CN110658459B - Lithium ion battery state of charge estimation method based on bidirectional cyclic neural network - Google Patents

Lithium ion battery state of charge estimation method based on bidirectional cyclic neural network Download PDF

Info

Publication number
CN110658459B
CN110658459B CN201910862338.0A CN201910862338A CN110658459B CN 110658459 B CN110658459 B CN 110658459B CN 201910862338 A CN201910862338 A CN 201910862338A CN 110658459 B CN110658459 B CN 110658459B
Authority
CN
China
Prior art keywords
value
battery
data
neural network
bidirectional
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
CN201910862338.0A
Other languages
Chinese (zh)
Other versions
CN110658459A (en
Inventor
杨顺昆
何霍亮
边冲
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.)
Beihang University
Original Assignee
Beihang 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 Beihang University filed Critical Beihang University
Priority to CN201910862338.0A priority Critical patent/CN110658459B/en
Publication of CN110658459A publication Critical patent/CN110658459A/en
Application granted granted Critical
Publication of CN110658459B publication Critical patent/CN110658459B/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]
    • G01R31/367—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
    • G01R31/3842—Arrangements for monitoring battery or accumulator variables, e.g. SoC combining voltage and current measurements
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods

Landscapes

  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Molecular Biology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Secondary Cells (AREA)

Abstract

The invention discloses a lithium ion battery state of charge estimation method based on a bidirectional circulation neural network, which utilizes data generated by a lithium ion battery in real time and the trained bidirectional circulation neural network to obtain a real-time state of charge value of the lithium ion battery. The invention belongs to a data driving method, does not need tedious electrochemistry related knowledge, can effectively extract historical data of a lithium ion battery, models the discharge characteristic of the lithium ion battery to obtain accurate charge state estimation, can process a complex nonlinear system with a large amount of data, does not need information in the battery field, and only needs the historical data of the lithium ion battery.

Description

Lithium ion battery state of charge estimation method based on bidirectional cyclic neural network
Technical Field
The invention relates to the technical field of battery management systems and deep learning, in particular to a lithium ion battery state of charge estimation method based on a bidirectional cyclic neural network.
Background
Lithium ion batteries are currently the fastest growing, most promising battery technology. Compared with the traditional battery, the lithium ion battery has the advantages of light weight, quick charge, high energy density, low self-discharge rate, long service life and the like.
State of Charge (SOC), one of the key states for battery monitoring, is defined as the percentage of the battery's remaining capacity to its maximum capacity. The reliable SOC estimation can accurately judge the current state of the battery, prevent possible dangers and ensure the safe and stable operation of the battery. However, due to the non-linear and time-varying characteristics of the SOC of the lithium ion battery, the SOC value cannot be directly observed, and the battery discharge characteristics are easily affected by the aging of the battery, the temperature variation and other factors, so that the SOC estimation is challenging.
The battery SOC estimation method is mainly classified into two types, one is an SOC estimation method based on the characteristics of the battery itself, and the other is a data-driven SOC estimation method. The latter draws great attention in recent two years, and only historical data of battery charge and discharge are needed to learn the end-to-end mapping relation between the battery characteristics and the SOC value without depending on an electrochemical method in the field of traditional batteries. The method mainly takes a support vector machine and a neural network as main parts, particularly the neural network, and has strong data fitting capacity and can process data of large number of levels, so that the method has good effect in the field of battery SOC prediction.
For example, the simplest fully-connected neural network can grasp the nonlinear relation between output and input, the manually selected battery characteristics are used as the input of the fully-connected neural network, the battery SOC value is used as the output of the fully-connected neural network, the fully-connected neural network can well fit the relation, and the obtained effect exceeds that of a support vector machine. However, the fully-connected neural network simply grasps the relationship between input and output, and the effect of the fully-connected neural network in the field of battery SOC prediction is still to be improved.
Disclosure of Invention
In view of this, the present invention provides a lithium ion battery state of charge estimation method based on a bidirectional cyclic neural network, so as to realize accurate estimation of the lithium ion battery state of charge.
Therefore, the invention provides a lithium ion battery state of charge estimation method based on a bidirectional cyclic neural network, which comprises the following steps:
s1: acquiring a battery voltage value, a battery current value and a battery surface temperature value of the lithium ion battery at the current moment;
s2: performing data sampling processing, data standardization processing and data dimension change processing on the acquired data at the current moment;
s3: and inputting the processed data at the current moment into the trained bidirectional circulation neural network to obtain the charge state value of the lithium ion battery at the current moment.
In a possible implementation manner, in the above method for estimating a state of charge of a lithium ion battery provided by the present invention, the training process of the bidirectional recurrent neural network includes the following steps:
s11: manually selecting characteristics input to the bidirectional recurrent neural network, including a battery voltage value, a battery current value and a battery surface temperature value;
s12: under different working conditions, collecting a battery voltage value, a battery current value, a battery surface temperature value and a charge state value;
s13: carrying out data sampling processing on the acquired battery voltage value, battery current value, battery surface temperature value and state of charge value, and carrying out data standardization processing and data dimension change processing on the battery voltage value, battery current value and battery surface temperature value after the data sampling processing;
s14: initializing the bidirectional cyclic neural network, inputting data under the same working condition with the data to be measured in the processed battery voltage value, battery current value and battery surface temperature value into the initialized bidirectional cyclic neural network, training by utilizing a time-based back propagation algorithm, and continuously adjusting network hyper-parameters to obtain the trained bidirectional cyclic neural network.
In a possible implementation manner, in the lithium ion battery state of charge estimation method provided by the present invention, in step S13, data sampling processing is performed on the acquired battery voltage value, battery current value, battery surface temperature value, and state of charge value, and data normalization processing and data dimension change processing are performed on the battery voltage value, battery current value, and battery surface temperature value after the data sampling processing, which specifically includes the following steps:
s131: under different working conditions, sampling the battery voltage value, the battery current value, the battery surface temperature value and the state of charge value again, and setting the data interval to be 1s to generate a data point;
s132: standardizing the battery voltage value, the battery current value and the battery surface temperature value after resampling to ensure that the battery voltage value, the battery current value and the battery surface temperature value are all distributed in a [0,1] interval, wherein the formula of the standardization is as follows:
Figure BDA0002200172290000031
wherein D represents any one of a battery voltage value, a battery current value and a battery surface temperature valuetData representing time t, DminRepresents the smallest data point, DmaxRepresents the largest data point;
s133: dimension change processing is carried out on the battery voltage value, the battery current value and the battery surface temperature value after the standardization processing, and the battery voltage value, the battery current value and the battery surface temperature value at each time point in the data after the standardization processing are connected as a vector [ V, I, T]Connecting the data points of k time steps as a sample input data [ V ] of the bidirectional cyclic neural networkt,It,Tt],[Vt+1,It+1,Tt+1],……,[Vt+k-1,It+k-1,Tt+k-1]]Finally obtaining all sample input data [ sample number, time step, characteristic number]。
In a possible implementation manner, in the method for estimating the state of charge of the lithium ion battery provided by the present invention, in step S14, the bidirectional cyclic neural network is initialized, data of the processed battery voltage value, battery current value, and battery surface temperature value, which is under the same working condition as the data to be measured, is input to the initialized bidirectional cyclic neural network, the training is performed by using a time-based back propagation algorithm, and the network hyper-parameter is continuously adjusted to obtain the trained bidirectional cyclic neural network, which specifically includes the following steps:
s141: initializing each parameter value of the bidirectional cyclic neural network, and randomly setting each parameter value as any numerical value in an interval of [0,1 ];
s142: inputting all sample input data [ sample number, time step and characteristic number ] into the bidirectional circulation neural network, calculating a predicted value of the SOC at the current moment through forward propagation of the bidirectional circulation neural network, calculating the distance between the predicted value of the SOC at the current moment and the actual value of the SOC at the current moment, and solving the distance between the predicted value of the SOC and the actual value of the SOC of all samples, wherein the calculation formula is as follows:
Figure BDA0002200172290000041
where y represents the true value of the state of charge,
Figure BDA0002200172290000042
representing a predicted value of the state of charge, m representing the number of samples, and i representing the serial number of the samples;
s143: updating each parameter value of the bidirectional cyclic neural network by using a gradient descent and back propagation algorithm;
and repeating the step S142 to the step S143 until the bidirectional circulation neural network is converged, and finishing the training.
In a possible implementation manner, in the method for estimating the state of charge of the lithium ion battery provided by the present invention, the step S2 of performing data sampling processing, data normalization processing, and data dimension change processing on the acquired data at the current time specifically includes the following steps:
s21: sampling the battery voltage value, the battery current value and the battery surface temperature value again, and setting the data interval to be 1s to generate a data point;
s22: carrying out standardization processing on the sampled data to enable the battery voltage value, the battery current value and the battery surface temperature value to be distributed in a [0,1] interval, wherein the formula of the standardization processing is as follows:
Figure BDA0002200172290000043
wherein D represents any one of a battery voltage value, a battery current value and a battery surface temperature valuetIndicates the time tData of (D)minRepresents the smallest data point, DmaxRepresents the largest data point;
s23: performing dimension change processing on the normalized data, and connecting the battery voltage value, the battery current value and the battery surface temperature value of each time point in the normalized data into a vector [ V, I, T ]]Connecting the data points of k time steps as a sample input data [ V ] of the bidirectional cyclic neural networkt,It,Tt],[Vt+1,It+1,Tt+1],……,[Vt+k-1,It+k-1,Tt+k-1]]Finally obtaining all sample input data [ sample number, time step, characteristic number]。
According to the lithium ion battery state-of-charge estimation method provided by the invention, the real-time state-of-charge value of the lithium ion battery is obtained by utilizing the data generated by the lithium ion battery in real time and the trained bidirectional cyclic neural network model, the state-of-charge value can be estimated in real time after the training of the bidirectional cyclic neural network model is completed, the method is very convenient and fast, the bidirectional cyclic neural network can fully consider the characteristics of time sequence data, the effect of the method is more accurate than that of a unidirectional cyclic neural network by utilizing the data before the current result and the data after the current result, the method has great potential for estimating the state-of-charge value of the lithium ion battery, and is very suitable for being applied to the field of estimating the state-of-charge value of the lithium ion battery. The invention belongs to a data driving method, does not need tedious electrochemistry related knowledge, starts from data completely, can effectively extract information expressed by historical data of a lithium ion battery, models the discharge characteristic of the lithium ion battery, obtains an accurate charge state estimation result, can process a complex nonlinear system with a large amount of data, does not need information in the battery field, and only needs the historical data of the lithium ion battery.
Drawings
Fig. 1 is a flowchart of a lithium ion battery state of charge estimation method based on a bidirectional recurrent neural network according to the present invention;
fig. 2 is a schematic structural diagram of a bidirectional recurrent neural network in a lithium ion battery state of charge estimation method based on the bidirectional recurrent neural network provided by the invention;
fig. 3 is one of the flow charts of the bidirectional cyclic neural network training process in the lithium ion battery state of charge estimation method based on the bidirectional cyclic neural network according to the present invention;
fig. 4 is a second flowchart of a two-way recurrent neural network training process in the lithium ion battery state of charge estimation method based on a two-way recurrent neural network according to the present invention;
fig. 5 is a third flowchart of a bidirectional cyclic neural network training process in the lithium ion battery state of charge estimation method based on the bidirectional cyclic neural network according to the present invention;
fig. 6 is a second flowchart of a lithium ion battery state of charge estimation method based on a bidirectional recurrent neural network according to the present invention;
FIG. 7 is a graph showing the effect of the bidirectional recurrent neural network of the present invention applied to the US06 data set at 45 ℃;
FIG. 8 is a graph of the mean absolute error of the two-way recurrent neural network of the present invention applied to the US06 data set at 45 ℃;
FIG. 9 is a diagram showing the effect of the bi-directional recurrent neural network applied to the BJST data set at 45 deg.C in the present invention;
FIG. 10 is a graph of the mean absolute error of a BJST data set at 45 ℃ using a bi-directional recurrent neural network of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only illustrative and are not intended to limit the present invention.
The invention selects a data set of an INR 18650-20R type lithium ion battery, the data set is data acquired under three temperature conditions of 0 ℃, 25 ℃ and 45 ℃, loads of different simulated automobile driving states are respectively applied to the lithium ion battery under the same temperature condition, the loads comprise US06, FUDS, DST and BJDT, the detailed information of the lithium ion battery is shown in table 1, and the detailed information of the data set is shown in table 2.
TABLE 1 INR 18650-20R lithium ion batteries
Figure BDA0002200172290000061
TABLE 2 INR 18650-20R lithium ion Battery data set
Figure BDA0002200172290000071
Each data set comprises a complete charging process and a complete discharging process of the lithium ion battery, and two experiments are carried out under each data set, wherein one experiment is that a simulated load is applied from 80% of the capacity of the lithium ion battery to the end of discharging, the other experiment is that a simulated load is applied from 50% of the capacity of the battery to the end of discharging, the former is used as training data, and the latter is used as real-time data to test the effect of the invention.
The invention provides a lithium ion battery state of charge estimation method based on a bidirectional cyclic neural network, as shown in fig. 1, comprising the following steps:
s1: acquiring a battery voltage value, a battery current value and a battery surface temperature value of the lithium ion battery at the current moment;
s2: performing data sampling processing, data standardization processing and data dimension change processing on the acquired data at the current moment;
s3: and inputting the processed data at the current moment into the trained bidirectional circulation neural network to obtain the charge state value of the lithium ion battery at the current moment.
According to the lithium ion battery state-of-charge estimation method provided by the invention, the real-time state-of-charge value of the lithium ion battery is obtained by utilizing the data generated by the lithium ion battery in real time and the trained bidirectional cyclic neural network model, the state-of-charge value can be estimated in real time after the training of the bidirectional cyclic neural network model is completed, the method is very convenient and fast, the bidirectional cyclic neural network can fully consider the characteristics of time sequence data, the effect of the method is more accurate than that of a unidirectional cyclic neural network by utilizing the data before the current result and the data after the current result, the method has great potential for estimating the state-of-charge value of the lithium ion battery, and is very suitable for being applied to the field of estimating the state-of-charge value of the lithium ion battery. The invention belongs to a data driving method, does not need tedious electrochemistry related knowledge, starts from data completely, can effectively extract information expressed by historical data of a lithium ion battery, models the discharge characteristic of the lithium ion battery, obtains an accurate charge state estimation result, can process a complex nonlinear system with a large amount of data, does not need information in the battery field, and only needs the historical data of the lithium ion battery.
In specific implementation, in the lithium ion battery state of charge estimation method provided by the present invention, the specific structure of the bidirectional recurrent neural network may be: an input layer, a hidden layer and an output layer; wherein, the hidden layer adopts a bidirectional Long-Short Term Memory network (LSTM) layer, and a Full Connection layer (FC) is connected behind the LSTM layer; wherein, the bidirectional LSTM structure can be superimposed, that is, the bidirectional LSTM layer can include a plurality of bidirectional LSTM network structures interconnected, the specific structure is shown in fig. 2, X in fig. 2t、Xt+1…Xt+kTo input data, Yt、Yt+1…Yt+kTo output data. The specific parameter configuration of the bidirectional recurrent neural network is different due to data under different working conditions, and the invention is explained by taking the example that the time step is set to be 40-50, the number of hidden units is set to be 64, and a bidirectional LSTM structure is stacked into two layers.
In specific implementation, in the method for estimating the state of charge of the lithium ion battery provided by the present invention, the trained bidirectional recurrent neural network can be obtained by using the bidirectional recurrent neural network to train by using the historical data of the lithium ion battery, and specifically, the training process of the bidirectional recurrent neural network, as shown in fig. 3, may include the following steps:
s11: manually selecting characteristics input into the bidirectional cyclic neural network, wherein the characteristics comprise a battery voltage value, a battery current value and a battery surface temperature value;
s12: under different working conditions, collecting a battery voltage value, a battery current value, a battery surface temperature value and a charge state value;
specifically, the different working conditions may include conditions of different temperatures, different loads, and the like; the collection of the battery voltage value, the battery current value, the battery surface temperature value and the state of charge value needs to be respectively collected by using special equipment;
s13: carrying out data sampling processing on the acquired battery voltage value, battery current value, battery surface temperature value and state of charge value, and carrying out data standardization processing and data dimension change processing on the battery voltage value, battery current value and battery surface temperature value after the data sampling processing; thus, the processed data can be directly input into the bidirectional cyclic neural network for training;
s14: initializing a bidirectional cyclic neural network, inputting data under the same working condition with the data to be measured in the processed battery voltage value, battery current value and battery surface temperature value into the initialized bidirectional cyclic neural network, training by utilizing a time-based back propagation algorithm, and continuously adjusting network hyper-parameters to obtain the trained bidirectional cyclic neural network.
In a specific implementation, in the above method for estimating the state of charge of the lithium ion battery provided by the present invention, in step S13, data sampling processing is performed on the acquired battery voltage value, battery current value, battery surface temperature value, and state of charge value, and data normalization processing and data dimension change processing are performed on the battery voltage value, battery current value, and battery surface temperature value after the data sampling processing, as shown in fig. 4, the method specifically includes the following steps:
s131: under different working conditions, sampling the battery voltage value, the battery current value, the battery surface temperature value and the state of charge value again, and setting the data interval to be 1s to generate a data point;
s132: standardizing the battery voltage value, the battery current value and the battery surface temperature value after resampling to ensure that the battery voltage value, the battery current value and the battery surface temperature value are all distributed in a [0,1] interval, wherein the formula of the standardization is as follows:
Figure BDA0002200172290000091
wherein D represents any one of a battery voltage value, a battery current value and a battery surface temperature valuetData representing time t, DminRepresents the smallest data point, DmaxRepresents the largest data point;
specifically, the voltage value, the current value and the surface temperature value of the battery are distributed in a [0,1] interval, so that the training of a bidirectional circulation neural network is facilitated; meanwhile, a standardized reference value can be reserved so that the same standardized mode can be adopted in real-time estimation, and the consistency of data distribution can be ensured;
s133: dimension change processing is carried out on the battery voltage value, the battery current value and the battery surface temperature value after the standardization processing, and the battery voltage value, the battery current value and the battery surface temperature value at each time point in the data after the standardization processing are connected as a vector [ V, I, T]Connecting the data points of k time steps into a sample input data [ V ] of the bidirectional cyclic neural networkt,It,Tt],[Vt+1,It+1,Tt+1],……,[Vt+k-1,It+k-1,Tt+k-1]]Finally obtaining all sample input data [ sample number, time step, characteristic number]So that the two-way circulation neural network can be directly input for training.
In specific implementation, in the method for estimating the state of charge of the lithium ion battery provided by the present invention, step S14 is to initialize a bidirectional cyclic neural network, input data of the processed battery voltage value, battery current value, and battery surface temperature value, which is under the same working condition as the data to be measured, into the initialized bidirectional cyclic neural network, train the data by using a time-based back propagation algorithm, and continuously adjust the network hyper-parameters to obtain the trained bidirectional cyclic neural network, as shown in fig. 5, which specifically includes the following steps:
s141: initializing each parameter value of the bidirectional cyclic neural network, and randomly setting each parameter value as any numerical value in a [0,1] interval;
s142: inputting all sample input data [ sample number, time step and characteristic number ] into a bidirectional circulation neural network, calculating a predicted value of the state of charge at the current moment through forward propagation of the bidirectional circulation neural network, calculating the distance between the predicted value of the state of charge at the current moment and the actual value of the state of charge at the current moment, and solving the distance between the predicted value of the state of charge and the actual value of the state of charge of all samples, wherein the calculation formula is as follows:
Figure BDA0002200172290000101
where y represents the true value of the state of charge,
Figure BDA0002200172290000102
representing a predicted value of the state of charge, m representing the number of samples, and i representing the serial number of the samples;
specifically, calculating the distance between the predicted value of the current time SOC and the true value of the current time SOC, namely, making a difference and solving the square; the true value of the current time SOC is the resampled SOC value;
s143: updating each parameter value of the bidirectional cyclic neural network by using a gradient descent and back propagation algorithm;
and (5) repeating the step (S142) to the step (S143) until the bidirectional recurrent neural network converges, and finishing the training.
Preferably, in order to accelerate the training speed of the bidirectional cyclic neural network and improve the effect precision of the bidirectional cyclic neural network, an Adam optimization method and a mini-batch optimization method may be adopted, the size of the batch may be set to 128, and the training times on the whole training set may be set to 2000.
In a specific implementation, in the method for estimating a state of charge of a lithium ion battery provided by the present invention, in step S2, the data sampling process, the data normalization process, and the data dimension change process are performed on the acquired data at the current time, as shown in fig. 6, the method may specifically include the following steps:
s21: sampling the battery voltage value, the battery current value and the battery surface temperature value again, and setting the data interval to be 1s to generate a data point;
s22: carrying out standardization processing on the sampled data to enable the battery voltage value, the battery current value and the battery surface temperature value to be distributed in a [0,1] interval, wherein the formula of the standardization processing is as follows:
Figure BDA0002200172290000111
wherein D represents any one of a battery voltage value, a battery current value and a battery surface temperature valuetData representing time t, DminRepresents the smallest data point, DmaxRepresents the largest data point;
specifically, a standardized reference value reserved during network training can be used, so that the same standardized mode can be adopted for network training and real-time estimation, and the consistency of data distribution can be ensured;
s23: performing dimension change processing on the normalized data, and connecting the battery voltage value, the battery current value and the battery surface temperature value of each time point in the normalized data into a vector [ V, I, T ]]Connecting the data points of k time steps into a sample input data [ V ] of the bidirectional cyclic neural networkt,It,Tt],[Vt+1,It+1,Tt+1],……,[Vt+k-1,It+k-1,Tt+k-1]]Finally obtaining all sample input data [ sample number, time step, characteristic number]. In summary, the acquired data at the current moment is processed by using the same processing method as the historical data during network training, so that the distribution of the data input into the trained bidirectional cyclic neural network can be kept consistent, and the accuracy of the estimation of the state of charge of the lithium ion battery can be ensured.
In Table 3 are the effects of the four data sets US06, FUDS, DST and BJSDT at three temperatures of 0 deg.C, 25 deg.C and 45 deg.C, and the evaluation criterion is Mean Absolute Error (MAE).
TABLE 3 Effect of four data sets at three temperatures
Figure BDA0002200172290000121
As can be seen from table 3, the estimation method for the state of charge of the lithium ion battery provided by the present invention has an excellent effect, and under the conditions of 45 ℃ and 25 ℃, the MAE values of the four data sets US06, FUDS, DST and BJDST are all less than 1%, and especially under the condition of 45 ℃, the MAE values of the four data sets US06, FUDS, DST and BJDST are all the smallest, and the effect is the best. As shown in fig. 7 and 9, the predicted values of the SOC of the two data sets US06 and BJDST almost coincide with the true values, and as shown in fig. 8 and 10, the MAE values of the two data sets US06 and BJDST are both less than 1%, which illustrates that the above-mentioned estimation method of the state of charge of the lithium ion battery provided by the present invention has very accurate effect on the two data sets US06 and BJDST, which has reached a quite high standard in practical application, and proves the applicability of the bi-directional cyclic neural network in the field of predicting the SOC of the lithium ion battery.
According to the lithium ion battery state-of-charge estimation method provided by the invention, the real-time state-of-charge value of the lithium ion battery is obtained by utilizing the data generated by the lithium ion battery in real time and the trained bidirectional cyclic neural network model, the state-of-charge value can be estimated in real time after the training of the bidirectional cyclic neural network model is completed, the method is very convenient and fast, the bidirectional cyclic neural network can fully consider the characteristics of time sequence data, the effect of the method is more accurate than that of a unidirectional cyclic neural network by utilizing the data before the current result and the data after the current result, the method has great potential for estimating the state-of-charge value of the lithium ion battery, and is very suitable for being applied to the field of estimating the state-of-charge value of the lithium ion battery. The invention belongs to a data driving method, does not need tedious electrochemistry related knowledge, starts from data completely, can effectively extract information expressed by historical data of a lithium ion battery, models the discharge characteristic of the lithium ion battery, obtains an accurate charge state estimation result, can process a complex nonlinear system with a large amount of data, does not need information in the battery field, and only needs the historical data of the lithium ion battery.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.

Claims (1)

1. A lithium ion battery state of charge estimation method based on a bidirectional cyclic neural network is characterized by comprising the following steps:
s1: acquiring a battery voltage value, a battery current value and a battery surface temperature value of the lithium ion battery at the current moment;
s2: performing data sampling processing, data standardization processing and data dimension change processing on the acquired data at the current moment;
s3: inputting the processed data at the current moment into the trained bidirectional circulation neural network to obtain the charge state value of the lithium ion battery at the current moment;
the training process of the bidirectional cyclic neural network comprises the following steps:
s11: manually selecting characteristics input to the bidirectional recurrent neural network, including a battery voltage value, a battery current value and a battery surface temperature value;
s12: under different working conditions, collecting a battery voltage value, a battery current value, a battery surface temperature value and a charge state value;
s13: carrying out data sampling processing on the acquired battery voltage value, battery current value, battery surface temperature value and state of charge value, and carrying out data standardization processing and data dimension change processing on the battery voltage value, battery current value and battery surface temperature value after the data sampling processing;
s14: initializing the bidirectional cyclic neural network, inputting data under the same working condition with the data to be detected in the processed battery voltage value, battery current value and battery surface temperature value into the initialized bidirectional cyclic neural network, training by using a time-based back propagation algorithm, and continuously adjusting network parameters to obtain a trained bidirectional cyclic neural network;
step S14, initializing the bidirectional circulation neural network, inputting the data of the processed battery voltage value, battery current value and battery surface temperature value under the same working condition as the data to be tested into the initialized bidirectional circulation neural network, training by using a time-based back propagation algorithm, and continuously adjusting network parameters to obtain the trained bidirectional circulation neural network, which specifically comprises the following steps:
s141: initializing each parameter value of the bidirectional cyclic neural network, and randomly setting each parameter value as any numerical value in an interval of [0,1 ];
s142: inputting all sample input data [ sample number, time step and characteristic number ] into the bidirectional circulation neural network, calculating a predicted value of the SOC at the current moment through forward propagation of the bidirectional circulation neural network, calculating the distance between the predicted value of the SOC at the current moment and the actual value of the SOC at the current moment, and solving the distance between the predicted value of the SOC and the actual value of the SOC of all samples, wherein the calculation formula is as follows:
Figure FDA0003163561830000021
where y represents the true value of the state of charge,
Figure FDA0003163561830000022
representing a predicted value of the state of charge, m representing the number of samples, and i representing the serial number of the samples;
s143: updating each parameter value of the bidirectional cyclic neural network by using a gradient descent and back propagation algorithm; repeating the step S142 to the step S143 until the bidirectional recurrent neural network converges to finish the training;
step S13, performing data sampling processing on the acquired battery voltage value, battery current value, battery surface temperature value, and state of charge value, and performing data standardization processing and data dimension change processing on the battery voltage value, battery current value, and battery surface temperature value after the data sampling processing, specifically including the following steps:
s131: under different working conditions, sampling the battery voltage value, the battery current value, the battery surface temperature value and the state of charge value again, and setting the data interval to be 1s to generate a data point;
s132: standardizing the battery voltage value, the battery current value and the battery surface temperature value after resampling to ensure that the battery voltage value, the battery current value and the battery surface temperature value are all distributed in a [0,1] interval, wherein the formula of the standardization is as follows:
Figure FDA0003163561830000023
wherein D represents any one of a battery voltage value, a battery current value and a battery surface temperature value, Dt represents data at time t, Dmin represents a minimum data point, and Dmax represents a maximum data point;
s133: performing dimension change processing on the battery voltage value, the battery current value and the battery surface temperature value after the standardization processing, connecting the battery voltage value, the battery current value and the battery surface temperature value at each time point in the data after the standardization processing into a vector [ V, I, T ], connecting data points of k time steps into one sample input data [ [ Vt, It, Tt ], [ Vt +1, It +1, Tt +1], … …, [ Vt + k-1, It + k-1, Tt + k-1] ] of the bidirectional cyclic neural network, and finally obtaining all sample input data [ sample number, time step, characteristic number ];
in the step S2, the data sampling process, the data normalization process, and the data dimension changing process are performed on the acquired data at the current time, and the method specifically includes the following steps:
s21: sampling the battery voltage value, the battery current value and the battery surface temperature value again, and setting the data interval to be 1s to generate a data point;
s22: carrying out standardization processing on the sampled data to enable the battery voltage value, the battery current value and the battery surface temperature value to be distributed in a [0,1] interval, wherein the formula of the standardization processing is as follows:
Figure FDA0003163561830000031
wherein D represents any one of a battery voltage value, a battery current value and a battery surface temperature value, Dt represents data at time t, Dmin represents a minimum data point, and Dmax represents a maximum data point;
s23: performing dimension change processing on the normalized data, connecting a battery voltage value, a battery current value and a battery surface temperature value at each time point in the normalized data into a vector [ V, I, T ], connecting data points of k time steps into one sample input data [ [ Vt, It, Tt ], [ Vt +1, It +1, Tt +1], … …, [ Vt + k-1, It + k-1, Tt + k-1] ] of the bidirectional cyclic neural network, and finally obtaining all sample input data [ sample number, time step, characteristic number ];
the specific structure of the bidirectional circulation neural network is as follows: an input layer, a hidden layer and an output layer; wherein, the hidden layer adopts a bidirectional long-short term memory network layer and is connected with a full connection layer at the back; wherein, the bidirectional LSTM structure is overlapped, that is, the bidirectional LSTM layer comprises a plurality of bidirectional LSTM network structures which are connected with each other.
CN201910862338.0A 2019-09-12 2019-09-12 Lithium ion battery state of charge estimation method based on bidirectional cyclic neural network Active CN110658459B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910862338.0A CN110658459B (en) 2019-09-12 2019-09-12 Lithium ion battery state of charge estimation method based on bidirectional cyclic neural network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910862338.0A CN110658459B (en) 2019-09-12 2019-09-12 Lithium ion battery state of charge estimation method based on bidirectional cyclic neural network

Publications (2)

Publication Number Publication Date
CN110658459A CN110658459A (en) 2020-01-07
CN110658459B true CN110658459B (en) 2021-10-15

Family

ID=69037234

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910862338.0A Active CN110658459B (en) 2019-09-12 2019-09-12 Lithium ion battery state of charge estimation method based on bidirectional cyclic neural network

Country Status (1)

Country Link
CN (1) CN110658459B (en)

Families Citing this family (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111413622B (en) * 2020-04-03 2022-04-15 重庆大学 Lithium battery life prediction method based on stacking noise reduction automatic coding machine
CN111537888A (en) * 2020-05-09 2020-08-14 国网福建省电力有限公司莆田供电公司 Data-driven echelon battery SOC prediction method
CN112269134B (en) * 2020-09-10 2022-12-02 杭州电子科技大学 Battery SOC and SOH joint estimation method based on deep learning
CN113391225B (en) * 2021-05-19 2022-06-03 北京航空航天大学 Lithium battery state-of-charge estimation method considering capacity degradation
CN114089190B (en) * 2021-11-11 2023-08-22 深圳力高新能技术有限公司 Battery SOC estimation method based on neural network
CN114966436A (en) * 2022-01-06 2022-08-30 湖北理工学院 Lithium battery state of charge prediction method, device, equipment and readable storage medium
CN115097296A (en) * 2022-06-22 2022-09-23 云南电网有限责任公司电力科学研究院 Relay reliability evaluation method and device

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1890574A (en) * 2003-12-18 2007-01-03 株式会社Lg化学 Apparatus and method for estimating state of charge of battery using neural network
CN108519556A (en) * 2018-04-13 2018-09-11 重庆邮电大学 A kind of lithium ion battery SOC prediction techniques based on Recognition with Recurrent Neural Network
CN109143105A (en) * 2018-09-05 2019-01-04 上海海事大学 A kind of state-of-charge calculation method of lithium ion battery of electric automobile
CN109459699A (en) * 2018-12-25 2019-03-12 北京理工大学 A kind of lithium-ion-power cell SOC method of real-time
KR101965832B1 (en) * 2017-11-27 2019-04-05 (주) 페스코 Battery SOC estimation system and battery SOC estimation method using the same

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11171498B2 (en) * 2017-11-20 2021-11-09 The Trustees Of Columbia University In The City Of New York Neural-network state-of-charge estimation

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1890574A (en) * 2003-12-18 2007-01-03 株式会社Lg化学 Apparatus and method for estimating state of charge of battery using neural network
KR101965832B1 (en) * 2017-11-27 2019-04-05 (주) 페스코 Battery SOC estimation system and battery SOC estimation method using the same
CN108519556A (en) * 2018-04-13 2018-09-11 重庆邮电大学 A kind of lithium ion battery SOC prediction techniques based on Recognition with Recurrent Neural Network
CN109143105A (en) * 2018-09-05 2019-01-04 上海海事大学 A kind of state-of-charge calculation method of lithium ion battery of electric automobile
CN109459699A (en) * 2018-12-25 2019-03-12 北京理工大学 A kind of lithium-ion-power cell SOC method of real-time

Also Published As

Publication number Publication date
CN110658459A (en) 2020-01-07

Similar Documents

Publication Publication Date Title
CN110658459B (en) Lithium ion battery state of charge estimation method based on bidirectional cyclic neural network
Chang et al. Lithium-ion battery state of health estimation using the incremental capacity and wavelet neural networks with genetic algorithm
Wang et al. Lithium battery state-of-health estimation via differential thermal voltammetry with Gaussian process regression
Chen et al. State of health estimation for lithium-ion batteries based on fusion of autoregressive moving average model and elman neural network
CN107957562B (en) Online prediction method for residual life of lithium ion battery
Zhang et al. An on-line estimation of battery pack parameters and state-of-charge using dual filters based on pack model
CN110850298B (en) Lithium battery SOH estimation method and system based on data driving
CN111323719A (en) Method and system for online determination of health state of power battery pack of electric automobile
CN106405433A (en) Extended Kalman particle filtering based SOC (State Of Charge) estimation method and system
CN112067998A (en) Lithium ion battery state of charge estimation method based on deep neural network
CN113253116A (en) Lithium ion battery state of charge estimation method and storage medium
CN108872869B (en) Lithium ion battery degradation classification method based on BP neural network
CN111579993B (en) Lithium battery capacity online estimation method based on convolutional neural network
CN112986830A (en) Lithium battery capacity estimation method based on convolution time memory neural network
Wei et al. Lyapunov-based state of charge diagnosis and health prognosis for lithium-ion batteries
Li et al. A novel state estimation approach based on adaptive unscented Kalman filter for electric vehicles
CN112782594B (en) Method for estimating SOC (state of charge) of lithium battery by data-driven algorithm considering internal resistance
Li et al. CNN and transfer learning based online SOH estimation for lithium-ion battery
CN112083333A (en) Power battery pack state of charge estimation method based on machine learning model
CN116224085A (en) Lithium battery health state assessment method based on data driving
CN112946481A (en) Based on federation H∞Filtering sliding-mode observer lithium ion battery SOC estimation method and battery management system
CN206161820U (en) System based on extension kalman particle filter
CN110232432B (en) Lithium battery pack SOC prediction method based on artificial life model
CN112946480B (en) Lithium battery circuit model simplification method for improving SOC estimation real-time performance
Chen et al. A hybrid DNN-KF model for real-time SOC estimation of lithium-ion batteries under different ambient temperatures

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