Battery's state of charge estimation apparatus

0Associated
Cases 
0Associated
Defendants 
0Accused
Products 
0Forward
Citations 
0
Petitions 
1
Assignment
First Claim
1. A battery'"'"'s state of charge estimation apparatus comprising:
 a charge and discharge current detection unit for detecting a charge current and a discharge current of a battery;
a terminal voltage detection unit for detecting a terminal voltage of the battery;
a first processing unit for providing an open circuit voltage method state of charge estimate, the first processing unit configured to estimate an open circuit voltage of the battery based on the charge current and the discharge current detected by the charge and discharge current detection unit and the terminal voltage detected by the terminal voltage detection unit, and configured to estimate, based on the open circuit voltage, the open circuit voltage method state of charge from an open circuit voltagestate of charge characteristic of the battery;
a second processing unit for providing a current integration method state of charge, the second processing unit configured to use as a current integration model, a discrete state space model of a spreading system that uses a state variable of the spreading system in consideration of a current fluctuation Δ
i expressed by the following equation;
1 Assignment
0 Petitions
Accused Products
Abstract
A battery'"'"'s state of charge estimation apparatus will be provided.
The battery'"'"'s state of charge estimation apparatus includes: a charge and discharge current detection unit; a terminal voltage detection unit; an open circuit voltage method state of charge estimation unit for estimating an open circuit voltage of the battery and an open circuit voltage method state of charge; a current integration method state of charge estimation unit for obtaining a current integration method state of charge as a state variable; and an error correction value calculation unit for calculating an error correction value for correcting the current integration method state of charge. The current integration method state of charge estimation unit corrects the current integration method state of charge by using the error correction value.
3 Citations
No References
THROUGHTIME RADIAL GRAPPA CALIBRATION  
Patent #
US 20110093233A1
Filed 10/21/2009

Current Assignee
Case Western Reserve University

Sponsoring Entity
Case Western Reserve University

Method and system for joint battery state and parameter estimation  
Patent #
US 20060111870A1
Filed 11/23/2004

Current Assignee
LG Chem Limited

Sponsoring Entity
LG Chem Limited

Simple optimal estimator for PbA state of charge  
Patent #
US 20050046388A1
Filed 08/28/2003

Current Assignee
GM Global Technology Operations LLC

Sponsoring Entity
GM Global Technology Operations Incorporated

2 Claims
 1. A battery'"'"'s state of charge estimation apparatus comprising:
a charge and discharge current detection unit for detecting a charge current and a discharge current of a battery; a terminal voltage detection unit for detecting a terminal voltage of the battery; a first processing unit for providing an open circuit voltage method state of charge estimate, the first processing unit configured to estimate an open circuit voltage of the battery based on the charge current and the discharge current detected by the charge and discharge current detection unit and the terminal voltage detected by the terminal voltage detection unit, and configured to estimate, based on the open circuit voltage, the open circuit voltage method state of charge from an open circuit voltagestate of charge characteristic of the battery; a second processing unit for providing a current integration method state of charge, the second processing unit configured to use as a current integration model, a discrete state space model of a spreading system that uses a state variable of the spreading system in consideration of a current fluctuation Δ
i expressed by the following equation; View Dependent Claims (2)
1 Specification
The present invention relates to a battery'"'"'s state of charge estimation apparatus for estimating a battery'"'"'s state of charge that is used for an electronic vehicle and the like.
In an electric vehicle and a hybrid electric vehicle, for example, in order to supply power to an electric motor for driving these vehicles, or in order to store electronic energy by charging power from the electric motor for using energy generated during braking as a generator or from a power source installed on the ground, a rechargeable battery (a secondary battery) is used.
In this case, in order to maintain an optimum battery status for a long time, it is necessary to monitor a battery status, a state of charge (SOC) in particular, at all times as battery management.
As conventional SOC detection methods, there are known a current integration method (also called as a Coulomb counting method or a book keeping method) and an open circuit voltage method. According to the current integration method, data about the voltage and the current of the battery therein/thereout are all recorded in a chronological order, and an electric charge at the present time is obtained by time integration of the current by using the data, and thus the SOC is obtained by using an initial value of the electric charge of the battery and a full charge capacity. According to the open circuit voltage method, an input current value and a terminal voltage value of the battery are input and, by using a battery equivalent circuit model, an open circuit voltage value serving as a model state quantity is sequentially estimated, thereby estimating the SOC from the open circuit voltage value.
There are advantages and disadvantages to each method. That is, the current integration method of the former, although being capable of estimating the SOC in a short time more accurately than the open circuit voltage method of the latter for estimating the SOC by using the open circuit voltage value, the current integration method requires monitoring at all times and, due to accumulation of errors over time, an accuracy deteriorates. On the other hand, although the open circuit voltage method of the latter does not require the monitoring at all times, due to a small fluctuation of the open circuit voltage associated with a change in the SOC, the open circuit voltage method is no better than the current integration method of the former in estimating a variation quantity of the SOC in a short time.
As such, there has conventionally known a method that, by correcting an estimation error of the SOC obtained by the above methods, attempts to improve the accuracy of the estimation of the SOC.
As one of conventional SOC estimating apparatuses employing such a method, there is known a SOC estimation apparatus including: a first electrical quantity calculating means for calculating a SOC from an open circuit voltage estimation value estimated by employing the open circuit voltage method using an adaptive digital filter based on a discharge current and a terminal voltage of the battery; a second electrical quantity calculating means for carrying out time integration of a charge current and the discharge current of the battery by employing the current integration method and, based on thus obtained integrated value, calculating a changing quantity of the electricity charged in the battery; and an offset quantity estimating means for estimating, from a difference between the changing quantities obtained by the first electrical quantity calculating means and the second electrical quantity calculating means, an offset quantity serving as a measured value error of a discharge current measuring instrument, wherein correction for the offset quantity is carried out to reduce the measured value error of the discharge current measuring instrument and to improve the accuracy of a measured current value, thereby improving the accuracy of the estimation of an internal state of the battery such as the SOC (for example, see Patent Document 1).
Also, as another conventional SOC estimation method, there is known a method including: a battery information obtainment step of measuring battery information (the current value, a voltage value, and temperature); a current correction step of calculating a current integrated value by carrying out integration of a corrected current value corrected by using the voltage value; an integrated capacity calculation step of calculating a current integrated value by carrying out integration of the corrected current value and also calculating a current integrated capacity reflecting the current integrated value and a battery charge and discharge efficiency; a correction determination step of determining, based on a battery forward voltage capacity calculated from the battery information, whether to correct the current integrated capacity; and an integration correction step of obtaining a remaining capacity of the battery with or without correcting the current integrated capacity according to the determination, thereby attempting to improve accuracy of the calculation of the SOC by correcting a measurement error of the battery information (for example, see Patent Document 2).
Patent Document 1: Japanese Patent Application LaidOpen Publication No. 2010203854
Patent Document 2: Japanese Patent Application LaidOpen Publication No. 2009250970
However, both of the inventions described above have the following problems.
That is, the conventional SOC estimation apparatus of the former compares between a changing quantity of electricity estimated by the current integration method and the changing quantity of electricity estimated by the open circuit voltage method using an adaptive digital filter and, from a difference therebetween, estimates an error value of the current detection unit (an offset quantity of the measured current value).
However, since each of the estimated values of the quantities of electricity and the like obtained by the above methods have different statistical characteristics, a simple comparison between the changing quantities of the electricity obtained by the above methods cannot avoid an error caused by such statistical characteristics. Therefore, an error of a measured value of the discharge current measuring instrument may not be estimated with high accuracy. Also, there is a fluctuation of the current within the battery due to uncertainty of chemical reaction therein. As a result, there is another problem that the accuracy of the estimation of the SOC is also lowered.
Also, in the conventional SOC estimation method of the latter, the characteristics of the discharge current measuring instrument changes according to time, condition and environment during the use and, further, variations between each discharge current measuring instrument cannot be avoided. Nonetheless, this conventional method estimates as described above assuming that all of the discharge current measuring instruments have identical characteristics. Therefore, this conventional method cannot handle an error of the measured value caused by a difference from actual characteristics of the discharge current measuring instrument, and thus is incapable of estimating the error of the measured value of the discharge current measuring instrument with high accuracy. Further, there is the fluctuation of the current within the battery due to the uncertainty of the chemical reaction therein. As a result, there is a problem that this conventional method also has a difficulty in estimating the SOC with high accuracy.
The present invention in view of the above problems aims to provide a SOC estimation apparatus that, regardless of the changes and variations caused by the characteristics of the charge and discharge current detection unit and the fluctuation of the current within the battery, is capable of estimating the SOC with high accuracy.
In order to achieve the above object, a battery'"'"'s state of charge estimation apparatus according to claim 1 of the present invention includes:
a charge and discharge current detection unit for detecting a charge current and a discharge current of a battery;
a terminal voltage detection unit for detecting a terminal voltage of the battery;
an open circuit voltage method state of charge estimation unit for estimating an open circuit voltage of the battery based on the charge current and the discharge current detected by the charge and discharge current detection unit and the terminal voltage detected by the terminal voltage detection unit, and estimating, based on the open circuit voltage, an open circuit voltage method state of charge from an open circuit voltagestate of charge characteristic of the battery;
a current integration method state of charge estimation unit, by using, in a current integration model, a current in consideration of fluctuation as an input and the open circuit voltage method state of charge and the charge current and the discharge current obtained as observed amounts as outputs, for obtaining a current integration method state of charge as a state variable; and an error correction value calculation unit, by comparing between the current integration method state of charge obtained by the current integration method state of charge estimation unit and the open circuit voltage method state of charge estimated by the open circuit voltage method state of charge estimation unit, for calculating an error correction value for correcting the current integration method state of charge, wherein the current integration method state of charge estimation unit corrects the current integration method state of charge by using the error correction value calculated by the error correction value calculation unit.
A battery'"'"'s state of charge estimation apparatus according to claim 2 is the state of charge estimation apparatus according to claim 1, wherein the error correction value calculation unit has Kalman filter and, by multiplying a difference between the current integration method state of charge and the open circuit voltage method state of charge by Kalman gain, obtains the error correction value.
The SOC estimation apparatus according to claim 1, regardless of changes and variations caused by characteristics of the charge and discharge current detection unit and a fluctuation of the current within the battery, is capable of estimating the SOC with high accuracy.
The SOC estimation apparatus according to claim 2, by using the Kalman filter, is capable of easily correcting the current integration method SOC.
Hereinafter, an embodiment of the present invention will be described in detail based on an example illustrated in the accompanying drawings.
A SOC estimation apparatus according to the present embodiment is mounted in a hybrid vehicle (HEV: Hybrid Electric Vehicle) having an internal combustion engine and an electric motor for driving. The SOC estimation apparatus is connected to a battery that supplies power for operating the electric motor for driving and an electronic device and estimates a SOC.
First, an entire configuration of the SOC estimation apparatus according to the present embodiment will be described with reference to
As illustrated in the figure, the SOC estimation apparatus connected to a battery B includes a charge and discharge current detection unit 1, a terminal voltage detection unit 2, a current integration method SOC estimation unit (a current integration method SOC estimation unit) 3, an open circuit voltage method SOC estimation unit (an open circuit voltage method SOC estimation unit) 4, a subtraction unit 5, and a Kalman filter (Kalman gain) 6.
The battery B is a rechargeable battery and, according to the present embodiment, may be a lithium ion battery, for example. Note that according to the present embodiment, as a matter of course, the battery B does not need to be the lithium ion battery but may be of another type such as a nickelmetal hydride battery and the like.
The charge and discharge current detection unit 1 detects a magnitude of a discharge current i when the power is supplied to an electric motor or the like (not shown) from the battery B. The charge and discharge current detection unit 1 also detects the magnitude of a charge current i when the electric motor is made to function as a generator during braking and a portion of braking energy is collected by the battery B, or when the battery B is charged from a power supply equipment installed on the ground. The charge and discharge current detection unit 1, by using, for example, a shunt resistor or the like, detects the charge and discharge current i flowing inside the battery B.
The charge and discharge current i that is detected is input, as an input signal, to both the current integration method SOC estimation unit 3 and the open circuit voltage method SOC estimation unit 4.
Note that, the charge and discharge current detection unit 1 of any type having various configurations may be appropriately employed.
The terminal voltage detection unit 2 detects the voltage between terminals of the battery B. A terminal voltage v detected by the terminal voltage detection unit 2 is input to the open circuit voltage method SOC estimation unit 4.
The terminal voltage detection unit 2 of any type having various configurations may be appropriately employed.
To the current integration method SOC estimation unit 3, the charge and discharge current i detected by the charge and discharge current detection unit 1 and an error correction value obtained by the Kalman filter 6 are input. The current integration method SOC estimation unit 3 uses a current taking the fluctuation into account in a current integration model as an input and thus obtains the current integration method SOC serving as a state variable. Also, the current integration method SOC estimating section 3 corrects the current integration method SOC based on the error correction value and thus obtains a current integration method state of charge SOCi.
Here, the current integration method state of charge SOC_{i }is a value including a true state of charge SOC_{true }having an error n_{i }added thereon.
The current integration method state of charge SOC_{i }is input to the subtracting unit 5.
Note that the above processing and calculation carried out by the current integration method SOC estimation unit 3 will be described in detail below.
The open circuit voltage method SOC estimation unit 4, based on the charge and discharge current i obtained from the charge and discharge current detection unit 1 and the terminal voltage v obtained from the terminal voltage detection unit 2, by using the Kalman filter (not illustrated) with a battery equivalent circuit model of the battery B illustrated in
The open circuit voltage method SOC estimation unit 4 stores relationship data: open circuit voltage of the battery B [V]−SOC [%], which is preliminarily obtained in an experiment and illustrated in
The open circuit voltage method state of charge SOC_{v }is input to the subtraction unit 5.
In the Kalman filter of the open circuit voltage method SOC estimation unit 4, to the battery equivalent circuit model of the battery B, an input the same as that to the actual battery B (such as charge and discharge current and battery temperature) is input. Then, the Kalman filter compares between outputs (terminal voltages) of the battery equivalent circuit model and the actual battery B. When there is a difference therebetween, the Kalman filter performs feedback by multiplying the difference by Kalman gain and corrects the values of a resistor and a capacitor forming the battery equivalent circuit model such that an error is minimized Through sequential repetition of this processing, the open circuit voltage OCV indicating a true internal state quantity is estimated.
On the other hand, the battery equivalent circuit model illustrated in
This circuit is made up of four parallel circuits connected in series to a resistance r_{0 }for setting DC components such as an electrolyte resistance of the battery B and an ohmic resistance of a connection, and the four parallel circuits include two parallel circuits representing fast response and respectively including charge transfer resistances r_{1 }and r_{2 }and capacitors c_{1 }and c_{2}, and two parallel circuits representing slow response and respectively including diffusion process resistances r_{3 }and r_{4 }and capacitors c_{3 }and c_{4}.
As illustrated in
The subtraction unit 5 subtracts, from the open circuit voltage method state of charge SOC, obtained by the open circuit voltage method SOC estimation unit 4, the current integration method state of charge SOC_{i }obtained by the current integration method SOC estimation unit 3, and thus obtains a difference between the errors (n_{v}−n_{i}).
This difference between the errors is input to the Kalman filter 6.
Kalman filter 6 multiplies a difference between the errors (n_{v}−n_{i}) input from the subtraction unit 5 by the Kalman gain, and thus calculates the error correction value. Then, the Kalman filter 6 inputs the error correction value to the current integration method SOC estimation unit 3.
Note that the Kalman filter 6 performs feedback by adjusting the Kalman gain in such a manner that the error n_{i }of the current integration method state of charge SOC_{i }becomes zero.
Also, the subtraction unit 5 and the Kalman filter 6 correspond to an error correction value calculation unit of the present invention.
Next, the processing carried out by the current integration method SOC estimation section 3 will be described below.
Here, correction of the calculation of the current integration method by using the state of charge SOC, estimated by the open circuit voltage method allows a simple configuration overall. A Kalman filter model that achieves such a configuration will be considered.
First, a current integration model such as Equation (1):
is considered.
Here, SOC, i, FCC, and Δt represent the SOC, the charge and discharge current, a full charge capacity, and a sampling period, respectively. Also, an index k represents discrete time.
In the foregoing Equation (1), a state variable x, an input u, and an output y are defined as follows:
[Equation 2]
x=SOC (2)
u=i (3)
y=[SOCi]^{Y} (4)
In the above equation, in particular, the output is separated into the state of charge SOC and the charge and discharge current i.
That is, both the state of charge SOC estimated by the open circuit voltage method and the charge and discharge current i detected by the charge and discharge current detection unit 1 are considered as observed quantities respectively having regular white sensor noise added thereon.
At this time, a discrete time state space model, providing that v and w represent process noise and sensor noise, respectively, is expressed by the following equation:
Here, the state variable x and the input u represent the SOC and the charge and discharge current, respectively.
Next, in order to consider the process noise associated with the current, this model is converted into a spreading system.
That is, provided that the state variable z of the spreading system is expressed by the following equation:
the discrete time state space model of the spreading system is expressed by the following equation:
Note that ζ represents the process noise associated with the current, i.e. a fluctuation of the current.
Application of the Kalman filter to the model including the fluctuation of the current in addition allows better estimation of the SOC.
Here, to describe the meaning of the process noise ζ associated with the current, a current flowing outside the battery and a current flowing inside the battery are not necessarily equal to each other.
Since complex electrochemical reaction is taken place in the battery and the current is biased locally within the battery, it is considered that there is apparent generation and disappearance of the charge.
Although it is difficult to express such a phenomenon by using a model, in this case it may be considered that, for example, the current flowing inside the battery is considered as one of the state variables having regular whiteness process noise (an average value: 0) added thereon.
That is, it may be formulated as the following equation:
[Equation 6]
i_{k+1}=i_{k}+ζ_{k} (10)
Since the current flowing inside the battery and the electrochemical reaction are closely related to each other, it fits reality to assume that there is process noise as expressed by Equation (8), rather than having the current as an input as expressed by Equation (5) and assuming that the process noise has no influence on the system.
Note that in actual design the process noise of the current will be determined as an adjustment parameter through trial and error.
Next, results of simulations for confirming the validity of the SOC estimating apparatus of the above embodiment will be described.
First, the simulations used a charge and discharge current and a terminal voltage obtained by inputting, to a battery model simulating a HEV battery with a full charge capacity of 6.5 Ah, a waveform of an actual driving current of the HEV.
Here, the battery model simulating the HEV battery is the equivalent circuit model illustrated in
Noted that the unit Ah of the full charge capacity is a unit of the electric charge quantity and satisfies the following equation: 1 Ah=3600 C (Coulomb).
The waveform of the charge and discharge current, the waveform of the terminal voltage, and the waveform of a true value of the SOC are illustrated in
Note that the HEV is characteristic in such a way that the SOC thereof by the internal combustion engine in addition to the charging by charging regenerative braking is controlled to be around 50%, which greatly differs from an EV that is charged by the regenerative braking alone during running.
In this simulations, the following three estimation methods are compared to one another. That is,
(a) Current integration method,
(b) Open circuit voltage method using the Kalman filter, and
(c) What is called a sensor fusion method using a combination of the above two methods of the present invention, i.e., the present embodiment are compared to one another.
Further, seven conditions described below are given. Condition 1 assumes an ideal condition, and all of the other conditions assume the worst values.
<Condition 1> There is no quantization error or other errors other than the sensor noise.
<Condition 2> The current sensor (the charge and discharge current detection unit) has an offset error of +5 A.
<Condition 3> The current sensor has the offset error of −5 A.
<Condition 4> The open circuit voltage has an initial value error of +0.2 V.
<Condition 5> The open circuit voltage has the initial value error of −0.2 V.
<Condition 6> The voltage sensor (the terminal voltage detection unit) has an offset error of +40 mV.
<Condition 7> The voltage sensor has the offset error of −40 mV.
Further, processing conforming to an actual HEV/EV sensor such as the quantization error and the sensor noise is added to the waveforms of the currents and the terminal voltage. For example, the current sensor is applied with the noise of an average value of 0 A and a variance of 0.5 A and quantized with a quantization width of 0.7 A. Also, the voltage sensor is applied with the noise of the average value of 0 V and the variance of 20 mV and quantized with the quantization width of 20 mV. A sampling period was set to 0.1 second.
Furthermore, four items are set: a variance Q of the process noise of the Kalman filter, a variance r of the sensor noise, an initial value of the estimated value of the state variable, and a covariance P_{0}.
Settings of the Kalman filter used by the open circuit voltage method are as follows:
[Equation 7]
Q=diag(0.1,10−7,10^{−7},10^{−13}) (11)
r=3·10^{−3} (12)
P_{0}=10^{5}I_{4×4} (13)
{circumflex over (x)}_{0}=[0.996.1·10^{−3}−1.2·10^{−2}6.1·10^{−3}]T (14)
Next, settings of the Kalman filter used by the sensor fusion method are as follows:
[Equation 8]
Q=diag(10^{−5},10^{−5}) (15)
r=3·10^{−2} (16)
P_{0}=10^{5}I_{2×2} (17)
{circumflex over (x)}_{0}=[SOC_{initial}0]^{T} (18)
Note that SOC_{initial }represents an initial value of a measured SOC. Accordingly, under the conditions 4 to 7, for example, a value including an error is given as the initial value.
Based on results of the simulations, the results of the estimation methods are compared as described below.
Results of the estimation of the SOC of the simulation under the condition 1 are illustrated in
Note that horizontal axes in
Also, horizontal axes in
In this case, it can be seen that the SOCs estimated by the sensor fusion method substantially converge on the true value.
Results of the estimation of the SOC of the simulation under the condition 2 are illustrated in
In
As can be seen from the results, a divergence of the SOC occurs in the current integration method, and the SOC is accompanied by noise while staying close to the true value in the open circuit voltage method. However, the sensor fusion method, by integrating these results, allows the estimation of the SOC with high accuracy.
Results of the estimation of the SOC of the simulation under the condition 6 are illustrated in
In
Under this condition, neither one of the methods, while estimating the SOC close to the true value, estimates with high accuracy compared with those under the conditions 1 and 2 described above.
That is, the offset error occurs in the current integration method. This is because, since the initial value is obtained by using the voltage sensor, the offset error of the voltage sensor has an impact thereon. Also, in the open circuit voltage method, similarly to other methods, the estimated SOC contains a great amount of noise. In the sensor fusion method, the offset error at the same level as that of the current integration method occurs after convergence.
However, as will be described below, in this case deterioration of the accuracy of the estimation remains at a level causing no problem in practical use.
In other words, a deterioration in the estimation error under the conditions 6 is caused by the open circuit voltage method. The impact of the offset error of the voltage sensor appears in the form of the offset error of the estimation of the SOC of the open circuit voltage method. Therefore, in combination with the current integration method causing the offset error, no correction will be made.
Under the conditions 1 and 2, since the open circuit voltage method generates the estimation value which contains the noise but has no offset error, the sensor fusion method may be used for correction. According to the present invention, i.e., the present embodiment, since it is assumed that the SOC estimated by the open circuit voltage method originally has the regular whiteness noise alone added thereon, the open circuit voltage method cannot deal with the offset error in addition.
However, as can be seen from the table in
Note that in the table in
It can be seen from the table that, even when the accuracy of the SOC of the current integration method is extremely deteriorated, the sensor fusion method may estimate the SOC with high accuracy.
Although the present invention has been illustrated with reference to the embodiment as described above, the present invention is not limited thereto but includes design modifications thereof without departing from the scope of the present invention.
For example, the SOC estimation apparatus according to the present invention is also applicable to, in addition to a hybrid electric vehicle, an electric vehicle and other products using the battery for estimating the SOC.
 B battery
 1 charge and discharge current detection unit
 2 terminal voltage detection unit
 3 current integration method SOC estimation unit
 4 open circuit voltage method SOC estimation unit
 5 subtraction unit (error correction value calculation unit)
 6 Kalman filter (error correction value calculation unit)