description 17
- 238000005457 optimization Methods 0.000 abstract description 6
- 238000001914 filtration Methods 0.000 description 5
- 241001481710 Cerambycidae Species 0.000 description 3
- 230000008901 benefit Effects 0.000 description 3
- 238000012804 iterative process Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- WHXSMMKQMYFTQS-UHFFFAOYSA-N Lithium Chemical compound [Li] WHXSMMKQMYFTQS-UHFFFAOYSA-N 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 229910052744 lithium Inorganic materials 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 239000002245 particle Substances 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 230000003213 activating effect Effects 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000005311 autocorrelation function Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000005314 correlation function Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000003487 electrochemical reaction Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
- G01R31/382—Arrangements for monitoring battery or accumulator variables, e.g. SoC
- G01R31/3842—Arrangements for monitoring battery or accumulator variables, e.g. SoC combining voltage and current measurements
-
- 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
-
- 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
- H01M10/48—Accumulators combined with arrangements for measuring, testing or indicating the condition of cells, e.g. the level or density of the electrolyte
-
- 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)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Electrochemistry (AREA)
- General Chemical & Material Sciences (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Chemical & Material Sciences (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Manufacturing & Machinery (AREA)
- Secondary Cells (AREA)
Abstract
The present disclosure provides a method and a system for autoregressive estimation of a charge state of a storage battery based on longhorn beetle whisker identification, wherein a nonlinear autoregressive neural network based on external input is constructed as a charge state estimation model of the storage battery in consideration of state time sequence correlation; and carrying out parameter identification on the established state of charge estimation model by adopting a heuristic group intelligent algorithm based on longhorn beetle whisker search, so as to realize the estimation of the real-time state of charge of the storage battery. The method and the device comprehensively consider the weighted influence of the model topology parameters such as the input/output time sequence correlation, the hidden layer neuron number and the like, realize the collaborative identification optimization of the model topology parameters and the neural network weight parameters, and improve the estimation precision of the state of charge of the lead-acid storage battery while reducing the complexity of the regression estimation identification model of the state of charge of the storage battery.
Description
Technical Field
The disclosure relates to the technical field of storage batteries, in particular to a method and a system for autoregressive estimation of a state of charge of a storage battery based on longhorn beetle whisker identification.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
The lead-acid storage battery is widely applied to echelon utilization scenes such as power systems, communication standby power supplies and the like by virtue of the advantages of long cycle life, strong adaptability, secondary utilization and the like. With the increasingly wide application of the power supply system of the lead-acid storage battery, the running safety problem of the power supply system of the lead-acid storage battery is increasingly outstanding, and the core problem is the safety management of the lead-acid storage battery pack. The state of charge of the lead-acid storage battery is an important index for evaluating the state of health of the lead-acid storage battery, and the accurate evaluation of the state of charge of the lead-acid storage battery is important for ensuring safe operation of the lead-acid storage battery, prolonging the service life of the lead-acid storage battery and avoiding overcharge/overdischarge of a battery pack. However, since the electrochemical reaction of the lead-acid storage battery is very complex, the running temperature is continuously changed, so that the change of the charge state of the lead-acid storage battery is nonlinear characteristics with complex lines, and the accurate estimation of the charge state of the lead-acid storage battery still faces challenges.
The traditional state of charge estimation is based on an ampere-hour metering method, and although the ampere-hour metering method has the advantages of simple and convenient calculation, easy realization and the like, errors can be gradually accumulated along with time, so that the state of charge estimation accuracy is obviously affected. Therefore, in order to improve the estimation precision of the state of charge of the lead-acid storage battery, scholars at home and abroad research and put forward various state of charge estimation methods including a Kalman filtering method and a particle filtering method based on model driving, a support vector machine method based on data driving, a neural network method and the like. In the model driving method, the Kalman filtering method can estimate the state of a dynamic system from a series of incomplete and noise-containing measurements, but the method has higher requirements on the statistical distribution characteristics of the measured noise. This is also a major problem in state of charge estimation by kalman filtering and particle filtering. In the data driving method, the support vector machine has high requirements on the quality of working condition data and poor generalization capability because of the inherent large sample dependence. In recent years, the neural network method has the advantage of independent battery model and mathematical relationship, so that the lead-acid storage battery state-of-charge estimation method based on the neural network is researched as a current hot spot. The state of charge estimation adopting the neural network method can properly solve the problems of low convergence speed, data overfitting, easy sinking into local optimum and the like, and the performance of the state of charge estimation is limited due to incomplete consideration factors.
Disclosure of Invention
In order to solve the problems, the present disclosure provides a method and a system for autoregressive estimation of a battery state of charge based on longhorn beetle whisker identification, which comprehensively consider the influence of state time sequence correlation on the performance of estimating the state of charge based on a neural network, and construct a model for estimating the state of charge of the battery based on a nonlinear autoregressive neural network with external input. On the basis, the established state of charge estimation model is subjected to parameter identification by adopting a heuristic group intelligent algorithm based on improved longhorn beetle whisker search, so that the accuracy of state of charge estimation is improved.
In order to achieve the above purpose, the present disclosure adopts the following technical scheme:
the first aspect of the disclosure provides a method for autoregressive estimation of a state of charge of a storage battery based on longhorn beetle whisker identification, which is used for constructing a nonlinear autoregressive neural network based on external input as a model for estimating the state of charge of the storage battery by considering state time sequence correlation; and carrying out parameter identification on the established state of charge estimation model by adopting a heuristic group intelligent algorithm based on longhorn beetle whisker search, so as to realize the estimation of the real-time state of charge of the storage battery.
A second aspect of the present disclosure provides a battery state of charge autoregressive estimation system based on longhorn beetle whisker identification, comprising:
a delay initial value determining module: an input delay initial value configured to determine a battery state of charge estimate based on a correlation of the parameter data of the battery with the state of charge;
an initial model building module: the method comprises the steps of setting up a state of charge regression estimation initial model of a storage battery based on a single hidden layer nonlinear autoregressive neural network with an external band input by taking a determined delay initial value into consideration;
model identification module: the method comprises the steps of using an improved longhorn beetle whisker algorithm to identify model parameters of a state of charge regression estimation initial model of the storage battery, and obtaining a state of charge regression estimation identification model of the storage battery;
the state of charge estimation module: the method is used for acquiring parameter data of the storage battery to be estimated, taking measured parameter data as input, taking the state of charge as output feedback input, and estimating the real-time state of charge of the storage battery through a state of charge regression estimation identification model of the storage battery.
A third aspect of the present disclosure provides a battery state of charge autoregressive estimation system based on longhorn beetle whisker identification, comprising: the parameter data acquisition device of the storage battery is used for acquiring parameter data of the storage battery, and the processor is configured to execute the storage battery state-of-charge autoregressive estimation method based on the longhorn beetle whisker identification.
A fourth aspect of the present disclosure provides an electronic device comprising a memory and a processor and computer instructions stored on the memory and running on the processor, which when executed by the processor, perform the steps of the method described above.
Compared with the prior art, the beneficial effects of the present disclosure are:
according to the method provided by the disclosure, on the basis of the adaptability evaluation function of the traditional longhorn beetle whisker search algorithm, the weighted influence of the model topology parameters such as the input/output time sequence correlation, the hidden layer neuron number and the like is comprehensively considered, the cooperative identification optimization of the model topology parameters and the neural network weight parameters is realized, and the complexity of the storage battery state-of-charge regression estimation identification model can be reduced, and meanwhile, the estimation precision of the storage battery state-of-charge is improved.
Additional aspects of the disclosure will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the disclosure.
Drawings
The accompanying drawings, which are included to provide a further understanding of the disclosure, illustrate and explain the exemplary embodiments of the disclosure and together with the description serve to explain and do not limit the disclosure.
FIG. 1 is a flow chart of a method of embodiment 1 of the present disclosure;
fig. 2 is a system block diagram of embodiment 2 of the present disclosure.
The specific embodiment is as follows:
the disclosure is further described below with reference to the drawings and examples.
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the present disclosure. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments in accordance with the present disclosure. As used herein, the singular is also intended to include the plural unless the context clearly indicates otherwise, and furthermore, it is to be understood that the terms "comprises" and/or "comprising" when used in this specification are taken to specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof. It should be noted that, without conflict, the various embodiments and features of the embodiments in the present disclosure may be combined with each other. The embodiments will be described in detail below with reference to the accompanying drawings.
Example 1
In the technical scheme disclosed by one or more embodiments, a non-linear autoregressive neural network based on out-of-band input is constructed as a storage battery state of charge estimation model by considering state time sequence correlation based on a storage battery state of charge autoregressive estimation method identified by longhorn whiskers; and carrying out parameter identification on the established state of charge estimation model by adopting a heuristic group intelligent algorithm based on improved longhorn beetle whisker search, so as to realize the estimation of the real-time state of charge of the storage battery.
The implementation flow of the automatic regression estimation method of the charge state of the storage battery is shown in a figure 1, and the method specifically comprises the following steps:
s1, determining an input delay initial value of battery state of charge estimation according to the correlation between parameter data of a storage battery and the state of charge;
s2, based on a single hidden layer nonlinear autoregressive neural network with source input, taking the determined initial delay value into consideration, and establishing a state of charge regression estimation initial model of the storage battery;
s3, identifying model parameters of the initial model of the state of charge regression estimation of the storage battery by adopting an improved longhorn beetle whisker algorithm to obtain an identification model of the state of charge regression estimation of the storage battery;
s4, acquiring parameter data of the storage battery to be estimated, taking the actually measured parameter data as input, taking the state of charge as output and feedback input, and estimating the real-time state of charge of the storage battery through a state of charge regression estimation identification model of the storage battery.
The embodiment considers the state time sequence correlation, and can improve the estimation precision of the state of charge of the storage battery while reducing the complexity of the regression estimation identification model of the state of charge of the storage battery.
The storage battery in the embodiment is a lead-acid storage battery or a lithium battery.
In step S1, determining that the input delay initial value of the battery state of charge estimation includes a battery parameter input delay initial value and an output feedback input delay initial value;
specifically, the battery parameter data comprise voltage, current and temperature; the output feedback input is the state of charge of the storage battery.
Taking voltage, current and temperature as input, taking a state of charge as output feedback input, performing correlation analysis according to voltage, current, temperature and state of charge data obtained by a charge-discharge test, and determining a battery parameter input delay initial value and an output feedback input delay initial value of a state of charge regression estimation initial model; specifically, the method comprises the following steps:
s101, performing charge and discharge test on the storage battery according to charge and discharge test conditions, and recording battery parameter data and corresponding charge states; namely recording experimental data of current, voltage, temperature and state of charge;
the working conditions of the charge and discharge test comprise a dynamic stress test working condition and a light vehicle test circulation working condition;
s102, carrying out normalization processing on the recorded battery parameter data and the corresponding state of charge;
s103, performing unit root verification on the battery parameter data and the corresponding state of charge to evaluate the stability of the experimental data, if the battery parameter data and the corresponding state of charge do not have the stability, executing the step 104, otherwise, executing the step S105;
s104, performing differential operation on the battery parameter data and the corresponding state of charge data, and executing step S103;
s105, carrying out correlation analysis on the battery parameter data with stability and the corresponding state of charge to obtain an input delay initial value.
The method specifically comprises the steps of performing cross correlation analysis on the voltage, current and temperature experimental data to determine an input delay initial value of the storage battery state-of-charge regression estimation initial model; and carrying out autocorrelation analysis on the state of charge experimental data to determine an exogenous output feedback input delay initial value of the storage battery state of charge regression estimation initial model.
In this embodiment, the exogenous output feedback input is the estimated state of charge of the initial model of the regression estimation of the state of charge of the storage battery.
In this embodiment, the model uses NARX, which is collectively referred to as a nonlinear autoregressive neural network with external input, and the external output feedback input refers to its output feedback as the input of the neural network.
The correlation analysis method specifically comprises the following steps: carrying out unit root verification on the voltage, current and temperature experimental data, if the voltage, current and temperature experimental data do not meet the data stability requirement, carrying out differential calculation until the data stability requirement is met, and further carrying out cross correlation analysis based on a cross correlation function on the verified differential experimental data; and carrying out unit root verification on the charge state experimental data, and if the charge state experimental data does not meet the data stability requirement, carrying out differential calculation until the charge state experimental data meets the stability requirement, and further carrying out autocorrelation analysis based on an autocorrelation function on the verified differential experimental data.
In step S2, specifically, voltage, current, and temperature are used as inputs, a state of charge is used as an output feedback input, and a state of charge regression estimation initial model of the storage battery is built based on a single hidden layer nonlinear autoregressive neural network with source input, including a single hidden layer neuron initial model and an initial output model of an output layer neuron, as follows:
s201, obtaining a single hidden layer neuron initial model at the time t as shown in a formula (1) based on a single hidden layer nonlinear autoregressive neural network model with source input:
wherein n is 0 The initial number of hidden layer neurons;represents the nth 0 The output of the individual hidden layer neurons; />Represents the nth 0 Activating functions corresponding to the hidden layer neurons; />Represents the nth 0 Initial bias of individual hidden layer neurons;representing model input, ++>An input delay initial value, < >, entered for said model>Representing an initial connection weight of the model input and an nth hidden layer neuron;output feedback input representing model, +.>Feeding back an input delay for an initial output of the model output,/->Representing the initial connection weights of the model outputs and the nth hidden layer neurons.
The initial output model of the output layer neuron at the time t is shown as a formula (2) based on a single hidden layer nonlinear autoregressive neural network model with the external input:
wherein,representing the output of a state of charge regression estimation initial model of the storage battery; f (f) g Representing an activation function corresponding to the output layer neuron; b g Representing an initial bias of the output layer neurons; />Representing initial connection weights between the neurons of the single hidden layer and the neurons of the output layer.
In step S3, based on the improved longhorn beetle whisker algorithm, parameters of the initial model of the state of charge regression estimation of the storage battery are identified, and a model of the state of charge regression estimation identification of the storage battery is obtained.
Wherein, the parameters of the initial model of the state of charge regression estimation of the storage battery can comprise: input delay, output delay, hidden layer neuron number, neural network bias parameter and weight parameter.
The model parameter identification method specifically comprises the following steps:
s301, improving a longhorn beetle whisker algorithm, adding quantitative calculation of the influence of the topological parameter of the state-of-charge regression estimation identification model of the storage battery on the fitness evaluation function, namely adding a weighting part of network topological parameters, and calculating the fitness evaluation function, wherein the quantitative calculation comprises the following steps:
wherein,outputting a model at the time t of estimating and identifying the model for the regression of the state of charge of the storage battery at the time of the s-th iteration, and y r (t) experimental state of charge data at time t of the storage battery; d, d 1 ，d 2 Separate tableShowing the input and output layer delay orders. n is the number of hidden layer neurons.
The improved fitness evaluation function can realize the compromise of the charge state estimation precision and the charge state regression estimation identification model complexity, and reduces the charge state regression estimation identification model complexity on the premise of meeting the charge state estimation precision.
S302, improving a longhorn beetle whisker search algorithm, and determining a search range as shown in a formula (4):
wherein zl s The position coordinates of the left beard of the longicorn in the s-th iteration; zr s The position coordinates of the right beards of the longicorn in the s-th iteration; d, d s Is the distance between the two whiskers at the s-th iteration;centroid coordinates at the s-th iteration, and; />Is a unit direction vector.
S303, improving a longhorn beetle whisker search algorithm, wherein an iteration model is shown in a formula (5):
wherein delta s Step length at the s-th iteration; sign (·) is a sign function.
S304, performing iterative computation according to the determined search range of the longhorn beetle whisker search algorithm and the iterative model, and performing iterative search on model parameters until the iterative termination condition is reached, so as to obtain the state of charge regression estimation identification model of the storage battery.
The iterative calculation process is as follows:
specifically, parameter optimization is performed by a two-stage parameter searching method:
(1) One stage is an outer ring, and performs the longhorn beetle whisker search of the network topology parameters, namely the network topology parameters n and d 1 And d 2 Is a search of (2);
(2) The two stages are inner loops, and LM search of network weights and offsets is carried out, namely the search of the network weights and the offsets is carried out based on a Levenberg-Marquardt iterative algorithm, and each complete iterative process consists of the two stages.
In each iteration process, the connection weight of the storage battery state of charge regression estimation identification model is calculatedAnd offset value->Performing parameter searching based on a Levenberg-Marquardt iterative algorithm; the termination conditions for the parameter search iteration are: the number of parameter search iterations is equal to the specified maximum number of iterations.
(3) Calculating Fitness evaluation function value Fitness according to the Fitness evaluation function, and outputting network topology parameters and connection weights when the Fitness evaluation function value Fitness is smaller than a specified threshold or the iteration number s is equal to the set maximum iteration number, and ending the iterationAnd offset value->The method comprises the steps of carrying out a first treatment on the surface of the Otherwise, executing the step (1) to perform the next iterative computation.
When the iteration of the improved longhorn beetle whisker search algorithm is terminated, the barycenter coordinate zs of the longhorn beetle whisker and the connection weight value are used for obtaining the barycenter coordinate zs of the longhorn beetle whiskerAnd offset value->Updating the saidAnd obtaining the state of charge regression estimation identification model of the storage battery by using the state of charge regression estimation model parameters of the storage battery.
The model identification process of the present embodiment includes two nested iterative processes: one stage is that the outer ring improves the longhorn beetle whisker search to carry out the topological parameter n, d 1 And d 2 The two stages are that the inner loop searches for network weights and offsets based on a Levenberg-Marquardt iterative algorithm, and each complete iterative process consists of the two stages.
The state of charge estimation method provided by the embodiment increases optimization of parameters such as input delay and exogenous feedback input delay of network topology; compared with the state of charge estimation method of the storage battery or the lithium battery based on the neural network in the prior art, the state of charge estimation method only performs search optimization on the network weight, and can improve the accuracy of the estimation model on state of charge identification.
In step S4, the real-time state of charge of the storage battery is estimated by using a state of charge regression estimation identification model of the storage battery, specifically, the real-time state of charge of the storage battery is estimated by using measured voltage, current and temperature as input and using state of charge as output feedback input, based on the state of charge regression estimation identification model, the method may include the following steps:
s401, determining the quantity and the acquisition time of input quantities according to the obtained longicorn centroid coordinates of the storage battery state-of-charge regression estimation identification model, and constructing the input quantities of battery parameters and output feedback input quantities;
the input quantity comprises voltage, current and temperature, and the output feedback input quantity is state of charge output feedback.
(1) The barycenter coordinates of the longhorn beetles when the iteration of the improved longhorn beetle whisker search algorithm is terminatedThe voltage, current and temperature input at the moment t of the constructed storage battery state-of-charge autoregressive estimation identification model are the voltage, current and temperature input phasors at the moment t, and the specific steps are as follows:
(2) According to the longhorn beetle whisker centroid coordinate zm when the iteration of the improved longhorn beetle whisker search algorithm is terminated, the constructed estimated state of charge output feedback input of the storage battery state of charge autoregressive estimated identification model is as follows:
s402, estimating model parameters of an identification model according to the regression of the charge state of the storage battery, obtaining an identification model of each neuron of a single hidden layer and an identification output model of the neuron of an output layer, and calculating the identification model output y of the neuron of the output layer m And (t) is the t-moment state of charge of the storage battery.
Based on the obtained storage battery state-of-charge regression estimation identification model, according to the connection weightAnd offset value->Calculating a single hidden layer neuron identification model at the t moment as shown in the formula (8):
based on the obtained storage battery state-of-charge regression estimation identification model, according to the connection weightAnd offset value->Calculating an identification output model of the output layer neuron at the time t as shown in the formula (9):
identification model output y of neuron of output layer at time t m And (t) is the t-moment state of charge of the storage battery.
Compared with the existing storage battery state of charge estimation method, the method provided by the embodiment comprehensively considers the weighted influence of the model topology parameters such as the input/output time sequence correlation, the hidden layer neuron number and the like on the basis of the adaptability evaluation function of the traditional longhorn beetle whisker search algorithm, realizes the collaborative identification optimization of the model topology parameters and the neural network weight parameters, and can improve the estimation precision of the state of charge of the lead-acid storage battery while reducing the complexity of the storage battery state of charge regression estimation identification model.
Example 2
Based on embodiment 1, as shown in fig. 2, the present embodiment provides a battery state of charge autoregressive estimation system based on longhorn beetle whisker identification, including:
a delay initial value determining module: an input delay initial value configured to determine a battery state of charge estimate based on a correlation of the parameter data of the battery with the state of charge;
an initial model building module: the method comprises the steps of setting up a state of charge regression estimation initial model of a storage battery based on a single hidden layer nonlinear autoregressive neural network with an external band input by taking a determined delay initial value into consideration;
model identification module: the method comprises the steps of using an improved longhorn beetle whisker algorithm to identify model parameters of a state of charge regression estimation initial model of the storage battery, and obtaining a state of charge regression estimation identification model of the storage battery;
the state of charge estimation module: the method is used for acquiring parameter data of the storage battery to be estimated, taking measured parameter data as input, taking the state of charge as output feedback input, and estimating the real-time state of charge of the storage battery through a state of charge regression estimation identification model of the storage battery.
Example 3
The embodiment provides a storage battery state of charge autoregressive estimation system based on longhorn beetle whisker identification, which comprises the following steps: the device for acquiring parameter data of the storage battery in communication connection and the processor are used for acquiring the parameter data of the storage battery, and the processor is configured to execute the method for automatically estimating the state of charge of the storage battery based on the longhorn beetle whisker identification in the embodiment 1.
Example 4
The present embodiment provides an electronic device comprising a memory and a processor, and computer instructions stored on the memory and running on the processor, which when executed by the processor, perform the steps described in the method of embodiment 1 above.
The foregoing description of the preferred embodiments of the present disclosure is provided only and not intended to limit the disclosure so that various modifications and changes may be made to the present disclosure by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present disclosure should be included in the protection scope of the present disclosure.
While the specific embodiments of the present disclosure have been described above with reference to the drawings, it should be understood that the present disclosure is not limited to the embodiments, and that various modifications and changes can be made by one skilled in the art without inventive effort on the basis of the technical solutions of the present disclosure while remaining within the scope of the present disclosure.
Claims (6)
1. The method for autoregressive estimation of the charge state of the storage battery based on the identification of the longhorn beetles is characterized by comprising the following steps of: taking state time sequence correlation into consideration, constructing a nonlinear autoregressive neural network based on external input as a storage battery state of charge estimation model; carrying out parameter identification on the established state of charge estimation model by adopting a heuristic group intelligent algorithm based on longhorn beetle whisker search, so as to realize the estimation of the real-time state of charge of the storage battery;
determining an input delay initial value of battery state of charge estimation according to the correlation between the parameter data of the storage battery and the state of charge;
based on a single hidden layer nonlinear autoregressive neural network with a belt-based input, a determined initial delay value is considered, and a state-of-charge regression estimation initial model of the storage battery is established;
identifying model parameters of the initial model of the state of charge regression estimation of the storage battery by adopting an improved longhorn beetle whisker algorithm to obtain an identification model of the state of charge regression estimation of the storage battery;
the improved longhorn beetle whisker algorithm comprises the improvement of the fitness evaluation function, and the quantitative calculation of the influence of the topological parameter of the state-of-charge regression estimation identification model of the storage battery on the fitness evaluation function is increased;
the model parameter identification method specifically comprises the following steps:
adding quantitative calculation of the influence of topological parameters of a state-of-charge regression estimation identification model of the storage battery on the fitness evaluation function, and determining the fitness evaluation function;
determining a search range and an iteration model;
performing iterative computation according to the determined search range of the longhorn beetle whisker search algorithm and the iterative model, performing iterative search on model parameters until reaching an iterative termination condition, and obtaining a state-of-charge regression estimation identification model of the storage battery;
the iterative calculation is carried out according to the determined search range of the longhorn beetle whisker search algorithm and the iterative model, and the method for carrying out iterative search on model parameters comprises the following steps:
step (1) performing longhorn beetle whisker search of network topology parameters;
and (2) searching for network weights and bias values based on a Levenberg-Marquardt iterative algorithm, wherein the search iteration termination conditions are as follows: the search iteration number is equal to the specified maximum iteration number;
calculating an fitness evaluation function value, and outputting a network topology parameter, a connection weight and a bias value when the fitness evaluation function value is smaller than a specified threshold or the iteration number s is equal to the set maximum iteration number, and ending the iteration; otherwise, executing the step (1);
and acquiring parameter data of the storage battery to be estimated, taking the actually measured parameter data as input, taking the state of charge as output feedback input, and estimating the real-time state of charge of the storage battery through a state of charge regression estimation identification model of the storage battery.
2. The method for autoregressive estimation of battery state of charge based on longhorn beetle whisker identification according to claim 1, wherein the method for determining the initial value of the input delay of the battery state of charge estimation comprises the steps of:
s101, performing charge and discharge test on the storage battery according to charge and discharge test working conditions, and recording experimental data of current, voltage, temperature and state of charge of the storage battery;
s102, carrying out normalization processing on recorded data;
s103, performing unit root verification on the battery parameter data and the corresponding state of charge to evaluate the stability of the experimental data, if the battery parameter data and the corresponding state of charge do not have the stability, executing the step 104, otherwise, executing the step S105;
s104, performing differential operation on the battery parameter data and the corresponding state of charge data, and executing step S103;
s105, carrying out correlation analysis on the battery parameter data with stability and the corresponding state of charge to obtain an input delay initial value.
3. The method for autoregressive estimation of the charge state of a storage battery based on the identification of the tendrils of claim 1, wherein the real-time charge state of the storage battery is estimated by a charge state regression estimation identification model of the storage battery, comprising the following steps:
determining the quantity and the acquisition time of the input quantity according to the obtained longhorn beetle whisker centroid coordinates of the storage battery state-of-charge regression estimation identification model, and constructing the battery parameter input quantity and the output feedback input quantity;
and estimating model parameters of the identification model according to the regression of the charge state of the storage battery, obtaining an identification model of each neuron of a single hidden layer and an identification output model of the neuron of an output layer, calculating the identification model output of the neuron of the output layer, and obtaining the predicted charge state of the storage battery.
4. The utility model provides a battery state of charge autoregressive estimation system based on longhorn beetle whisker discernment which characterized in that includes:
a delay initial value determining module: an input delay initial value configured to determine a battery state of charge estimate based on a correlation of the parameter data of the battery with the state of charge;
an initial model building module: the method comprises the steps of setting up a state of charge regression estimation initial model of a storage battery based on a single hidden layer nonlinear autoregressive neural network with an external band input by taking a determined delay initial value into consideration;
model identification module: the method comprises the steps of using an improved longhorn beetle whisker algorithm to identify model parameters of a state of charge regression estimation initial model of the storage battery, and obtaining a state of charge regression estimation identification model of the storage battery;
the improved longhorn beetle whisker algorithm comprises the improvement of the fitness evaluation function, and the quantitative calculation of the influence of the topological parameter of the state-of-charge regression estimation identification model of the storage battery on the fitness evaluation function is increased;
the model parameter identification method specifically comprises the following steps:
adding quantitative calculation of the influence of topological parameters of a state-of-charge regression estimation identification model of the storage battery on the fitness evaluation function, and determining the fitness evaluation function;
determining a search range and an iteration model;
performing iterative computation according to the determined search range of the longhorn beetle whisker search algorithm and the iterative model, performing iterative search on model parameters until reaching an iterative termination condition, and obtaining a state-of-charge regression estimation identification model of the storage battery;
the iterative calculation is carried out according to the determined search range of the longhorn beetle whisker search algorithm and the iterative model, and the method for carrying out iterative search on model parameters comprises the following steps:
step (1) performing longhorn beetle whisker search of network topology parameters;
and (2) searching for network weights and bias values based on a Levenberg-Marquardt iterative algorithm, wherein the search iteration termination conditions are as follows: the search iteration number is equal to the specified maximum iteration number;
calculating an fitness evaluation function value, and outputting a network topology parameter, a connection weight and a bias value when the fitness evaluation function value is smaller than a specified threshold or the iteration number s is equal to the set maximum iteration number, and ending the iteration; otherwise, executing the step (1);
the state of charge estimation module: the method is used for acquiring parameter data of the storage battery to be estimated, taking measured parameter data as input, taking the state of charge as output feedback input, and estimating the real-time state of charge of the storage battery through a state of charge regression estimation identification model of the storage battery.
5. The utility model provides a battery state of charge autoregressive estimation system based on longhorn beetle whisker discernment which characterized in that includes: a parameter data acquisition device of a communicatively connected battery for acquiring parameter data of the battery, and a processor configured to perform the longhorn beetle whisker identification-based battery state of charge autoregressive estimation method of any of claims 1-3.
6. An electronic device comprising a memory and a processor and computer instructions stored on the memory and running on the processor, which when executed by the processor, perform the steps of the method of any one of claims 1-3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111308153.9A CN114062944B (en) | 2021-11-05 | 2021-11-05 | Method and system for autoregressive estimation of state of charge of storage battery based on longhorn beetle whisker identification |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111308153.9A CN114062944B (en) | 2021-11-05 | 2021-11-05 | Method and system for autoregressive estimation of state of charge of storage battery based on longhorn beetle whisker identification |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114062944A CN114062944A (en) | 2022-02-18 |
| CN114062944B true CN114062944B (en) | 2024-03-26 |
Family
ID=80274458
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111308153.9A Active CN114062944B (en) | 2021-11-05 | 2021-11-05 | Method and system for autoregressive estimation of state of charge of storage battery based on longhorn beetle whisker identification |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114062944B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20160002336A (en) * | 2014-06-30 | 2016-01-07 | 숭실대학교산학협력단 | Battery status estimation method, system and recording medium for performing the method |
| CN109782176A (en) * | 2018-12-20 | 2019-05-21 | 上海交通大学 | Battery remaining power On-line Estimation method based on NARX model |
| CN110082682A (en) * | 2019-03-12 | 2019-08-02 | 浙江大学 | A kind of lithium battery charge state estimation method |
| CN110646737A (en) * | 2019-09-20 | 2020-01-03 | 广州市香港科大霍英东研究院 | Battery SOC dynamic estimation method and system based on multiple models and storage medium |
| US10969436B1 (en) * | 2019-10-25 | 2021-04-06 | The Florida International University Board Of Trustees | Systems and methods for forecasting battery state of charge |
-
2021
- 2021-11-05 CN CN202111308153.9A patent/CN114062944B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20160002336A (en) * | 2014-06-30 | 2016-01-07 | 숭실대학교산학협력단 | Battery status estimation method, system and recording medium for performing the method |
| CN109782176A (en) * | 2018-12-20 | 2019-05-21 | 上海交通大学 | Battery remaining power On-line Estimation method based on NARX model |
| CN110082682A (en) * | 2019-03-12 | 2019-08-02 | 浙江大学 | A kind of lithium battery charge state estimation method |
| CN110646737A (en) * | 2019-09-20 | 2020-01-03 | 广州市香港科大霍英东研究院 | Battery SOC dynamic estimation method and system based on multiple models and storage medium |
| US10969436B1 (en) * | 2019-10-25 | 2021-04-06 | The Florida International University Board Of Trustees | Systems and methods for forecasting battery state of charge |
Non-Patent Citations (3)
| Title |
|---|
| IoT System for Pellet Proportioning Based on BAS Intelligent Recommendation Model;Aimin Yang;IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS;第17卷(第2期);934-942 * |
| 基于BAS 算法优化的感应电机参数估计;金逸珲;现代电子技术;第43卷(第3期);112-120 * |
| 基于IAF-NARX的车载锂电池SOC实时估算;陈德海;江西理工大学学报;第41卷(第3期);96-102 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114062944A (en) | 2022-02-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109061506A (en) | Lithium-ion-power cell SOC estimation method based on Neural Network Optimization EKF | |
| CN113253116A (en) | Lithium ion battery state of charge estimation method and storage medium | |
| CN106443480B (en) | A kind of lithium-ion battery systems SOC estimation method | |
| CN112331941B (en) | Cloud auxiliary battery management system and method | |
| CN108594135A (en) | A kind of SOC estimation method for the control of lithium battery balance charge/discharge | |
| CN105954679B (en) | A kind of On-line Estimation method of lithium battery charge state | |
| CN103515661B (en) | Method for prolonging residual recycling service life of in-parallel lithium ion battery cupboard | |
| CN112881914B (en) | Lithium battery health state prediction method | |
| CN110554324A (en) | SOC and SOH joint estimation method | |
| Li et al. | A novel state estimation approach based on adaptive unscented Kalman filter for electric vehicles | |
| CN112798961B (en) | Method for predicting remaining service life of power battery of electric automobile | |
| CN211702078U (en) | Battery manager based on cloud-end mixed state estimation | |
| CN114660464A (en) | Lithium ion battery state of charge estimation method | |
| CN109655751A (en) | A kind of method and system using Gaussian process regression estimates battery charging state | |
| CN112415391A (en) | Marine lithium ion battery parameter identification method based on improved particle swarm optimization | |
| CN109507598A (en) | The lithium battery SOC prediction technique of the LM-BP neural network of Bayesian regularization | |
| CN109346787B (en) | Self-adaptive optimized charging method for power battery of electric automobile | |
| CN112163372A (en) | SOC estimation method of power battery | |
| CN115986874A (en) | Battery charging safety monitoring and early warning method and system | |
| CN111537887B (en) | Hybrid power system battery open-circuit voltage model optimization method considering hysteresis characteristic | |
| CN113466710A (en) | SOC and SOH collaborative estimation method for energy storage battery in receiving-end power grid containing new energy | |
| CN116953547A (en) | Energy storage battery health evaluation method, system, equipment and storage medium | |
| CN114062944B (en) | Method and system for autoregressive estimation of state of charge of storage battery based on longhorn beetle whisker identification | |
| TWI527337B (en) | The method of determining the charging current of multi - stage lithium battery using Taguchi orthogonal table and fuzzy algorithm | |
| CN110932300B (en) | Power distribution network optimal scheduling method considering echelon utilization energy storage |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |