CN115575840A - Early prediction method for remaining service life of lithium ion battery - Google Patents
Early prediction method for remaining service life of lithium ion battery Download PDFInfo
- Publication number
- CN115575840A CN115575840A CN202211016487.3A CN202211016487A CN115575840A CN 115575840 A CN115575840 A CN 115575840A CN 202211016487 A CN202211016487 A CN 202211016487A CN 115575840 A CN115575840 A CN 115575840A
- Authority
- CN
- China
- Prior art keywords
- battery
- service life
- lithium ion
- data
- temperature difference
- 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
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]
- G01R31/392—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]
- G01R31/385—Arrangements for measuring battery or accumulator variables
- G01R31/387—Determining ampere-hour charge capacity or SoC
- G01R31/388—Determining ampere-hour charge capacity or SoC involving voltage measurements
-
- 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]
- G01R31/396—Acquisition or processing of data for testing or for monitoring individual cells or groups of cells within a battery
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Charge And Discharge Circuits For Batteries Or The Like (AREA)
Abstract
The invention relates to the field of battery capacity detection, and discloses a method for early predicting the remaining service life of a lithium ion battery, which comprises the following steps: selecting the battery with the longest service life from the plurality of battery capacity aging data sets as a reference battery, and taking the temperature data of the first 20 circles of charging as the most healthy temperature state; respectively calculating the temperature difference of the first 20 cycles of circulation, and arranging the temperature difference data among different battery monomers in a high-dimensional space to meet the input requirement of ConvLSTM; extracting the space-time characteristics of different single battery temperature difference matrixes in a high-dimensional space; embedding an attention mechanism into a compact feature map, extracting key space-time features, and filtering signal noise; capturing the time sequence correlation of a low-dimensional feature space by adopting a 4-layer LSTM network, further generating a representative compact feature, and capturing the change rule of the feature along with the increase of circulation; and finally, modeling the residual service life of the battery through a multilayer sensor to generate an early life prediction result.
Description
Technical Field
The invention relates to the field of battery capacity detection, in particular to a method for early predicting the remaining service life of a lithium ion battery.
Background
Lithium ion batteries are widely used to power electronic devices and mobile devices due to their high energy density, high discharge current rate, no memory effect, and the like. In recent years, with the popularization of the low-carbon life concept, new energy automobiles taking lithium ion batteries as main power sources are rapidly developed. However, real-time monitoring of the state of health and prediction of the remaining service life of lithium ion batteries have been a great challenge to address mileage anxiety and to ensure battery reliability and safety. The capacity fade of commercial lithium batteries initially exhibits a strong linear relationship, but near the end of life, a highly nonlinear trend occurs. Moreover, different factors such as charging and discharging configuration, ambient temperature, anode and cathode materials of the battery, and battery capacity regeneration phenomena can also cause differentiated battery capacity attenuation trends. The complexity and non-linearity of the battery capacity fade makes accurate life prediction still a formidable challenge. And accurate life prediction is the guarantee of failure early warning and replacement of the lithium ion battery, and has important practical significance.
Most current research on lithium battery life prediction is based on specific cycling conditions, such as specific charging and discharging strategies and constant temperature environments. In practical application of the lithium battery, the lithium battery has complex ambient temperature and charging and discharging habits with distinct user differences, and a discharging curve is influenced by random output power. The prediction model based on the internal resistance of the battery is very complicated in measuring the internal resistance, and almost impossible to be used for on-line monitoring and prediction, and practical application thereof is hindered. In addition, the circulation capacity is used as long-term data, the problems of long-term dependence and gradient explosion of the data are difficult to solve by a general data-driven model, and a data set is not enough to establish a prediction model with strong generalization performance. This presents a serious test for the practicability and versatility of the battery remaining service life prediction model.
In summary, the current lithium ion battery remaining service life prediction model has key challenges of being interfered by random workload, having a small data set, being incapable of capturing long sequence correlation and the like, and key problems of low prediction accuracy of remaining service life, weak robustness, poor generalization performance of the model and the like are caused, so that a lithium ion battery remaining service life early prediction method is provided.
Disclosure of Invention
Technical problem to be solved
Aiming at the defects of the prior art, the invention provides the early prediction method for the remaining service life of the lithium ion battery, which uses a directly measurable variable, namely temperature, in the charging process as a monitoring variable, eliminates the influence of random working load and ensures that a model has stronger robustness. The charging temperature difference between different battery monomers is used as HI (Health Indicator), which is beneficial to online acquisition and real-time prediction of data and enables the model to have stronger practicability. The method adopts ConvLSTM (Convolutional Long Short-Term Memory) and LSTM (Long Short-Term Memory) to capture the time sequence characteristics between temperature difference data, and solves the problems of Long-Term dependence and gradient explosion in Long sequence data prediction. The addition of an AM (Attention Mechanism) enables the model to focus on the extraction of key aging characteristics, information noise is filtered, and the prediction accuracy of the model is improved. And finally, model output is constructed through a multilayer sensor, and early accurate prediction of the residual service life of the lithium ion battery is realized.
(II) technical scheme
In order to achieve the above purpose, the invention provides the following technical scheme:
a method for early predicting the residual service life of a lithium ion battery comprises the following steps:
the first step is as follows: selecting the battery with the longest service life from the plurality of battery capacity aging data sets as a reference battery, and taking the temperature data of the first 20 circles of charging as the most healthy temperature state;
the second step is that: respectively calculating the temperature difference between each single battery and the reference battery in the first 20 cycles of circulation to serve as HI for measuring the aging degree of the battery, and arranging temperature difference data between different single batteries in a high-dimensional space to meet the input requirement of ConvLSTM;
the third step: adopting a 3-layer ConvLSTM network to consider the difference between different circulation states and extracting the space-time characteristics of different single battery temperature difference matrixes in a high-dimensional space;
the fourth step: embedding an attention mechanism into the compact feature mapping, further extracting key space-time features, and filtering signal noise;
the fifth step: capturing the time sequence correlation of the low-dimensional feature space by adopting a 4-layer LSTM network, further generating representative compact features, and capturing the change rule of the features along with the increase of cycles;
and a sixth step: and finally, modeling the residual service life of the battery through a multilayer sensor to generate an early service life prediction result.
Preferably, the second step comprises the following specific steps: the charging process data and the cycle life of each battery monomer are obtained through a public data set, the person with the longest life is screened out to be used as the most healthy state reference, the temperature difference of each battery and the reference battery in the first 20 cycles is calculated, and interpolation is carried out on the temperature difference data through a completion missing value and an Akima interpolation method, so that the data are ensured to have the same matrix shape.
Preferably, the spatio-temporal feature extraction method in the third step is as follows:
setting the temperature difference matrix as U, taking a submatrix with the shape of the convolution kernel size, and recording as X t Then X t Will be used as input and the current cell state (C) t ) Co-determining the future state C of the cell t+1 The role of the gate function in ConvLSTM cells and its screening process for data flow are as follows:
f t =σ(W Xf *X t +W hf *h t-1 +b f );
i t =σ(W Xi *X t +W hg *h t-1 +b i );
g t =tanh(W Xg *X t +W hg *h t-1 +b g );
o t =σ(W Xo *X t +W ho *h t-1 +b o );
C t =f t ⊙C t-1 +i t ⊙g t ;
h t =o t ⊙tanh(C t );
wherein f represents the output of the forgetting gate, i represents the output of the input gate, o represents the output of the output gate, g represents the candidate memory, C represents the cell state of ConvLSTM, h is the hidden layer output, X represents the input, W represents the output of the input gate, and g represents the candidate memory X~ And W h~ Is a 2-D convolution kernel, the subscript t indicates the corresponding time, the symbol "+" indicates a convolution operation, and "" indicates a hadamard product.
Preferably, the content of the fourth step is as follows: representing the output volume of the previous convolutional layer as F, generating an attention weight matrix through two fully-connected layers, the elements A in the matrix i,j Can be expressed as:
where δ and ω are weights, b and c are both offsets, n Fc The lower corner, the number of neurons, is the index of the element in the matrix. g (-) and f (-) represent hyperbolic tangent function and Sigmoid function, respectively, as follows:
the Sigmoid function controls the elements of the weight matrix within 0-1, controls the size of information flow entering the next layer of network through the Hadamard product of the two matrixes, and embeds the characteristic output L after the attention mechanism l Can be expressed as:
L l =A⊙F l 。
preferably, the specific content of the fifth step is as follows:
using the hadamard product operation, as follows:
f t =σ(W f ⊙[h t-1 ,x t ]+b f );
i t =σ(W i ⊙[h t-1 ,x t ]+b i );
g t =tanh(W g ⊙[h t-1 ,x t ]+b g );
C t =f t ⊙C t-1 +i t ⊙g t ;
o t =σ(W Xo ⊙[h t-1 ,x t ]+b o );
h t =o t ⊙tanh(C t )。
preferably, the sixth step comprises the following specific steps: define the output of LSTM asAn early prediction of the remaining useful life of a lithium ion battery is generated by a multilayer perceptron as follows:
(III) advantageous effects
Compared with the prior art, the early prediction method for the remaining service life of the lithium ion battery provided by the invention has the following beneficial effects:
1. the method for early predicting the remaining service life of the lithium ion battery is based on the data modeling of the battery charging process, and eliminates the interference of random workload on the model, so that the model has stronger robustness, practicability and generalization performance.
2. According to the method for early predicting the remaining service life of the lithium ion battery, the charging temperature difference of different battery monomers in the same cycle state is used as HI, the data is simple and easy to obtain, strong correlation is formed between the charging temperature difference and capacity fading, and the prediction accuracy of a model is improved.
3. According to the early prediction method for the remaining service life of the lithium ion battery, convLSTM and LSTM extract high-dimensional and low-dimensional space-time characteristics of a temperature difference matrix, and the problems of long-term dependence and gradient explosion of long-sequence data prediction are solved. The addition of the attention mechanism improves the extraction capability of the model to the key features, effectively filters useless data noise and further improves the prediction precision of the model.
Drawings
FIG. 1 is a schematic flow chart of a lithium ion battery remaining service life early-stage prediction method based on battery cell temperature difference and ConvLSTM-AM-LSTM in an embodiment of the present invention;
FIG. 2 is a graph of capacity fade for 124 battery data sets used in an embodiment of the present invention;
FIG. 3 is a temperature curve of a battery cell with different battery life under the same cycling condition according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of a ConvLSTM network structure according to an embodiment of the present invention;
FIG. 5 is a graph of predicted performance in early prediction of remaining useful life for an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Examples
Referring to fig. 1-5, an embodiment of the invention provides a method for early predicting remaining service life of a lithium ion battery, which includes the following steps:
The step 1 is implemented according to the following steps:
step 1.1, an experimental data set consists of 124 commercial lithium iron phosphate/graphite batteries, the rated capacity is 1.1Ah, and the rated voltage is 3.3V. These cells were cycled to failure in a 48-channel Abin charge and discharge cabinet and 30 ℃ incubator under 72 charge strategies and a fixed discharge rate. The data set included 96700 cycle period data, with the longest-lived battery having 2338 cycle lifetimes.
Fig. 2 shows a capacity fade graph for 124 battery data sets.
And 2, processing the temperature difference data by a completion missing value and-Akima interpolation method to ensure that all temperature matrixes have the same shape. And respectively calculating the temperature difference of each battery cell and the reference battery in the first 20 cycles as HI for measuring the aging degree of the battery. The data is arranged as a high-dimensional tensor to satisfy the processing requirements of ConvLSTM.
The step 2 is implemented according to the following steps:
and 2.1, extracting the charging temperature data of the first 20 circles of each single battery, and standardizing the charging temperature data of each circle to be 1024 data length by complementing the missing value and an Akima interpolation method. And calculating the difference value between the normalized temperature matrix and the temperature matrix of the reference battery, and arranging the temperature difference matrix into a 5-dimensional tensor which is used as the input of the ConvLSTM network.
Fig. 3 shows the charging temperature curves of the first cycle of the battery cells with different lifetimes. As can be seen from fig. 3, the temperatures of the batteries with different lifetimes in the same cycle state are regularly distributed, and the shorter the remaining service life is, the higher the temperature of the battery cell during charging is, and the strong correlation therebetween is obtained.
And 3, inputting the high-dimensional data into ConvLSTM to extract compact space-time characteristics.
Step 3 is specifically implemented according to the following steps:
step 3.1, setting the temperature difference matrix as U, taking a submatrix with the shape of the convolution kernel size, and recording the submatrix as X t Then X t Will be used as input to the current cell state (C) t ) Co-determining the future state of the cell C t+1 . The gate function effect in ConvLSTM cells and their screening process for data flow are as follows:
f t =σ(W Xf *X t +W hf *h t-1 +b f ) (1)
i t =σ(W Xi *X t +W hg *h t-1 +b i ) (2)
g t =tanh(W Xg *X t +W hg *h t-1 +b g ) (3)
o t =σ(W Xo *X t +W ho *h t-1 +b o ) (4)
C t =f t ⊙C t-1 +i t ⊙g t (5)
h t =o t ⊙tanh(C t ) (6)
wherein f represents the output of the forgetting gate; i represents the output of the input gate; o represents the output of the output gate; g represents a candidate memory; c represents the cellular status of ConvLSTM; h is hidden layer output; x represents an input; w X~ And W h~ Is a 2-D convolution kernel; the lower subscript t indicates the corresponding time instant. The symbol "+" represents convolution operation, "" represents hadamard product, "σ" represents sigmoid function. The Sigmoid function controls output elements of the three gates within a range of 0-1, and then Hadamard products are carried out, wherein 0 is completely forgotten to be input, and 1 is completely stored to be input so as to complete inheritance and forgetting of historical information.
Fig. 4 shows a schematic diagram of the network structure of ConvLSTM.
And 4, embedding the attention mechanism into a compact feature map, further extracting key space-time features, and filtering signal noise.
Step 4 is specifically implemented according to the following steps:
step 4.1, represent the output volume of the previous convolutional layer as F. Generating an attention weight matrix by two fully connected layers, element A within the matrix i,j Can be expressed as:
where δ and ω are weights, b and c are both offsets, n Fc The number of neurons. The lower corner is labeled the index of the element in the matrix. g (-) and f (-) represent hyperbolic tangent function and Sigmoid function, respectively, as follows:
the Sigmoid function controls the elements of the weight matrix within 0-1, controls the size of information flow entering the next layer of network through the Hadamard product of the two matrixes, and embeds the characteristic output L after the attention mechanism l Can be expressed as:
L l =A⊙F l (10)
and 5, inputting the features of the embedded attention mechanism into the LSTM, extracting low-dimensional time sequence features, and capturing the long-term dependence relationship between the residual service life and the time sequence features.
Step 5 is specifically implemented according to the following steps:
step 5.1, taking a submatrix with the shape of the size of a convolution kernel from the feature mapping after the attention mechanism is embedded, and recording the submatrix as x t . Similar to ConvLSTM, LSTM also controls the transmission of critical information streams by means of gate functions, with the difference that the former have partial convolution operations, while the latter are all hadamard product operations, as followsShown in the specification:
f t =σ(W f ⊙[h t-1 ,x t ]+b f ) (11)
i t =σ(W i ⊙[h t-1 ,x t ]+b i ) (12)
g t =tanh(W g ⊙[h t-1 ,x t ]+b g ) (13)
C t =f t ⊙C t-1 +i t ⊙g t (13)
o t =σ(W Xo ⊙[h t-1 ,x t ]+b o ) (14)
h t =o t ⊙tanh(C t ) (15)
and 6, constructing model output through the two multilayer sensors, and predicting the residual service life of the lithium ion battery.
Step 6 is implemented according to the following steps:
step 6.1, for the sake of brevity, define the output of LSTM asFinally, an early prediction value of the remaining service life of the lithium ion battery is generated by two multilayer perceptrons as follows:
the notations are similar to those of the formula (7).
FIG. 5 shows the predicted performance of the disclosed method in early prediction of remaining useful life.
The method for early predicting the remaining service life of the lithium ion battery provided by the embodiment of the invention is characterized in that the battery with the longest cycle life in a data set is taken as a reference battery, and the temperature data of the first 20 cycles of the battery is extracted as the reference of the most healthy temperature state; calculating the temperature difference between each battery and the reference battery in the first 20 cycles as HI; identifying the capacity attenuation degree in HI through ConvLSTM, and capturing the space-time characteristics of the characteristics in a high-dimensional data space; by embedding an attention mechanism, the capability of the model for extracting the key information of the battery aging is improved, and useless information and data noise are filtered; the time sequence characteristics in a low-dimensional space are extracted through the LSTM, inheritance and abandonment of an original data stream are determined, and long-term dependence of temperature difference data among different battery cells can be well learned; and finally, constructing an output layer through a plurality of layers of sensors, and realizing early and accurate prediction of the residual service life of the lithium ion battery.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (6)
1. A method for early predicting the residual service life of a lithium ion battery is characterized by comprising the following steps:
the first step is as follows: selecting the battery with the longest service life from the plurality of battery capacity aging data sets as a reference battery, and taking the temperature data of the first 20 circles of charging as the most healthy temperature state;
the second step is that: respectively calculating the temperature difference between each single battery and the reference battery in the first 20 cycles of circulation to serve as HI for measuring the aging degree of the battery, and arranging temperature difference data between different single batteries in a high-dimensional space to meet the input requirement of ConvLSTM;
the third step: adopting a 3-layer ConvLSTM network to consider the difference between different circulation states and extracting the space-time characteristics of different single battery temperature difference matrixes in a high-dimensional space;
the fourth step: embedding an attention mechanism into the compact feature mapping, further extracting key space-time features, and filtering signal noise;
the fifth step: capturing the time sequence correlation of the low-dimensional feature space by adopting a 4-layer LSTM network, further generating representative compact features, and capturing the change rule of the features along with the increase of cycles;
and a sixth step: and finally, modeling the residual service life of the battery through a multilayer sensor to generate an early life prediction result.
2. The early prediction method for the remaining service life of the lithium ion battery according to claim 1, characterized in that: the second step comprises the following specific steps: the charging process data and the cycle life of each battery monomer are obtained through a public data set, the person with the longest life is screened out to be used as the most healthy state reference, the temperature difference of each battery and the reference battery in the first 20 cycles is calculated, and interpolation is carried out on the temperature difference data through a completion missing value and an Akima interpolation method, so that the data are ensured to have the same matrix shape.
3. The early prediction method for the remaining service life of the lithium ion battery according to claim 1, characterized in that: the space-time feature extraction method in the third step is as follows:
setting the temperature difference matrix as U, taking a submatrix with the shape of convolution kernel size as X t Then X t Will be used as input to the current cell state (C) t ) Co-determining the future state C of the cell t+1 The role of the gate function in ConvLSTM cells and its screening process for data flow are as follows:
f t =σ(W Xf *X t +W hf *h t-1 +b f );
i t =σ(W Xi *X t +W hg *h t-1 +b i );
g t =tanh(W Xg *X t +W hg *h t-1 +b g );
o t =σ(W Xo *X t +W ho *h t-1 +b o );
C t =f t ⊙C t-1 +i t ⊙g t ;
h t =o t ⊙tanh(C t );
where f represents the output of the forgetting gate, iRepresents the output of the input gate, o represents the output of the output gate, g represents the candidate memory, C represents the cellular state of ConvLSTM, h is the hidden layer output, X represents the input, W X~ And W h~ Is a 2-D convolution kernel, the subscript t indicates the corresponding time, the symbol "+" indicates a convolution operation, and "" indicates a hadamard product.
4. The early prediction method for the remaining service life of the lithium ion battery according to claim 1, characterized in that: the content of the fourth step is as follows: representing the output volume of the previous convolutional layer as F, generating an attention weight matrix through two fully-connected layers, the elements A in the matrix i,j Can be expressed as:
where δ and ω are weights, b and c are both offsets, n Fc The lower corner, the number of neurons, is the index of the element in the matrix. g (-) and f (-) represent hyperbolic tangent function and Sigmoid function, respectively, as follows:
the Sigmoid function controls the elements of the weight matrix within 0-1, controls the size of information flow entering the next layer of network through the Hadamard product of the two matrixes, and embeds the characteristic output L after the attention mechanism l Can be expressed as:
L l =A⊙F l 。
5. the early prediction method for the remaining service life of the lithium ion battery according to claim 1, characterized in that: the concrete content of the fifth step is as follows:
the hadamard product operation is used, as follows:
f t =σ(W f ⊙[h t-1 ,x t ]+b f );
i t =σ(W i ⊙[h t-1 ,x t ]+b i );
g t =tanh(W g ⊙[h t-1 ,x t ]+b g );
C t =f t ⊙C t-1 +i t ⊙g t ;
o t =σ(W Xo ⊙[h t-1 ,x t ]+b o );
h t =o t ⊙tanh(C t )。
6. the early prediction method for the remaining service life of the lithium ion battery according to claim 1, characterized in that: the sixth step comprises the following specific steps: define the output of LSTM asAn early prediction of the remaining useful life of a lithium ion battery is generated by a multilayer perceptron as follows:
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211016487.3A CN115575840A (en) | 2022-08-24 | 2022-08-24 | Early prediction method for remaining service life of lithium ion battery |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211016487.3A CN115575840A (en) | 2022-08-24 | 2022-08-24 | Early prediction method for remaining service life of lithium ion battery |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115575840A true CN115575840A (en) | 2023-01-06 |
Family
ID=84580134
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211016487.3A Pending CN115575840A (en) | 2022-08-24 | 2022-08-24 | Early prediction method for remaining service life of lithium ion battery |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115575840A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116298947A (en) * | 2023-03-07 | 2023-06-23 | 中国铁塔股份有限公司黑龙江省分公司 | Storage battery nuclear capacity monitoring device |
-
2022
- 2022-08-24 CN CN202211016487.3A patent/CN115575840A/en active Pending
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116298947A (en) * | 2023-03-07 | 2023-06-23 | 中国铁塔股份有限公司黑龙江省分公司 | Storage battery nuclear capacity monitoring device |
CN116298947B (en) * | 2023-03-07 | 2023-11-03 | 中国铁塔股份有限公司黑龙江省分公司 | Storage battery nuclear capacity monitoring device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Gao et al. | Machine learning toward advanced energy storage devices and systems | |
Khan et al. | Batteries state of health estimation via efficient neural networks with multiple channel charging profiles | |
Li et al. | Battery thermal runaway fault prognosis in electric vehicles based on abnormal heat generation and deep learning algorithms | |
CN107797067A (en) | Lithium ion battery life migration prediction method based on deep learning | |
CN115656857A (en) | Multi-scale fusion prediction method for remaining service life of lithium ion battery | |
CN109375116B (en) | Battery system abnormal battery identification method based on self-encoder | |
CN113406521B (en) | Lithium battery health state online estimation method based on feature analysis | |
CN113109715A (en) | Battery health condition prediction method based on feature selection and support vector regression | |
CN115575840A (en) | Early prediction method for remaining service life of lithium ion battery | |
CN115510612A (en) | Method for predicting remaining service life of lithium ion battery | |
Zhang et al. | Improved particle swarm optimization-extreme learning machine modeling strategies for the accurate lithium-ion battery state of health estimation and high-adaptability remaining useful life prediction | |
CN115980584A (en) | Lithium battery RUL estimation method based on multi-feature fusion LSTM network | |
CN117031310A (en) | Method for predicting residual service life of power battery of electric automobile | |
Pang et al. | A novel hybrid model for lithium-ion batteries lifespan prediction with high accuracy and interpretability | |
CN114578234A (en) | Lithium ion battery degradation and capacity prediction model considering causality characteristics | |
CN116774086B (en) | Lithium battery health state estimation method based on multi-sensor data fusion | |
Ma et al. | A novel health index for battery RUL degradation modeling and prognostics | |
CN117034762A (en) | Composite model lithium battery life prediction method based on multi-algorithm weighted sum | |
CN114137430A (en) | Estimation method for residual life of retired lithium battery | |
Guo et al. | Future degradation trajectory prediction of lithium-ion battery based on a three-step similarity evaluation criterion for battery selection and transfer learning | |
Xu et al. | Research on residual useful life prediction of li-ion battery based on GA-LM-BPNN | |
Zhou et al. | Residual Life Prediction of Lithium Battery Based on Small Sample Data Sets | |
Hu et al. | Health factor analysis and remaining useful life prediction for batteries based on a cross-cycle health factor clustering framework | |
Wang et al. | A life-prediction method for lithium-ion batteries based on a fusion model and an attention mechanism | |
CN117519448B (en) | Service time warning system and method for tablet personal computer |
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 |