CN114167295B - Lithium ion battery SOC estimation method and system based on multi-algorithm fusion - Google Patents
Lithium ion battery SOC estimation method and system based on multi-algorithm fusion Download PDFInfo
- Publication number
- CN114167295B CN114167295B CN202111438921.2A CN202111438921A CN114167295B CN 114167295 B CN114167295 B CN 114167295B CN 202111438921 A CN202111438921 A CN 202111438921A CN 114167295 B CN114167295 B CN 114167295B
- Authority
- CN
- China
- Prior art keywords
- layer
- soc
- neuron
- value
- output
- 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]
- G01R31/382—Arrangements for monitoring battery or accumulator variables, e.g. SoC
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/60—Other road transportation technologies with climate change mitigation effect
- Y02T10/70—Energy storage systems for electromobility, e.g. batteries
Abstract
The invention provides a lithium ion battery SOC estimation method and system based on multi-algorithm fusion, which overcome the difficulty that an ideal model capable of completely describing the external characteristics of a battery is difficult to construct in the implementation process of the existing SOC estimation method based on a model, and aim at the traditional BP layer networkThe network is easy to fall into local optimum, and the complex iterative computation process can obviously reduce the network convergence speed, and the optimization algorithm based on the particle swarm optimization is provided for improvement. Respectively introducing adaptive extended Kalman filtering and adaptive H to eliminate noise influence as much as possible ∞ The filtering is organically combined with the fusion algorithm, thereby providing good suppression effect for the situations of white Gaussian noise and colored noise.
Description
Technical Field
The invention belongs to the technical field of power battery management, and particularly relates to a method and a system for realizing joint estimation of SOC of a lithium ion battery.
Background
In the prior art, a model-based method is adopted for estimating the SOC of the lithium ion battery, which is a currently preferred solution, however, because the state change inside the battery is complex and has strong nonlinear characteristics, it is very difficult to find an ideal model to completely describe the external characteristics of the battery, and the practicability and precision of the battery model are limited by practical factors such as calculation cost, so that the defect that the internal characteristics of the battery cannot be truly and effectively reflected is generally existed. With the continuous development of artificial intelligence and big data technology, the research method based on data driving can obtain the internal characteristics of the measured object at the current moment by directly linking with the acquired basic data without depending on establishing an accurate mathematical model. The deep learning method can organize a plurality of neurons with simple processing capability, so that a complex nonlinear network has strong generalization capability. Therefore, if the nonlinear characteristics of the lithium ion battery can be accurately simulated by using the method, the accuracy and efficiency of real-time estimation of the SOC are expected to be remarkably improved.
Disclosure of Invention
Aiming at the technical problems in the prior art, the invention provides a lithium ion battery SOC estimation method based on multi-algorithm fusion, which specifically comprises the following steps:
carrying out a verification experiment aiming at a lithium ion battery, taking voltage, current and temperature data of a working condition cyclic discharge stage as a training input set, calculating the SOC of each moment of the working condition cyclic discharge stage based on an ampere-hour integration method according to the charge capacity of a CC-CV (constant current-constant voltage) charge stage, and taking the SOC as a training output set;
establishing a deep confidence network consisting of an input layer, an RBM layer, a BP layer and an output layer, and training the deep confidence network by using the training input set and the training output set; wherein a restricted Boltzmann machine training process is employed for the RBM layer;
determining the length of particles based on the number of neurons in the BP layer network structure, and calculating the optimal weight bias of the BP layer of the deep belief network by utilizing a particle swarm optimization algorithm;
step four, taking the voltage, the current and the temperature as input and the SOC as output, on one hand, establishing an SOC estimation model based on an Adaptive Extended Kalman Filter (AEKF) algorithm, and on the other hand, establishing an SOC estimation model based on an adaptive H ∞ An SOC estimation model of a filtering (AHIFF) algorithm, and establishing an SOC fusion estimator according to a weight distribution method by using estimation results of the two SOC estimation models; updating the initial values of the two SOC estimation models respectively by using the deep belief network;
estimating the SOC of the lithium ion battery by using the SOC fusion estimator; and comparing the estimation result with the actual SOC value, and updating each estimation model periodically.
Further, the voltage, current and temperature data of the condition cycle discharge stage in the step one are obtained according to Dynamic Stress Test (DST) conditions or U.S. urban road cycle (UDDS) conditions.
Further, the second step specifically includes the following steps:
determining the network structures and the number of neurons of an input layer, an RBM layer, a BP layer and an output layer in the deep belief network; determining an activation function and an evaluation function, and performing optimization calculation on each RBM structure by using a random gradient descent method with root mean square error as a target function;
the restricted boltzmann machine training process specifically comprises:
firstly, setting a weight matrix W, a hidden layer bias vector b and a visible layer bias vector c of an RBM layer:
wherein m and n are the neuron serial numbers of the visible layer and the hidden layer respectively;
the following steps are performed in sequence:
1) assigning the input data x to the visible layer unit, and calculating the conditional probability value P of each hidden layer neuron:
wherein h is a hidden layer neuron; v is a visible layer neuron; j is 1,2, …, n;is Sigmoid function, e is exponential constant;
2) reconstructing the corresponding value of the hidden layer neuron by using one Gibbs sampling to generate [0,1 ]]Random number r of j And then:
3) reconstructing visible layer neurons with hidden layer neurons, v for each visible layer neuron i (i ═ 1,2, …, m) its conditional probability value is calculated:
in the formula, superscript denotes reconstruction update;
4) reconstructing the corresponding value of the neuron in the visible layer by using the Gibbs sampling once again to generate a [0,1 ]]Random number s of i Then:
5) and reconstructing hidden layer neurons by using the visible layer neurons, and calculating a conditional probability value P of each hidden layer neuron:
6) updating a weight matrix W, a hidden layer bias vector b and a visible layer bias vector c of the RBM layer according to a predetermined learning rate lambda and iteration times:
W * =W+λ[P(h=1|v)v T -P(h * =1|v * )v *T ]
b * =b+λ(v-v * )
c * =c+λ[P(h=1|v)-P(h*=1|v*)]。
further, the third step specifically includes the following steps:
firstly, determining a weight matrix W of a BP layer to see a layer offset vector c:
wherein m and n are the neuron serial numbers of the visible layer and the hidden layer respectively;
the following steps are performed in sequence:
1) calculate the value for each hidden layer neuron:
h (l) =c+W (l) v
wherein h is a hidden layer neuron; v is a visible layer neuron; l is the layer number index of the neural network;
2) using Sigmoid function, a normalized hidden value σ (h) is obtained (l) And calculating the output value Y of the output layer:
Y (l) =σ(h) (l) =σ(c+W (l) v);
3) constructing a cost function based on the root mean square error as a judgment criterion for evaluating the output value of the output layer:
wherein E is the root mean square error in the training process, N is the number of samples,X i respectively representing the output value and the ideal output value of the output layer;
4) updating the weight matrix W and the visible layer bias vector c of the BP layer according to the determined particle length, particle scale and evolution times and in combination with the predetermined learning rate lambda and iteration times:
the updating process comprises the following steps:
in a search space defining a dimension D, n particles together form a population X ═ X (X) 1 ,X 2 ,…,X n ) Wherein the position of the ith (i ═ 1,2, …, m) particle in the D-dimensional search space is X i =[x i1 ,x i2 ,…,x iD ] T ,x id Represents D (D ═ 1,2, …, D) dimensional coordinates; velocity of the ith particle is V i =[V i1 ,V i2 ,…,V iD ] T With an individual extremum of P i =[P i1 ,P i2 ,…,P iD ] T Global extremum of race is P g =[P g1 ,P g2 ,…,P gD ] T ;
Updating the particle moving speed respectively:
wherein, V id Is the moving speed of the particles; omega is the inertial weight; k is the current iteration number; alpha (alpha) ("alpha") 1 And alpha 2 Is an acceleration factor; r is 1 And r 2 Is distributed in [0,1 ]]A random number in between; x gd Is the extreme value of the particle position;
and updating the particle position:
correspondingly, the invention also provides a lithium ion battery SOC estimation system based on multi-algorithm fusion, and the lithium ion battery SOC is calculated by executing the method.
The method and the system provided by the invention overcome the difficulty that an ideal model capable of completely describing the external characteristics of the battery is difficult to construct in the implementation process of the existing SOC estimation method based on the model, and provide the optimization algorithm based on the particle swarm optimization for improvement aiming at the defects that the traditional BP layer network is easy to fall into local optimization and the network convergence speed is obviously reduced in the complex iterative computation process. Respectively introducing adaptive extended Kalman filtering and adaptive H to eliminate noise influence as much as possible ∞ The filtering is organically combined with the fusion algorithm, thereby providing good suppression effect for the situations of white Gaussian noise and colored noise.
Drawings
FIG. 1 is a flow chart of deep belief network training based on particle swarm optimization;
FIG. 2 is a general flow chart of the method provided by the present invention.
Detailed Description
The technical solutions of the present invention will be described clearly and completely with reference to the accompanying drawings, and it should be understood that the described embodiments are some, but not all embodiments of the present invention. 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.
The invention provides a lithium ion battery SOC estimation method based on multi-algorithm fusion, which specifically comprises the following steps as shown in FIG. 2:
carrying out a verification experiment aiming at a lithium ion battery, taking voltage, current and temperature data of a working condition cyclic discharge stage as a training input set, calculating the SOC of each moment of the working condition cyclic discharge stage based on an ampere-hour integration method according to the charge capacity of a CC-CV (constant current-constant voltage) charge stage, and taking the SOC as a training output set;
establishing a deep confidence network consisting of an input layer, an RBM layer, a BP layer and an output layer, and training the deep confidence network by using the training input set and the training output set; wherein a restricted Boltzmann machine training process is employed for the RBM layer;
determining the length of particles based on the number of neurons in the BP layer network structure, and calculating the optimal weight bias of the BP layer of the deep belief network by utilizing a particle swarm optimization algorithm;
step four, taking voltage, current and temperature as input and SOC as output, on one hand, establishing an SOC estimation model based on an Adaptive Extended Kalman Filter (AEKF) algorithm, and on the other hand, establishing an SOC estimation model based on an adaptive H algorithm ∞ An SOC estimation model of a filtering (AHIFF) algorithm, and establishing an SOC fusion estimator according to a weight distribution method by using estimation results of the two SOC estimation models; updating the initial values of the two SOC estimation models respectively by using the deep belief network;
estimating the SOC of the lithium ion battery by using the SOC fusion estimator; and comparing the estimation result with the actual SOC value, and updating each estimation model periodically.
In a preferred embodiment of the present invention, the voltage, current and temperature data of the duty cycle discharge phase in step one are obtained according to Dynamic Stress Test (DST) duty or us urban road cycle (UDDS) duty. The data recorded in the DST working condition data set comprises voltage, current, temperature and sampling time, verification experiment data of a test battery with the same capacity point at four temperature points and experiment data of a specific lithium ion battery at DST working condition cyclic discharge stages at 0 ℃, 10 ℃ and 40 ℃ are selected as training samples, and on the basis of the charge capacity of a CC-CV charge stage, the calculation of SOC at each time of the working condition cyclic discharge stage is completed in an ampere-hour integration mode to serve as output of model training. And taking the DST and UDDS verification working condition data of the specific battery at 25 ℃ as test samples respectively for verifying the effectiveness and the applicability of the algorithm.
In a preferred embodiment of the present invention, the second step specifically comprises the following steps:
finally determining 6 layers and 5 RBM structures in the network pre-training stage through testing, wherein the number of neurons in an input layer is 3, the number of neurons in a first layer is set to be 1000, the number of neurons in a middle four layers is set to be 100, and the number of neurons in a last layer is 20 and is used as the input of a BP layer; the BP layer consists of an input layer, a hidden layer and an output layer, the number of the neurons of each layer is respectively 20, 6 and 1, and the output of the BP layer is used as the output of the whole DBN network. For each RBM structure, the root mean square error is taken as an objective function, and a random gradient descent method is adopted as an optimization algorithm (the learning rate is set to be 0.1, and the minimum batch is 50). The learning rate of the parameter fine-tuning stage is set to be 0.1, and the iteration times are set to be 100; the learning rate was set to 0.1 and the minimum batch was 50.
The limited Boltzmann machine training process specifically comprises the following steps:
firstly, setting a weight matrix W, a hidden layer bias vector b and a visible layer bias vector c of an RBM layer:
wherein m and n are the neuron serial numbers of the visible layer and the hidden layer respectively;
the following steps are performed in sequence:
1) assigning the input data x to the visible layer unit, and calculating the conditional probability value P of each hidden layer neuron:
wherein h is a hidden layer neuron; v is a visible layer neuron; j is 1,2, …, n;is Sigmoid function, e is exponential constant;
2) reconstructing the corresponding value of the hidden layer neuron by using one Gibbs sampling to generate [0,1 ]]Random number r of j And then:
3) reconstructing visible layer neurons with hidden layer neurons, v for each visible layer neuron i (i ═ 1,2, …, m) its conditional probability value is calculated:
in the formula, superscript denotes reconstruction update;
4) reconstructing the corresponding value of the neuron in the visible layer by using the Gibbs sampling once again to generate a [0,1 ]]Random number s of i And then:
5) and reconstructing hidden layer neurons by using the visible layer neurons, and calculating a conditional probability value P of each hidden layer neuron:
6) updating a weight matrix W, a hidden layer bias vector b and a visible layer bias vector c of the RBM layer according to a predetermined learning rate lambda and iteration times:
W * =W+λ[P(h=1|v)v T -P(h * =1|v * )v *T ]
b * =b+λ(v-v * )
c * =c+λ[P(h=1|v)-P(h*=1|v*)]。
in a preferred embodiment of the present invention, the step three specifically includes the following processes:
firstly, determining a weight matrix W of a BP layer to see a layer offset vector c:
wherein m and n are the neuron serial numbers of the visible layer and the hidden layer respectively;
the following steps are performed in sequence:
1) calculate the value for each hidden layer neuron:
h (l) =c+W (l) v
in the formula, h is hidden layer neuron; v is a visible layer neuron; l is the layer number index of the neural network;
2) using Sigmoid function, a normalized hidden value σ (h) is obtained (l) And calculating the output value Y of the output layer:
Y (l) =σ(h) (l) =σ(c+W (l) v);
3) constructing a cost function based on the root mean square error as a judgment criterion for evaluating the output value of the output layer:
wherein E is the root mean square error in the training process, N is the number of samples,X i respectively representing the output value and the ideal output value of the output layer;
4) according to the determined particle length, particle size and evolution frequency, for example, for the last BP layer, the network structure is 20-6-1, so that 20 × 6+6 × 1 equals 126 weights, 6+1 equals 7 thresholds, the particle length of the PSO algorithm is 126+7 equals 133, the particle size is set to be 130, and the evolution frequency is set to be 50. And updating the weight matrix W and the visible layer offset vector c of the BP layer by combining the predetermined learning rate lambda and the iteration times:
as shown in fig. 1, the updating process specifically includes:
in a search space defining a dimension D, n particles together form a population X ═ X (X) 1 ,X 2 ,…,X n ) Wherein the position of the ith (i ═ 1,2, …, m) particle in the D-dimensional search space is X i =[x i1 ,x i2 ,…,x iD ] T ,x id Represents D (D ═ 1,2, …, D) dimensional coordinates; velocity of the ith particle is V i =[V i1 ,V i2 ,…,V iD ] T With an individual extremum of P i =[P i1 ,P i2 ,…,P iD ] T Global extremum of race is P g =[P g1 ,P g2 ,…,P gD ] T ;
Updating the particle moving speed respectively:
wherein, V id Is the moving speed of the particles; omega is the inertial weight; k is the current iteration number; alpha is alpha 1 And alpha 2 Is an acceleration factor; r is 1 And r 2 Is distributed in [0,1 ]]A random number in between; x gd Is the extreme value of the particle position;
and updating the particle position:
in the fourth step, the calculation process of the SOC estimation method based on the (PSO-DBN) -AHIFF or AHIFF fusion algorithm is as follows:
model discretization equation:
(1) initialization
(AEKF) sets the initial value of the state observer: x is the number of 0 ,P 0 ,Q 0 ,R 0
(AHIFF) setting an initial value of the state observer: x is the number of 0 ,P 0 ,Q 0 ,R 0 ,L k ,S k ,1/γ
(2) A priori estimate-predict: (k-1) + →k -
(3) a posteriori estimation-correction: k is a radical of - →k +
And correcting the estimated values of the system state and the error covariance through the measured value at the moment k:
(AEKF) filter gain:
(AHIFF) filter gain:
adaptive noise covariance matching:
(AKEF) error covariance matrix correction:
(AHIFF) error covariance matrix correction:
(4) and (3) updating the time scale: k is k +1, and a state estimate at time (k +1) is prepared.
When the estimation error of the AEKF is large, the accuracy requirement cannot be met, and the SOC fusion estimator starts to function at the moment: when the estimation error of the AEKF is in the boundary value period, the SOC fusion estimation value is positioned between the AEKF and the AHIFF; and when the estimation error of the AEKF is large, the AEKF has lost the filtering performance, and the estimation result at this time is represented by AHIFF. The SOC estimation error after fusion can be basically maintained within 2.8%, so that the fusion algorithm provided by the invention can effectively improve the estimation precision under the condition of complex noise.
The estimation method provided by the invention has the advantages that the error is basically within 1% under the noise-free condition, the estimation error of the AEKF part is basically controlled within 1.8%, and the estimation error of the overall algorithm is basically controlled within 2.8% under the colored noise condition, so that the characteristics that the fusion algorithm can improve the estimation precision and the robustness are proved.
It should be understood that, the sequence numbers of the steps in the embodiments of the present invention do not mean the execution sequence, and the execution sequence of each process should be determined by the function and the inherent logic of the process, and should not constitute any limitation on the implementation process of the embodiments of the present invention.
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 (4)
1. The lithium ion battery SOC estimation method based on multi-algorithm fusion is characterized by comprising the following steps: the method specifically comprises the following steps:
carrying out a verification experiment aiming at a lithium ion battery, taking voltage, current and temperature data of a working condition cyclic discharge stage as a training input set, calculating the SOC of each moment of the working condition cyclic discharge stage based on an ampere-hour integration method according to the charge capacity of a constant-current constant-voltage charge stage, and taking the SOC as a training output set;
establishing a deep confidence network consisting of an input layer, an RBM layer, a BP layer and an output layer, and training the deep confidence network by using the training input set and the training output set; wherein a restricted Boltzmann machine training process is employed for the RBM layer; the method specifically comprises the following steps:
determining the network structures and the number of neurons of each layer of an input layer, an RBM layer, a BP layer and an output layer in the deep belief network; determining an activation function and an evaluation function, and performing optimization calculation on each RBM structure by using a random gradient descent method with root mean square error as a target function;
the limited Boltzmann machine training process specifically comprises the following steps:
firstly, setting a weight matrix W, a hidden layer bias vector b and a visible layer bias vector c of an RBM layer:
wherein m and n are the neuron serial numbers of the visible layer and the hidden layer respectively;
the following steps are performed in sequence:
1) assigning the input data x to the visible layer unit, and calculating the conditional probability value P of each hidden layer neuron:
wherein h is a hidden layer neuron; v is a visible layer neuron; j is 1,2, …, n;is Sigmoid function, e is exponential constant;
2) reconstructing the corresponding value of the hidden layer neuron by using one Gibbs sampling to generate [0,1 ]]Random number r of j And then:
3) reconstructing visible layer neurons with hidden layer neurons, v for each visible layer neuron i (i-1, 2, …, m) calculating its conditional probabilityThe value:
in the formula, superscript denotes reconstruction update;
4) reconstructing the corresponding value of the neuron in the visible layer by using the Gibbs sampling once again to generate a [0,1 ]]Random number s of i And then:
5) and reconstructing hidden layer neurons by using the visible layer neurons, and calculating the conditional probability value P of each hidden layer neuron:
6) updating a weight matrix W, a hidden layer bias vector b and a visible layer bias vector c of the RBM layer according to a predetermined learning rate lambda and iteration times:
W * =W+λ[P(h=1|v)v T -P(h * =1|v * )v *T ]
b * =b+λ(v-v * )
c * =c+λ[P(h=1|v)-P(h*=1|v*)];
determining the length of particles based on the number of neurons in the BP layer network structure, and calculating the optimal weight bias of the BP layer of the deep belief network by utilizing a particle swarm optimization algorithm;
step four, taking voltage, current and temperature as input and SOC as output, on one hand, establishing an SOC estimation model based on the adaptive extended Kalman filtering algorithm, and on the other hand, establishing an SOC estimation model based on the adaptive H ∞ An SOC estimation model of a filtering algorithm, and an SOC fusion estimator is established according to the estimation results of the two SOC estimation models and a weight distribution method; using the deep belief networkRespectively updating the initial values of the two SOC estimation models;
estimating the SOC of the lithium ion battery by using the SOC fusion estimator; and comparing the estimation result with the actual SOC value, and updating each estimation model periodically.
2. The method of claim 1, wherein: and acquiring voltage, current and temperature data in the working condition cycle discharge stage in the step one according to a Dynamic Stress Test (DST) working condition or a U.S. urban road cycle UDDS working condition.
3. The method of claim 1, wherein: the third step specifically comprises the following steps:
firstly, determining a weight matrix W of a BP layer to see a layer offset vector c:
wherein m and n are the neuron serial numbers of the visible layer and the hidden layer respectively;
the following steps are performed in sequence:
1) calculate the value for each hidden layer neuron:
h (l) =c+W (l) v
wherein h is a hidden layer neuron; v is a visible layer neuron; l is the layer number index of the neural network;
2) using Sigmoid function, a normalized hidden value σ (h) is obtained (l) And calculating the output value Y of the output layer:
Y (l) =σ(h) (l) =σ(c+W (l) v);
3) constructing a cost function based on the root mean square error as a judgment criterion for evaluating the output value of the output layer:
wherein E is the root mean square error in the training process, N is the number of samples,X i respectively representing the output value and the ideal output value of the output layer;
4) updating the weight matrix W and the visible layer bias vector c of the BP layer according to the determined particle length, particle scale and evolution times and in combination with the predetermined learning rate lambda and iteration times:
the updating process comprises the following steps:
in a search space defining a dimension D, n particles together form a population X ═ X 1 ,X 2 ,…,X n ) Wherein the position of the ith (i ═ 1,2, …, m) particle in the D-dimensional search space is X i =[x i1 ,x i2 ,…,x iD ] T ,x id Represents D (D ═ 1,2, …, D) dimensional coordinates; velocity of the ith particle is V i =[V i1 ,V i2 ,…,V iD ] T With an individual extremum of P i =[P i1 ,P i2 ,…,P iD ] T Global extremum of race is P g =[P g1 ,P g2 ,…,P gD ] T ;
Updating the particle moving speed respectively:
wherein, V id Is the moving speed of the particles; omega is the inertial weight; k is whenThe number of previous iterations; alpha is alpha 1 And alpha 2 Is an acceleration factor; r is 1 And r 2 Is distributed in [0,1 ]]A random number in between; x gd Is the extreme value of the particle position;
and updating the particle position:
4. lithium ion battery SOC estimation system based on multi-algorithm fusion, its characterized in that: performing the method of any of claims 1-3 to calculate a lithium ion battery SOC.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111438921.2A CN114167295B (en) | 2021-11-30 | 2021-11-30 | Lithium ion battery SOC estimation method and system based on multi-algorithm fusion |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111438921.2A CN114167295B (en) | 2021-11-30 | 2021-11-30 | Lithium ion battery SOC estimation method and system based on multi-algorithm fusion |
Publications (2)
Publication Number | Publication Date |
---|---|
CN114167295A CN114167295A (en) | 2022-03-11 |
CN114167295B true CN114167295B (en) | 2022-08-30 |
Family
ID=80481598
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111438921.2A Active CN114167295B (en) | 2021-11-30 | 2021-11-30 | Lithium ion battery SOC estimation method and system based on multi-algorithm fusion |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114167295B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116627028B (en) * | 2023-07-21 | 2023-09-29 | 阳谷新太平洋电缆有限公司 | Control method for crosslinked cable production line |
CN117236137B (en) * | 2023-11-01 | 2024-02-02 | 龙建路桥股份有限公司 | Winter continuous construction control system for deep tunnel in high and cold area |
Family Cites Families (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9366451B2 (en) * | 2010-12-24 | 2016-06-14 | Commonwealth Scientific And Industrial Research Organisation | System and method for the detection of faults in a multi-variable system utilizing both a model for normal operation and a model for faulty operation |
CN106679672A (en) * | 2017-01-15 | 2017-05-17 | 吉林大学 | AGV (Automatic Guided Vehicle) location algorithm based on DBN (Dynamic Bayesian Network) and Kalman filtering algorithm |
CN108414923A (en) * | 2018-02-05 | 2018-08-17 | 武汉大学 | A kind of analog-circuit fault diagnosis method based on the extraction of depth confidence network characterization |
CN108537337A (en) * | 2018-04-04 | 2018-09-14 | 中航锂电技术研究院有限公司 | Lithium ion battery SOC prediction techniques based on optimization depth belief network |
CN108896924B (en) * | 2018-07-09 | 2020-08-04 | 哈尔滨工业大学 | Lithium ion battery state-of-charge estimation method based on fusion of deep belief network and Kalman filtering |
CN111220920B (en) * | 2019-11-22 | 2023-04-25 | 国网浙江省电力有限公司台州供电公司 | Retired lithium ion battery state of charge calculation method based on H-infinity unscented Kalman filtering algorithm |
CN112327182B (en) * | 2020-08-02 | 2021-11-16 | 西北工业大学 | Adaptive H-infinity filtering SOC estimation method based on measurement value residual sequence |
-
2021
- 2021-11-30 CN CN202111438921.2A patent/CN114167295B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN114167295A (en) | 2022-03-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111638465B (en) | Lithium battery health state estimation method based on convolutional neural network and transfer learning | |
CN114167295B (en) | Lithium ion battery SOC estimation method and system based on multi-algorithm fusion | |
CN110888059B (en) | Charge state estimation algorithm based on improved random forest combined volume Kalman | |
CN110659722B (en) | Electric vehicle lithium ion battery health state estimation method based on AdaBoost-CBP neural network | |
CN113917337A (en) | Battery health state estimation method based on charging data and LSTM neural network | |
CN111047085B (en) | Hybrid vehicle working condition prediction method based on meta-learning | |
WO2022077767A1 (en) | Traffic flow prediction method and apparatus, computer device, and readable storage medium | |
CN110888058A (en) | Algorithm based on power battery SOC and SOH joint estimation | |
CN111310965A (en) | Aircraft track prediction method based on LSTM network | |
CN113253116A (en) | Lithium ion battery state of charge estimation method and storage medium | |
CN113064093A (en) | Energy storage battery state of charge and state of health joint estimation method and system | |
CN114184962B (en) | Multi-algorithm fusion lithium ion battery SOC and SOH joint estimation method | |
CN112269134A (en) | Battery SOC and SOH joint estimation method based on deep learning | |
CN113702843B (en) | Lithium battery parameter identification and SOC estimation method based on suburb optimization algorithm | |
CN110133524B (en) | Battery charge state calculation method, device, server and medium | |
CN113189490B (en) | Lithium battery health state estimation method based on feature screening and Gaussian process regression | |
CN115856678A (en) | Lithium ion battery health state estimation method | |
CN114819068A (en) | Hybrid target track prediction method and system | |
CN115015760A (en) | Lithium battery health state evaluation method based on neural network and migration integrated learning | |
CN110765706B (en) | Aerofoil unsteady stall aerodynamic coefficient modeling method based on OHNGBM (1, 1) | |
CN116052254A (en) | Visual continuous emotion recognition method based on extended Kalman filtering neural network | |
CN116106761A (en) | Real-time lithium ion battery electric quantity estimation method based on typical correlation analysis | |
CN115219906A (en) | Multi-model fusion battery state of charge prediction method and system based on GA-PSO optimization | |
CN109253727B (en) | Positioning method based on improved iteration volume particle filter algorithm | |
CN113033104B (en) | Lithium battery state of charge estimation method based on graph convolution |
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 |