CN112083346B - LSTM-based parallel battery pack internal current distribution estimation method and system - Google Patents

LSTM-based parallel battery pack internal current distribution estimation method and system Download PDF

Info

Publication number
CN112083346B
CN112083346B CN202010768177.1A CN202010768177A CN112083346B CN 112083346 B CN112083346 B CN 112083346B CN 202010768177 A CN202010768177 A CN 202010768177A CN 112083346 B CN112083346 B CN 112083346B
Authority
CN
China
Prior art keywords
current
battery pack
lstm
internal resistance
discharging
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
Application number
CN202010768177.1A
Other languages
Chinese (zh)
Other versions
CN112083346A (en
Inventor
崔纳新
崔忠瑞
杜艺
王春雨
张承慧
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shandong University
Original Assignee
Shandong University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shandong University filed Critical Shandong University
Priority to CN202010768177.1A priority Critical patent/CN112083346B/en
Publication of CN112083346A publication Critical patent/CN112083346A/en
Application granted granted Critical
Publication of CN112083346B publication Critical patent/CN112083346B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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]
    • G01R31/396—Acquisition or processing of data for testing or for monitoring individual cells or groups of cells within a battery
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
    • G01R31/367—Software therefor, e.g. for battery testing using modelling or look-up tables
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
    • G01R31/382—Arrangements for monitoring battery or accumulator variables, e.g. SoC
    • 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]
    • G01R31/389—Measuring internal impedance, internal conductance or related variables

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Secondary Cells (AREA)
  • Tests Of Electric Status Of Batteries (AREA)

Abstract

The invention provides an LSTM-based method and system for estimating current distribution in a parallel battery pack. The method for estimating the internal current distribution of the parallel battery pack based on the LSTM comprises the steps of obtaining terminal voltage, terminal current, SOC and direct-current internal resistance parameters of the parallel battery pack in the charging and discharging process and using the terminal voltage, the terminal current, the SOC and the direct-current internal resistance parameters as input data sets to carry out normalization packaging and slicing processing; and estimating the current distribution in the battery pack by the processed data set through an LSTM network operated at fixed time intervals.

Description

LSTM-based parallel battery pack internal current distribution estimation method and system
Technical Field
The invention belongs to the field of battery current estimation, and particularly relates to an LSTM-based parallel battery pack internal current distribution estimation method and system.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
The current Battery Management System (BMS) treats the entire parallel battery pack as a battery, can only collect terminal voltages and total currents of the entire parallel battery pack, and then performs corresponding management. Obviously, this method cannot acquire the current distribution condition in the parallel battery pack, and thus cannot effectively manage the single battery in the battery pack. If the current sensor is added to obtain the current of each battery, the cost is greatly increased, and the narrow space in the battery pack does not allow the method. Meanwhile, in the design aspect of the battery pack, in order to avoid current distribution unevenness and accelerated aging caused by battery inconsistency, a battery screening mode is generally adopted, that is, electric cores with consistent monomer parameters such as capacity and internal resistance are screened out to be assembled in parallel, but the problem of current distribution unevenness cannot be completely solved by the mode, because of the limitation of the production process, the difference among the single batteries is inevitable, even though the batteries in the same batch are used. Also, as the battery pack is put into service, the current distribution and battery differences within the pack cannot be discovered in time.
In order to meet the requirements of high-capacity battery applications, such as electric vehicles and energy storage systems, single batteries need to be connected in parallel before being used. In this case, the capacity of the battery after parallel connection is increased, and the battery is generally treated as a large-capacity battery in the specific application and battery management process. However, the inventor found that, since the performance of each single battery in the parallel battery pack is not completely consistent, the current distribution in the parallel battery pack is not uniform during the charging and discharging processes, that is, the charging and discharging currents of the single batteries are different. The current distribution is generally related to the internal resistance difference, the connection resistance, and the open-circuit voltage between the unit cells, and the larger the difference is, the more uneven the current distribution is. If the current distribution in a parallel battery pack is not uniform, the inconsistency among the single batteries is further increased along with the use of the whole battery pack, so that the early aging of the single batteries is caused, and even the safety problem of the battery is caused.
Disclosure of Invention
In order to solve the above problems, a first aspect of the present invention provides an LSTM-based method for estimating internal current distribution of a parallel battery pack, which estimates internal current distribution of the parallel battery pack by using information of the battery pack, including terminal voltage, total current, battery electrical quantity (SOC), and battery internal resistance, collected and calculated by an existing BMS, without increasing hardware cost, and can obtain current of each battery, thereby effectively managing each battery cell and timely predicting various faults in the battery pack.
In order to achieve the purpose, the invention adopts the following technical scheme:
an LSTM-based parallel battery pack internal current distribution estimation method comprises the following steps:
acquiring terminal voltage, terminal current, SOC and direct current internal resistance parameters of a parallel battery pack in the charging and discharging process, and performing normalized packing and slicing processing by using the terminal voltage, the terminal current, the SOC and the direct current internal resistance parameters as input data sets;
and estimating the current distribution in the battery pack by the processed data set through an LSTM network operated at fixed time intervals.
In order to solve the above problems, a second aspect of the present invention provides an LSTM-based internal current distribution estimation system for a parallel battery pack, which estimates internal current distribution of the parallel battery pack by using information of the battery pack, including terminal voltage, total current, battery electrical quantity (SOC), and battery internal resistance, collected and calculated by an existing BMS, without increasing hardware cost, and can obtain current of each battery, thereby effectively managing each battery cell and timely predicting various faults in the battery pack.
In order to achieve the purpose, the invention adopts the following technical scheme:
an LSTM-based parallel battery pack internal current distribution estimation system, comprising:
the parameter acquisition module is used for acquiring terminal voltage, terminal current, SOC (system on chip) and direct current internal resistance parameters of the parallel battery pack in the charging and discharging process;
the input data processing module is used for carrying out normalization, packaging and slicing processing on the input data group;
and the current distribution estimation module is used for estimating the current distribution in the battery pack by the LSTM network which operates the processed data set at fixed time intervals.
A third aspect of the invention provides a computer-readable storage medium.
A computer-readable storage medium, on which a computer program is stored, which when executed by a processor implements the steps in the LSTM-based parallel battery pack internal current distribution estimation method as described above.
A fourth aspect of the invention provides a computer apparatus.
A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the steps in the LSTM-based parallel battery internal current distribution estimation method as described above when executing the program.
Compared with the prior art, the invention has the beneficial effects that:
aiming at the problems that the current battery management can not obtain the current distribution in the parallel battery module and can not effectively manage the single battery, the invention estimates the current distribution in the battery by utilizing the parallel battery pack information which is acquired and calculated by the prior BMS and comprises terminal voltage, total current, battery electric quantity (SOC) and battery internal resistance on the premise of not increasing the hardware cost. Therefore, the current of each battery can be obtained, and then each single battery can be effectively managed, and various faults in the battery pack can be timely forecasted.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, are included to provide a further understanding of the invention, and are incorporated in and constitute a part of this specification, illustrate exemplary embodiments of the invention and together with the description serve to explain the invention and not to limit the invention.
FIG. 1 is a parallel cell current distribution model of an embodiment of the present invention;
FIG. 2 is an LSTM network structure for current distribution estimation according to an embodiment of the present invention;
FIG. 3 is a flow chart of current distribution prediction according to an embodiment of the present invention;
FIG. 4(a) is a stream discharge data set of an embodiment of the present invention;
FIG. 4(b) is a pulsed discharge data set of an embodiment of the present invention;
FIG. 4(c) is a UDDS operating condition discharge dataset of an embodiment of the present invention;
fig. 5(a) is a constant current discharge current distribution estimation result of the embodiment of the present invention;
fig. 5(b) is a result of estimating the distribution of the pulse discharge current 1 according to the embodiment of the present invention;
fig. 5(c) is a result of estimating the distribution of the pulse discharge current 2 according to the embodiment of the present invention;
FIG. 5(d) shows the result of estimating the distribution of UDDS discharge current 1 according to the embodiment of the present invention;
fig. 5(e) is a UDDS discharge current 2 distribution estimation result according to an embodiment of the present invention.
Detailed Description
The invention is further described with reference to the following figures and examples.
It is to be understood that the following detailed description is exemplary and is intended to provide further explanation of the invention as claimed. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments according to the invention. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
After parallel batteries are grouped, the current distribution in the group is mainly related to the direct current internal resistance, SOC, battery capacity and open-circuit voltage of the batteries. The direct current internal resistance of the battery directly determines the output current of each single battery in the group, and as the charging or discharging process is carried out, the open-circuit voltage of each battery is also different due to uneven current distribution, and the difference of the open-circuit voltage further causes the inconsistency of the current distribution. Therefore, for the estimation of the current distribution, sufficient accuracy can be achieved by using the internal resistance model. Because the number of the single batteries in the parallel battery pack is different according to actual requirements without loss of generality, the scheme takes the parallel connection of two batteries as an example for introduction.
Example one
In FIG. 1, two batteries are shown in parallel, where ItRepresenting the total current of the whole battery pack, the current of each battery branch is I1And I2，UOCVRepresents the open circuit voltage of the battery, which changes with the change of SOC, R1And R2Indicating the dc internal resistance of the battery. According to kirchhoff's law, the method comprises the following steps:
Ibat＝I1+I2
OCV1-I1R1＝OCV2-I2R2
total current IbatAnd terminal voltage UtIs data directly measurable by BMS, total internal resistance R of parallel batterytAnd the SOC of the parallel battery pack can be calculated by an ampere-hour integration method.
The scheme utilizes the terminal voltage U of the parallel battery packtTotal current IbatTotal internal resistance RtAnd the residual electric quantity SOC is obtained by predicting the current I of each battery branch through an LSTM network1And I2. The specific implementation process is introduced in three parts, namely the network structure design of the LSTM, the training of the LSTM network and the prediction of the branch current distribution.
The long-short term memory neural network LSTM is an improved algorithm for the recurrent neural network RNN. The method overcomes the problem of explosion or disappearance of the traditional RNN gradient, is suitable for processing and predicting important events with relatively long intervals and delays in a time sequence, and has a network structure as shown in figure 2.
The network consists of an input layer, an LSTM layer and an output layer. The network input is the terminal voltage, total current, SOC and internal resistance of the parallel module. The LSTM layer is provided with 32 cells, and the larger the number of cells, the better the fitting effect, but the more the calculation amount is increased. The time step is set to 60s, that is, the current distribution state of the current state is predicted by the data of the past 60 seconds. The output layer is two battery branch currents.
The LSTM network requires enough data sets to train to improve its predictive power. And collecting data by adopting a battery experimental test mode under a specific working condition. The data collected are as follows
Figure BDA0002615487550000061
Wherein, Ut，It，I1，I2The recording can be measured directly.
The SOC is obtained by adopting an ampere-hour integration method:
Figure BDA0002615487550000062
where SOC0 is an initial value, ItIs the module current.
Here, it should be noted that the SOC of the parallel battery pack may also be obtained by using kalman filtering, extended kalman filtering, or the like.
RtThe response of the terminal voltage to the total current is determined according to the charging and discharging process. The current is constant in the charging process, the polarization state in the battery is stable, and the internal resistance is obtained by adopting an overpotential method: rt＝(Ut-OCV)/It. And in the discharging process, obtaining the internal resistance by adopting a least square method.
Training data set test conditions: the test conditions used for training data set collection should include various scenarios for parallel battery application. The selection is carried out according to the charging working condition, the discharging working condition and the internal resistance change. The charging condition selects a constant current charging mode, and the charging current is the maximum charging current IchargeMaxOne fifth of IchargeMax(ii) a/5 increment; the discharge working conditions are divided into three types, namely constant current discharge, pulse discharge and UDDS (ultra dynamic metering Driving Schedule, FTP-72 test) working condition discharge; the internal resistance changes in a mode of external series resistance, and the series resistance is gradually increased from zero to the initial internal resistance of the battery and is used for simulating the internal resistance change of the battery caused by aging and temperature change. The FTP-72 loop simulates urban road conditions of 12.07km and comprises frequent parking situations. The maximum vehicle speed was 91.2km/h and the average vehicle speed was 31.5 km/h. The FTP-72 cycle includes two phases: the first stage is 505 seconds, 5.78km and the average speed is 41.2 km/h; the second stage is 864 seconds. The first cycle starts with a cold start。
The operating conditions are tabulated below:
Figure BDA0002615487550000071
Figure BDA0002615487550000081
wherein, IdischargeMaxIs the maximum discharge current;
Rinitis the initial internal resistance of the battery.
Network training: normalizing the data obtained by the test, and selecting Ut，It，SOC，RtAs training input data, I1，I2As training validation data, the root mean square error was chosen as a loss function. And continuously optimizing the network node weight through training.
As shown in fig. 3, the LSTM network after training can be used for current distribution prediction, as follows:
and (6) collecting data. Because the LSTM is suitable for processing the time series prediction problem, the scheme adopts a batch processing mode when predicting the current distribution, namely, when charging or discharging the parallel battery pack, data is recorded, and the LSTM network is operated at fixed time intervals for estimating the current distribution. The time interval may be set according to the data storage capability and the computing capability of the computing platform, where one charge or discharge cycle is selected as one processing cycle. Recording real-time information collected by the BMS along with the progress of the charging and discharging process, including terminal voltage UtAnd total current It。
And (6) data processing. Calculating SOC and R in the charging and discharging process according to the acquired voltage and current informationt。
SOC still employs ampere-hour integration, i.e.
Figure BDA0002615487550000082
RtIn two casesObtaining: charging and discharging.
The charging process current is invariable, and the polarization state in the battery is stable, adopts following mode to obtain the internal resistance:
Rt＝(Ut-OCV)/It
wherein OCV is the open circuit voltage of the cell, obtainable using a low current discharge or HPPC discharge test;
and (3) dynamically changing the current in the discharging process, and acquiring real-time internal resistance by adopting a recursive least square algorithm:
Figure BDA0002615487550000083
Figure BDA0002615487550000091
Figure BDA0002615487550000092
wherein, KkGain matrix, P, for an iterative process of least squareskFor covariance matrices in the iterative process, thetakIs the identified parameter. The recursive least square algorithm can obtain the current model parameters in real time through gradual iterative update, in the formula, k is the time of each recursive calculation,
Figure BDA0002615487550000093
the regression vector, which is calculated for each iteration, includes voltage information and current information of the battery model,
Figure BDA0002615487550000094
to represent
Figure BDA0002615487550000095
λ is a forgetting factor for ensuring tracking performance, ykRepresenting the model output, i.e. the terminal voltage of the battery; i denotes an identity matrix.
The first order RC or the second order RC equivalent circuit model can be selected for parameter identification, and the direct current internal resistance RtAnd extracting the identified parameter theta.
Thus, data of a charge-discharge period is obtained, and then input data for current distribution prediction is obtained through normalization processing: [ U ]t N It N SOCN Rt N]Where N is the data length, i.e. the data used for prediction is two-dimensional data of N x 4.
And (6) predicting. In order to obtain the current distribution at each moment and improve the operation speed of the LSTM network, the input data set is first packed and sliced. Setting the prediction history data length step _ size to be 60s, and the prediction target length to be the current time, namely obtaining 3-dimensional data used for LSTM network input, wherein the data size is (N-step _ size) step _ size 4. And inputting the data into a network to obtain an estimated value of the internal current in the whole charging and discharging process.
Through the steps, the method of the embodiment can estimate the current of each single battery branch in the parallel battery pack on the premise of not increasing the hardware cost of a current sensor and the like. Through the estimation of the current distribution, the single batteries inside the parallel batteries can be effectively monitored and managed, for example, connection fault detection, single battery capacity estimation, single battery fault detection in the battery pack, single battery aging estimation and the like are carried out according to the current distribution condition, and the method has great improvement compared with the current black box processing mode of taking the whole parallel battery pack as one battery. The scheme can be applied to real-time management, cloud data management and diagnosis of high-capacity battery packs such as electric vehicles and energy storage systems, and can also be used in a plurality of fields such as offline quality detection of battery pack production.
In order to verify the effectiveness of the invention, battery data is generated through a battery model and then used for training and predicting the LSTM network, and the verification process is as follows:
firstly, data for network training is collected, and data of constant current discharge, pulse discharge and UDDS working condition discharge are collected as training data sets respectively, as shown in fig. 4(a) -4 (c).
The parallel battery packs were discharged under constant current 4.4A, pulsed discharge, UDDS conditions, and the results of discharge current prediction are shown in fig. 5(a) -5 (e).
Example two
The present embodiment provides an LSTM-based system for estimating internal current distribution of parallel battery packs, including:
(1) and the parameter acquisition module is used for acquiring terminal voltage, terminal current, SOC (system on chip) and direct current internal resistance parameters of the parallel battery pack in the charging and discharging process.
Because the LSTM is suitable for processing the time series prediction problem, the scheme adopts a batch processing mode when predicting the current distribution, namely, when charging or discharging the parallel battery pack, data is recorded, and the LSTM network is operated at fixed time intervals for estimating the current distribution. The time interval may be set according to the data storage capability and the computing capability of the computing platform, where one charge or discharge cycle is selected as one processing cycle. Recording real-time information collected by the BMS along with the progress of the charging and discharging process, including terminal voltage UtAnd total current It。
(2) The input data processing module is used for carrying out normalization, packaging and slicing processing on the input data group;
calculating SOC and R in the charging and discharging process according to the acquired voltage and current informationt。
SOC still employs ampere-hour integration, i.e.
Figure BDA0002615487550000111
RtThere are two cases: charging and discharging.
The charging process current is invariable, and the polarization state in the battery is stable, adopts following mode to obtain the internal resistance:
Rt＝(Ut-OCV)/It
wherein OCV is the open circuit voltage of the cell, obtainable using a low current discharge or HPPC discharge test;
and (3) dynamically changing the current in the discharging process, and acquiring real-time internal resistance by adopting a recursive least square algorithm:
Figure BDA0002615487550000112
Figure BDA0002615487550000113
Figure BDA0002615487550000114
wherein, KkGain matrix, P, for an iterative process of least squareskFor covariance matrices in the iterative process, thetakIs the identified parameter.
The first order RC or the second order RC equivalent circuit model can be selected for parameter identification, and the direct current internal resistance RtAnd extracting the identified parameter theta.
Thus, data of a charge-discharge period is obtained, and then input data for current distribution prediction is obtained through normalization processing: [ U ]t N It N SOCN Rt N]Where N is the data length, i.e. the data used for prediction is two-dimensional data of N x 4.
(3) And the current distribution estimation module is used for estimating the current distribution in the battery pack by the LSTM network which operates the processed data set at fixed time intervals.
The scheme utilizes the terminal voltage U of the parallel battery packtTotal current IbatTotal internal resistance RtAnd the residual electric quantity SOC is obtained by predicting the current I of each battery branch through an LSTM network1And I2. The specific implementation process is introduced in three parts, namely the network structure design of the LSTM, the training of the LSTM network and the prediction of the branch current distribution.
The long-short term memory neural network LSTM is an improved algorithm for the recurrent neural network RNN. The method overcomes the problem of explosion or disappearance of the traditional RNN gradient, is suitable for processing and predicting important events with relatively long intervals and delays in a time sequence, and has a network structure as shown in figure 2.
The network consists of an input layer, an LSTM layer and an output layer. The network input is the terminal voltage, total current, SOC and internal resistance of the parallel module. The LSTM layer is provided with 32 cells, and the larger the number of cells, the better the fitting effect, but the more the calculation amount is increased. The time step is set to 60s, that is, the current distribution state of the current state is predicted by the data of the past 60 seconds. The output layer is two battery branch currents.
The LSTM network requires enough data sets to train to improve its predictive power. And collecting data by adopting a battery experimental test mode under a specific working condition. The data collected are as follows
Figure BDA0002615487550000121
Wherein, Ut，It，I1，I2The recording can be measured directly.
The SOC is obtained by adopting an ampere-hour integration method:
Figure BDA0002615487550000122
where SOC0 is an initial value, ItIs the module current.
Here, it should be noted that the SOC of the parallel battery pack may also be obtained by using kalman filtering, extended kalman filtering, or the like.
RtAnd obtaining the response of the terminal voltage to the total current in the charging and discharging process by adopting kirchhoff law. The current is constant in the charging process, the polarization state in the battery is stable, and the internal resistance is obtained by adopting an overpotential method: rt＝(Ut-OCV)/It。
Training data set test conditions: the test conditions used for training data set collection should include various scenarios for parallel battery application. The selection is carried out according to the charging working condition, the discharging working condition and the internal resistance change. The charging condition selects a constant current charging mode, and the charging current is the maximum charging current IchargeMaxOne fifth of IchargeMax(ii) a/5 increment; discharge of electricityThe working conditions are divided into three types, namely constant current discharge, pulse discharge and UDDS working condition discharge; the internal resistance changes in a mode of external series resistance, and the series resistance is gradually increased from zero to the initial internal resistance of the battery and is used for simulating the internal resistance change of the battery caused by aging and temperature change. The operating conditions are tabulated below:
Figure BDA0002615487550000131
wherein, IdischargeMaxIs the maximum discharge current;
Rinitis the initial internal resistance of the battery.
Network training: normalizing the data obtained by the test, and selecting Ut，It，SOC，RtAs training input data, I1，I2As training validation data, the root mean square error was chosen as a loss function. And continuously optimizing the network node weight through training.
In order to obtain the current distribution at each moment and improve the operation speed of the LSTM network, the input data set is first packed and sliced. Setting the prediction history data length step _ size to be 60s, and the prediction target length to be the current time, namely obtaining 3-dimensional data used for LSTM network input, wherein the data size is (N-step _ size) step _ size 4. And inputting the data into a network to obtain an estimated value of the internal current in the whole charging and discharging process.
Through the steps, the current of each single battery branch in the parallel battery pack can be estimated on the premise of not increasing the hardware cost of a current sensor and the like. Through the estimation of the current distribution, the single batteries inside the parallel batteries can be effectively monitored and managed, for example, connection fault detection, single battery capacity estimation, single battery fault detection in the battery pack, single battery aging estimation and the like are carried out according to the current distribution condition, and the method has great improvement compared with the current black box processing mode of taking the whole parallel battery pack as one battery. The scheme can be applied to real-time management, cloud data management and diagnosis of high-capacity battery packs such as electric vehicles and energy storage systems, and can also be used in a plurality of fields such as offline quality detection of battery pack production.
EXAMPLE III
The present embodiment provides a computer-readable storage medium, on which a computer program is stored, which when executed by a processor implements the steps in the LSTM-based parallel battery pack internal current distribution estimation method as described in the first embodiment above.
Example four
The present embodiment provides a computer device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and when the processor executes the program, the processor implements the steps in the LSTM-based parallel battery pack internal current distribution estimation method according to the first embodiment.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of a hardware embodiment, a software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), or the like.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (8)

1. An LSTM-based method for estimating the internal current distribution of parallel batteries is characterized by comprising the following steps:
acquiring terminal voltage, total current, SOC and direct current internal resistance parameters of a parallel battery pack in the charging and discharging process, and performing normalized packing and slicing processing as an input data set; calculating the SOC and the direct current internal resistance in the charge and discharge process according to the acquired terminal voltage and total current information, and using the terminal voltage, the total current, the direct current internal resistance and the SOC of the parallel battery pack as input variables; estimating the current distribution in the battery pack by the LSTM network which operates at fixed time intervals after processing the data set, wherein the fixed time intervals of the LSTM network operation are charging or discharging cycle periods, and a batch processing mode is adopted when predicting the current distribution, namely, when charging or discharging the parallel battery packs, recording the data, and operating the LSTM network at the fixed time intervals to estimate the current distribution; the network input is the terminal voltage, the total current, the SOC and the direct current internal resistance of the parallel battery pack, and the output layer is the branch current of the two batteries;
in the training of the LSTM network, the test working conditions comprise a charging working condition, a discharging working condition and direct current internal resistance change; the charging working condition selects a constant current charging mode, and the discharging working condition is divided into three working conditions of constant current discharging, pulse discharging and UDDS discharging; the UDDS is an abbreviation of an Urban Dynamometer Driving Schedule, the direct current internal resistance changes in an external series resistor mode, and the series resistor is gradually increased from zero to the initial direct current internal resistance of the battery pack to simulate the direct current internal resistance change of the battery pack caused by aging and temperature change.
2. The LSTM-based method for estimating internal current distribution of parallel battery packs according to claim 1, wherein the charging process current is constant and the polarization state in the battery is stable, and the overpotential method is used to obtain the internal dc resistance of the parallel battery packs.
3. The LSTM-based method for estimating internal current distribution of parallel battery packs according to claim 1, wherein during discharge, the dc internal resistance parameters of the parallel battery packs are obtained using a recursive least square algorithm.
4. An LSTM-based system for estimating internal current distribution of parallel connected battery packs, comprising:
the parameter acquisition module is used for acquiring terminal voltage, total current, SOC (system on chip) and direct current internal resistance parameters of the parallel battery pack in the charging and discharging process;
the current distribution estimation module is used for estimating the current distribution in the battery pack by the LSTM network which operates the processed data set at fixed time intervals; the network input is the terminal voltage, the total current, the SOC and the direct current internal resistance of the parallel battery pack, and the output layer is the branch current of the two batteries; in the training of the LSTM network, the test working conditions comprise a charging working condition, a discharging working condition and direct current internal resistance change; the charging working condition selects a constant current charging mode, and the discharging working condition is divided into three working conditions of constant current discharging, pulse discharging and UDDS discharging; the UDDS is an abbreviation of an Urban Dynamometer Driving Schedule, the direct current internal resistance changes in an external series resistor mode, and the series resistor is gradually increased from zero to the initial direct current internal resistance of the battery pack to simulate the direct current internal resistance change of the battery pack caused by aging and temperature change.
5. The LSTM-based system for estimating internal current distribution of parallel battery packs according to claim 4, wherein in the parameter obtaining module, the charging process current is constant and the polarization state in the battery is stable, and the overpotential method is used to obtain the internal dc resistance of the parallel battery packs.
6. The LSTM-based system for estimating internal current distribution of parallel battery packs according to claim 4, wherein in the parameter obtaining module, the dc internal resistance parameters of the parallel battery packs are obtained by using a recursive least square algorithm during discharging.
7. A computer readable storage medium, on which a computer program is stored, which when executed by a processor implements the steps in the LSTM-based parallel battery pack internal current distribution estimation method according to any one of claims 1 to 3.
8. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor when executing the program implements the steps in the LSTM-based parallel battery internal current distribution estimation method according to any of claims 1-3.
CN202010768177.1A 2020-08-03 2020-08-03 LSTM-based parallel battery pack internal current distribution estimation method and system Active CN112083346B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010768177.1A CN112083346B (en) 2020-08-03 2020-08-03 LSTM-based parallel battery pack internal current distribution estimation method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010768177.1A CN112083346B (en) 2020-08-03 2020-08-03 LSTM-based parallel battery pack internal current distribution estimation method and system

Publications (2)

Publication Number Publication Date
CN112083346A CN112083346A (en) 2020-12-15
CN112083346B true CN112083346B (en) 2021-11-09

Family

ID=73735764

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010768177.1A Active CN112083346B (en) 2020-08-03 2020-08-03 LSTM-based parallel battery pack internal current distribution estimation method and system

Country Status (1)

Country Link
CN (1) CN112083346B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113362912B (en) * 2021-04-29 2023-04-28 中南大学 Alumina concentration secondary simulation method, system and storage medium
CN113884905B (en) * 2021-11-02 2022-06-14 山东大学 Power battery state of charge estimation method and system based on deep learning
CN114624603A (en) * 2022-03-15 2022-06-14 哈尔滨工业大学（威海） Battery system branch current estimation method based on machine learning
CN117637044A (en) * 2023-10-07 2024-03-01 武汉理工大学 Battery voltage prediction method and device, electronic equipment and storage medium

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103935260A (en) * 2014-05-08 2014-07-23 山东大学 Battery managing method based on battery safety protection
CN107422272A (en) * 2017-07-07 2017-12-01 淮阴工学院 A kind of electric automobile power battery SOC intellectualized detection devices
CN107436409A (en) * 2017-07-07 2017-12-05 淮阴工学院 A kind of electric automobile power battery SOC intelligent predicting devices
CN110165248A (en) * 2019-05-27 2019-08-23 湖北工业大学 Fault-tolerant control method for air supply system of fuel cell engine

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103935260A (en) * 2014-05-08 2014-07-23 山东大学 Battery managing method based on battery safety protection
CN107422272A (en) * 2017-07-07 2017-12-01 淮阴工学院 A kind of electric automobile power battery SOC intellectualized detection devices
CN107436409A (en) * 2017-07-07 2017-12-05 淮阴工学院 A kind of electric automobile power battery SOC intelligent predicting devices
CN110165248A (en) * 2019-05-27 2019-08-23 湖北工业大学 Fault-tolerant control method for air supply system of fuel cell engine

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
Charging Current Estimation of Electric Vehicle DC Fast Chargers using Long Short-Term Memory Networks;Wenwen Zhou, et al.;《2019 IEEE Vehicle Power and Propulsion Conference (VPPC)》;20200109;第1-6页 *
并联电池组电流分布及寿命一致性演变规律研究;魏学哲 等;《机电一体化》;20181215;第24卷(第Z1期);第3-11页 *
并联锂离子电池组的模型化与电流分配;施宝昌 等;《计算机测量与控制》;20171025;第25卷(第10期);第189-193页 *
蓄电池可接受充电电流预测;王丽 等;《电源技术》;20120720;第36卷(第07期);第962-965页 *
面向电动汽车的电池管理系统关键技术研究;李维嘉;《中国优秀硕士学位论文全文数据库 工程科技II辑》;20160415(第04期);第C035-182页 *

Also Published As

Publication number Publication date
CN112083346A (en) 2020-12-15

Similar Documents

Publication Publication Date Title
CN112083346B (en) LSTM-based parallel battery pack internal current distribution estimation method and system
Farmann et al. A comprehensive review of on-board State-of-Available-Power prediction techniques for lithium-ion batteries in electric vehicles
Chiang et al. Online estimation of internal resistance and open-circuit voltage of lithium-ion batteries in electric vehicles
CN108333528B (en) SOC and SOT united state estimation method based on power battery electric-thermal coupling model
Fotouhi et al. Lithium–sulfur battery state-of-charge observability analysis and estimation
Wang et al. Probability based remaining capacity estimation using data-driven and neural network model
Xu et al. Improving the state of charge estimation of reused lithium-ion batteries by abating hysteresis using machine learning technique
CN106772064B (en) A kind of health state of lithium ion battery prediction technique and device
CN103472398B (en) Based on the electrokinetic cell SOC method of estimation of spreading kalman particle filter algorithm
Sassi et al. Comparative study of ANN/KF for on-board SOC estimation for vehicular applications
CN111448467B (en) Method and system for modeling and estimating battery capacity
CN105021994B (en) The method and apparatus of cell uniformity in a kind of detection battery pack
CN109001640B (en) Data processing method and device for power battery
CN110850298B (en) Lithium battery SOH estimation method and system based on data driving
CN112213643B (en) Method, system and equipment for predicting initial capacity and state of health of battery
CN107167743A (en) Charge state estimation method and device based on electric vehicle
CN111142027A (en) Lithium iron phosphate battery state-of-charge monitoring and early warning method based on neural network
Wang et al. Battery available power prediction of hybrid electric vehicle based on improved Dynamic Matrix Control algorithms
Zhu et al. The SOH estimation of LiFePO4 battery based on internal resistance with Grey Markov Chain
Guo et al. A data-model fusion method for online state of power estimation of lithium-ion batteries at high discharge rate in electric vehicles
Sheikhan et al. State of charge neural computational models for high energy density batteries in electric vehicles
CN115542186B (en) Method, device, equipment and medium for evaluating state and consistency of energy storage battery
CN113777510A (en) Lithium battery state of charge estimation method and device
Farag et al. Li-ion battery SOC estimation using non-linear estimation strategies based on equivalent circuit models
Wong et al. Li-ion batteries state-of-charge estimation using deep lstm at various battery specifications and discharge cycles

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