CN107301266B - LOC estimation method and system for lithium iron phosphate battery - Google Patents
LOC estimation method and system for lithium iron phosphate battery Download PDFInfo
- Publication number
- CN107301266B CN107301266B CN201710339846.1A CN201710339846A CN107301266B CN 107301266 B CN107301266 B CN 107301266B CN 201710339846 A CN201710339846 A CN 201710339846A CN 107301266 B CN107301266 B CN 107301266B
- Authority
- CN
- China
- Prior art keywords
- network
- loc
- wavelet
- iron phosphate
- lithium iron
- 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.)
- Expired - Fee Related
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
Abstract
A lithium iron phosphate battery LOC estimation method comprises the following steps: establishing a wavelet neural network model; constructing a wavelet neural network; acquiring input parameters related to LOC of the lithium iron phosphate battery; and inputting the input parameters into a wavelet neural network for data processing, thereby outputting the LOC value of the lithium iron phosphate battery. The invention also discloses a LOC estimation system of the lithium iron phosphate battery corresponding to the estimation method. The method can accurately, effectively and reliably estimate the LOC of the lithium iron phosphate battery, can be applied to charge and discharge management of the battery pack of the electric automobile and related battery industries, is easy to realize, and has good application prospect and market value.
Description
Technical Field
The invention relates to the field of storage battery LOC (available time) estimation methods, in particular to a lithium iron phosphate battery LOC estimation method which can be applied to charge and discharge management of an electric automobile battery pack and related battery industries.
Background
The electric automobile is favored by people as a development trend of future vehicles with the characteristics of environmental protection, energy conservation, portability and the like, and the performance of the battery as a key technical component directly influences the performance of the whole automobile. The storage battery is a power source of the electric automobile, the service life and the price of the storage battery are related to the cost of the electric automobile, the energy storage capacity and the power density of the storage battery determine the driving mileage and the speed of the electric automobile, and the factors directly influence whether the electric automobile can really go to the market, so that the development of the storage battery with long service life, high performance and low cost is a necessary way for developing the electric automobile.
The power battery is used as an energy storage power source and is a core component of the electric automobile, and the super capacitor and the lithium ion battery are the main energy storage power sources applied to the electric automobile at present. The super capacitor is influenced by the energy density and price, and the application of the super capacitor to a pure electric vehicle is greatly limited. Compared with other storage batteries, the lithium battery with the anode material of LiFePO4 has the advantages of high specific energy, high voltage, low self-discharge rate, long charge-discharge service life, no memory effect, wide working temperature range, no pollution, safety, reliability and the like.
As an important issue in the control and management of a battery system, it is also a key to the anxiety of electric vehicle driving that how much time remains for the battery to be used. The invention provides the LOC available time parameter, and has very important practical significance and economic value.
In the actual running process of the electric automobile, the working condition of the battery is very complex, the relations between measurable parameters such as voltage, current, temperature and the like and LOC are complex nonlinear relations, particularly, in the starting, accelerating or climbing stage of the electric automobile, the battery is in a heavy-current discharging state, the voltage and current change is severe, the fluctuation is large, and meanwhile, different environmental temperatures have great influence on the performance of the battery,
the invention provides a method for predicting the available time of a lithium iron phosphate battery by using a scalable dimension wavelet neural network method by analyzing various factors influencing the LOC of a storage battery.
Disclosure of Invention
In order to solve the technical problems and realize accurate, effective and reliable estimation of the LOC of the lithium iron phosphate battery, the invention provides a method and a system for estimating the LOC of the lithium iron phosphate battery, and the technical scheme adopted by the invention for solving the problems is as follows:
the invention provides a LOC estimation method for a lithium iron phosphate battery, which comprises the following steps of:
s1: establishing a wavelet neural network model;
s2: constructing a wavelet neural network;
s3: acquiring input parameters related to LOC of the lithium iron phosphate battery;
s4: and inputting the input parameters into a wavelet neural network for data processing, thereby outputting the LOC value of the lithium iron phosphate battery.
Further, the topology of the wavelet neural network model includes an input layer, a hidden layer, and an output layer.
Further, the node transfer function of the hidden layer is a wavelet basis function.
Further, the wavelet basis function adopts a Morlet function, and the mathematical formula of the Morlet function is as follows:
further, the step S2 includes:
(1) establishing an initial network, wherein the initial network is formed by a wavelet frame with the dimension j;
(2) training, judging whether the error is increased in the training process, recording the weight before the error is increased and taking the weight as the final weight of the current network, judging whether the precision meets the requirement, stopping training if the precision meets the requirement, merging the sub-network with the dimension of j +1 if the precision does not meet the requirement, and enabling the expected output of the current-level network to be the error of the previous-level network, namely learning the error of the previous-level network;
(3) and by analogy, when the wavelet neural network is merged into the mth sub-network, the training is finished when the precision meets the requirement, the weight of the trained whole network is obtained at the moment, and the wavelet neural network is constructed.
Further, the input parameters include current, voltage, temperature, discharge rate, and cycle number.
In addition, the present invention also provides a system for implementing the lithium iron phosphate battery LOC estimation method, the system comprising:
the modeling unit is used for establishing a wavelet neural network model;
the network construction unit is used for constructing a wavelet neural network;
an acquisition unit for acquiring input parameters related to the LOC;
and the processing unit is used for inputting the input parameters into the wavelet neural network for data processing, so that the LOC value of the lithium iron phosphate battery is output.
Further, the topology of the wavelet neural network model includes an input layer, a hidden layer, and an output layer.
Further, the node transfer function of the hidden layer is a wavelet basis function, the wavelet basis function adopts a Morlet function, and the mathematical formula of the Morlet function is as follows:
further, the input parameters include current, voltage, temperature, discharge rate, and cycle number.
The invention has the beneficial effects that:
1. and comprehensively selecting the scale parameters and the translation parameters. In general, a multi-dimensional wavelet network based on frames adopts different wavelet frames to approximate the nonlinear mapping between each dimension of input signals to output signals, which not only results in the possibility of large redundancy among frames, but also the number of hidden nodes of the network is multiplied as the input dimension is increased. The invention carries out comprehensive time domain analysis on each dimension of input signals, selects proper displacement parameters under the condition of determining scale parameters to enable a wavelet expansion system, namely a wavelet tight frame under a single scale to cover each dimension of input vectors, thereby approximating the nonlinear mapping from each dimension of input signals to output signals by using one wavelet tight frame. In this way, not only is the redundancy between different wavelet frameworks reduced, but the number of hidden layer nodes of the network does not multiply with the increase in input dimension.
2. Adaptive determination of hidden layer nodes. The self-adaptive first-order self-feedback wavelet neural network constructed by the invention based on the single-scale multi-dimensional tight-support radial wavelet framework comprises an initial network and a sub-network which is self-adaptively incorporated according to the requirement of precision in the training process. The initial network and each sub-network are forward neural networks with single hidden layers, the hidden layers of each network are composed of multi-dimensional tight-branch radial wavelet frames under a single scale, the scales are increased step by step, each network learns the error of the network at the previous stage, and the training of the parameters of the newly incorporated sub-network does not affect the parameters of the trained network until the precision requirement is met. Therefore, the number of hidden layer nodes of the network can be determined in a self-adaptive manner according to the requirement of precision, so that the determination of the number of the hidden layer nodes can be based on the requirement, and the structure of the network is optimized to a great extent.
3. Application of a first order self-feedback structure. The hidden layer node of the network adopts a first-order self-feedback structure, and the dynamic characteristic of the system is reflected by introducing a first-order self-feedback link, so that the dynamic characteristic of the system can be reflected without performing m-order and n-order time delay on the input and the output of the network respectively and feeding back the output of the network to the input end like the traditional multilayer dynamic forward network, and the number of input nodes of the network is reduced to a great extent. Because the input of the hidden layer node consists of the input at the current moment and the output at the previous moment, the output at the previous moment is a function of the input at the moment, and the input at the previous moment comprises the output at the next previous moment, infinite recursion is formed, the memory of the hidden layer node to the information is infinite, and the dynamic characteristic of the system can be better reflected on the premise of not increasing network input nodes.
Drawings
FIG. 1 is a block diagram of a step flow of a LOC estimation method for a lithium iron phosphate battery according to the present invention;
FIG. 2 is a topological structure of a wavelet neural network model according to a preferred embodiment of the present invention;
FIG. 3 is a flow chart of the construction of a wavelet neural network according to the preferred embodiment of the present invention;
fig. 4 is a schematic structural block diagram of a LOC estimation system of a lithium iron phosphate battery according to an embodiment of the present invention.
Detailed Description
The invention is described in detail below with reference to the figures and examples, wherein the specific examples should not be construed as limiting the scope of the invention.
The LOC of the battery is a direct basis for the sustainable driving of the electric vehicle, and therefore, the electric vehicle puts higher demands on the accuracy and reliability of the estimation. First, it is necessary to understand the theoretical formula of LOC for lithium iron phosphate battery: LOC ═ QnX (SOC-DZSOC)/V. Service life LOC of single battery in single cycle, state of charge SOC of battery, dead zone capacity DZSOC and rated capacity Q of batterynAnd the discharge rate V is related, and meanwhile, the weighting coefficient is adopted to correct the above formula in consideration of various working conditions and stress in the actual model.
In order to estimate LOC of the lithium iron phosphate battery more accurately, effectively and reliably, the invention provides a method for estimating LOC of the lithium iron phosphate battery, which refers to a flow chart of steps shown in fig. 1, and in one embodiment, the method comprises the following steps:
s1: establishing a wavelet neural network model;
s2: constructing a wavelet neural network;
s3: acquiring input parameters related to LOC of the lithium iron phosphate battery;
s4: and inputting the input parameters into a wavelet neural network for data processing, thereby outputting the LOC value of the lithium iron phosphate battery.
Further, the wavelet neural network model established in step 1 is based on the BP neural network model, the node transfer function of the hidden layer in the BP neural network model is replaced by the wavelet basis function, and the replaced BP neural network model forms the wavelet neural network model. The advantage of this is that the wavelet transform performs multi-scale analysis on the signal by translation and scale expansion, which can more effectively extract the local information of the signal; for the neural network, the neural network has the characteristics of good self-adaptability, good self-learning capability, strong fault tolerance and the like, and can be used for approximating most functions. The predictive model preferably employs a three-layer WNN predictive structure, i.e., comprising an input layer, a hidden layer, and an output layer. Because a prediction model with less than three-layer structure can not be accurately predicted; prediction models with structures larger than three layers will correspondingly increase the amount of computation. Therefore, in a preferred embodiment, the topological structure of the wavelet neural network model is shown in fig. 2, wherein the wavelet basis function preferably adopts a Morlet function, and the mathematical formula thereof is as follows:
referring to fig. 2, the mathematical model of the established wavelet neural network can be expressed as:
wherein x isi(t) is the ith input of the network, Ij(t) is the sum of the jth wavelet basis function inputs, Sj(t) is the output of the jth wavelet. The psi is a function of a wavelet,
and yk(t) is the kth output of the networkNode, ek(t) is the model error, the weight from the input layer to the hidden layer is
The weight from the hidden layer to the output layer is
The hidden layer node reflects the dynamic characteristic of the system by introducing a first-order self-feedback link, so that the dynamic characteristic of the system can be reflected without using the traditional multilayer dynamic forward network which respectively carries out m-order and n-order time delay on the input and the output of the network and feeds the output of the network back to the input end, and the number of input nodes of the network is reduced to a great extent. The input of the hidden layer node is composed of the input of the current moment and the output of the previous moment, the output of the previous moment is a function of the input of the moment, the input of the previous moment comprises the output of the previous moment, and therefore infinite recursion is formed, and the memory of the hidden layer node on the information is infinite. Therefore, compared with a multilayer dynamic forward network, the dynamic characteristic of the system can be better reflected by adopting a first-order delay structure, and the number of input nodes of the network is not increased.
Further, in a preferred embodiment, the flow of constructing the wavelet neural network in step 2 is shown in fig. 3, and may specifically include:
(1) an initial network is established, wherein the initial network is composed of a wavelet frame with a dimension j, wherein j is a smaller integer. When j takes 1 or 2, the training accuracy is usually low, so the scale of the initial network can take an integer of 3, 4 or more;
(2) training, judging whether the error is increased or not in the training process, gradually reducing the error in the training process until the error is kept basically constant, when the error is increased, indicating that the network is over-trained, if the training is continued, continuously increasing the error of the network, recording a weight before the error is increased and taking the weight as a final weight of the current network, meanwhile, judging whether the precision meets the requirement or not, if the precision meets the requirement, stopping the training, if the precision does not meet the requirement, merging the weight into a sub-network with the size of j +1, and enabling the expected output of the current-level network to be the error of the previous-level network, namely learning the error of the previous-level network;
(3) and by analogy, when the wavelet neural network is merged into the mth sub-network, the training is finished when the precision meets the requirement, the weight of the trained whole network is obtained at the moment, and the wavelet neural network is constructed.
The self-adaptive first-order self-feedback wavelet neural network constructed by the invention based on the single-scale multi-dimensional tight-support radial wavelet framework comprises an initial network and a sub-network which is self-adaptively incorporated according to the requirement of precision in the training process. The initial network and each sub-network are forward neural networks with single hidden layers, the hidden layers of each network are composed of multi-dimensional tight-branch radial wavelet frames under a single scale, the scales are increased step by step, each network learns the error of the network at the previous stage, and the training of the parameters of the newly incorporated sub-network does not affect the parameters of the trained network until the precision requirement is met. Therefore, the number of hidden layer nodes of the network can be determined in a self-adaptive manner according to the requirement of precision, so that the determination of the number of the hidden layer nodes can be based on the requirement, and the structure of the network is optimized to a great extent.
Further, in step 3, acquiring the input parameters related to the lithium iron phosphate battery LOC may include: current, voltage, temperature, discharge rate, and cycle number.
Further, in step 4, the input parameters related to the LOC of the lithium iron phosphate battery obtained in step 3 are input into the wavelet neural network constructed in step 2 for data processing, so as to output the LOC value of the lithium iron phosphate battery.
In addition, the present invention further provides a lithium iron phosphate battery LOC estimation system, as shown in fig. 4, which is a schematic structural block diagram of a specific embodiment of the system, and the system includes:
the modeling unit is used for establishing a wavelet neural network model;
the network construction unit is used for constructing a wavelet neural network;
an acquisition unit for acquiring input parameters related to the LOC;
and the processing unit is used for inputting the input parameters into the wavelet neural network for data processing, so that the LOC value of the lithium iron phosphate battery is output.
Further, the topological structure of the wavelet neural network model established by the modeling unit comprises an input layer, a hidden layer and an output layer.
Further, the node transfer function of the hidden layer is a wavelet basis function, the wavelet basis function adopts a Morlet function, and the mathematical formula of the Morlet function is as follows:
further, the process of constructing the wavelet neural network by the network construction unit is described in step 2.
Further, the acquiring unit acquires input parameters related to the lithium iron phosphate battery LOC, and may include: current, voltage, temperature, discharge rate, and cycle number.
Further, the processing unit inputs the input parameters related to the LOC of the lithium iron phosphate battery acquired by the acquisition unit into the wavelet neural network for data processing, so as to output the LOC value of the lithium iron phosphate battery.
The above description is only a preferred embodiment of the present invention, and the present invention is not limited to the above embodiment, and the present invention shall fall within the protection scope of the present invention as long as the technical effects of the present invention are achieved by the same means. The invention is capable of other modifications and variations in its technical solution and/or its implementation, within the scope of protection of the invention.
Claims (9)
1. A LOC estimation method for a lithium iron phosphate battery is characterized by comprising the following steps:
s1: establishing a wavelet neural network model, wherein the mathematical model of the wavelet neural network is expressed as:
wherein x isi(t) is the ith input of the network, Ij(t) is the sum of the jth wavelet basis function inputs, Sj(t) is the output of the jth wavelet, ψ is the wavelet function,
and yk(t) is the kth output node of the network, ek(t) is the error of the model,
for the weights of the input layer to the hidden layer,
the weights for the hidden layer to the output layer,
is 1;
s2: constructing a wavelet neural network, comprising:
(1) establishing an initial network, wherein the initial network is formed by a wavelet frame with the dimension j;
(2) training, judging whether the error is increased in the training process, recording the weight before the error is increased and taking the weight as the final weight of the current network, judging whether the precision meets the requirement, stopping training if the precision meets the requirement, merging the sub-network with the dimension of j +1 if the precision does not meet the requirement, and enabling the expected output of the current-level network to be the error of the previous-level network, namely learning the error of the previous-level network;
(3) by analogy, when the wavelet neural network is merged into the mth sub-network, the training is finished when the precision meets the requirement, the weight of the trained whole network is obtained at the moment, and the wavelet neural network is constructed;
s3: acquiring input parameters related to LOC of the lithium iron phosphate battery;
s4: and inputting the input parameters into a wavelet neural network for data processing, thereby outputting the LOC value of the lithium iron phosphate battery.
2. The lithium iron phosphate battery LOC estimation method according to claim 1, wherein the topological structure of the wavelet neural network model comprises an input layer, a hidden layer and an output layer.
3. The lithium iron phosphate battery LOC estimation method according to claim 2, wherein the node transfer function of the hidden layer is a wavelet basis function.
4. The LOC estimation method for lithium iron phosphate battery according to claim 3, wherein the wavelet basis function adopts a Morlet function, and the mathematical formula is as follows:
5. the lithium iron phosphate battery LOC estimation method according to claim 1, wherein said input parameters include current, voltage, temperature, discharge rate and cycle number.
6. A lithium iron phosphate battery LOC estimation system, the system comprising:
the modeling unit is used for establishing a wavelet neural network model, wherein the mathematical model of the wavelet neural network is expressed as:
wherein x isi(t) is the ith input of the network, Ij(t) is the sum of the jth wavelet basis function inputs, Sj(t) is the output of the jth wavelet, ψ is the wavelet function,
and yk(t) is the kth output node of the network, ek(t) is the error of the model,
for the weights of the input layer to the hidden layer,
the weights for the hidden layer to the output layer,
is 1;
the network construction unit is used for constructing the wavelet neural network and comprises the following components:
(1) establishing an initial network, wherein the initial network is formed by a wavelet frame with the dimension j;
(2) training, judging whether the error is increased in the training process, recording the weight before the error is increased and taking the weight as the final weight of the current network, judging whether the precision meets the requirement, stopping training if the precision meets the requirement, merging the sub-network with the dimension of j +1 if the precision does not meet the requirement, and enabling the expected output of the current-level network to be the error of the previous-level network, namely learning the error of the previous-level network;
(3) by analogy, when the wavelet neural network is merged into the mth sub-network, the training is finished when the precision meets the requirement, the weight of the trained whole network is obtained at the moment, and the wavelet neural network is constructed;
an acquisition unit for acquiring input parameters related to the LOC;
and the processing unit is used for inputting the input parameters into the wavelet neural network for data processing, so that the LOC value of the lithium iron phosphate battery is output.
7. The lithium iron phosphate battery LOC estimation system of claim 6, wherein the topology of the wavelet neural network model comprises an input layer, a hidden layer and an output layer.
8. The lithium iron phosphate battery of claim 6The LOC estimation system is characterized in that the node transfer function of the hidden layer is a wavelet basis function, the wavelet basis function adopts a Morlet function, and the mathematical formula of the Morlet function is as follows:
9. the lithium iron phosphate battery LOC estimation system of claim 6, wherein said input parameters include current, voltage, temperature, discharge rate and cycle number.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710339846.1A CN107301266B (en) | 2017-05-15 | 2017-05-15 | LOC estimation method and system for lithium iron phosphate battery |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710339846.1A CN107301266B (en) | 2017-05-15 | 2017-05-15 | LOC estimation method and system for lithium iron phosphate battery |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107301266A CN107301266A (en) | 2017-10-27 |
| CN107301266B true CN107301266B (en) | 2020-11-10 |
Family
ID=60137219
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710339846.1A Expired - Fee Related CN107301266B (en) | 2017-05-15 | 2017-05-15 | LOC estimation method and system for lithium iron phosphate battery |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107301266B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110007235A (en) * | 2019-03-24 | 2019-07-12 | 天津大学青岛海洋技术研究院 | A kind of accumulator of electric car SOC on-line prediction method |
| CN112818594B (en) * | 2021-01-28 | 2023-05-30 | 温州大学 | Multi-objective battery pack structure optimizing method based on neural network |
| CN113049961B (en) * | 2021-02-26 | 2022-07-19 | 佛山职业技术学院 | DZSOC algorithm of lithium iron phosphate battery |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8760115B2 (en) * | 2009-08-20 | 2014-06-24 | GM Global Technology Operations LLC | Method for charging a plug-in electric vehicle |
| US10664562B2 (en) * | 2013-02-24 | 2020-05-26 | Fairchild Semiconductor Corporation and University of Connecticut | Battery state of charge tracking, equivalent circuit selection and benchmarking |
| CN103716177A (en) * | 2013-11-18 | 2014-04-09 | 国家电网公司 | Security risk assessment method and apparatus |
| CN106443453A (en) * | 2016-07-04 | 2017-02-22 | 陈逸涵 | Lithium battery SOC estimation method based on BP neural network |
| CN106650060B (en) * | 2016-12-08 | 2020-05-15 | 国网青海省电力公司 | Photovoltaic cell internal resistance attenuation coefficient prediction method |
-
2017
- 2017-05-15 CN CN201710339846.1A patent/CN107301266B/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| CN107301266A (en) | 2017-10-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112331941B (en) | Cloud auxiliary battery management system and method | |
| CN110286332B (en) | Electric vehicle power battery SOC estimation method based on multiple innovation theory | |
| CN110398697B (en) | Lithium ion health state estimation method based on charging process | |
| KR20200140093A (en) | Prediction Method and Prediction System for predicting Capacity Change according to Charging / Discharging Cycle of Battery | |
| Zhou et al. | Battery health prognosis using improved temporal convolutional network modeling | |
| CN111832825B (en) | Wind power prediction method and system integrating long-term memory network and extreme learning machine | |
| CN107301266B (en) | LOC estimation method and system for lithium iron phosphate battery | |
| Zhang et al. | Intelligent state of charge estimation of battery pack based on particle swarm optimization algorithm improved radical basis function neural network | |
| Zhang et al. | State-of-charge estimation of lithium-ion battery pack based on improved RBF neural networks | |
| AU2021101964A4 (en) | Artificial intelligence based smart electric vehicle battery management system | |
| CN112305423A (en) | Lithium ion power battery state of charge estimation method, device, medium and equipment | |
| Wang et al. | Battery full life cycle management and health prognosis based on cloud service and broad learning | |
| De Sousa et al. | Comparison of different approaches for lead acid battery state of health estimation based on artificial neural networks algorithms | |
| CN113791351B (en) | Lithium battery life prediction method based on transfer learning and difference probability distribution | |
| Cai et al. | Remaining useful life prediction method of EV power battery for DC fast charging condition | |
| CN112881914B (en) | Lithium battery health state prediction method | |
| CN113721151B (en) | Battery capacity estimation model and method based on double-tower deep learning network | |
| CN105676129A (en) | Lithium ion battery state of charge prediction method based on grey theory | |
| CN113093014B (en) | Online collaborative estimation method and system for SOH and SOC based on impedance parameters | |
| Yu et al. | Remaining useful life prediction based on multi-scale residual convolutional network for aero-engine | |
| Channegowda et al. | An attention‐based synthetic battery data augmentation technique to overcome limited dataset challenges | |
| Wang et al. | Joint prediction of Li-ion battery state of charge and state of health based on the DRSN-CW-LSTM model | |
| Mu et al. | Estimation of state of charge of lithium-ion batteries based on wide and deep neural network model | |
| CN111882119A (en) | Battery SOH prediction optimization method based on SA-BP neural network | |
| Shi et al. | Research on DC-DC power health management technology based on degradation test and deep learning |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20201110 |