CN112684346A - Lithium battery health state estimation method based on genetic convolutional neural network - Google Patents
Lithium battery health state estimation method based on genetic convolutional neural network Download PDFInfo
- Publication number
- CN112684346A CN112684346A CN202011434822.2A CN202011434822A CN112684346A CN 112684346 A CN112684346 A CN 112684346A CN 202011434822 A CN202011434822 A CN 202011434822A CN 112684346 A CN112684346 A CN 112684346A
- Authority
- CN
- China
- Prior art keywords
- battery
- network structure
- cnn
- network
- data
- 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.)
- Granted
Links
Images
Abstract
The invention discloses a lithium battery health state estimation method based on a genetic convolution neural network, which specifically comprises the following steps: charging and discharging different types of lithium batteries under a constant current condition until the end of the service life of the batteries is recorded, and forming a constant current charging voltage curve of the lithium batteries; after each charging of the battery, determining the current capacity of the battery as a true value of the CNN model; characterizing a voltage curve by using the characteristic points of the recorded voltage curve, and using the voltage curve as input data of a CNN model; initializing a network structure and parameters; grouping the processed training set data, and training each CNN network; and inputting the processed test set data into a group of CNN network structures, and selecting the network structure with the minimum mean square error between a real value and a predicted value as a final prediction model.
Description
Technical Field
The invention belongs to the technical field of battery management, and relates to a lithium battery health state estimation method based on a genetic convolutional neural network.
Background
Lithium Ion Batteries (LIBs) have been widely used in the fields of electric vehicles, electric tools, base station backup power supplies, etc. because of their advantages of high energy density, long life, strong stability, and little impact on the environment. In practical application, a series of irreversible chemical reactions occur inside the battery along with one-time charge and discharge, so that the battery is gradually aged and is expressed as capacity decline, power loss and the like. Therefore, it is necessary to estimate the health state of the battery in advance during the use of the battery, and it can send out an early warning message when the battery life reaches the end, and prompt the user or the equipment provider to replace the battery in time.
The state of degradation (aging) of a battery is generally described by the state of health (SOH) of the battery. The SOH can be obtained by directly measuring or indirectly calculating the ratio of the current value to the initial value of a certain characteristic parameter, the current SOH of the battery needs to be measured by an instrument after each charge and discharge of the battery, and in actual life, the battery cannot be detached and measured anytime and anywhere, so that the method is not suitable.
Disclosure of Invention
The invention aims to provide a lithium battery health state estimation method based on a genetic convolutional neural network, which can be used for rapidly and accurately predicting the battery health state on the basis of automatically learning a network structure.
The technical scheme adopted by the invention is that the lithium battery health state estimation method based on the genetic convolution neural network is implemented according to the following steps:
step 9, grouping the processed training set data, wherein each 5 training sets form a group, circularly inputting each group of training sets into a CNN network structure, and training each CNN network;
step 10, automatically learning the network structure by population evolution through a genetic algorithm to obtain a group of good CNN network structures;
step 11, inputting the processed test set data into a group of CNN network structures, and selecting the network structure with the minimum mean square error between a real value and a predicted value as a final prediction model; when the voltage data of another battery is input again, the residual capacity of the battery can be predicted through the trained model, and the SOH of the battery can be calculated quickly and accurately.
The specific steps of the step 5 treatment are as follows: according to 135 voltage curves, 8 characteristic points are respectively extracted from each voltage curve to form a matrix with 135 rows and 8 columns, 8 data of each row are converted into a matrix with 2 rows and 4 columns, and the matrix is input into a CNN model as an image.
In step 6, the data set is divided into 80% training set and 20% testing set.
In step 8, a group of CNN network structures is initialized randomly, the group of CNN network structures is regarded as a population, each individual in the population represents one CNN network structure, and the size and generation number of the population are initialized.
The step 10 specifically comprises:
step 10.1, encoding the network structure, dividing the network structure into different stages for each network structure, namely an individual, by taking the pooling layer as a boundary, encoding each stage, combining the stages into a fixed binary character string, and expressing each network structure by using a genotype;
step 10.2, evaluating the fitness value of the individual, inputting voltage data into the individual for training, and taking MSE (mean square error) between the real value and the predicted value as the fitness value of the individual;
step 10.3: generating a new network structure through selection, intersection and variation, evaluating the fitness value of the new network structure, reserving individuals with low fitness value in each iteration, deleting individuals with high fitness value, always keeping the population size N until the maximum generation number is reached, finishing the iteration, and outputting a group of good CNN network structures.
The invention has the beneficial effects that: according to the SOH estimation method, the SOH estimation model is established through off-line data, the SOH value is estimated on line in real time, and the estimation of the SOH value can be conveniently realized; the CNN network structure is automatically designed through binary coding, and a group of network structures with the best performance are automatically searched by using a genetic algorithm, so that the search space is enlarged, and the operation efficiency is improved; according to the method, through dichotomy sampling, the optimal health characteristics are selected from the charging voltage curve and are converted into the matrix to be used as the input of the model, so that the optimal health characteristics are effectively utilized, and the calculation complexity is reduced.
Drawings
FIG. 1 is a dichotomy sampling diagram in the lithium battery health state estimation method based on the genetic convolutional neural network of the present invention;
FIG. 2 is a row of transformed data image in the method for estimating the health status of a lithium battery based on a genetic convolutional neural network according to the present invention;
FIG. 3 is a diagram of a single stage structure in the method for estimating the health status of a lithium battery based on a genetic convolutional neural network according to the present invention;
FIG. 4 is a diagram of network structure coding operation in the lithium battery health state estimation method based on the genetic convolutional neural network of the present invention;
FIG. 5 is a flow chart of an evolutionary network structure in the method for estimating the health state of a lithium battery based on a genetic convolutional neural network according to the present invention;
fig. 6 is a general flowchart of the lithium battery health state estimation method based on the genetic convolutional neural network according to the present invention.
Detailed Description
The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.
As shown in fig. 6, a method for estimating the health state of a lithium battery based on a genetic convolutional neural network is specifically implemented according to the following steps:
step 9, grouping the processed training set data, wherein each 5 training sets form a group, circularly inputting each group of training sets into a CNN network structure, and training each CNN network;
step 10, automatically learning the network structure by population evolution through a genetic algorithm to obtain a group of good CNN network structures;
step 11, inputting the processed test set data into a group of CNN network structures, and selecting the network structure with the minimum mean square error between a real value and a predicted value as a final prediction model; when the voltage data of another battery is input again, the residual capacity of the battery can be predicted through the trained model, and the SOH of the battery can be calculated quickly and accurately.
The specific steps of the step 5 treatment are as follows: according to 135 voltage curves, 8 characteristic points are respectively extracted from each voltage curve to form a matrix with 135 rows and 8 columns, 8 data of each row are converted into a matrix with 2 rows and 4 columns, and the matrix is input into a CNN model as an image. The specific transformation is shown in FIG. 2.
In step 6, the data set is divided into 80% training set and 20% testing set.
In step 8, a group of CNN network structures is initialized randomly, the group of CNN network structures is regarded as a population, each individual in the population represents one CNN network structure, and the size and generation number of the population are initialized.
The step 10 specifically comprises:
step 10.1, encoding the network structure, dividing the network structure into different stages for each network structure, namely an individual, by taking the pooling layer as a boundary, encoding each stage, combining the stages into a fixed binary character string, and expressing each network structure by using a genotype;
step 10.2, evaluating the fitness value of the individual, inputting voltage data into the individual for training, and taking MSE (mean square error) between the real value and the predicted value as the fitness value of the individual;
step 10.3: generating a new network structure through selection, intersection and variation, evaluating the fitness value of the new network structure, reserving individuals with low fitness value in each iteration, deleting individuals with high fitness value, always keeping the population size N until the maximum generation number is reached, finishing the iteration, and outputting a group of good CNN network structures.
Stage and node division:
the possible stage structure in a part of single network is shown in fig. 3, each CNN network structure is composed of S stages randomly, and the symbol Vs，ksDenotes the S-th stage (S ═ 1,2,3 … S), and Ks denotes Vs，ksThe number of nodes included in the phase. Each node represents a convolution operation, adjacent stages are connected by a pooling operation, and an average pooling operation is performed at the end of the last stage.
A default node:
two default nodes are set in each stage, and the default input nodes represent Vs,0Accepting data from a previous stage, performing a convolution operation, and sending an output to each node without a preceding node, e.g. Vs,1。Vs,Ks+1Is a default output node that accepts data such as V from all nodes without subsequent nodess,ksThe convolution is performed on them in summary and the output is sent to the pooling layer, the connection between the normal node and the default node does not need to be encoded.
And (4) common nodes.
The common nodes except the default node in one stage have unique and ordered serial numbers, each node represents a convolution operation, the Ks node numbers are ordered from small to large, and the symbol Vs,ksIndicating that the s-th stage contains Ks nodes.
And (4) performing operation among the nodes.
After summing the values of all input nodes (connected to its lower numbered nodes), a convolution is performed followed by batch normalization, the Relu operation.
And (3) encoding rules:
use of
Bit-to-code directed edges between nodes within a phase, the first bit representing (V)s,1,Vs,2) Whether there is a directed edge connection between nodesNext, the next two bits represent (V)s,1,Vs,3)，(Vs,2,Vs,3) If there is a directed edge connection, the code is 1 if there is a directed edge connection, otherwise it is 0. Adding a bit at the end of the encoding indicates skipping the connection and forwarding the input information directly to the output, bypassing the entire block.
Special cases are as follows:
and if the isolated node exists, the isolated node is ignored and does not participate in coding, and the default node is not connected with the isolated node. If there is no connection in a phase, i.e. all bits of the phase are 0, the convolution operation is performed only once.
Specific examples are as follows:
(1) the method is carried out by using a lithium battery which is a repository of the American aerospace agency and is 18650, the rated capacity is 2Ah, and B5 batteries and B6 batteries are selected.
(2) The method comprises the steps of carrying out cyclic charge and discharge on a B5 lithium battery under the condition of constant current 4A, obtaining a constant-current charging voltage change curve after each charge, obtaining 135 charging voltage curves after 135 cyclic charge and discharge until the battery is scrapped, and only recording charging voltage data of the B6 lithium battery.
(3) Before each time of charging the battery, measuring the residual capacity of the battery before charging, and after the battery is fully charged (the voltage is not changed), measuring the current capacity of the battery, wherein the current battery capacity is equal to the sum of the charging amount and the residual capacity required by the current charging, the current capacity of the battery is used as the true value of the CNN model, the output of the CNN model is used as the predicted value, and the optimal CNN model is trained by continuously reducing the error between the true value and the predicted value.
(4) For each voltage curve recorded in step 2, 8 points are sampled by using a bisection method as feature points, 8 feature points are used for representing one voltage curve and are used as input data of the CNN model, the sampling result is shown in fig. 1, and the specific sampling process is as follows:
in fig. 1, the horizontal axis represents time variation, the vertical axis represents voltage variation, the curve in the figure is a charging voltage curve, a characteristic point is extracted from the charging voltage curve, the midpoint extracting operation is performed through the voltage variation of the vertical axis, and then whether the midpoint extracting is in the front half part or the back half part of the current voltage midpoint is determined through the proportion of time of the horizontal axis. The starting voltage of 2.8V and the terminal voltage of 4.2V are known at present.
Taking the terminal voltage of 4.2V as a 1 st characteristic point F1;
a middle point is taken between the starting voltage and the terminal voltage, the obtained voltage value is 3.5V and serves as a 2 nd characteristic point F2, 2.8V-3.5V is newly divided into a first half part and 3.5V-4.2V is newly divided into a second half part.
After the second characteristic point is determined, the proportion of the corresponding time on the voltage curve in the whole time is observed, the first half part is 0-1700t, the second half part is 1700t-12300t, and the second half part time interval is larger, so that the middle point is taken in the second half part of the voltage from 3.5V to 4.2V, the obtained voltage value is 3.85V as the 3 rd characteristic point F3, the voltage value from 3.5V to 3.85V is newly divided into the first half part and the voltage value from 3.85V to 4.2V is newly divided into the second half part.
The first half part of the time interval corresponding to the 3 rd characteristic point is 1700t-8000t, the second half part is 8000t-14000t, and the time intervals are not different much, so that the middle points are respectively taken in the two voltage intervals.
The specific process of extracting the characteristic points of the first half part 1700t-8000t is as follows:
the voltage corresponding to the first half part is 3.5V-3.85V, the 3.675V obtained from the middle point is taken as the 4 th characteristic point F4, the 3.5V-3.675V is newly divided into the first half part and the 3.675V-3.85V is taken as the second half part.
The first half part of the time interval corresponding to the 4 th characteristic point is 1700t-2900t, the second half part is 2900t-8000t, and the second half part is comparatively large, so that the middle point is taken in the voltage of 3.675V-3.85V in the second half part, obtained 3.7625V is taken as the 5 th characteristic point F5, 3.675V-3.7625V is newly divided as the first half part, and 3.7625V-3.85V is taken as the second half part.
The first half of the time interval corresponding to the 5 th characteristic point is 2900t-4050t, the second half is 4050t-6300t, and the time of the second half is relatively large, so that the middle point is taken from 3.7625V-3.85V in the second half, 3.80625V is obtained as the 6 th characteristic point F6, 3.7625V-3.80625V is newly divided into the first half and 3.80625V-3.85V is adopted as the second half.
The specific process of extracting the feature points in the second half 8000t-14000t is as follows:
the second half is 3.85V-4.2V, the middle point is taken to obtain 4.025V as the 7 th characteristic point F7, the middle point is subdivided into 3.85V-4.025V as the first half, and the middle point is subdivided into 4.025V-4.2V as the second half.
The 4.025V time interval has the first half of 8000t-10300t and the second half of 10300t-12300t, and the first half accounts for a relatively large amount, so that the middle point is taken at the first half of 3.85V-4.025V, and the obtained 3.9375 is used as the 8 th characteristic point.
(5) According to 135 voltage curves, 8 characteristic points are respectively extracted from each voltage curve to form a matrix with 135 rows and 8 columns, wherein the data of each row corresponds to a current capacity value of the battery, and the 8 data of each row are converted into a matrix with 2 rows and 4 columns to be used as the input of the CNN model, and the specific conversion mode is shown in FIG. 2.
(6) The data set was divided, and 135 pieces of data were divided into 80% of training set and 20% of test set.
(7) The SOH is calculated. The current state of health of the lithium battery is estimated by calculating the ratio of the current capacity of the battery to the rated capacity of the battery when leaving the factory, the current SOH of the battery can be calculated by obtaining the current capacity value of the battery because the rated capacity of the battery when leaving the factory is given, and the SOH is estimated by fitting the current capacity of the battery.
The expression for calculating SOH is as follows:
wherein, CcurrentRepresenting the current capacity, C, of the batterynewIndicating the rated capacity of the battery when it leaves the factory.
(8) The network model and the parameters are initialized. Randomly generating training weights, setting parameters (such as a learning rate of 0.01, a batch size of 5, a round epoch of 50, a network stage number S of 3, and an intra-stage node number ks of 6) required in network training, and randomly initializing a group of CNN network structures, wherein possible stage structures in a part of single networks are shown in FIG. 3, and one CNN network structure is formed by combining several stages. Considering a group of network structures as a population, one individual in the population represents one CNN network structure, and the initial population size is 40 and the generation number is 20.
(9) Grouping the processed training set data, wherein each 5 training sets form a group, circularly inputting each group of training sets into a CNN network structure, and training each CNN network, wherein the specific steps are as follows.
a. Inputting a set of data to each CNN network;
b. calculating forward propagation to obtain a loss function loss;
c. calculating gradients through back propagation;
d. iteratively updating the weight parameters of the network with the portion of the gradient;
and continuously circulating the above four steps until all the rounds of training are finished.
(10) And (3) evolving the population through a genetic algorithm to realize automatic learning of the network structure, so as to obtain a group of good CNN network structures.
The initialized network structure is encoded as shown in fig. 4.
Here, the number of stages S is set to 3, and the nodes in each stage are (K1, K2, K3) to (6, 6, 6). Placing a maximum pooling layer with the step length of 2 after the first stage and the second stage, extracting main features, setting a global average pooling layer after the last stage, carrying out average sampling, encoding each stage, and combining each encoded stage into a fixed binary character string. Each coded network architecture is called an individual, all the network architectures form a group, the size of the group is N, and each generation is kept unchanged. Each binary string has a length L of 45, meaning 245A total of 1200 network architectures are searched for, because of the possible individuals (network architectures).
Fitness values of individuals are evaluated.
Inputting training set data into each network structure (individual) in the population for training, and performing iterative training on the network structure on the training set by using a standard random gradient descent (SGD) back propagation algorithm and a cosine annealing learning rate scheme until a proper weight is obtained, so that a loss function is reduced, and mse between a predicted value and a true value obtained in the training process is used as an adaptability value of the network structure (individual).
And generating a new network structure through selection, intersection and variation, evaluating the fitness value of the network structure in the population, adopting the average value of the historical fitness of the old network structure as the fitness, and calculating mse as the fitness value of the newly generated network structure. And (3) retaining the individuals with low fitness value in each iteration, deleting the individuals with high fitness value, always keeping the population size N until the maximum generation number is reached, ending the iteration, and outputting a group of good CNN network structures.
(11) And inputting the processed data of the test set into a group of CNN network structures obtained currently, and selecting a network structure with the minimum mse (mean square error) between a real value and a predicted value as a final prediction model. When voltage data of the B6 battery is input, the SOH of the battery can be quickly and accurately calculated through the capacity data predicted by the model. Therefore, in practical application, the trained model can be directly used for estimating the current SOH of the battery only by recording the charging voltage data of the constant current of the battery.
Claims (5)
1. A lithium battery health state estimation method based on a genetic convolution neural network is characterized by comprising the following steps:
step 1, calculating the rated capacity of lithium batteries when the lithium batteries leave a factory according to different types of lithium batteries;
step 2, charging and discharging different types of lithium batteries under a constant current condition, recording voltage data under charging work in real time until the end of the service life of the batteries is recorded, forming a constant current charging voltage curve of the lithium batteries according to the recorded data, and acquiring aging characteristics of the batteries from the voltage curve;
step 3, after the battery in the step 2 is charged every time, determining the current capacity of the battery as a true value of the CNN model, and continuously reducing the error between the true value and a predicted value to train the optimal CNN model;
step 4, sampling 8 points of each voltage curve recorded in the step 2 by using a dichotomy as characteristic points, representing one voltage curve by using 8 characteristic points, and using the voltage curve as input data of the CNN model;
step 5, processing the characteristic points in the step 4;
step 6, dividing a data set;
step 7, estimating the SOH of the current lithium battery according to the ratio of the current capacity of the battery to the rated capacity of the battery when the battery leaves a factory;
step 8, initializing a network structure and parameters;
step 9, grouping the processed training set data, wherein each 5 training sets form a group, circularly inputting each group of training sets into a CNN network structure, and training each CNN network;
step 10, automatically learning the network structure by population evolution through a genetic algorithm to obtain a group of good CNN network structures;
step 11, inputting the processed test set data into a group of CNN network structures, and selecting the network structure with the minimum mean square error between a real value and a predicted value as a final prediction model; when the voltage data of another battery is input again, the residual capacity of the battery can be predicted through the trained model, and the SOH of the battery can be calculated quickly and accurately.
2. The lithium battery health state estimation method based on the genetic convolutional neural network as claimed in claim 1, wherein the specific steps processed in step 5 are as follows: according to 135 voltage curves, 8 characteristic points are respectively extracted from each voltage curve to form a matrix with 135 rows and 8 columns, 8 data of each row are converted into a matrix with 2 rows and 4 columns, and the matrix is input into a CNN model as an image.
3. The method as claimed in claim 1, wherein in step 6, the data set is divided into 80% of training set and 20% of testing set.
4. The method for estimating the health state of the lithium battery based on the genetic convolutional neural network as claimed in claim 1, wherein in the step 8, a group of CNN network structures is initialized randomly, the group of network structures is regarded as a population, and each individual in the population represents one CNN network structure, the size of the population is initialized, and the generation number is initialized.
5. The method for estimating the health state of the lithium battery based on the genetic convolutional neural network as claimed in claim 1, wherein the step 10 specifically comprises:
step 10.1, encoding the network structure, dividing the network structure into different stages for each network structure, namely an individual, by taking the pooling layer as a boundary, encoding each stage, combining the stages into a fixed binary character string, and expressing each network structure by using a genotype;
step 10.2, evaluating the fitness value of the individual, inputting voltage data into the individual for training, and taking MSE (mean square error) between the real value and the predicted value as the fitness value of the individual;
step 10.3: generating a new network structure through selection, intersection and variation, evaluating the fitness value of the new network structure, reserving individuals with low fitness value in each iteration, deleting individuals with high fitness value, always keeping the population size N until the maximum generation number is reached, finishing the iteration, and outputting a group of good CNN network structures.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011434822.2A CN112684346B (en) | 2020-12-10 | 2020-12-10 | Lithium battery health state estimation method based on genetic convolutional neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011434822.2A CN112684346B (en) | 2020-12-10 | 2020-12-10 | Lithium battery health state estimation method based on genetic convolutional neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112684346A true CN112684346A (en) | 2021-04-20 |
| CN112684346B CN112684346B (en) | 2023-06-20 |
Family
ID=75446612
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011434822.2A Active CN112684346B (en) | 2020-12-10 | 2020-12-10 | Lithium battery health state estimation method based on genetic convolutional neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112684346B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113376540A (en) * | 2021-05-18 | 2021-09-10 | 西安理工大学 | LSTM battery health state estimation method based on evolution attention mechanism |
| CN114114052A (en) * | 2021-11-25 | 2022-03-01 | 福州大学 | Method for rapidly estimating SOH and SOC of battery based on shock response characteristic |
| CN114184861A (en) * | 2021-11-28 | 2022-03-15 | 辽宁石油化工大学 | Fault diagnosis method for oil-immersed transformer |
| CN115616415A (en) * | 2022-12-06 | 2023-01-17 | 北京志翔科技股份有限公司 | Method, device and equipment for evaluating state of battery pack and storage medium |
| CN117074964A (en) * | 2023-10-13 | 2023-11-17 | 快电动力(北京)新能源科技有限公司 | Method, device, system and component for monitoring battery health status |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110687452A (en) * | 2019-09-05 | 2020-01-14 | 南京理工大学 | Lithium battery capacity online prediction method based on K-means clustering and Elman neural network |
| CN110824364A (en) * | 2019-10-24 | 2020-02-21 | 重庆邮电大学 | Lithium battery SOH estimation and RUL prediction method based on AST-LSTM neural network |
| CN111638465A (en) * | 2020-05-29 | 2020-09-08 | 浙大宁波理工学院 | Lithium battery health state estimation method based on convolutional neural network and transfer learning |
| WO2020191800A1 (en) * | 2019-03-27 | 2020-10-01 | 东北大学 | Method for predicting remaining service life of lithium-ion battery employing wde-optimized lstm network |
-
2020
- 2020-12-10 CN CN202011434822.2A patent/CN112684346B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020191800A1 (en) * | 2019-03-27 | 2020-10-01 | 东北大学 | Method for predicting remaining service life of lithium-ion battery employing wde-optimized lstm network |
| CN110687452A (en) * | 2019-09-05 | 2020-01-14 | 南京理工大学 | Lithium battery capacity online prediction method based on K-means clustering and Elman neural network |
| CN110824364A (en) * | 2019-10-24 | 2020-02-21 | 重庆邮电大学 | Lithium battery SOH estimation and RUL prediction method based on AST-LSTM neural network |
| CN111638465A (en) * | 2020-05-29 | 2020-09-08 | 浙大宁波理工学院 | Lithium battery health state estimation method based on convolutional neural network and transfer learning |
Non-Patent Citations (1)
| Title |
|---|
| 邓伟锋;李振璧;: "基于GA优化BP神经网络的微电网蓄电池健康状态评估", 电测与仪表, no. 21 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113376540A (en) * | 2021-05-18 | 2021-09-10 | 西安理工大学 | LSTM battery health state estimation method based on evolution attention mechanism |
| CN113376540B (en) * | 2021-05-18 | 2023-07-07 | 西安理工大学 | LSTM battery health state estimation method based on evolutionary attention mechanism |
| CN114114052A (en) * | 2021-11-25 | 2022-03-01 | 福州大学 | Method for rapidly estimating SOH and SOC of battery based on shock response characteristic |
| CN114184861A (en) * | 2021-11-28 | 2022-03-15 | 辽宁石油化工大学 | Fault diagnosis method for oil-immersed transformer |
| CN115616415A (en) * | 2022-12-06 | 2023-01-17 | 北京志翔科技股份有限公司 | Method, device and equipment for evaluating state of battery pack and storage medium |
| CN117074964A (en) * | 2023-10-13 | 2023-11-17 | 快电动力(北京)新能源科技有限公司 | Method, device, system and component for monitoring battery health status |
| CN117074964B (en) * | 2023-10-13 | 2024-01-16 | 快电动力(北京)新能源科技有限公司 | Method, device, system and component for monitoring battery health status |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112684346B (en) | 2023-06-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112684346B (en) | Lithium battery health state estimation method based on genetic convolutional neural network | |
| CN110687452B (en) | Lithium battery capacity online prediction method based on K-means clustering and Elman neural network | |
| CN110082682B (en) | Lithium battery state of charge estimation method | |
| CN113267733B (en) | Automatic configuration method for lithium battery health state estimation based on Gaussian process regression | |
| CN112633316A (en) | Load prediction method and device based on boundary estimation theory | |
| CN113176518B (en) | Lithium battery health state estimation method | |
| CN112733417A (en) | Abnormal load data detection and correction method and system based on model optimization | |
| CN113820604B (en) | Lithium battery SOH estimation method based on temperature prediction | |
| CN114726045B (en) | Lithium battery SOH estimation method based on IPEA-LSTM model | |
| CN113138344B (en) | SOC estimation method based on fractional order extended Kalman filtering algorithm | |
| CN113376540B (en) | LSTM battery health state estimation method based on evolutionary attention mechanism | |
| CN114757340A (en) | Lithium battery health state prediction method and system based on neural network integration | |
| CN112287597B (en) | Lead-acid storage battery SOH estimation method based on VPGA-GPR algorithm | |
| CN117289167A (en) | Battery remaining life prediction method, device and medium based on multiple neural network | |
| CN112580211A (en) | Lead-acid storage battery SOH estimation method based on SA and ANN algorithm | |
| CN117007974A (en) | Solid-state battery SOC estimation method based on model fusion | |
| CN115993537A (en) | Lithium battery capacity prediction method based on correlation analysis and WOA-LSTM | |
| CN115327389A (en) | Lithium battery SOC estimation method based on genetic algorithm improved double-Kalman filtering | |
| CN114252797B (en) | Uncertainty estimation-based lithium battery remaining service life prediction method | |
| CN113627655B (en) | Method and device for simulating and predicting pre-disaster fault scene of power distribution network | |
| CN115201684A (en) | Method for evaluating service life of lithium ion battery in different degradation modes | |
| CN115481796A (en) | Method for predicting remaining service life of battery based on Bayesian hybrid neural network | |
| CN115032544A (en) | Method for predicting residual life of power lithium battery under condition of incomplete data | |
| CN115409317A (en) | Transformer area line loss detection method and device based on feature selection and machine learning | |
| CN114239463A (en) | Battery cluster state of charge correction method based on big data |
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 |