A method for estimating and correcting current of parallel branches of power batteries based on machine learning
Through the dual BP neural network, the branch current of the lithium-ion battery parallel system in new energy vehicles is estimated and corrected, which solves the problem of large current estimation error in the battery pack and improves the service life and safety of the battery pack.
Patent Information
- Application Number
- CN202210422955.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-04-21
- Publication Date
- 2024-12-06
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
In new energy vehicles, branch current estimation errors in the parallel system of lithium-ion batteries due to the differences in characteristics of each battery cell, which affects the service life and safety of the battery pack. Research on branch current estimation and correction in the prior art Fewer and unstable errors.
The branch current of the parallel battery system is estimated and corrected by using the dual BP neural network, and the branch current, branch voltage and state of charge under dynamic operating conditions are used to estimate the branch current, branch voltage and state of charge under dynamic operating conditions, and the error is adjusted through time series. Achieve correction.
It significantly reduces the estimation error of branch current and improves the service life and safety of the battery pack.
Smart Images

Figure CN114740365B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of battery technology, specifically the field of new energy vehicles and energy storage power battery technology, and in particular to a method for estimating and correcting parallel branch currents of power batteries based on machine learning. Background Art
[0002] Lithium-ion batteries are widely used in electric vehicles due to their high energy density, long cycle life and light weight. In the use of automotive lithium-ion batteries, hundreds or thousands of cells are usually used in series and parallel to meet the needs of high power output. However, due to the inevitable differences in the manufacturing process of each battery cell, the capacity, temperature, degree of aging and other characteristics of each battery cell will vary during use. These differences will cause different branch currents of parallel battery packs, which will in turn cause differences in the depth of charge and discharge of different battery cells. As time goes by, the differences between parallel batteries will further increase, thus affecting the service life of the battery pack and posing safety risks. Therefore, it is very important to estimate the branch current and make appropriate corrections to the estimated value.
[0003] At present, there are few studies on branch current estimation, and therefore even fewer studies on correcting the estimated value of branch current. Although neural networks are used in the prior art to realize branch current estimation of battery systems, there are still some shortcomings. For example, due to the sudden change of the main current, the branch current estimation error cannot be stably maintained at a low level, which affects the development of subsequent research. Summary of the invention
[0004] The purpose of the present invention is to provide a method for estimating and correcting the parallel branch current of a power battery based on machine learning. The branch current of the parallel battery system is estimated and corrected respectively by a dual BP neural network. The method used can significantly reduce the estimation error of the branch current.
[0005] To achieve the above object, the present invention adopts the following technical scheme:
[0006] A method for estimating and correcting current of parallel branches of a power battery based on machine learning comprises the following steps:
[0007] S1, obtaining the main circuit I, two branch currents I1, I2 and branch voltage V of the parallel battery pack under dynamic working conditions;
[0008] S2. Use the ampere-hour integration method to calculate the main circuit current to obtain the state of charge SOC of the parallel battery pack at each moment I ;
[0009] S3, training the first BP neural network;
[0010] S4. The branch current of the parallel battery pack is estimated by the first BP neural network to obtain the estimated value of the branch current of the parallel battery pack under dynamic conditions. and the estimated errors EOB1 and EOB2 of the dynamic conditions other than the target condition;
[0011] S5, training the second BP neural network;
[0012] S6. The error of the estimated current values of the two branches of the parallel battery pack under the target working condition is estimated by the second BP neural network to obtain the error of the estimated current values of the two branches under the target working condition. and
[0013] S7. Subtract the estimated error from the estimated value of the branch current under the target operating condition to obtain a corrected estimated value of the branch current.
[0014] S3 uses the following method to train the first BP neural network:
[0015] S3.1, set the parameters of BP neural network;
[0016] S3.2, the main current I, branch voltage V, and state of charge SOC I As input, the branch current data sets I1 and I2 are used as output, and the BP neural network is trained to obtain a trained BP neural network.
[0017] S5 uses the following method to train the second BP neural network:
[0018] S5.1. Set the parameters of BP neural network;
[0019] S5.2. The time series T, the main current I, and the estimated branch current As input, the branch current estimation errors EOB1 and EOB2 are used as output to train the BP neural network to obtain a trained BP neural network.
[0020] The expression of the ampere-hour integration method in S2 is:
[0021]
[0022] In the formula Indicates the SOC value of the parallel battery pack at the initial moment.
[0023] S4 is obtained by the following method EOB1 and EOB2;
[0024] The main current I, branch voltage V, and state of charge SOC IAs input, the trained first BP neural network is used to estimate the branch current and obtain the estimated branch current value under the corresponding working condition.
[0025] Branch current estimation error under corresponding working conditions
[0026] S6 is obtained by the following method:
[0027] The time series T, the main current I, and the estimated branch current As input, the error of the estimated branch current is estimated using the trained second BP neural network to obtain the error of the estimated branch current under the target working condition. and
[0028] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention estimates and corrects the branch current of the parallel battery system respectively through a dual BP neural network, and the method used can significantly reduce the estimation error of the branch current. BRIEF DESCRIPTION OF THE DRAWINGS
[0029] Attached Figure 1 It is a parallel battery pack model of the present invention;
[0030] Attached Figure 2 It is the structure of the BP neural network of the present invention;
[0031] Attached Figure 3 is a flow chart of branch current estimation and correction of the present invention;
[0032] Attached Figure 4 It is the correction result diagram of the present invention. DETAILED DESCRIPTION
[0033] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0034] The present invention discloses a method for correcting the estimated value of branch current in parallel of an energy storage battery system. Taking two single cells in parallel as an example, the branch current estimation and correction includes the following steps:
[0035] S1, obtaining the main circuit I, two branch currents I1, I2 and branch voltage V of the parallel battery pack under dynamic working conditions;
[0036] S2. Use the ampere-hour integration method to calculate the main circuit current to obtain the state of charge SOC of the parallel battery pack at each moment I , the expression of the ampere-hour integration method is:
[0037]
[0038] In the formula Indicates the SOC value of the parallel battery pack at the initial moment.
[0039] S3. Train the first BP neural network. Specifically, the following method is used to train the first BP neural network:
[0040] S3.1, set the parameters of BP neural network;
[0041] S3.2, the main current I, branch voltage V, and state of charge SOC I As input, the branch current data sets I1 and I2 are used as output, and the BP neural network is trained to obtain a trained BP neural network.
[0042] S4. The branch current of the parallel battery pack is estimated by the first BP neural network to obtain the estimated value of the branch current of the parallel battery pack under dynamic conditions. And the estimated errors EOB1 and EOB2 of the dynamic conditions other than the target condition are obtained by the following method: EOB1 and EOB2;
[0043] The main current I, branch voltage V, and state of charge SOC I As input, the trained first BP neural network is used to estimate the branch current and obtain the estimated branch current value under the corresponding working condition.
[0044] Branch current estimation error under corresponding working conditions
[0045] S5. Train the second BP neural network. Specifically, the following method is used to train the second BP neural network:
[0046] S5.1. Set the parameters of BP neural network;
[0047] S5.2. The time series T, the main current I, and the estimated branch current As input, the branch current estimation errors EOB1 and EOB2 are used as output to train the BP neural network to obtain a trained BP neural network.
[0048] S6. The error of the estimated current values of the two branches of the parallel battery pack under the target working condition is estimated by the second BP neural network to obtain the error of the estimated current values of the two branches under the target working condition. and The specific method is as follows:
[0049] The time series T, the main current I, and the estimated branch current As input, the error of the estimated branch current is estimated using the trained second BP neural network to obtain the error of the estimated branch current under the target working condition. and
[0050] S7. Subtract the estimated error from the estimated value of the branch current under the target operating condition to obtain a corrected estimated value of the branch current.
[0051] In order to more clearly understand the method of the present invention, the following is a detailed description with reference to the accompanying drawings:
[0052] like Figure 1 As shown, the parallel battery pack uses two lithium-ion batteries Cell1 with a nominal capacity of 22.53Ah and Cell2 with a nominal capacity of 27.08Ah. In the experiment, current sensors are connected in series to the main circuit and the two branches, and voltage sensors are connected in parallel at both ends of the branch to obtain the main circuit current of the parallel battery pack and the current and branch voltage of the two branches.
[0053] BP neural network uses BP algorithm, its structure is as shown in the following Figure 2 As shown in the figure, the structure of the BP neural network is divided into three parts: an input layer, one or more hidden layers, and an output layer. The number of nodes in different types of layers can be set as needed. i , x k , x m represents input features, f1, f2 represent activation functions. y1, y k ,y l Represents output parameters. i,j , w k,i Represents the weights between different nodes. 1,1 ,θ 1,2 ,θ 1,i ,θ 1,j ,θ 1,q ,θ 2,1 ,θ 2,k ,θ 2,lRepresents the threshold of different nodes. The following sets the parameters of the BP neural network for branch current estimation. The number of input layer nodes is 3, the number of hidden layer nodes is 10, the number of output layer nodes is 2, the activation function of the hidden layer is logsig, the activation function between the output layers is purelin, the training algorithm is trainlm, the number of network iterations is set to 1000, the network training accuracy is 0.000027, the learning rate is set to 0.05, and the training ends when the error of the verification sample does not decrease for 20 consecutive times.
[0054] Figure 3 Shown is the branch current estimation and correction process.
[0055] S1. First, conduct offline tests under DST (Dynamic Stress Test), UDDS (Urban Dynamometer Driving Schedule), HPPC (Hybrid Pulse Power Characterization), and FUDS (Federal Urban Driving Schedule) conditions to obtain the main circuit I, two branch currents I1, I2, and branch voltage V of the parallel battery pack;
[0056] S2. Use the ampere-hour integration method to calculate the main circuit current to obtain the state of charge SOC of the parallel battery pack at each moment I ; The expression of the ampere-hour integration method is:
[0057]
[0058] In the formula Indicates the SOC value of the parallel battery pack at the initial moment.
[0059] S3, train the first BP neural network, set the parameters of the BP neural network; transform the main current I, branch voltage V, state of charge SOC I As input, the branch current data sets I1 and I2 are used as output, and the BP neural network is trained to obtain a trained first BP neural network;
[0060] S4. The branch currents of the parallel battery packs under four working conditions are estimated by the first BP neural network to obtain the estimated branch currents of the parallel battery packs under four dynamic working conditions. and the estimated errors EOB1 and EOB2 of the other dynamic working conditions except the target working condition; the main current I, branch voltage V, SOC IAs input, branch current I1, I2 as output, train the BP neural network to obtain the trained first BP neural network, and use the trained first BP neural network to estimate the branch current estimation value under DST, UDDS, HPPC, and FUDS working conditions. And the branch current estimation error under the corresponding working conditions
[0061] S5, train the second BP neural network, set the parameters of the BP neural network; transform the time series T, the main current I, the branch current estimation value As input, the branch current estimation errors EOB1 and EOB2 are used as outputs to train the BP neural network to obtain a trained second BP neural network.
[0062] S6. The error of the estimated current values of the two branches of the parallel battery pack under the target working condition is estimated by the second BP neural network to obtain the error of the estimated current values of the two branches under the target working condition. and Specifically, the time series T, the main current I, and the estimated branch current As input, the error of the estimated branch current is estimated using the trained second BP neural network to obtain the error of the estimated branch current under the target working condition. and
[0063] S7. Subtract the estimated error from the estimated value of the branch current under the target operating condition to obtain a corrected estimated value of the branch current.
[0064] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.
Claims
1. A method for estimating and correcting the current of parallel branches of a power battery based on machine learning, characterized in that: The following steps are involved: S1, obtaining the main circuit I, two branch currents I1, I2 and branch voltage V of the parallel battery pack under dynamic working conditions; S2. Use the ampere-hour integration method to calculate the main circuit current to obtain the state of charge SOC of the parallel battery pack at each moment I ; S3, training the first BP neural network; S4. The branch current of the parallel battery pack is estimated by the first BP neural network to obtain the estimated value of the branch current of the parallel battery pack under dynamic conditions. and the estimated errors EOB1 and EOB2 of the dynamic conditions other than the target condition; S5, training the second BP neural network; S6. The error of the estimated current values of the two branches of the parallel battery pack under the target working condition is estimated by the second BP neural network to obtain the error of the estimated current values of the two branches under the target working condition. and S7. Subtract the estimated error from the estimated value of the branch current under the target operating condition to obtain a corrected estimated value of the branch current.
2. A method for estimating and correcting parallel branch current of a power battery based on machine learning according to claim 1, characterized in that: S3 uses the following method to train the first BP neural network: S3.1, set the parameters of BP neural network; S3.2, the main current I, branch voltage V, and state of charge SOC I As input, and the branch current data sets I1 and I2 as output, the BP neural network is trained to obtain a trained first BP neural network.
3. The method for estimating and correcting the parallel branch current of a power battery based on machine learning according to claim 1, characterized in that: S5 uses the following method to train the second BP neural network: S5.
1. Set the parameters of BP neural network; S5.
2. The time series T, the main current I, and the estimated branch current As input, the branch current estimation errors EOB1 and EOB2 are used as output to train the BP neural network to obtain a trained BP neural network.
4. The method for estimating and correcting the parallel branch current of a power battery based on machine learning according to claim 1, characterized in that: The expression of the ampere-hour integration method in S2 is: In the formula Indicates the SOC value of the parallel battery pack at the initial moment.
5. The method for estimating and correcting the parallel branch current of a power battery based on machine learning according to claim 1, characterized in that: S4 is obtained by the following method EOB1 and EOB2; The main current I, branch voltage V, and state of charge SOC I As input, the trained first BP neural network is used to estimate the branch current and obtain the estimated branch current value under the corresponding working condition. Branch current estimation error under corresponding working conditions 6. The method for estimating and correcting current of a parallel branch circuit of a power battery based on machine learning according to claim 1, characterized in that: S6 is obtained by the following method: The time series T, the main current I, and the estimated branch current As input, the error of the estimated branch current is estimated using the trained second BP neural network to obtain the error of the estimated branch current under the target working condition. and
Citation Information
Patent Citations
Low-power-consumption acquisition system life evaluation method based on accelerated life test
CN103837772A
Power battery external short circuit fault diagnosing and temperature rise prediction method and system based on BP neural networks
CN106526493A
Fusion method for improving state-of-charge estimation precision of lithium ion battery
CN112083331A
Lithium ion battery charge state estimation method and system
CN114280490A
Method of calculating state variables of secondary battery and apparatus for estimating state variables of secondary battery
US20080234956A1
