CN116643190A - Real-time monitoring method and system for lithium battery health state - Google Patents
Real-time monitoring method and system for lithium battery health state Download PDFInfo
- Publication number
- CN116643190A CN116643190A CN202310465215.XA CN202310465215A CN116643190A CN 116643190 A CN116643190 A CN 116643190A CN 202310465215 A CN202310465215 A CN 202310465215A CN 116643190 A CN116643190 A CN 116643190A
- Authority
- CN
- China
- Prior art keywords
- lithium battery
- data
- parameter data
- real
- current
- 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
- WHXSMMKQMYFTQS-UHFFFAOYSA-N Lithium Chemical compound [Li] WHXSMMKQMYFTQS-UHFFFAOYSA-N 0.000 title claims abstract description 259
- 229910052744 lithium Inorganic materials 0.000 title claims abstract description 259
- 230000036541 health Effects 0.000 title claims abstract description 76
- 238000000034 method Methods 0.000 title claims abstract description 63
- 238000012544 monitoring process Methods 0.000 title claims abstract description 34
- 238000001914 filtration Methods 0.000 claims abstract description 21
- 238000004364 calculation method Methods 0.000 claims abstract description 18
- 238000012545 processing Methods 0.000 claims abstract description 18
- 238000003062 neural network model Methods 0.000 claims abstract description 17
- 238000009499 grossing Methods 0.000 claims description 35
- 238000012549 training Methods 0.000 claims description 26
- 238000010606 normalization Methods 0.000 claims description 17
- 238000001514 detection method Methods 0.000 claims description 16
- 239000011159 matrix material Substances 0.000 claims description 16
- 239000003016 pheromone Substances 0.000 claims description 15
- 230000003862 health status Effects 0.000 claims description 14
- 210000002569 neuron Anatomy 0.000 claims description 14
- 102000008297 Nuclear Matrix-Associated Proteins Human genes 0.000 claims description 12
- 108010035916 Nuclear Matrix-Associated Proteins Proteins 0.000 claims description 12
- 210000000299 nuclear matrix Anatomy 0.000 claims description 12
- 238000004891 communication Methods 0.000 claims description 9
- 239000013598 vector Substances 0.000 claims description 9
- 238000004458 analytical method Methods 0.000 claims description 7
- 230000003750 conditioning effect Effects 0.000 claims description 6
- GNFTZDOKVXKIBK-UHFFFAOYSA-N 3-(2-methoxyethoxy)benzohydrazide Chemical compound COCCOC1=CC=CC(C(=O)NN)=C1 GNFTZDOKVXKIBK-UHFFFAOYSA-N 0.000 claims description 5
- 238000000354 decomposition reaction Methods 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 4
- 241000257303 Hymenoptera Species 0.000 claims description 3
- 108010063499 Sigma Factor Proteins 0.000 claims description 3
- 230000007704 transition Effects 0.000 claims description 3
- 238000005259 measurement Methods 0.000 description 6
- 238000013528 artificial neural network Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000000691 measurement method Methods 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 238000009529 body temperature measurement Methods 0.000 description 1
- 210000004027 cell Anatomy 0.000 description 1
- 238000007599 discharging Methods 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000008569 process Effects 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]
- G01R31/392—Determining battery ageing or deterioration, e.g. state of health
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
- G01R31/367—Software therefor, e.g. for battery testing using modelling or look-up tables
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
- G01R31/382—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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/10—Pre-processing; Data cleansing
- G06F18/15—Statistical pre-processing, e.g. techniques for normalisation or restoring missing data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/27—Regression, e.g. linear or logistic regression
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Probability & Statistics with Applications (AREA)
- Secondary Cells (AREA)
Abstract
The invention relates to the technical field of lithium batteries, in particular to a method and a system for monitoring the health state of a lithium battery in real time, which are characterized in that a plurality of parameters of the battery are collected, and the collected data are processed by utilizing weighted moving average filtering, so that the real-time performance of the collected parameter data of the lithium battery is better, the noise is smaller, and the calculation precision is improved; and then, the real-time state of health information of the current lithium battery is obtained by carrying out real-time calculation on the state of health of the lithium battery by using a generalized regression neural network model, so that on one hand, the calculation difficulty of the state of health information of the battery is reduced, the processing efficiency of the real-time state of health information of the lithium battery is improved, on the other hand, the large-scale data is not relied on, the convergence rate is improved, the calculation difficulty is reduced, and the accuracy, the stability and the reliability of the calculation result are also improved.
Description
Technical Field
The invention relates to the technical field of lithium batteries, in particular to a method and a system for monitoring the health state of a lithium battery in real time.
Background
Because the lithium battery has the characteristics of large capacity, high energy density, environmental protection, safety and the like, the lithium battery becomes a key component of the electric automobile and is important for the safe operation and stability of the electric automobile, so that the lithium battery is monitored and managed at all times.
The state of health (SOH) is an important index for evaluating the performance of a battery, and has been receiving attention in recent years. The reasonable estimation of SOH can accurately predict the service life of the battery, and early warning information is timely sent out before the lithium battery fails, so that safe running of the automobile and life safety of a driver are effectively guaranteed, the accurate estimation of the SOH is not only beneficial to management of a battery system, but also accuracy of other parameters such as a state of charge and a power state can be improved.
The current lithium battery SOH prediction method is roughly divided into an electrochemical mechanism model-based equivalent circuit model and a data driving method.
The method based on the electrochemical mechanism model has higher interpretation degree, but has high modeling and calculating difficulty, and improves the prediction difficulty of SOH of the lithium battery.
The method based on the equivalent circuit model needs to select a proper equivalent circuit model, and the parameter identification of the equivalent circuit model needs to be carried out, so that the process is complex, and the SOH prediction difficulty of the lithium battery is increased.
Based on the data driving method, the SOH can be predicted without knowing the complex mechanism inside the battery, and the method has higher accuracy, but needs to rely on large-scale data, has low convergence speed, and further increases the difficulty of SOH prediction of the lithium battery.
Disclosure of Invention
The invention aims to provide a method and a system for monitoring the health state of a lithium battery in real time, which solve the problem of high difficulty in detecting the health degree of the lithium battery at present.
The invention solves the technical problems as follows:
the method for monitoring the health state of the lithium battery in real time is characterized by comprising the following steps of:
s1, acquiring lithium battery parameter data in real time, and processing the lithium battery parameter data acquired in real time through weighted moving average filtering;
s2, carrying out normalization pretreatment on the lithium battery parameter data processed in the step S1;
s3, extracting the lithium battery parameter data subjected to normalization pretreatment in the step S2 through a kernel entropy component analysis algorithm to obtain feature principal component information;
and S4, inputting the characteristic principal element information into a generalized regression neural network model to obtain the real-time information of the health state of the lithium battery.
Further defined, said step S1 comprises the steps of:
s11, continuously collecting n lithium battery parameter data of each data type according to a collection period in a charge-discharge period of the current lithium battery, storing the n lithium battery parameter data of the same data type into a buffer area of the corresponding data type according to a collection sequence, wherein n is more than 2 and less than or equal to 10, and n is a positive integer; the data types comprise battery temperature, lithium battery voltage, lithium battery current, lithium battery discharge time and lithium battery charge and discharge times;
s12, processing the n lithium battery parameter data in the same data type buffer area through weighted moving average filtering to obtain a weighted average value of the current corresponding data type
Wherein x is i The ith lithium battery parameter data, w, in the n lithium battery parameter data in the same data type buffer area i Is x i I is more than or equal to 1 and less than or equal to n, i is a positive integer, and n is the number of lithium battery parameter data in each data type cache region;
s13, continuously acquiring primary lithium battery parameter data according to an acquisition period, storing the primary lithium battery parameter data into corresponding data type cache areas according to an acquisition sequence, removing the lithium battery parameter data with earliest acquisition time in each cache area, and updating the lithium battery parameter data in each cache area;
s14, repeatedly executing the step S12 and the step S13, and obtaining the weighted average value of the lithium battery parameter data of each data type in real time.
Further defined, the step S2 specifically includes:
receiving a weighted average of each data type obtained in step S12And carrying out normalization preprocessing on the weighted average value of each data type to obtain preprocessed data D of the current data type Data type ，D Data type ∈[0,1]。
Further defined, the step S3 specifically includes the following steps:
s31, obtaining the target product in real time through Parzen window pairsIs defined by the respective pre-processed data D of (a) Data type Establishing an N-dimensional nuclear matrix and carrying out feature decomposition to obtain a feature value of the current N-dimensional nuclear matrix and a feature vector of the current N-dimensional nuclear matrix, wherein N is the number of acquired lithium battery parameter data types;
s32, calculating Renyi entropy of the current N-dimensional kernel matrix, and sorting the eigenvalues of the current N-dimensional kernel matrix and the eigenvectors of the current N-dimensional kernel matrix according to the calculated entropy value;
s33, selecting the first B principal elements with the contribution value of Renyi entropy of the current N-dimensional kernel matrix more than 95%, and selecting the feature vectors of the first B principal elements as feature principal element information B, wherein 0< B < N.
Further defined, said step S4 comprises the steps of:
s41, establishing a sample database, wherein the sample database comprises M sample data, the sample data comprises parameter data of a lithium battery and battery health state values Y obtained by detection when corresponding to the parameter data, and M sample data are selected from the M sample data to serve as training data, wherein m=0.6M-0.8M, and M is rounded to an integer;
s42, establishing a generalized regression neural network model, and taking characteristic principal element information as an input layer;
calculating Gauss function value between the input layer and each training data to obtain output P of the mode layer a ：
Wherein a is the a training data in m training data, e is a natural constant, sigma is a smoothing factor, sigma epsilon (0, 15), and B is characteristic principal element information;
s43, calculating the output of each neuron node of the summation layer respectively, wherein the number of the neuron nodes of the summation layer is 2:
the output S of the first neuron node D The method comprises the following steps:
the output S of the second neuron node d The method comprises the following steps:
wherein Y is a The battery state of health value in the a-th training data, m is the number of training data;
s44, calculating output y of the output layer:
wherein y is the current real-time information of the lithium battery health state.
Further defined, the determination of the smoothing factor comprises the steps of:
a. initializing ant colony, determining ant quantity as C, maximum iteration cycle number K, setting smoothing factor sigma, and setting initial pheromone concentration;
b. constructing a solution space, dividing the solution space into Q solution intervals, and randomly dispersing C ants in the Q solution intervals to obtain the initial position of each ant;
c. each ant moves according to a transition probability formula, and each ant obtains a moving path;
d. calculating the length of each moving path, updating the pheromone concentration by combining with a pheromone iteration formula, simultaneously recording the optimal pheromone concentration, and taking the optimal pheromone concentration as the current optimal smoothing factor sigma k K is the current cycle number, k=1 to K;
e. for any sample data by the current optimal smoothing factor sigma k Calculating to obtain the health state predictive value y (sigma) k ) Using the obtained calculated state of health predictor y (σ k ) Calculating a prediction error according to the actual battery state of health value Y corresponding to the sample data to obtain a current optimal smoothness factor sigma k Is calculated error of (a);
f. judging whether K is less than K, if so, clearing all moving paths, re-executing the step c, and if not, executing the step g;
g. determining a corresponding smoothing factor according to the minimum calculation error of the smoothing factor, and taking the smoothing factor as an optimal smoothing factor;
and (c) in the step S42, the smoothing factor takes a value according to the optimal smoothing factor obtained in the step g.
Further defined, the lithium battery health state real-time monitoring method further comprises:
and S5, displaying the real-time information of the health state of the lithium battery.
A lithium battery state of health real-time monitoring system, comprising:
the lithium battery monitoring module is used for collecting parameter data of the lithium battery in real time;
the battery management module is used for processing the lithium battery parameter data acquired in real time through weighted moving average filtering; the method is used for carrying out normalization pretreatment on the processed lithium battery parameter data; the method comprises the steps of extracting characteristic principal component information from normalized and preprocessed lithium battery parameter data through a kernel entropy component analysis algorithm; the method comprises the steps of inputting characteristic principal element information into a generalized regression neural network model to obtain real-time information of the health state of the lithium battery;
and the display module is used for receiving and displaying the real-time information of the health state of the lithium battery.
Further defined, the lithium battery monitoring module includes:
temperature sensor, voltage detection device, current sensor, charge-discharge timer and charge-discharge cycle number detection device;
the battery management module includes:
the signal conditioning unit is used for amplifying and filtering the lithium battery temperature signal acquired by the temperature sensor, the lithium battery current signal acquired by the current sensor, the lithium battery voltage signal acquired by the voltage detection device, the lithium battery charge-discharge time signal acquired by the charge-discharge time and the lithium battery charge-discharge cycle number signal acquired by the charge-discharge cycle number detection device, respectively, converting the amplified and filtered lithium battery temperature signal, lithium battery current signal, lithium battery voltage signal, lithium battery charge-discharge time signal and lithium battery charge-discharge cycle number signal into corresponding digital signals, and transmitting the lithium battery parameter data acquired in real time to the storage unit through the communication unit;
the main control unit is used for taking the digital signal converted by the signal conditioning unit as lithium battery parameter data acquired in real time; processing the lithium battery parameter data acquired in real time through weighted moving average filtering; the method is used for carrying out normalization pretreatment on the processed lithium battery parameter data; the method comprises the steps of extracting characteristic principal component information from normalized and preprocessed lithium battery parameter data through a kernel entropy component analysis algorithm; the method comprises the steps of inputting characteristic principal element information into a generalized regression neural network model to obtain real-time information of the health state of the lithium battery, and sending lithium battery parameter data acquired in real time to a storage unit through a communication unit;
the communication unit is used for sending the obtained lithium battery health state real-time information to the display module and the storage unit, and sending the lithium battery parameter data acquired in real time to the storage unit;
and the storage unit is used for receiving and storing the lithium battery parameter data and the lithium battery health state real-time information acquired in real time.
A computer readable storage medium storing one or more programs, wherein one or more of the programs comprise instructions, which when executed by a computing device, cause the computing device to perform any of the methods described above.
The invention has the beneficial effects that:
1. according to the invention, through collecting various parameters of the battery and processing the collected data by utilizing weighted moving average filtering, the real-time performance of the collected lithium battery parameter data is better, the noise is smaller, and the calculation precision is improved; and then, the real-time state of health information of the current lithium battery is obtained by carrying out real-time calculation on the state of health of the lithium battery by using a generalized regression neural network model, so that on one hand, the calculation difficulty of the state of health information of the battery is reduced, the processing efficiency of the real-time state of health information of the lithium battery is improved, on the other hand, the large-scale data is not relied on, the convergence rate is improved, the calculation difficulty is reduced, and the accuracy, the stability and the reliability of the calculation result are also improved.
2. Compared with the traditional RBF network, the method has the advantages that the battery health state information is calculated by utilizing the generalized regression neural network model, the method is faster in convergence, only the smooth factor sigma is required to be set, and meanwhile, in order to eliminate the influence of artificial smooth factor setting on the neural network, the generalized regression neural network model is optimized by adopting an ant colony algorithm (IACA), so that the optimal smooth factor is obtained, the battery health state information is calculated by utilizing the generalized regression neural network model more accurately and reliably, and the actual use requirement is met.
Drawings
FIG. 1 is a flow chart of a method for monitoring the health status of a lithium battery in real time according to the present invention;
FIG. 2 is a flow chart of the method for monitoring the state of health of a lithium battery in real time according to the invention;
FIG. 3 is a diagram showing the relationship between the number of principal elements and the contribution value of the real-time monitoring method for the health status of the lithium battery;
FIG. 4 is a diagram showing a GRNN neural network structure of the method for monitoring the health status of a lithium battery in real time;
FIG. 5 is a flow chart of a method for acquiring a smoothing factor in a real-time monitoring method of the health status of a lithium battery according to the present invention;
fig. 6 is a schematic diagram of a real-time monitoring system for the health status of a lithium battery according to the present invention.
Detailed Description
It should be understood that the terms "comprises" and "comprising," when used in this specification and the appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
Example 1
Referring to fig. 1, the invention provides a method for monitoring the health status of a lithium battery in real time, which comprises the following steps:
s1, acquiring lithium battery parameter data in real time, and processing the lithium battery parameter data acquired in real time through weighted moving average filtering;
s2, carrying out normalization pretreatment on the lithium battery parameter data processed in the step S1;
s3, extracting the lithium battery parameter data subjected to normalization pretreatment in the step S2 through a KECA algorithm to obtain feature principal component information;
and S4, inputting the characteristic principal element information into a Generalized Regression Neural Network (GRNN) model to obtain the real-time information of the health state of the lithium battery.
Referring to fig. 2, step S1 includes the steps of:
s11, continuously collecting n lithium battery parameter data of each data type according to a collection period in a charge-discharge period of the current lithium battery, storing the n lithium battery parameter data of the same data type into a buffer area of the corresponding data type according to a collection sequence, wherein n is more than 2 and less than or equal to 10, and n is a positive integer; the data types comprise battery temperature, lithium battery voltage, lithium battery current, lithium battery discharge time and lithium battery charge and discharge times; the data types comprise battery temperature, lithium battery voltage, lithium battery current, lithium battery discharge time and lithium battery charge and discharge times;
in actual operation, continuous lithium battery parameter data acquisition is required to be performed in a charging and discharging period of a lithium battery to be detected or a current lithium battery, an acquisition interval of the acquisition period can be determined according to requirements, for example, a real-time battery health state is required to be concerned, an acquisition frequency can be increased, and if energy consumption is reduced, the acquisition frequency can be reduced.
The type of the data to be collected is determined according to the pearson correlation coefficient method, namely, the characteristic physical quantity of the lithium battery to be collected is determined according to the pearson correlation coefficient method, and the pearson correlation coefficient can well select the characteristic parameters with larger correlation degree with the SOH of the lithium battery, so that the problem of data redundancy caused by more characteristic parameters is solved.
The lithium battery parameter data acquired each time are lithium battery temperature data, lithium battery voltage data, lithium battery current data, lithium battery discharge time data and lithium battery charge and discharge times data which are five data types, namely the data types comprise battery temperature, lithium battery voltage, lithium battery current, lithium battery discharge time and lithium battery charge and discharge times.
When data acquisition is carried out for the first time, n times of lithium battery parameter data are required to be continuously acquired to obtain 5 types of data types, the number of lithium battery parameters of each data type is n, then the n lithium battery parameter data of each data type are sequentially stored in a buffer area of the data type according to the sequence of acquisition time, for example, the acquired n=10 lithium battery temperature data are firstly acquired and placed at the rightmost side, and then the rest 9 lithium battery temperature data are sequentially arranged from right to left to obtain a lithium battery temperature parameter buffer area; in this way, the lithium battery parameter data of the other four data types are stored in corresponding data type buffer areas, and 10 lithium battery parameter data of corresponding data types are placed in each buffer area.
S12, processing the n lithium battery parameter data in the same data type buffer area through weighted moving average filtering to obtain a weighted average value of the current corresponding data type
Wherein x is i The ith lithium battery parameter data, w, in the n lithium battery parameter data in the same data type buffer area i Is x i I is more than or equal to 1 and less than or equal to n, i is a positive integer, and n is the number of lithium battery parameter data in each data type cache region;
specifically, each data type is cachedThe n=10 lithium battery parameter data in the data type is subjected to weighted moving average filtering processing to obtain a weighted average value of the current data typeThe current time is determined by the collection time of the leftmost lithium battery parameter data in any data type buffer zone, and because the parameter data of five data types are collected simultaneously, the latest collection time according to any buffer zone can be used as the current time, and usually, the time difference exists between the collection time and the calculated result time, but the actual reading of the lithium battery health state information can be considered that the health state information calculated by the data of the latest collection time is the health state real-time information.
Each lithium battery parameter data is processed in the manner of carrying out weighted moving average filtering processing in the above formula to obtain the weighted average value of each type of lithium battery parameter data at the current momentFor example a weighted average of battery voltage dataWeighted average of battery current data +.>And a weighted average of the battery temperature data +.>Etc.
S13, continuously acquiring primary lithium battery parameter data according to an acquisition period, storing the primary lithium battery parameter data into corresponding data type cache areas according to an acquisition sequence, removing the lithium battery parameter data with earliest acquisition time in each cache area, and updating the lithium battery parameter data in each cache area;
specifically, as time passes, the lithium battery needs to be continuously collected once according to the collection period, the latest collected lithium battery parameter data is stored at the leftmost end of the data type buffer area, meanwhile, the lithium battery parameter data with the earliest collection time needs to be removed, the updating of the lithium battery parameter data in the data type buffer area is completed, and n, for example 10, lithium battery parameter data in each buffer area are ensured.
S14, repeatedly executing the step S12 and the step S13, and obtaining the weighted average value of the lithium battery parameter data of each data type in real time.
Then, the updated 10 lithium battery parameter data with the same data type in each buffer area are subjected to weighted moving average filtering treatment to obtain the latest dataAnd the weighted average value of the lithium battery parameter data after the weighted moving average filtering treatment can be obtained in real time according to the acquisition frequency by circulating in this way.
The step S2 specifically comprises the following steps:
the weighted average of the receiving step S12Carrying out normalization preprocessing on weighted averages of different data type cache areas to obtain preprocessed data D of the current data type Data type ：
Wherein,,for a weighted average of the corresponding data types +.>And->Determining from large to small by a common sorting algorithm, D Data type ∈[0,1]。
For example, a weighted average of each obtained lithium battery parameter dataThe weighted average value +.>Normalized preprocessed data D Data type 。
The step S3 specifically comprises the following steps:
s31, performing Parzen window on each piece of preprocessed data D obtained in real time Data type Establishing an N-dimensional nuclear matrix and carrying out feature decomposition to obtain a feature value of the current N-dimensional nuclear matrix and a feature vector of the current N-dimensional nuclear matrix, wherein N is the number of acquired lithium battery parameter data types;
wherein five groups of preprocessing data D obtained each time Data type And establishing a five-dimensional nuclear matrix through a Parzen window and carrying out feature decomposition to obtain the feature value of the current five-dimensional nuclear matrix and the feature vector of the current five-dimensional nuclear matrix.
S32, calculating Renyi entropy of the current N-dimensional kernel matrix, and sorting the eigenvalues of the current N-dimensional kernel matrix and the eigenvectors of the current N-dimensional kernel matrix according to the calculated entropy value;
specifically, the Renyi entropy of the five-dimensional kernel matrix is calculated, and the eigenvalue of the current N-dimensional kernel matrix and the eigenvector of the current N-dimensional kernel matrix are arranged according to the sequence from large to small according to the obtained entropy value.
S33, selecting the first B principal elements with the contribution value of Renyi entropy of the current N-dimensional kernel matrix more than 95%, and selecting the feature vectors of the first B principal elements as feature principal element information B, wherein 0< B < N.
Referring to fig. 3, as a calculation result of a Kernel Entropy Component Analysis (KECA) dimension reduction algorithm, the abscissa is the number of principal elements, and the ordinate is a contribution value, the calculation result can be obtained through experiments, and when 4 principal elements are selected as shown in the figure, the contribution rate has reached 99%, so that principal elements with lower contribution rate are removed, and the first four principal elements are selected as inputs of a neural network, so that data redundancy can be effectively reduced, calculation speed is increased, and accuracy is improved.
When the first 4 principal components exist, the requirement that the entropy contribution value of the current five-dimensional kernel matrix feature vector Renyi is more than 95% can be met, and at the moment, the feature vectors of the first 4 principal components are selected as feature principal component information B, so that dimension reduction is realized, and the processing complexity is reduced.
Referring to fig. 4, the step S4 includes the steps of:
s41, establishing a sample database, wherein the sample database comprises M sample data, the sample data comprises parameter data of a lithium battery and battery health state values Y obtained by detection when corresponding to the parameter data, and M sample data are selected from the M sample data to serve as training data, wherein m=0.6M-0.8M, and M is rounded to an integer;
the method comprises the steps that a sample database is required to be established, the sample database comprises battery health state values Y which are obtained by measuring different lithium battery voltages, lithium battery currents and/or battery charge and discharge times of lithium batteries under the same or different temperatures of the same type of batteries in a charge and discharge cycle period, each sample data comprises battery current data, battery voltage data, battery temperature data and battery capacitance data of the lithium batteries at any moment of the same type of lithium batteries, the battery health state at the moment is obtained by measuring the ratio of the current capacity of the battery to the rated capacity of the type of battery, the current capacity can be directly obtained by measuring, namely, the sample data can comprise parameters of four data types and the health state of the battery in the current parameter state, and the sample database comprises M sample data, for example, M is 2000.
M training data are then selected from the sample database, m=0.6m to 0.8M, e.g. 1400 sample data are randomly selected from the sample database as training data.
S42, establishing a generalized regression neural network model, and taking characteristic principal element information as input of an input layer;
calculating Gauss function value between the input layer and each training data to obtainOutput P of mode layer a ：
Wherein x is a For the a-th training data in m training data, e is a natural constant, sigma is a smoothing factor, sigma epsilon (0, 15)]B is characteristic principal element information;
the generalized regression neural network model is built to obtain an input layer, a mode layer, a summation layer and an output layer, and the input layer is calculated first.
Taking the feature principal element information obtained each time as an input layer, namely the feature principal element information B as an input layer, wherein each feature principal element information comprises 4 principal elements, calculating Gauss function values between the input layer and each training data at the moment, and taking the obtained result as an output P of a mode layer a And sent to the mode layer.
Referring to fig. 5, the value of the smoothing factor is selected as an optimal value according to an operation algorithm, and the calculation of the optimal smoothing factor is specifically as follows:
a. initializing ant colony, determining ant quantity as C, maximum iteration cycle number K, setting smoothing factor sigma, and setting initial pheromone concentration;
b. constructing a solution space, dividing the solution space into Q solution intervals, and randomly dispersing C ants in the Q solution intervals to obtain the initial position of each ant;
c. each ant moves according to a transition probability formula, and each ant obtains a moving path;
d. calculating the length of each moving path, updating the pheromone concentration by combining with a pheromone iteration formula, simultaneously recording the optimal pheromone concentration, and taking the optimal pheromone concentration as the current optimal smoothing factor sigma k K is the current cycle number, k=1 to K;
e. for the acquired lithium battery parameter data, the current optimal smoothness factor sigma is adopted k Calculating to obtain the health state predictive value y (sigma) k ) Using the obtained calculated state of health predictor y (σ k ) Corresponds to the sample dataCalculating a prediction error by using an actual battery state of health value Y corresponding to sample data of lithium battery parameter data to obtain a current optimal smoothness factor sigma k Is calculated error of (a);
wherein the current optimal smoothing factor sigma is adopted k Calculating any sample data to obtain a state of health prediction value y (sigma k )：
The input layer respectively inputs the ith sample data in the sample data, namely, the battery current data, the battery voltage data, the battery temperature data and the battery capacitance data in the current sample in the ith sample data as the input R of the input layer, and the output P of the mode layer a (σ k ) The method comprises the following steps:
then, the output of the first neuron node is obtainedThe method comprises the following steps:
the output S of the second neuron node d (σ k ) The method comprises the following steps:
calculating the output y (sigma) of the output layer k )：
R is input of an input layer, Y u The battery state of health value in the (u) th training data, m is the number of training data and x is the number of training data a The a-th training data in the m training data.
After obtainingBattery state of health prediction value y (sigma) calculated using current optimal smoothness factor k ) After that, through Calculating a predicted value y (sigma) k ) An actual battery state of health value Y corresponding to the sample data u And calculating a prediction error, thereby obtaining an error value of the current optimal smoothing factor.
f. Judging whether K is less than K, if so, clearing all moving paths, re-executing the step c, and if not, executing the step g;
g. determining a corresponding smoothing factor according to the minimum calculation error of the smoothing factor, taking the smoothing factor as an optimal smoothing factor, and;
and obtaining a final smoothing factor according to the steps, and substituting the obtained final smoothing factor into the step S42 for calculation to obtain the output of the mode layer.
S43, calculating the output of each neuron node of the summation layer respectively, wherein the number of the neuron nodes of the summation layer is 2:
the output S of the first neuron node D The method comprises the following steps:
the output S of the second neuron node d The method comprises the following steps:
wherein Y is a The battery state of health value in the a-th training data, m is the number of training data;
s44, calculating output y of the output layer:
wherein y is current real-time information of the lithium battery health state;
only one output layer is needed as the battery state of health is calculated only; from this, the lithium battery parameter data is collected in real time according to the step S1, and then the current real-time information of the lithium battery health state is obtained by calculating the lithium battery parameter data collected each time according to the steps S2 to S4.
Example 2
Referring to fig. 6, based on embodiment 1, this embodiment provides a real-time monitoring system for health status of a lithium battery, including:
the lithium battery monitoring module is used for collecting parameter data of the lithium battery in real time;
the battery management module is used for processing the lithium battery parameter data acquired in real time through weighted moving average filtering; the method is used for carrying out normalization pretreatment on the processed lithium battery parameter data; extracting the lithium battery parameter data subjected to normalization pretreatment through a KECA algorithm to obtain feature principal element information; the method comprises the steps of inputting characteristic principal element information into a generalized regression neural network model to obtain real-time information of the health state of the lithium battery;
and the display module is used for receiving and displaying the real-time information of the health state of the lithium battery.
Wherein, lithium cell monitoring module includes:
temperature sensor, voltage detection device, current sensor, charge-discharge timer and charge-discharge cycle number detection device;
temperature is an important influencing factor of the health state of a lithium battery, so that high requirements are placed on the accuracy of temperature measurement. The system adopts a PT100 temperature sensor, the precision of which can reach within +/-0.1 ℃, and the requirement of the system on the precision is met. PT100 can directly measure the temperature of liquid, gas and various mediums in the range of-50 ℃ to +600 ℃ in the industrial process, converts the temperature signal into a 4-20 mA current output signal which is linear with the temperature signal, and the power supply voltage is 24V, and an AD converter with 12-bit resolution is used for measurement.
Measuring voltage is an important step for monitoring the health state of a lithium battery, and has higher precision requirements for voltage measurement. The system adopts a differential mode measurement method to measure the battery voltage. Differential mode measurement is the determination of the differential voltage between two independent points in a circuit. Measuring the voltage across a single resistor requires measuring across the resistor. The voltage difference is the terminal voltage across the resistor. Therefore, when the number of series-parallel batteries is large and the requirement on measurement accuracy is high, a differential mode measurement method is generally adopted.
The current measurement of the lithium battery generally has two modes, namely split-flow detection, namely current measurement is carried out by a series resistor; the other is to monitor the current with a hall sensor. The system adopts a series resistance method to measure the current.
And monitoring the charge-discharge cycle times and the discharge time, monitoring the charge-discharge cycle times and the discharge time of the lithium battery by the main control unit, and storing data.
The battery management module includes:
the signal conditioning unit is used for amplifying and filtering the lithium battery temperature signal acquired by the temperature sensor, the lithium battery current signal acquired by the current sensor, the lithium battery voltage signal acquired by the voltage detection device, the lithium battery charge-discharge time signal acquired by the charge-discharge time and the lithium battery charge-discharge cycle number signal acquired by the charge-discharge cycle number detection device, respectively, converting the amplified and filtered lithium battery temperature signal, lithium battery current signal, lithium battery voltage signal, lithium battery charge-discharge time signal and lithium battery charge-discharge cycle number signal into corresponding digital signals, and transmitting the lithium battery parameter data acquired in real time to the storage unit through the communication unit;
the main control unit is used for taking the digital signal converted by the signal conditioning unit as lithium battery parameter data acquired in real time; processing the lithium battery parameter data acquired in real time through weighted moving average filtering; the method is used for carrying out normalization pretreatment on the processed lithium battery parameter data; extracting the lithium battery parameter data subjected to normalization pretreatment through a KECA algorithm to obtain feature principal element information; the method comprises the steps of inputting characteristic principal element information into a generalized regression neural network model to obtain real-time information of the health state of the lithium battery, and sending lithium battery parameter data acquired in real time to a storage unit through a communication unit;
the communication unit is used for sending the obtained lithium battery health state real-time information to the display module and the storage unit, and sending the lithium battery parameter data acquired in real time to the storage unit;
and the storage unit is used for receiving and storing the lithium battery parameter data and the lithium battery health state real-time information acquired in real time.
The present embodiment also provides a computer-readable storage medium storing one or more programs, wherein one or more of the programs comprise instructions, which when executed by a computing device, cause the computing device to perform any of the methods according to embodiment 1.
The present embodiment also provides a computing device comprising a memory, one or more processors, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by one or more of the processors, the one or more programs comprising instructions for performing any of the methods according to embodiment 1.
Claims (10)
1. The method for monitoring the health state of the lithium battery in real time is characterized by comprising the following steps of:
s1, acquiring lithium battery parameter data in real time, and processing the lithium battery parameter data acquired in real time through weighted moving average filtering;
s2, carrying out normalization pretreatment on the lithium battery parameter data processed in the step S1;
s3, extracting the lithium battery parameter data subjected to normalization pretreatment in the step S2 through a kernel entropy component analysis algorithm to obtain feature principal component information;
and S4, inputting the characteristic principal element information into a generalized regression neural network model to obtain the real-time information of the health state of the lithium battery.
2. The method for monitoring the health status of a lithium battery according to claim 1, wherein the step S1 comprises the steps of:
s11, continuously collecting n lithium battery parameter data of each data type according to a collection period in a charge-discharge period of the current lithium battery, storing the n lithium battery parameter data of the same data type into a buffer area of the corresponding data type according to a collection sequence, wherein n is more than 2 and less than or equal to 10, and n is a positive integer; the data types comprise battery temperature, lithium battery voltage, lithium battery current, lithium battery discharge time and lithium battery charge and discharge times;
s12, processing the n lithium battery parameter data in the same data type buffer area through weighted moving average filtering to obtain a weighted average value of the current corresponding data type
Wherein x is i The ith lithium battery parameter data, w, in the n lithium battery parameter data in the same data type buffer area i Is x i I is more than or equal to 1 and less than or equal to n, i is a positive integer, and n is the number of lithium battery parameter data in each data type cache region;
s13, continuously acquiring primary lithium battery parameter data according to an acquisition period, storing the primary lithium battery parameter data into corresponding data type cache areas according to an acquisition sequence, removing the lithium battery parameter data with earliest acquisition time in each cache area, and updating the lithium battery parameter data in each cache area;
s14, repeatedly executing the step S12 and the step S13, and obtaining the weighted average value of the lithium battery parameter data of each data type in real time.
3. The method for monitoring the health status of a lithium battery according to claim 2, wherein the step S2 specifically comprises:
receiving a weighted average of each data type obtained in step S12And carrying out normalization preprocessing on the weighted average value of each data type to obtain preprocessed data D of the current data type Data type ，D Data type ∈[0,1]。
4. The method for monitoring the health status of a lithium battery according to claim 3, wherein the step S3 specifically comprises the following steps:
s31, performing Parzen window on each piece of preprocessed data D obtained in real time Data type Establishing an N-dimensional nuclear matrix and carrying out feature decomposition to obtain a feature value of the current N-dimensional nuclear matrix and a feature vector of the current N-dimensional nuclear matrix, wherein N is the number of acquired lithium battery parameter data types;
s32, calculating Renyi entropy of the current N-dimensional kernel matrix, and sorting the eigenvalues of the current N-dimensional kernel matrix and the eigenvectors of the current N-dimensional kernel matrix according to the calculated entropy value;
s33, selecting the first B principal elements with the contribution value of Renyi entropy of the current N-dimensional kernel matrix more than 95%, and selecting the feature vectors of the first B principal elements as feature principal element information B, wherein 0< B < N.
5. The method for monitoring the health status of a lithium battery according to claim 4, wherein the step S4 comprises the steps of:
s41, establishing a sample database, wherein the sample database comprises M sample data, the sample data comprises parameter data of a lithium battery and battery health state values Y obtained by detection when corresponding to the parameter data, and M sample data are selected from the M sample data to serve as training data, wherein m=0.6M-0.8M, and M is rounded to an integer;
s42, establishing a generalized regression neural network model, and taking characteristic principal element information as an input layer;
calculating Gauss function value between the input layer and each training data to obtain output P of the mode layer a ：
Wherein x is a For the a-th training data in m training data, e is a natural constant, sigma is a smoothing factor, sigma epsilon (0, 15)]B is characteristic principal element information;
s43, calculating the output of each neuron node of the summation layer respectively, wherein the number of the neuron nodes of the summation layer is 2:
the output S of the first neuron node D The method comprises the following steps:
the output S of the second neuron node d The method comprises the following steps:
wherein Y is a The battery state of health value in the a-th training data, m is the number of training data;
s44, calculating output y of the output layer:
wherein y is the current real-time information of the lithium battery health state.
6. The method for real-time monitoring of the health status of a lithium battery according to claim 5, wherein the determination of the smoothness factor comprises the steps of:
a. initializing ant colony, determining ant quantity as C, maximum iteration cycle number K, setting smoothing factor sigma, and setting initial pheromone concentration;
b. constructing a solution space, dividing the solution space into Q solution intervals, and randomly dispersing C ants in the Q solution intervals to obtain the initial position of each ant;
c. each ant moves according to a transition probability formula, and each ant obtains a moving path;
d. calculating the length of each moving path, updating the pheromone concentration by combining with a pheromone iteration formula, simultaneously recording the optimal pheromone concentration, and taking the optimal pheromone concentration as the current optimal smoothing factor sigma k K is the current cycle number, k=1 to K;
e. for any sample data by the current optimal smoothing factor sigma k Calculating to obtain the health state predictive value y (sigma) k ) Using the obtained calculated state of health predictor y (σ k ) Calculating a prediction error according to the actual battery state of health value Y corresponding to the sample data to obtain a current optimal smoothness factor sigma k Is calculated error of (a);
f. judging whether K is less than K, if so, clearing all moving paths, re-executing the step c, and if not, executing the step g;
g. determining a corresponding smoothing factor according to the minimum calculation error of the smoothing factor, taking the smoothing factor as an optimal smoothing factor, and;
and (c) in the step S42, the smoothing factor takes a value according to the optimal smoothing factor obtained in the step g.
7. The method for monitoring the health status of a lithium battery in real time according to claim 6, further comprising:
and S5, displaying the real-time information of the health state of the lithium battery.
8. A lithium battery state of health real-time monitoring system, comprising:
the lithium battery monitoring module is used for collecting parameter data of the lithium battery in real time;
the battery management module is used for processing the lithium battery parameter data acquired in real time through weighted moving average filtering; the method is used for carrying out normalization pretreatment on the processed lithium battery parameter data; the method comprises the steps of extracting characteristic principal component information from normalized and preprocessed lithium battery parameter data through a kernel entropy component analysis algorithm; the method comprises the steps of inputting characteristic principal element information into a generalized regression neural network model to obtain real-time information of the health state of the lithium battery;
and the display module is used for receiving and displaying the real-time information of the health state of the lithium battery.
9. The lithium battery health status real-time monitoring system of claim 8, wherein the lithium battery monitoring module comprises:
temperature sensor, voltage detection device, current sensor, charge-discharge timer and charge-discharge cycle number detection device;
the battery management module includes:
the signal conditioning unit is used for amplifying and filtering the lithium battery temperature signal acquired by the temperature sensor, the lithium battery current signal acquired by the current sensor, the lithium battery voltage signal acquired by the voltage detection device, the lithium battery charge-discharge time signal acquired by the charge-discharge time and the lithium battery charge-discharge cycle number signal acquired by the charge-discharge cycle number detection device, respectively, converting the amplified and filtered lithium battery temperature signal, lithium battery current signal, lithium battery voltage signal, lithium battery charge-discharge time signal and lithium battery charge-discharge cycle number signal into corresponding digital signals, and transmitting the lithium battery parameter data acquired in real time to the storage unit through the communication unit;
the main control unit is used for taking the digital signal converted by the signal conditioning unit as lithium battery parameter data acquired in real time; processing the lithium battery parameter data acquired in real time through weighted moving average filtering; the method is used for carrying out normalization pretreatment on the processed lithium battery parameter data; the method comprises the steps of extracting characteristic principal component information from normalized and preprocessed lithium battery parameter data through a kernel entropy component analysis algorithm; the method comprises the steps of inputting characteristic principal element information into a generalized regression neural network model to obtain real-time information of the health state of the lithium battery, and sending lithium battery parameter data acquired in real time to a storage unit through a communication unit;
the communication unit is used for sending the obtained lithium battery health state real-time information to the display module and the storage unit, and sending the lithium battery parameter data acquired in real time to the storage unit;
and the storage unit is used for receiving and storing the lithium battery parameter data and the lithium battery health state real-time information acquired in real time.
10. A computer readable storage medium storing one or more programs, wherein one or more of the programs comprise instructions, which when executed by a computing device, cause the computing device to perform any of the methods of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310465215.XA CN116643190A (en) | 2023-04-26 | 2023-04-26 | Real-time monitoring method and system for lithium battery health state |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310465215.XA CN116643190A (en) | 2023-04-26 | 2023-04-26 | Real-time monitoring method and system for lithium battery health state |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116643190A true CN116643190A (en) | 2023-08-25 |
Family
ID=87642503
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310465215.XA Pending CN116643190A (en) | 2023-04-26 | 2023-04-26 | Real-time monitoring method and system for lithium battery health state |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116643190A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117648661A (en) * | 2024-01-30 | 2024-03-05 | 长春汽车工业高等专科学校 | Battery fault classification and identification method and system based on big data |
-
2023
- 2023-04-26 CN CN202310465215.XA patent/CN116643190A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117648661A (en) * | 2024-01-30 | 2024-03-05 | 长春汽车工业高等专科学校 | Battery fault classification and identification method and system based on big data |
| CN117648661B (en) * | 2024-01-30 | 2024-04-19 | 长春汽车工业高等专科学校 | Battery fault classification and identification method and system based on big data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Oji et al. | Data-driven methods for battery soh estimation: Survey and a critical analysis | |
| Fan et al. | A novel machine learning method based approach for Li-ion battery prognostic and health management | |
| CN113343633A (en) | Thermal runaway fault classification and risk prediction method and system for power lithium battery | |
| CN113917334B (en) | Battery health state estimation method based on evolution LSTM self-encoder | |
| CN112098873B (en) | Lithium battery health state estimation method based on charging voltage curve geometric characteristics | |
| Su et al. | Prediction of remaining useful life for lithium-ion battery with multiple health indicators | |
| Venugopal et al. | Analysis of optimal machine learning approach for battery life estimation of Li-ion cell | |
| CN113447828A (en) | Lithium battery temperature estimation method and system based on Bayesian neural network | |
| Xu et al. | State-of-charge estimation and health prognosis for lithium-ion batteries based on temperature-compensated Bi-LSTM network and integrated attention mechanism | |
| CN116400244B (en) | Abnormality detection method and device for energy storage battery | |
| CN115792637A (en) | Lithium battery health state estimation method | |
| CN116643190A (en) | Real-time monitoring method and system for lithium battery health state | |
| Jafari et al. | Optimizing battery RUL prediction of lithium-ion batteries based on Harris hawk optimization approach using random forest and LightGBM | |
| CN116125306A (en) | Power battery thermal runaway diagnosis method and system based on hybrid prediction framework | |
| CN114036647A (en) | Power battery safety risk assessment method based on real vehicle data | |
| Wang et al. | An efficient state-of-health estimation method for lithium-ion batteries based on feature-importance ranking strategy and PSO-GRNN algorithm | |
| Wang et al. | A conditional random field based feature learning framework for battery capacity prediction | |
| Song et al. | Capacity estimation method of lithium-ion batteries based on deep convolution neural network | |
| Shan et al. | Review of Various Machine Learning Approaches for Predicting Parameters of Lithium-Ion Batteries in Electric Vehicles | |
| Zhu et al. | State of Health Estimation of Lithium‐Ion Battery Using Time Convolution Memory Neural Network | |
| Chang et al. | Micro-short circuit fault diagnosis of lithium-ion battery based on voltage curve similarity ranking volatility | |
| Bamati et al. | Virtual Temperature Sensor in Battery Thermal Management System Using LSTM-DNN | |
| CN115389947B (en) | Lithium battery health state prediction method and device, electronic equipment and storage medium | |
| CN112485676B (en) | Battery energy storage system state estimation early warning method under digital mirror image | |
| CN118393364A (en) | Multi-fault diagnosis method, device, medium and product of power battery system |
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 |