CN110568359B - Lithium battery residual life prediction method - Google Patents
Lithium battery residual life prediction method Download PDFInfo
- Publication number
- CN110568359B CN110568359B CN201910832555.5A CN201910832555A CN110568359B CN 110568359 B CN110568359 B CN 110568359B CN 201910832555 A CN201910832555 A CN 201910832555A CN 110568359 B CN110568359 B CN 110568359B
- Authority
- CN
- China
- Prior art keywords
- lithium battery
- life prediction
- data
- residual life
- hidden layer
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
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]
- G01R31/367—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]
- G01R31/392—Determining battery ageing or deterioration, e.g. state of health
Abstract
The invention discloses a method for predicting the residual service life of a lithium battery, which comprises the steps of firstly carrying out multi-scale decomposition on dischargeable capacity by using empirical mode decomposition, then respectively predicting decomposed information by using different methods, and finally adding results to obtain the dischargeable capacity of the lithium battery so as to obtain the residual service life of the lithium battery. The method and the device can effectively predict the charge state and the residual service life of the battery, have better prediction efficiency and prediction precision, effectively judge the future working capacity, find problems in time and avoid unnecessary troubles and loss.
Description
Technical Field
The invention relates to the field of lithium battery life prediction, in particular to a lithium battery residual life prediction method.
Background
The lithium battery is a novel energy source, replaces the traditional batteries such as lead storage batteries, nickel cadmium batteries and the like due to the advantages of high working voltage, large specific energy, high charging and discharging efficiency, low self-discharging rate, no memory effect, long cycle life and the like, and is applied to various fields such as mobile phones, computers, electric vehicles and the like. However, in the process of long-term use of the lithium battery, the discharge capacity of the lithium iron phosphate battery gradually decreases due to a series of physicochemical changes occurring inside the lithium battery, that is, the State of Health (State of charge) of the battery gradually decreases, and the related equipment may be damaged, and in a serious case, the whole system may be rushed, and even property loss and casualties may be caused. In recent years, relevant researchers have set themselves to develop better batteries on the one hand and have conducted a great deal of research into the prediction of battery life on the other hand. At present, the material and the manufacturing level of the battery are greatly improved, but the problem of the reduction of the state of charge is not fundamentally solved.
Disclosure of Invention
The invention aims to provide a method for predicting the residual life of a lithium battery.
The invention provides a lithium battery life prediction method, which comprises the following steps:
the method comprises the following steps: extracting lithium battery capacity data, current, voltage and temperature and corresponding time data in the lithium battery operation data to serve as lithium battery residual life prediction data, and dividing the lithium battery residual life prediction data into two groups, wherein one group is a training set, and the other group is a testing set;
step two: performing empirical mode decomposition on the residual life prediction data of the lithium battery, decomposing the residual life prediction data into a plurality of eigenmode functions as the characteristics of the residual life prediction data of the lithium battery under different scales; the features under different scales at least comprise information features of global attenuation tendency, capacity regeneration data and local fluctuation;
step three: setting parameters of the long and short term memory network model, and inputting the eigenmode function obtained by decomposition into the long and short term memory network model for training;
step four: setting parameters of the deep neural network, and inputting the residual quantity after extracting the eigenmode function, the current voltage temperature in the lithium battery operation data and the corresponding time data into a deep neural network model for training;
step five: and respectively inputting an eigenmode function obtained by performing empirical mode decomposition on the lithium battery residual life prediction data, the current voltage temperature in the lithium battery operation data and corresponding time data into the trained long-short term memory network model and the trained deep neural network model, and adding results output by the models to obtain a lithium battery residual life prediction result.
The method for carrying out empirical mode decomposition on the battery capacity data of the lithium battery comprises the following steps:
the first step is as follows: finding all extreme points of the sequence x (t);
the second step is that: forming a lower envelope x for the minimum points by interpolationl(t) forming an upper envelope x for the maximau(t);
The third step: calculating the average value of the envelope on the lower envelope:
m(t)=[xl(t)+xu(t)]/2 (1)
the fourth step: extracting an eigenmode function:
h(t)=x(t)-m(t) (2)
the fifth step: judging whether the termination condition is satisfied, if so, outputting x (t) and rn(t) ending the empirical mode decomposition, otherwise, executing the sixth step;
wherein N is original battery capacity data, and delta is a preset termination condition; j represents the iteration times, and if the iteration formula meets the formula (3), the calculation is ended;
and a sixth step: taking h (t) as one of the eigenmode functions:
cj(t)=h(t) (4)
the seventh step: return r (t) instead of x (t) to the first step for calculation:
r(t)=x(t)-cj(t) (5)
after the step of performing empirical mode decomposition on the battery capacity data of the lithium battery, the method further comprises the step of initializing deep neural network training parameters, and specifically comprises the following steps of:
after the step of performing empirical mode decomposition on the battery capacity data of the lithium battery, the method further comprises the step of initializing deep neural network training parameters, and specifically comprises the following steps:
according to the input variable, the connection weight omega between the input layer and the hidden layerijAnd the deviation b calculates the hidden layer output HjThe calculation formula is as follows:
f is a hidden layer excitation function, and the calculation formula is as follows:
y=x (7)
wherein l is the number of nodes of the hidden layer;
outputting H from a hidden layerjConnection weight omega between hidden layer and output layerjkAnd b, calculating the prediction output O of the deep neural network by the deviation b, wherein the calculation formula is as follows:
f is a hidden layer excitation function, and the calculation formula is as follows:
y=x (9)
wherein m is the number of nodes of the output layer.
Wherein, the parameters of the deep neural network are set as follows: the hidden layer is set to be 2 layers, the neuron number of each layer is set to be 32 and 8, the neuron number of the output layer is set to be 1, the activation function of the hidden layer neurons is set to be y-x, the activation function of the output layer is set to be y-x, the loss function is set to be mean square error (mse), and the optimizer uses adam; the number of training times varies from one starting point to another.
Wherein, the parameters of the long-term and short-term memory network model are set as follows: 2 batchs are taken for each training, the size of each batch is 32, the size of a hidden layer is 200, and the training times are set to be 400 times; and evaluating the long-short term memory network model, and predicting backwards on the basis of the evaluation data, wherein the quantity of the prediction data is different along with the difference of the prediction starting point.
The method comprises the following steps of inputting current, voltage and temperature in lithium battery residual life prediction data and corresponding time data into a trained deep neural network model, wherein the steps comprise:
selecting the number of layers of the hidden layer and the number of neurons of the hidden layer in the deep neural network algorithm; selecting two hidden layers according to the root mean square error and the result graph, wherein the number of the neurons of the hidden layers is 32 and 8 respectively;
and according to the selected extracted current, voltage, temperature and time characteristics, dividing the current, voltage and time characteristics into 5 groups of different training sets and test sets, and respectively training to obtain different neural network models.
The method comprises the following steps of inputting an eigenmode function obtained by empirical mode decomposition of lithium battery residual life prediction data into a trained long-short term memory network model, wherein the steps comprise:
selecting the hidden layer size and the time window of the long and short term memory model; selecting two layers of hidden layers according to the root mean square error and a result graph, wherein the size of the hidden layers is set to be 200, and the size of a time window is set to be 32;
the EMD-decomposed imfs information is predicted using the LSTM model, and the number of predicted points is different depending on the starting point.
When the training completion degree of the LSTM and DNN models is judged, Absolute Error (AE) and Root Mean Square Error (RMSE) are used as the standards of model performance.
Wherein, AE represents the difference between the actual residual life of the lithium battery and the predicted residual life of the lithium battery, and represents the accuracy of the predicted residual life of the lithium battery; RMSE represents the dischargeable capacity of the battery, indicating the accuracy of the state of charge prediction of the battery. Therefore, the accuracy of the model is evaluated by using the following equations 10 and 11:
(1)AE
AE=|T-P| (10)
wherein T represents true RUL and P represents predicted RUL;
(2)RMSE
n is the predicted node count, T, P represents the predicted state of charge value, s represents the point at which prediction begins, Ti+sRepresenting the actual state of charge, Pi+sRepresenting the predicted state of charge.
Different from the prior art, the method provided by the invention firstly carries out multi-scale decomposition on the dischargeable capacity by using empirical mode decomposition, then predicts the decomposed information by using different methods from transverse and longitudinal angles, and finally predicts the residual service life of the lithium battery. The method and the device can effectively predict the charge state of the battery, have better prediction efficiency and prediction precision, effectively judge the future working capacity, find problems in time and avoid unnecessary troubles and loss.
Drawings
Fig. 1 is a schematic flow chart of a lithium battery life prediction method provided by the present invention.
Fig. 2 is a diagram of dischargeable capacity of lithium battery No. 5 in the lithium battery life prediction method provided in the present invention.
Fig. 3 is a schematic diagram of information of each scale obtained by empirical mode decomposition of residual life prediction data of lithium battery No. 5 in the lithium battery life prediction method provided by the present invention.
Fig. 4-8 are diagrams illustrating a residual life prediction of lithium battery No. 5 in the method for predicting the life of a lithium battery according to the present invention.
Detailed Description
The technical solution of the present invention will be further described in more detail with reference to the following embodiments. It is to be understood that the described embodiments are merely a few embodiments of the invention, and not all embodiments. All other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, shall fall within the scope of protection of the present invention.
The invention provides a method for predicting the residual life of a lithium battery, which comprises the following steps:
the method comprises the following steps: extracting lithium battery capacity data, current, voltage and temperature and corresponding time data in the lithium battery operation data to serve as lithium battery residual life prediction data, and dividing the lithium battery residual life prediction data into two groups, wherein one group is a training set, and the other group is a testing set;
step two: performing empirical mode decomposition on the residual life prediction data of the lithium battery, decomposing the residual life prediction data into a plurality of eigenmode functions as the characteristics of the residual life prediction data of the lithium battery under different scales; the features under different scales at least comprise information features of global attenuation tendency, capacity regeneration data and local fluctuation;
step three: setting parameters of the long and short term memory network model, and inputting the eigenmode function obtained by decomposition into the long and short term memory network model for training;
step four: setting parameters of the deep neural network, and inputting the residual quantity after extracting the eigenmode function, the current voltage temperature in the lithium battery operation data and the corresponding time data into a deep neural network model for training;
step five: and respectively inputting an eigenmode function obtained by performing empirical mode decomposition on the lithium battery residual life prediction data, the current voltage temperature in the lithium battery operation data and corresponding time data into the trained long-short term memory network model and the trained deep neural network model, and adding results output by the models to obtain a lithium battery residual life prediction result.
The method for carrying out empirical mode decomposition on the battery capacity data of the lithium battery comprises the following steps:
the first step is as follows: finding all extreme points of the sequence x (t);
the second step is that: forming a lower envelope x for the minimum points by interpolationl(t) forming an upper envelope x for the maximau(t);
The third step: calculating the average value of the envelope on the lower envelope:
m(t)=[xl(t)+xu(t)]/2 (1)
the fourth step: extracting an eigenmode function:
h(t)=x(t)-m(t) (2)
the fifth step: judging whether the termination condition is satisfied, if so, outputting x (t) and rn(t) ending the empirical mode decomposition, otherwise, executing the sixth step;
wherein N is original battery capacity data, and delta is a preset termination condition; j represents the iteration times, and if the iteration formula meets the formula (3), the calculation is ended;
and a sixth step: taking h (t) as one of the eigenmode functions:
cj(t)=h(t) (4)
the seventh step: return r (t) instead of x (t) to the first step for calculation:
r(t)=x(t)-cj(t) (5)
after the step of performing empirical mode decomposition on the battery capacity data of the lithium battery, the method further comprises the step of initializing deep neural network training parameters, and specifically comprises the following steps of:
after the step of performing empirical mode decomposition on the battery capacity data of the lithium battery, the method further comprises the step of initializing deep neural network training parameters, and specifically comprises the following steps:
according to the input variable, the connection weight omega between the input layer and the hidden layerijAnd the deviation b calculates the hidden layer output HjThe calculation formula is as follows:
f is a hidden layer excitation function, and the calculation formula is as follows:
y=x (7)
wherein l is the number of nodes of the hidden layer;
outputting H from a hidden layerjConnection weight omega between hidden layer and output layerjkAnd b, calculating the prediction output O of the deep neural network by the deviation b, wherein the calculation formula is as follows:
f is a hidden layer excitation function, and the calculation formula is as follows:
y=x (9)
wherein m is the number of nodes of the output layer.
The parameters of the deep neural network are set as follows: the hidden layer is set to be 2 layers, the neuron number of each layer is set to be 32 and 8, the neuron number of the output layer is set to be 1, the activation function of the hidden layer neurons is set to be y-x, the activation function of the output layer is set to be y-x, the loss function is set to be mean square error (mse), and the optimizer uses adam to initialize the deep neural network.
The parameters of the long-short term memory model are set as follows: 2 batchs are taken for each training, the size of each batch is 32, the size of the hidden layer is 200, and the training times are set to be 400 times. The model is then evaluated and predicted backwards on the basis of the evaluation data, the amount of prediction data varying from one prediction starting point to another.
Selecting the number of layers of the hidden layer and the number of neurons of the hidden layer in the deep neural network algorithm; selecting two layers of hidden layers according to the root mean square error and the result graph, wherein the number of the neurons of the hidden layers is 32 and 8 respectively;
according to the characteristics of the selected extracted current, voltage, temperature, time and the like, the neural network model is divided into 15 groups of different training sets and test sets, and different neural network models are obtained through training respectively.
Selecting the size of a hidden layer and a time window of the long-term and short-term memory model; selecting two layers of hidden layers according to the root mean square error and a result graph, wherein the size of the hidden layers is set to be 200, and the size of a time window is set to be 32;
the EMD-decomposed imfs information is predicted using the LSTM model, and the number of predicted points is different depending on the starting point.
And predicting by using the obtained EMD-LSTM-DNN model to obtain a predicted mean square error, and verifying effectiveness and accuracy.
We used Absolute Error (AE), and Root Mean Square Error (RMSE) as criteria for model performance.
AE represents the difference between the actual RUL and the predicted RUL, indicating the accuracy of the predicted RUL. RMSE represents the dischargeable capacity of the battery, indicating the accuracy of the state of charge prediction of the battery. Equation 10 and equation 10 are used to evaluate the accuracy of the model:
(3)AE
AE=|T-P| (10)
where T represents the true RUL and P represents the predicted RUL.
(4)RMSE
n is the predicted number of nodes, T, P represents the predicted state of charge value, s represents the point at which prediction beginsi+sRepresenting the actual state of charge, Pi+sRepresenting the predicted state of charge.
The experimental data used in this example was from battery number five in the NASA experimental data set. The relevant rated data of the lithium battery of the experimental model are as follows: rated capacity 2Ah, rated charge cut-off voltage 4.2V, and rated discharge cut-off voltage 2.7V. The input parameter is information of each scale of battery capacity after empirical mode decomposition, and the output parameter is available capacity of the battery pack.
After empirical mode decomposition, the dischargeable capacity and the remaining service life of the battery are predicted at 5 different starting points, and the comparison result of the model on the predicted situation and the actual situation of a test set is as follows:
fig. 2 is a graph of the dischargeable capacity of battery No. 5, and it can be seen that the capacity data shows a downward trend as the number of cycles increases but the intermediate process slightly increases. Fig. 3 is an exploded view of empirical mode, which is used to decompose the volume data into 3 pieces of information and a margin, fig. 4-8 are a comparison graph of prediction and reality using empirical mode and deep neural network algorithm and long-short term memory model, blue is used to indicate the prediction result, and yellow is used to indicate the actual volume. It can be found that the algorithm can effectively fit the trend of the dischargeable capacity of the lithium battery.
Through research, the number of training samples, the number of layers of hidden layers and the number of nodes of neurons of hidden layers have great influence on the performance of the trained deep neural network prediction model. Generally, input variables are well selected in advance by researchers according to professional knowledge and rich experience, but in practical application, the selection of the input variables is difficult to determine in advance, the prediction performance of the model is reduced, and therefore, the optimization of the input independent variable parameters in the process of training the prediction model has important significance.
The empirical mode decomposition algorithm can effectively decompose the capacity data into information with physical significance in a plurality of scales. After decomposition, the prediction was performed by different methods at 5 different starting electricity, and the average root mean square error of the prediction was 0.00096, which demonstrates the effectiveness of the method proposed herein.
Table 1 shows the absolute error and the root mean square error obtained by 5 different starting nodes in the method for predicting the service life of the lithium battery of the No. 5 battery provided by the present invention.
TABLE 15 result chart of battery prediction
Aiming at the problem of service life prediction of the lithium battery, the invention firstly carries out multi-scale decomposition on the capacity data based on an empirical mode decomposition algorithm, and multi-scale information obtained after decomposition is predicted by different methods, so that the lithium battery has excellent performance.
Different from the prior art, the method provided by the invention firstly carries out multi-scale decomposition on the dischargeable capacity by using empirical mode decomposition, then predicts the decomposed information by using different methods from transverse and longitudinal angles, and finally predicts the residual service life of the lithium battery. The method and the device can effectively predict the charge state of the battery, have better prediction efficiency and prediction precision, effectively judge the future working capacity, find problems in time and avoid unnecessary troubles and loss.
The above description is only an embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications of equivalent structures and equivalent processes, which are made by using the contents of the present specification and the accompanying drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.
Claims (9)
1. A method for predicting the residual life of a lithium battery is characterized by comprising the following steps:
the method comprises the following steps: extracting lithium battery capacity data, current, voltage and temperature and corresponding time data in the lithium battery operation data to serve as lithium battery residual life prediction data, and dividing the lithium battery residual life prediction data into two groups, wherein one group is a training set, and the other group is a testing set;
step two: performing empirical mode decomposition on the residual life prediction data of the lithium battery, decomposing the residual life prediction data into a plurality of eigenmode functions as the characteristics of the residual life prediction data of the lithium battery under different scales; the features under different scales at least comprise information features of global attenuation tendency, capacity regeneration data and local fluctuation;
step three: setting parameters of the long and short term memory network model, and inputting the eigenmode function obtained by decomposition into the long and short term memory network model for training;
step four: setting parameters of the deep neural network, and inputting the parameters into a deep neural network model for training by using the margin after the eigenmode function is extracted;
step five: inputting an eigenmode function obtained by empirical mode decomposition of lithium battery residual life prediction data into the long-short term memory network model, inputting current voltage temperature and corresponding time data in lithium battery operation data into the deep neural network model, and adding results output by the model to obtain a lithium battery residual life prediction result.
2. The lithium battery remaining life prediction method according to claim 1, characterized in that:
the method for carrying out empirical mode decomposition on the battery capacity data of the lithium battery comprises the following steps:
the first step is as follows: constructing a sequence x (t) by using the residual life prediction data of the lithium battery, and finding out all extreme points of the sequence x (t);
the second step is that: forming a lower envelope x for the minimum points by interpolationl(t) forming an upper envelope x for the maximau(t);
The third step: calculating the average value of the envelope on the lower envelope:
m(t)=[xl(t)+xu(t)]/2 (1)
the fourth step: extracting an eigenmode function:
h(t)=x(t)-m(t) (2)
the fifth step: judging whether formula 3 is satisfied, if so, outputting x (t) and margin rn(t) ending the empirical mode decomposition, otherwise, executing the sixth step;
wherein N is the cycle number, and delta is a preset threshold value; j represents the number of iterations;
and a sixth step: taking h (t) as one of the eigenmode functions:
cj(t)=h(t) (4)
the seventh step: return r (t) instead of x (t) to the first step for calculation:
r(t)=x(t)-cj(t) (5)。
3. the lithium battery remaining life prediction method according to claim 1, characterized in that:
after the step of performing empirical mode decomposition on the residual life prediction data of the lithium battery, the method further comprises the step of initializing deep neural network training parameters, and specifically comprises the following steps of:
according to the input variable, the connection weight omega between the input layer and the hidden layerijAnd the deviation b calculates the hidden layer output HjThe calculation formula is as follows:
f is a hidden layer excitation function, and the calculation formula is as follows:
y=x (7)
wherein l is the number of nodes of the hidden layer;
outputting H from a hidden layerjConnection weight omega between hidden layer and output layerjkAnd deviation b calculating the predicted output O of the deep neural networkkThe calculation formula is as follows:
f is a hidden layer excitation function, and the calculation formula is as follows:
y=x (9)
wherein m is the number of nodes of the output layer.
4. The lithium battery remaining life prediction method according to claim 1, characterized in that:
the parameters of the deep neural network are set as follows: the hidden layer is set to be 2 layers, the neuron number of each layer is set to be 32 and 8, the neuron number of the output layer is set to be 1, the activation function of the hidden layer neurons is set to be y-x, the activation function of the output layer is set to be y-x, the loss function is set to be mean square error (mse), and the optimizer uses adam; the number of training times varies from one starting point to another.
5. The lithium battery remaining life prediction method according to claim 1, characterized in that:
the parameters of the long-short term memory network model are set as follows: 2 batchs are taken for each training, the size of each batch is 32, the size of a hidden layer is 200, and the training times are set to be 400 times; and evaluating the long-short term memory network model, and predicting backwards on the basis of the evaluation data, wherein the quantity of the prediction data is different along with the difference of the prediction starting point.
6. The lithium battery remaining life prediction method according to claim 1, characterized in that:
inputting current, voltage and temperature in the lithium battery residual life prediction data and corresponding time data into a trained deep neural network model, wherein the method comprises the following steps of:
selecting the number of layers of the hidden layer and the number of neurons of the hidden layer in the deep neural network algorithm; selecting two hidden layers according to the root mean square error and the result graph, wherein the number of the neurons of the hidden layers is 32 and 8 respectively;
and according to the selected extracted current, voltage, temperature and time characteristics, dividing the current, voltage and time characteristics into 5 groups of different training sets and test sets, and respectively training to obtain different neural network models.
7. The lithium battery remaining life prediction method according to claim 1, characterized in that:
inputting an eigenmode function obtained by empirical mode decomposition of lithium battery residual life prediction data into a trained long-short term memory network model, wherein the steps comprise:
selecting the hidden layer size and the time window of the long and short term memory model; selecting two layers of hidden layers according to the root mean square error and a result graph, wherein the size of the hidden layers is set to be 200, and the size of a time window is set to be 32;
the EMD-decomposed imfs information is predicted using the LSTM model, and the number of predicted points is different depending on the starting point.
8. The lithium battery remaining life prediction method according to claim 1, characterized in that:
in judging the training completion degree of the LSTM and DNN models, Absolute Error (AE) and Root Mean Square Error (RMSE) are used as the standards of model performance.
9. The lithium battery remaining life prediction method of claim 8, characterized in that:
AE represents the difference between the actual residual life of the lithium battery and the predicted residual life of the lithium battery, and represents the accuracy of the predicted residual life of the lithium battery; RMSE represents the dischargeable capacity of the battery and represents the accuracy of the state of charge prediction of the battery; therefore, the accuracy of the model is evaluated by using the following equations 10 and 11:
(1)AE
AE=|T-P| (10)
wherein T represents true RUL and P represents predicted RUL;
(2)RMSE
n is the predicted number of nodes, s is the point at which the prediction starts, Ti+sRepresenting the true state of charge of the battery, Pi+sRepresenting the predicted state of charge of the battery.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910832555.5A CN110568359B (en) | 2019-09-04 | 2019-09-04 | Lithium battery residual life prediction method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910832555.5A CN110568359B (en) | 2019-09-04 | 2019-09-04 | Lithium battery residual life prediction method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110568359A CN110568359A (en) | 2019-12-13 |
CN110568359B true CN110568359B (en) | 2021-11-23 |
Family
ID=68777773
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910832555.5A Active CN110568359B (en) | 2019-09-04 | 2019-09-04 | Lithium battery residual life prediction method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110568359B (en) |
Families Citing this family (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111220921A (en) * | 2020-01-08 | 2020-06-02 | 重庆邮电大学 | Lithium battery capacity estimation method based on improved convolution-long-and-short-term memory neural network |
CN111175659B (en) * | 2020-01-21 | 2022-04-29 | 湖南大学 | Lithium ion battery state detection system and method based on continuous acoustic emission signals |
CN111274737A (en) * | 2020-02-25 | 2020-06-12 | 山东大学 | Method and system for predicting remaining service life of mechanical equipment |
CN111413622B (en) * | 2020-04-03 | 2022-04-15 | 重庆大学 | Lithium battery life prediction method based on stacking noise reduction automatic coding machine |
CN111426957B (en) * | 2020-05-19 | 2021-06-11 | 华南理工大学 | SOC estimation optimization method for power battery under simulated vehicle working condition |
CN111999648A (en) * | 2020-08-20 | 2020-11-27 | 浙江工业大学 | Lithium battery residual life prediction method based on long-term and short-term memory network |
CN112098874B (en) * | 2020-08-21 | 2023-09-22 | 杭州电子科技大学 | Lithium ion battery electric quantity prediction method considering aging condition |
WO2022198616A1 (en) * | 2021-03-26 | 2022-09-29 | 深圳技术大学 | Battery life prediction method and system, electronic device, and storage medium |
CN113203953B (en) * | 2021-04-02 | 2022-03-25 | 中国人民解放军92578部队 | Lithium battery residual service life prediction method based on improved extreme learning machine |
CN113283632B (en) * | 2021-04-13 | 2024-02-27 | 湖南大学 | Early-stage fault early-warning method, system, device and storage medium for battery |
CN113702836B (en) * | 2021-07-23 | 2023-08-18 | 国家电网有限公司西北分部 | Lithium ion battery state of charge estimation method based on EMD-GRU |
CN113884937B (en) * | 2021-11-22 | 2022-08-30 | 江南大学 | Lithium ion battery residual life prediction method and system based on decomposition integration strategy |
CN115616415B (en) * | 2022-12-06 | 2023-04-07 | 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 |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108256697A (en) * | 2018-03-26 | 2018-07-06 | 电子科技大学 | A kind of Forecasting Methodology for power-system short-term load |
CN108549036A (en) * | 2018-05-03 | 2018-09-18 | 太原理工大学 | Ferric phosphate lithium cell life-span prediction method based on MIV and SVM models |
CN109242569A (en) * | 2018-09-13 | 2019-01-18 | 西安建筑科技大学 | A kind of molybdenum concentrate Long-term Market price analysis and prediction technique and system |
CN110135637A (en) * | 2019-05-13 | 2019-08-16 | 武汉科技大学 | Micro-capacitance sensor short-term load forecasting method based on shot and long term memory and adaptive boosting |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20180005151A1 (en) * | 2016-06-29 | 2018-01-04 | General Electric Company | Asset health management framework |
-
2019
- 2019-09-04 CN CN201910832555.5A patent/CN110568359B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108256697A (en) * | 2018-03-26 | 2018-07-06 | 电子科技大学 | A kind of Forecasting Methodology for power-system short-term load |
CN108549036A (en) * | 2018-05-03 | 2018-09-18 | 太原理工大学 | Ferric phosphate lithium cell life-span prediction method based on MIV and SVM models |
CN109242569A (en) * | 2018-09-13 | 2019-01-18 | 西安建筑科技大学 | A kind of molybdenum concentrate Long-term Market price analysis and prediction technique and system |
CN110135637A (en) * | 2019-05-13 | 2019-08-16 | 武汉科技大学 | Micro-capacitance sensor short-term load forecasting method based on shot and long term memory and adaptive boosting |
Non-Patent Citations (2)
Title |
---|
Remaining useful life estimation of engineered systems using vanilla LSTM neural networks;Yuting Wu等;《Neurocomputing》;20170526;167-179 * |
基于机器学习的设备剩余寿命预测方法综述;裴洪等;《机械工程学报》;20190430;第55卷(第8期);1-13 * |
Also Published As
Publication number | Publication date |
---|---|
CN110568359A (en) | 2019-12-13 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110568359B (en) | Lithium battery residual life prediction method | |
CN108896914B (en) | Gradient lifting tree modeling and prediction method for health condition of lithium battery | |
CN106055775B (en) | A kind of service life of secondary cell prediction technique that particle filter is combined with mechanism model | |
CN108846227B (en) | Lithium ion battery capacity degradation prediction and evaluation method based on random forest and capacity self-recovery effect analysis | |
KR20200140093A (en) | Prediction Method and Prediction System for predicting Capacity Change according to Charging / Discharging Cycle of Battery | |
CN110082682B (en) | Lithium battery state of charge estimation method | |
CN110658460B (en) | Battery life prediction method and device for battery pack | |
CN103389471A (en) | Cycle life indirect prediction method for lithium ion battery provided with uncertain intervals on basis of GPR (general purpose register) | |
CN111812515A (en) | XGboost model-based lithium ion battery state of charge estimation | |
CN113702843B (en) | Lithium battery parameter identification and SOC estimation method based on suburb optimization algorithm | |
CN105574231A (en) | Storage battery surplus capacity detection method | |
CN111999648A (en) | Lithium battery residual life prediction method based on long-term and short-term memory network | |
CN111812519B (en) | Battery parameter identification method and system | |
CN114636932A (en) | Method and system for predicting remaining service life of battery | |
CN115684947A (en) | Battery model construction method and battery degradation prediction device | |
CN107340476A (en) | The electrical state monitoring system and electrical state monitoring method of battery | |
CN113917336A (en) | Lithium ion battery health state prediction method based on segment charging time and GRU | |
CN113687242A (en) | Lithium ion battery SOH estimation method for optimizing and improving GRU neural network based on GA algorithm | |
CN113408138B (en) | Lithium battery SOH estimation method and system based on secondary fusion | |
CN114545277A (en) | Power battery retirement prediction method based on capacity attenuation | |
CN111337833B (en) | Lithium battery capacity integrated prediction method based on dynamic time-varying weight | |
CN116774088A (en) | Lithium ion battery health state estimation method based on multi-objective optimization | |
CN110555226A (en) | method for predicting residual life of lithium iron phosphate battery based on EMD and MLP | |
CN110515001B (en) | Two-stage battery performance prediction method based on charging and discharging | |
CN113884936B (en) | ISSA coupling DELM-based lithium ion battery health state prediction method |
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 |