US20100280777A1

US20100280777A1 - Method for Measuring SOC of a Battery in a Battery Management System and the Apparatus Thereof

Info

Publication number: US20100280777A1
Application number: US12/812,184
Authority: US
Inventors: Shanshan Jin; Jae Hwan Lim; Jeonkeun Oh; Chonghun Han; Sungwoo Cho
Original assignee: SK Energy Co Ltd
Current assignee: SK Innovation Co Ltd
Priority date: 2008-01-11
Filing date: 2009-01-12
Publication date: 2010-11-04
Also published as: EP2233939A4; WO2009088272A3; KR101189150B1; US20100283471A1; KR20090077657A; JP5483236B2; CN101971043A; WO2009088272A2; EP2233939B1; ATE551612T1; US8548761B2; KR20090077656A; JP2011515652A; WO2009088271A3; CN101965522A; JP2011515651A; EP2233938A4; EP2233938A2; WO2009088271A2; EP2233939A2

Abstract

According to an embodiment of this invention, a method for measuring SOC (State Of Charge) of a battery comprises the steps of: obtaining current data, voltage data and temperature by measuring the current, the voltage and the temperature of a battery; calculating SOCi (State Of Charge based on current) by accumulating the current data; calculating open circuit voltage by using an equivalent circuit model which simply presents the current data, the voltage data and the battery through an electric circuit; calculating SOCv (State Of Charge based on voltage) by using the temperature data and the open circuit voltage; and choosing at least one of the SOCi and the SOCv as SOC of the battery by using the SOCi and the SOCv based on the judgment on the current state of the battery for a certain time interval.

Description

TECHNICAL FIELD

The present invention relates to a method for measuring SOC (State Of Charge) of a battery in a battery management system and an apparatus thereof, and more particularly, to a method for setting up SOCi (State Of Charge based on current) or SOCv (State Of Charge based on voltage) to SOC of a battery in a battery management system according to a desired condition using a simple equivalent circuit, and an apparatus thereof.

BACKGROUND ART

An automobile with an internal combustion engine using gasoline or heavy oil generally has a serious influence on the generation of pollution like atmospheric pollution. Therefore, in order to reduce the generation of pollution, there have been many efforts to develop a hybrid vehicle or an electric vehicle.
Recently, there has developed a high power secondary battery using a high energy density non-aqueous electrolyte. The high power secondary battery may be provided in plural and connected in series in order to form a high capacity secondary battery.
As described above, the high capacity secondary battery (hereinafter, called “battery”) is typically comprised of the plurality of secondary batteries connected in series. In case of the battery, particularly, an HEV battery, since a few or a few ten secondary batteries are alternately charged and discharged, there is a necessity of managing the battery to control the charging and discharging of the battery and maintain the battery in an appropriate operation state.
To this end, there is provided BMS (Battery Management System) for managing all the states of the battery. The BMS detects voltage, current, temperature or the like, estimates SOC through a calculating operation and controls the SOC so as to optimize the fuel consumption efficiency of a vehicle. In order to precisely control the SOC, it is necessary to exactly measure the SOC of the battery in the charging and discharging operations are carried out.
As a prior art, there is disclosed Korean Patent Application No. 2005-00611234 (filed on Jul. 7, 2005) entitled “Method for resetting SOC of secondary battery module”.
In order to precisely calculate the SOC of the battery, the above-mentioned prior art includes measuring a current value, a voltage value and a temperature vale of a battery module when turning on a switch, calculating initial SOC using the measured values, accumulating the current value, calculating actual SOC according to the accumulated current value, determining whether the battery module is in a no-load state, determining whether the actual SOC is within a setup range that can be measured by accumulating the current value if the battery module is in the no-load state, and calculating the SOC according to the voltage value by measuring the voltage value if the actual SOC is outside the setup range. However, the prior art does not disclose a method that applies a simple equivalent circuit to an actual battery and an apparatus thereof.
Generally, SOCi does not have errors in the short term, but, as shown in FIG. 1, there is a tendency that the errors are accumulated. Therefore, in case that the battery is operated for a long time, considerable error is occurred. Especially, the accumulative error is mostly generated when the charging or discharging of battery is completely achieved. This is caused by that the degree of precision is influenced by the errors occurred by reduction of SOC due to self-discharge, and omission of LBS digit of CPU for calculating the SOC. Further, since the precision degree of the SOC is largely dependent on a current measuring sensor, it is impossible to correct the errors when the sensor has a trouble.
However, as shown in FIG. 2, SOCv measures the SOC through an open circuit voltage. In this measuring method, it is possible to obtain very precise results when a current is not flowed. However, when the current is flowed, the precision degree of the SOCv is dependent on a charging and discharging pattern of a battery. Therefore, since the precision degree of the SOC is also dependent on the charging and discharging pattern, it is deteriorated. Furthermore, the charging and discharging pattern that deteriorates the precision degree of the SOCv is within a range that a typical battery is used. Thus, although only the SOCv is used, there is also a problem that has to accept the considerable error.

DISCLOSURE

Technical Problem

An object of the present invention is to provide a method for measuring SOC (State Of Charge) of a battery in a battery management system and an apparatus thereof, which uses a simple equivalent circuit model and an adaptive digital filter and thereby easily and precisely measure the SOC of the battery.
Another object of the present invention is to provide a method for measuring SOC (State Of Charge) of a battery in a battery management system and an apparatus thereof, which determines whether a low current state is maintained for a desired time period and then sets up SOCi (State Of Charge based on current) or SOCv (State Of Charge based on voltage) to SOC of the battery, thereby easily and precisely measuring the SOC of the battery.

Technical Solution

To achieve the object of the present invention, the present invention provides a method for measuring SOC of a battery, comprising obtaining current data, voltage data and temperature data by measuring current, voltage and temperature of the battery; calculating SOCi (State Of Charge based on current) by accumulating the current data; calculating OCV (Open Circuit Voltage) using an equivalent circuit model in which the current data, the voltage data and the battery are simply expressed by an electric circuit; calculating SOCv (State Of Charge based on voltage) using the temperature data and the OCV; and judging a current state of the battery for a desired period of time, and setting up the SOC of the battery using at least one of the SOCv and the SOCi.
Further, the present invention provides an apparatus for measuring SOC of a battery, comprising a battery information obtaining part that measures current, voltage and temperature of the battery and obtains current data, voltage data and temperature data; a current accumulating part that calculates SOCi by accumulating the current data; a OCV calculating part that calculates OCV using an equivalent circuit model in which the current data, the voltage data and the battery are simply expressed by an electric circuit; a SOCv estimating part that estimates SOCv using the temperature data and the OCV; and a SOC setting part that judges a current state of the battery for a desired period of time, and sets up the SOC of the battery using at least one of the SOCv and the SOCi.

ADVANTAGEOUS EFFECTS

The present invention easily and precisely measures the SOC of the battery by using the simple equivalent circuit model and the adaptive digital filter.
Further, the present invention determines whether the low current state is maintained for a desired time period and then sets up the SOC of the battery using at least one of the SOCi and the SOCv, thereby easily and precisely measuring the SOC of the battery.

DESCRIPTION OF DRAWINGS

The above and other objects, features and advantages of the present invention will become apparent from the following description of preferred embodiments given in conjunction with the accompanying drawings, in which:

FIG. 1 is a graph showing a case that SOC of a battery is set up by using conventional SOCi.

FIG. 2 is a schematic view showing a case that the SOC of the battery is set up by using conventional SOCv.

FIG. 3 is a block diagram of an apparatus for measuring SOC of a battery in accordance with an embodiment of the present invention.

FIG. 4 is a flow chart of a method for measuring the SOC of the battery in accordance with an embodiment of the present invention.

FIG. 5 is a graph shown a result of simulation in a case that offset of 1A occurs in accordance with an embodiment of the present invention.

FIG. 6 is a view showing an equivalent circuit model in accordance with an embodiment of the present invention.

FIG. 7 is a graph showing actual SOC and calculated BMS SOC in case of using a model providing an integration effect in accordance with an embodiment of the present invention.

FIG. 8 is a graph showing a compensation point in case of using the model providing the integration effect in accordance with an embodiment of the present invention.

FIG. 9 is a view showing an equivalent circuit model in accordance with another embodiment of the present invention.

FIG. 10 is a graph showing a proposed time criteria in accordance with an embodiment of the present invention.

FIG. 11 is a graph showing an error in case of performing a simulation with a time criteria of 20 seconds and the proposed time criteria in accordance with an embodiment of the present invention.

DETAILED DESCRIPTION OF MAIN ELEMENTS

- 100: current accumulating part
- 200: low pass filtering part
- 300: open circuit voltage calculating part
- 400: SOCv estimating part
- 500: SOC setting part

BEST MODE

Various terms used in the application are generally described in this field, but in a special case, some terms are optionally selected by the applicant. In this case, the meanings thereof are defined in the description of the present invention. Therefore, the invention should be understood with the meanings of the terms, but not names thereof.
Hereinafter, the embodiments of the present invention will be described in detail with reference to accompanying drawings.
FIG. 3 is a block diagram of an apparatus for measuring SOC of a battery in a battery management system in accordance with an embodiment of the present invention. Referring to FIG. 3, the apparatus for measuring SOC of a battery includes a battery information obtaining part (not shown), a current accumulating part 100, a low pass filtering part 200, an open circuit voltage calculating part 300, a SOCv estimating part 400 and a SOC setting part 500.
A process for calculating the SOC of a battery in the battery management system (hereinafter, called “BMS SOC”) includes six steps as follows:
First step: collection of current and voltage data
Second step: calculation of SOCi through current accumulation
Third step: low pass filtering
Fourth step: equivalent circuit model and adaptive digital filtering
Fifth step: calculation of SOCv through open circuit voltage and temperature
Sixth step: proper selection of SOC.
The battery information obtaining part carries out the first step. That is, the battery information obtaining part collects current data, voltage data, temperature data and the like from the battery management system (BMS). The collected current data is transferred to the current accumulating part 100. In the current accumulating part 100, the current data is accumulated and then added to SOC (SOC(k−1) in FIG. 3) calculated in the previous time interval, thereby calculating SOCi.
Further, the current data and the voltage data collected by the battery information obtaining part are transferred to the low pass filtering part 200. The low pass filtering part 200 filters the current data and the voltage data and then transfers them to the open circuit voltage calculating part 300. The open circuit voltage calculating part 300 calculates parameters used in the equivalent circuit model through the equivalent circuit model and an adaptive digital filtering, and then calculates open circuit voltage (OCV) using the parameters. The SOCv estimating part 400 estimates SOCv using the temperature data and the OCV, and then transfers the estimated SOCv to the SOC setting part 500. The SOC setting part 500 sets up the SOCi calculated in the current accumulating part 100 or the SOCv estimated in the SOCv estimating part 400 to the BMS SOC according to a predetermined criteria. The detailed process performed in each part of the SOC measuring apparatus will be described below with reference to FIGS. 5 to 10.
FIG. 4 is a flow chart of a method for measuring the SOC of the battery in accordance with an embodiment of the present invention. The SOC measuring method will be described using the SOC measuring apparatus in FIG. 3.
Referring to FIG. 4, the battery information obtaining part measures the current data, the voltage data, the temperature data and the like of a battery pack in real time from the BMS (S401). And the current accumulating part 100 accumulates the current data and calculates the SOCi (S402). Then, the low pass filtering part 200 filters the current data and the voltage data (S403).
The filtered current data and voltage data are transferred to the OCV calculating part 300, and the OCV calculating part 300 calculates the parameters used in the equivalent circuit model through the adaptive digital filtering (S404) and calculates the OCV Vo using the parameters (S405). And the SOCv estimating part 400 estimates the SOCv using the OCV (S406).
Then, the SOC setting part 500 determines whether a low current state is maintained. If the low current state is maintained (S407), the SOCv is set up to the BMS SOC (S408), and if the low current state is not maintained (S407), the SOCi is set up to the BMS SOC (S409). The SOC of a battery in the battery management system is calculated through the above mentioned process (S410). Hereinafter, each step in the BMS SOC measuring method will be described in detail.
A. First Step: Collection of Current Data, Voltage Data, Temperature Data and the Like
This step collects the current data, voltage data and the like from the BMS. In this step, the current data may be not measured precisely due to trouble of a current sensor. Particularly, in case that the current sensor does not measure precisely the current intensity, but measures only a rough value thereof, considerable error may occur in the current estimating process. However, in the SOC measuring method of the present invention, the error of SOCi is compensated with the SOCv. It was checked through an actual simulation whether the SOC measuring method of the invention exactly calculated the BMS SOC. As a result, since the error was gradually accumulated in the SOCi, but compensated with the SOCv, there was no problem in calculating the final BMS SOC. It was confirmed that an error between a calculated value and an actual value was within a target tolerance of 5.000% that was 1.502˜−4.170%. That is, although the current was inaccurately measured due to the trouble of the current sensor, there was not a large error in the SOC measuring method of the invention. Besides the trouble of the current sensor, other problems may occur. The current value may be offset owing to the trouble of the current sensor or trouble of CAN (controller Area Network), and then transferred.
FIG. 5 is a graph shown a result of simulation in a case that offset of 1A occurs in accordance with an embodiment of the present invention. As shown in FIG. 5, the error is accumulated in the SOCi. It is understood that the degree of accumulation is very large. The accumulation is caused by that 1A is offset and calculated. However, it is also understood that the SOCv compensation occurs appropriately and thus the error does not occur considerably.
According to the whole analysis, since the SOCv compensation does not occur at beginning and end parts of a pattern, in which charging and discharging operations occur, there may be generated a problem in the BMS SOC. However, if the SOCv compensation occurs after the charging and discharging operations, it is possible to secure reliability of the calculation. Also, in case that offset of −1A occurs, it is possible to secure the reliability thereof in the same way.
B. Second Step: Calculation of SOCi Through Current Accumulation
In this step, the current data collected in the first step is accumulated and then added to the SOC calculated in the previous time interval, thereby calculating the SOCi. The calculation is carried out by integrating the current over time. A calculated result is divided by a whole capacity, and then the rest capacity is expressed in percentage. This may be expressed by an equation 1 as followed:
$\begin{matrix} SOC (%) = \frac{Remaining Capacity (Ah)}{Nominal Capacity (Ah)} \cdot 100 {SOC}_{i} = \frac{\int i \partial t}{{Ah}_{nominal}} \cdot 100 & Equation 1 \end{matrix}$
In the SOC measuring method of the present invention, since the current is detected every second, the equation 1 may be expressed by an equation 2
$\begin{matrix} {SOC}_{i} (k) = SOC (k - 1) + \frac{1}{Q_{\max}} \cdot I (k) \cdot t & Equation 2 \end{matrix}$
That is, the calculation of SOCi in a step k is performed by accumulating the increased SOC to the SOC in a step k−1. The increased SOC is the current flowed in step k, multiplied by interval time t, and divided by the whole capacity Qmax.
C. Third Step: Low Pass Filtering
The current data and the voltage data collected in the first step are passed through a low pass filter. The present invention employs a third order low pass filter, and a filter constant f is 0.6. However, the present invention is not limited to the conditions, and may use other kind of filters and other filter constants. The filter used in the present invention may be expressed in the form of an equation 3 as follows:
gi(n)=f ² i+3(1−f)gi(n−1)+3(1−f)² gi(n−2)+(1−f)³ gi(n−3) Equation 3
In the SOC measuring method of the present invention, there are total six kinds of current data and voltage data that are needed in an equivalent circuit model. The current data is used as it is, and the voltage data uses a difference value from an initial value. A current value, a differential value of the current and a second differential value of the current form one set of the current data. A difference value between an initial voltage value and a present voltage value, a first differential value thereof and a second differential value thereof form one set of the voltage data. Reasons why the differential data is required and the difference value of voltage is required will be fully described in the description of the equivalent circuit model.
D. Fourth Step: Equivalent Circuit Model and Adaptive Digital Filtering
The equivalent circuit model may be embodied into two models according to the embodiments of the present invention. In other words, the equivalent circuit model may be embodied into a first equivalent circuit model and a second equivalent circuit model according to each embodiment of the present invention. Hereinafter, the first equivalent circuit model will be described with reference to FIGS. 6 to 8, and the second equivalent circuit model will be described with reference to FIG. 9.
1) First Equivalent Circuit Model
(1) First Equivalent Circuit Model
In this step, the current data and the voltage data collected in the third step are applied to a battery model so as to obtain the OCV (open circuit voltage). This is caused by that it is possible to obtain the SOCv through the OCV. As the battery model, there is a first-principle model that considers thermal behavior and electrochemical phenomenon in a battery. However, since excessive time and cost are required to develop the above-mentioned model, the battery model in the present invention is embodied by an equivalent circuit model which is simply expressed by an electric circuit.
A modeling object is a lithium polymer battery (LiPB), and a circuit model is comprised of a first model. FIG. 6 is a view showing an equivalent circuit model in accordance with an embodiment of the present invention. The SOC measuring method of the present invention uses the equivalent circuit model that is expressed by a simple electric circuit. Each element such as a resistor and a capacitor included in the equivalent circuit model has a meaning shown in Table 1 as follows:

TABLE 1

Element of equivalent circuit model

	Step	Process

	I	Current (charge: (+)/discharge (−))
	V	Terminal voltage
	V_o	Open circuit voltage (OCV)
	R₁	Lumped interfacial resistances
	R₂	Lumped series resistances
	C	Electric double layer capacitor

In FIG. 6, R₂is resistance in electrodes, and R₁and C are resistance and a capacitor that express an electric double layer generated at an interface between one electrode and the other electrode or a separator. Typically, a numerical value of each parameter is obtained through the first-principle model or an experiment. The equivalent circuit model of FIG. 6 may be expressed by an equation 4 as follows:
$\begin{matrix} Δ V = Δ V_{0} + (\frac{R_{1} + R_{2} + {CR}_{1} R_{2} s}{1 + {CR}_{1} s}) I & Equation 4 \end{matrix}$
In the equation 4, it can be understood that the OCV is obtained by obtaining the parameters corresponding to each element forming the equivalent circuit model. In other words, an object of the battery modeling according to the present invention is to obtain the OCV by obtaining each parameter and substituting the obtained parameter in the equation 4.
The equation 4 may be derived through a process as follows. In the equivalent circuit model of FIG. 6, the current may be expressed in the form of an equation 5 by virtue of Kirchhoff's law.
I+I ₂ +I ₃=0 Equation 5
Further, when a model is set up in consideration of values of the resistance and the capacitor in the entire circuit, it may be expressed by an equation 6.
$\begin{matrix} V = V_{0} + {IR}_{2} - I_{3} R_{1} V = V_{0} + {IR}_{2} - \frac{Q}{C} I_{2} = \frac{\partial Q}{\partial t} & Equation 6 \end{matrix}$
Herein, when the voltage and the OCV are expressed by difference from an initial value (t=0), it may be expressed by an equation 7.
ΔV=V(t)−V(0)
ΔV ₀ =V ₀(t)−V ₀(0) Equation 7
If the equation 7 is arranged in consideration of ΔV₀=V₀(t)−V₀(0), it may be expressed by an equation 8.
$\begin{matrix} Δ V = Δ V_{0} + {IR}_{2} - I_{3} R_{1} Δ V = Δ V_{0} + {IR}_{2} - \frac{Q}{C} & Equation 8 \end{matrix}$
If the equation 8 is differentiated over time and then arranged, it may be expressed by an equation 9.
$\begin{matrix} \frac{\partial (Δ V - Δ V_{0})}{\partial t} + \frac{1}{{CR}_{1}} (Δ V - Δ V_{0}) = \frac{R_{2}}{{CR}_{1}} I + R_{2} \frac{\partial I}{\partial l} + \frac{I}{C} & Equation 9 \end{matrix}$
Then, if it is converted using Laplace transform, it may be expressed by an equation 10.
$\begin{matrix} Δ V = Δ V_{0} + (\frac{R_{1} + R_{2} + {CR}_{1} R_{2} s}{1 + {CR}_{1} s}) I & Equation 10 \end{matrix}$
Herein, assuming that the quantity of change in the current is proportional to the quantity of change in the OCV, and a proportional constant is h, the following equation may be set up:
$Δ V_{0} = \frac{h}{s} I$
If the equation is substituted in the equation 10, it may be expressed by an equation 11
$\begin{matrix} Δ V = \frac{(R_{1} + R_{2}) s + {CR}_{1} hs + h + {CR}_{1} R_{2} s^{2}}{(1 + {CR}_{1} s) s} I & Equation 11 \end{matrix}$
Herein, each factor may be defined by an equation 12.
ΔV=V₁I=I₁
sΔV=V₂sI=I₂
s²ΔV=V₂s²I=I₃ Equation 12
If the defined factors are is substituted in the equation 11, it may be expressed by an equation 13
V ₂ =−CR ₁ V ₃ +CR ₁ R ₂ I ₃+(R ₁ +R ₂ +CR ₁ h)I ₂ +hI ₁ Equation 13
If the equation 13 is expressed in the form of a matrix, it may be expressed by an equation 14
$\begin{matrix} V_{2} = [\begin{matrix} V_{3} & I_{3} & I_{2} & I_{1} \end{matrix}] [\begin{matrix} - {CR}_{1} \\ {CR}_{1} R_{2} \\ R_{1} + R_{2} + {CR}_{1} h \\ h \end{matrix}] & Equation 14 \end{matrix}$
Herein, factors relevant to the current and voltage may be obtained through the current data and voltage data that are collected from the BMS and filtered in the third step. Each parameter R₁, R₂, C, h is obtained by substituting the obtained factors and using the adaptive digital filter. A method of using the adaptive digital filter will be described later. If the parameters related to each situation are obtained through the filter, they are substituted in an equation 15 that is a basic equation for calculating the OCV.
$\begin{matrix} Δ V_{0} + {CR}_{1} s Δ V_{0} = V_{1} + {CR}_{1} V_{2} - {CR}_{1} R_{2} I_{2} - (R_{1} + R_{2}) I_{1} Δ V_{0} = \frac{V_{1} + {CR}_{1} V_{2} - {CR}_{1} R_{2} I_{2} - (R_{1} + R_{2}) I_{1}}{1 + {CR}_{1} s} & Equation 15 \end{matrix}$
The OCV obtained by using the equation 15 is used for calculating the SOCv in a next step.
The SOC measuring method of the present invention uses the above-mentioned equivalent circuit model. However, if an equation derived from the model is further integrated by one step, an equation 16 may be obtained.
$\begin{matrix} Δ V = \frac{{CR}_{1} R_{2} s + (R_{1} + R_{2}) + h / s + {CR}_{1} h}{CR} I & Equation 16 \end{matrix}$
By dividing a denominator and a numerator of the original equation by s, it is possible to obtain integration effect. FIG. 7 shows actual SOC and calculated BMS SOC in case of using a model providing the integration effect, and FIG. 8 is a graph showing a compensation point in case of using the model providing the integration effect.
In case of using the equivalent circuit model of the present invention, the compensation occurs properly as a whole. If the model providing the integration effect is used, the compensation occurs further frequently. Also if the model providing the integration effect is used, noise is further generated, particularly, at a potion that the compensation occurs. This means that the data becomes unstable as a whole, when carrying out the integration. However, since the degree of unstable data is not high, it is possible to use the model providing the integration effect. Basically, it is the most preferable to use the equivalent circuit model of the present invention, but if necessary, the model providing the integration effect may be used.
(2) Adaptive Digital Filter
Like the equation 14, the equivalent circuit model may be expressed in the form of a matrix. In the equation 14, assuming that
$[\begin{matrix} V_{3} \\ I_{3} \\ I_{2} \\ I_{1} \end{matrix}]$
is w, and
$[\begin{matrix} - {CR}_{1} \\ {CR}_{1} R_{2} \\ R_{1} + R_{2} + {CR}_{1} h \\ h \end{matrix}]$
is θ, the equation 14 may be expressed by an equation 17.
V₂=w⁻¹θ Equation 17
In this matrix, w obtains through the third step in which the current data and the voltage data are passed through the low pass filter, and V₂also obtains through the same result. An object of the adaptive digital filter is to obtain the matrix θ through the two values and estimate the parameter values in real time through each element. Assuming that V₂passing through the low pass filter is gV₂, the equation 17 may be expressed by an equation 18.
gV₂=w⁻¹θ Equation 18
First of all, an initial value of the matrix θ is obtained by obtaining the parameter values in an initial state in which the current is not flowed and the voltage has an OCV value and then substituting the parameter values in the equation 18. A matrix at this time is indicated by θ_o. A square of the matrix is expressed by an equation 19.
P₀=θ₀ ² Equation 19
Herein, a matrix K required to continuously renew the matrix θ may be defined as an equation 20.
$\begin{matrix} K = \frac{P_{0} \cdot w}{R + w^{- 1} \cdot P_{0} \cdot w} & Equation 20 \end{matrix}$
wherein R is a value that is decided so as to prevent a denominator from being diverged to 0 by an initial value of gV₂, and the value is very small. The matrix may be arranged to an equation 21.
$\begin{matrix} K = \frac{P_{0} \cdot w}{R + w^{- 1} \cdot P_{0} \cdot w} = \frac{{gV}_{2} (n - 1) \cdot θ_{0}}{R + {{gV}_{2} (n - 1)}^{2}} & Equation 21 \end{matrix}$
wherein gV₂(n−1) is a just previous value of gV₂. The continuous renewal of the matrix θ is occurred by proportional relationship like an equation 22.
$\begin{matrix} \frac{θ}{{gV}_{2} (n)} = \frac{θ_{0}}{{gV}_{2} (n - 1)} & Equation 22 \end{matrix}$
The equation 22 may be arranged into an equation 23.
$\begin{matrix} θ = θ_{0} - [\frac{{gV}_{2} (n - 1) \cdot θ_{0}}{{{gV}_{2} (n - 1)}^{2}} \cdot {{gV}_{2} (n - 1) - {gV}_{2} (n)}] & Equation 23 \end{matrix}$
Since R is very small, the matrix K may be substituted. This is arranged into an equation 24.
θ=θ₀ −[K·{gV ₂(n−1)−gV ₂(n)}] Equation 24
If a relational expression is substituted in the equation 24, it may be expressed by an equation 25
θ=θ₀ −[K·{w ⁻¹·θ₀ −gV ₂(n)}] Equation 25
A value of θ in the equation 25 may be calculated through the current data, the voltage data and the previous matrix θ. Therefore, it is possible to continuously estimate each parameter. After the initial stage, the matrix θ and the matrix P are renewed with new calculated values. Then, the OCV is calculated through the obtained parameters.
2) Second Equivalent Circuit Model
(1) Second Equivalent Circuit Model
Hereinafter, a second equivalent circuit model and an adaptive digital filtering method in accordance with a second embodiment of the present invention will be described with reference to FIG. 9.
The second equivalent circuit model and the adaptive digital filtering method in accordance with the second embodiment of the present invention provides a discrete equivalent circuit modeling method by using characteristic that the BMS discretely receives a voltage value, a current value and a temperature value of a battery. The second equivalent circuit model according to the second embodiment of the present invention, which is a kind of a reduced model, simply expresses electrochemical characteristic in the battery. Therefore, since it is possible to easily design a model and also to minimize a time period required for a modeling operation, it is adaptively applied to the BMS.
FIG. 9 is a view showing an equivalent circuit model in accordance with another embodiment of the present invention.
The second equivalent circuit model according to the second embodiment of the present invention is provided as a primary model. Each element such as resistance, capacitor and the like of the model has its own meaning shown in Table 1 as follows:

TABLE 2

Element of equivalent circuit model

	Step	Process

	I	Current (charge: (+)/discharge (−))
	V	Terminal voltage
	V_o	Open circuit voltage (OCV)
	R₀	Lumped interfacial resistances
	R	Lumped series resistances
	C	Electric double layer capacitor

In FIG. 9, as shown in Table 2, R₀is resistance in electrodes, and R and C are resistance and a capacitor that express an electric double layer generated at an interface between one electrode and the other electrode or a separator, and V0 is the OCV of the battery.
A core idea applied in the second equivalent circuit model is that the current data and the voltage data are discretely input to the BMS at regular time intervals. Due to such core idea, it is possible to embody the model that expresses the behavior of the battery through the second equivalent circuit model. The parameters used in the model are calculated by the adaptive digital filter.
Equations 26 to 29 are provided through FIG. 9 and Table 2.
$\begin{matrix} I + I_{2} + I_{3} = 0 & Equation 26 \\ V = V_{0} + {IR}_{0} - I_{3} R & Equation 27 \\ V = V_{0} + {IR}_{0} - \frac{Q}{C} & Equation 28 \\ I_{2} = \frac{\partial Q}{\partial t} & Equation 29 \end{matrix}$
An equation 30 is provided by integrating the equation 28 over time and then arranging it.
$\begin{matrix} \frac{\partial}{\partial t} (V - V_{0}) + \frac{1}{RC} (V - V_{0}) = \frac{I}{C} (1 + \frac{R_{0}}{R}) + R_{0} \frac{\partial I}{\partial t} & Equation 30 \end{matrix}$
An equation 31 is provided by differentiating the equation 30 over time using an integrating factor.
$\begin{matrix} V = \frac{Q (0)}{C} e^{- t / RC} + V_{0} + {IR}_{0} + \frac{1}{C} \int_{ξ = 0}^{ξ = t} (I (ξ) e^{-- (t - ξ) / RC}) \partial ξ & Equation 31 \end{matrix}$
Assuming that a polarization phenomenon in the battery does not occur at an initial time of the BMS operation, Q(0)=0. Therefore, the equation 31 may be expressed by an equation 32.
$\begin{matrix} V = V_{0} + {IR}_{0} + \frac{I}{C} \int_{ξ = 0}^{ξ = t} (I (ξ) e^{-- (t - ξ) / RC}) \partial ξ & Equation 32 \end{matrix}$
In the equation 32, since the current data and the voltage data are input to the BMS at regular time intervals, it is possible to discretely express it A data input time interval is Δt.
In case of t≦0, assuming that I(t)=0 and T=0, it may be expressed as follows
∫₀ ⁰(I(ξ)e ^{−(t−ξ)IRC)} dξ=0,
Therefore, the equation 32 may be expressed by an equation 33
V−V ₀ −IR ₀|_t=0=0 Equation 33
Further, in case that t=Δt=t1, i.e., a period of time has passed by Δt, the equation 32 may be expressed into an equation 34 through partial integration.
$\begin{matrix} {V - V_{0} - {IR}_{0} \rangle}_{t = t_{1}} \approx \frac{e^{- t_{1} / RC}}{C} (I (t_{1}) Δ t \cdot e^{t_{1} / RC}) & Equation 34 \end{matrix}$
Furthermore, in case that t=2Δt=t2, i.e., a period of time has passed from t1 to Δt, the equation 32 may be expressed into an equation 35 through partial integration.
$\begin{matrix} {V - V_{0} - {IR}_{0} \rangle}_{t = t_{2}} \approx \frac{e^{- t_{2} / RC}}{C} [(I (t_{2}) Δ t \cdot e^{t_{2} / RC}) + (I (t_{1}) Δ t \cdot e^{t_{1} / RC})] & Equation 35 \end{matrix}$
The equations 34 and 35 may be combined into an equation 36.
$\begin{matrix} {V - V_{0} - {IR}_{0} \rangle}_{t = t_{2}} \approx \frac{I (t_{2}) Δ t}{C} + {e^{- Δ t / RC} [V - V_{0} - {IR}_{0} \rangle}_{t = t_{1}}] & Equation 36 \end{matrix}$
By repeating the above calculations, the equation 32 may be expressed into an equation 37 with respect to a time period t.
$\begin{matrix} {V - V_{0} - {IR}_{0} \rangle}_{t} \approx \frac{I Δ t}{C} + {e^{- Δ t / RC} [V - V_{0} - {IR}_{0} \rangle}_{t - Δ t}] & Equation 37 \end{matrix}$
Therefore, the equivalent circuit model is calculated through the equation 38.
$\begin{matrix} V = V_{0} + {IR}_{0} + \frac{1}{C} I Δ t + {e^{- Δ t / RC} [V - V_{0} - {IR}_{0} \rangle}_{t - Δ t}] & Equation 38 \end{matrix}$
In the equation 38, Δt is a data input time interval, and t−Δt is a parameter value in the previous time interval. Assuming that α=1/C and β=1/RC, the equation 38 may be simply arranged into an equation 39.
V=V ₀ +IR ₀ +αIΔt+exp(−βΔt)[V−V ₀ −IR ₀ |t−Δt] Equation 39
In the equation 39, β is a time constant, and τ is a reciprocal number and indicates a time when the battery arrives at a normal state. Generally, in case that a battery operation time of 3τ or more passes, it is estimated that the reaction in the battery arrives at the normal state. The factors of current and voltage are collected from the BMS and then calculated through the filtered current and voltage data. And the VOC is calculated by substituting each parameter.
R0 that indicates ohmic resistance of the battery itself decides a fixed value according to property of a material in the battery. Generally, the value is easily obtained through impedence. Variables α and β which are combined with the parameters R and C related to polarization are optimized through the adaptive digital filter.
(2) Adaptive Digital Filter
The equation 39 that mathematically indicates the equivalent circuit model of FIG. 9 may be expressed by an equation 40 as a determinant.
$\begin{matrix} V = [\begin{matrix} V_{0} & {IR}_{0} & I & {V - V_{0} - {IR}_{0} \rangle}_{n - 1} \end{matrix}] [\begin{matrix} 1 \\ 1 \\ αΔ t \\ \exp (- βΔ t) \end{matrix}] & Equation 40 \end{matrix}$
In the equation 40,
in case of
$[\begin{matrix} V_{0} \\ {IR}_{0} \\ I \\ {V - V_{0} - {IR}_{0} \rangle}_{n - 1} \end{matrix}] = w [\begin{matrix} 1 \\ 1 \\ αΔ t \\ \exp (- βΔ t) \end{matrix}] = θ,$
the equation 40 may be expressed into an equation 41.
V=w^Tθ Equation 41
In the equation 41, w is calculated through the current data and the voltage data, and v is also calculated through the current data and the voltage data. The adaptive digital filter calculates θ through the two values w and V, and the parameter value is estimated through each component of θ in real time.
Assuming that V passed through the low pass filter is gV, the equation 41 may be expressed into an equation 42.
gV=w^Tθ Equation 42
An initial value of θ is obtained by calculating a parameter in a state that t=0, i.e., the current is not flowed and the voltage has an OCV value. A matrix at this time is θ₀.
θ₀by θ₀is equal to an equation 43.
P ₀=θ₀·θ₀ ^T Equation 43
In the equation 43, a matrix k that is needed to continuously renew θ may be defined as an equation 44.
$\begin{matrix} K = \frac{P_{0} \cdot w}{r + w^{T} \cdot P_{0} \cdot w} & Equation 44 \end{matrix}$
In the equation 44, r is a constant that is defined to prevent a denominator from being diverged by V, and typically has a small value. The equation 44 may be changed into an equation 45 by arranging a matrix thereof.
$\begin{matrix} K = \frac{P_{0} \cdot w}{r + w^{T} \cdot P_{0} \cdot w} = \frac{gV (n - 1) \cdot θ_{0}}{r + {gV (n - 1)}^{2}} & Equation 45 \end{matrix}$
In the equation 45, gV(n−1) is a just previous value of gV. Assuming that the continuous renewal of θ is occurred by proportional relationship of the equation 46, the equation 46 may be arranged into an equation 47.
$\begin{matrix} θ = θ_{0} - [\frac{gV (n - 1) \cdot θ_{0}}{{gV (n - 1)}^{2}} \cdot {gV (n - 1) - gV (n)}] & Equation 47 \end{matrix}$
In the equation 47, since r is a very small value, it may be substituted by K. That is, the equations 45 and 47 may be combined into an equation 48.
θ=θ₀ −[K·{gV(n−1)−gV(n)}] Equation 48
By instituting the equation 42 to the equation 48, it is possible to obtain an equation 49.
θ=θ₀ −[K·{w ^T·θ₀ −gV(n)}] Equation 49
Main variables in the equation 49 are obtained through the current data, the voltage data and θ₀. Therefore, each parameter is continuously estimated in the same manner.
Further, after the initial stage, θ and P may be renewed into new values. Thus, the parameter values suitable to the equivalent circuit model may be continuously renewed. Also, it is possible to calculate the OCV using the parameters obtained in the same way.
E. Fifth Step: Calculation of SOCv Through Open Circuit Voltage and Temperature
A value of the OCV is calculated through the equivalent circuit model according to the present invention. Generally, SOCv is influenced by the OCV and the temperature and thus expressed by a function between them. In a room temperature, relation between OCV and SOCv is equal to an equation 50.
SOC _v(n)=−539.069·V ₀(n)⁴+7928.96·V ₀(n)³−43513.3·V ₀(n)²+105698·V ₀(n)−95954.8 Equation 50
As described above, it is possible to obtain SOCv through the relation between OCV and SOCv and the value of OCV at the room temperature. However, there is a problem in the equation 50. Since it is a model at the room temperature, errors occur when the temperature is changed. A maximum value and a minimum value of the error generated when a simulation is performed at a temperature of 45˜−10° C. besides a room temperature of 25° C. are described in Table 3.

TABLE 3

Maximum and minimum error of each temperature simulation

Error (%)

Temperature	Max	Min

25° C.	0.714	−1.936
45° C.	0.008	−2.692
−10° C.	0.000	−8.542

Therefore, using the relation between SOCv and OCV at the room temperature, it is possible to obtain precise values at a temperature of 45° C. However, it is understood that the values are incorrect at a temperature of −10° C. In other words, at the temperature of −10° C., it is necessary to use other relation or introduce a factor that considers the temperature. First of all, the relation between SOCv and OCV at the temperature of −10° C. is expressed by an equation 51.
SOC _v(n)=−425.6·V ₀(n)⁴+6207·V ₀(n)³−33740·V ₀(n)²+81113·V ₀(n)−72826 Equation 51
F. Sixth Step: Proper Selection of SOC
In the sixth step, it is decided which one is selected from SOCi and SOCv obtained in the previous step. In case of the low current state, it is known that SOCv has an exact value. Therefore, SOCv is used in the low current state. And the calculation is performed by accumulating the current value to the just previous SOC in other states.
In criteria of judgment of the low current state, if the current having a desired absolute value or less is continuously flowed for a desired period of time, it is determined as the low current state. Herein, the absolute value of current and the current flowing time are important criteria. Preferably, the absolute value and the time are 2A and 20˜60s, respectively.
In case that the absolute value of current intensity is more than 2A, the criteria of judgment of the low current state is relaxed considerably, and it is judged as a state that the current is being flowed even at a section that the current is not flowed. As a result, it is difficult to precisely estimate SOCv, and thus SOCv compensation is carried out too frequently. However, in case that the absolute value of current is less than 2A, it is impossible to recognize the low current when offset occurs. Therefore, it is preferable to set up the criteria to 2A.
It is more complicated to judge the criteria of current flowing time. Assuming that a period of time when the current corresponding to battery charging or discharging and having an intensity of 2A or less is continuously flowed is t, the criteria may be expressed as in Table 4.

TABLE 4

Time scale criteria for SOCv compensation

Continuous time criteria	Judgment

t < a	Use of SOCi
a ≦ t < b	Compensation with SOCv at the moment
	that low current section is ended
b ≦ t	SOCv compensation at a point of time
	of b seconds

FIG. 10 shows the time criteria. Referring to FIG. 10, if the low current is flowed for 20 seconds or less, SOC is calculated by the current accumulation. However, if the low current is flowed for 20 seconds or more, the SOCv compensation occurs. The SOCv compensation is carried out at a point of time when the low current section having an intensity of 2A or less is finished. However, if the low current is flowed for 60 seconds or more, the SOCv compensation is carried out at a point of time of 60 seconds. And at a moment that the compensation occurs, the continuous period of time when the low current is flowed is calculated again.
The criteria are decided by selecting the optimal time criteria from various simulations. The time criteria of minimum 20 seconds is decided based on a fact that the compensation is occurred even if the current sensor has a trouble. Actually, in case that the time criteria is changed into 10 seconds, an error of 8.3% occurs. Further, in case that the continuous period of time is set up to 60, the compensation does not occur properly at an up/down pattern, and the error is accumulated.
The main cause of allowing the time criteria to be movable is to prevent increase of the error due to improper compensation after the battery charging for a long time period. The present invention is not limited to the time criteria, and may include various time criteria.
FIG. 11( a) shows an error in case of performing a simulation with a time criteria of 20 seconds, and FIG. 11( b) shows an error in case of performing the simulation with a proposed time criteria.
Referring to FIG. 11, the utility of the proposed time criteria will be understood. In case of the simulation with a time criteria of 20 seconds, it was understood that considerable error occurred at beginning and end portions that are just after the charging. However, in case of the simulation with the proposed time criteria, the error did not occur at beginning and end portions. Entire error magnitude is hardly changed, but the error occurred at a point of time when the charging is finished is reduced by 1.5% or more.
The proposed time criteria is more effective in a low temperature state. In case of the low temperature state, it is generally effective, of cause, at the point of time when the charging is finished. In case of using the proposed time criteria, a maximum value of the error is reduced from 7,287% to 3.542%, and a minimum value thereof is reduced from −4.191% to −3.870%. This means that the proposed time criteria is more effective in a low temperature state.
If it is judged as the low current state on the basis of the time and current criteria as mentioned above, SOCv is set up to SOC. Otherwise, BMS SOC is precisely calculated by using the SOC measuring method of the present invention.
While the present invention has been described with respect to the specific embodiments, it will be apparent to those skilled in the art that various changes and modifications may be made without departing from the spirit and scope of the invention as defined in the following claims.

INDUSTRIAL APPLICABILITY

The SOC measuring method of the present invention may be embodied in the form of a program that is executed through various computing means, and then recorded on a computer-readable medium. The computer-readable medium may include a program order, a data file, a data structure and the like, or any combination thereof. The program order recorded on the medium may be specially designed or constructed to be used in the present invention, or used in a state that is provided to those skilled in the field of computer software. Further, the computer-readable medium may include a magnetic media such as a hard disk, a floppy disk and a magnetic tape, an optical media like DVD, a magnetro-optical media like a floptical disk, and a hardware device for storing and executing a program order, such as ROM, RAM and a flash memory. The medium may be a transmission media of a wave guide, a metal wire or light including a carrier wave for transmitting a signal indicating a program order, a data structure and the like. The program order includes a machine language code formed by a compiler as well as a high level language code that is formed using an interpreter so as to be executed by a computer. In order to perform an operation of the present invention, the hardware device may be constructed to be operated by one or more software modules, and the reverse thereof is the same.

Claims

1. A method for measuring SOC of a battery, comprising:

obtaining current data, voltage data and temperature data by measuring current, voltage and temperature of the battery;

calculating SOCi (State Of Charge based on current) by accumulating the current data;

calculating OCV (Open Circuit Voltage) using an equivalent circuit model in which the current data, the voltage data and the battery are simply expressed by an electric circuit;

calculating SOCv (State Of Charge based on voltage) using the temperature data and the OCV; and

judging a current state of the battery for a desired period of time, and setting up the SOC of the battery using at least one of the SOCv and the SOCi.

2. The method of claim 1, wherein the calculating of OCV using an equivalent circuit model in which the current data, the voltage data and the battery are simply expressed by an electric circuit comprises:

Filtering the current data and the voltage data using a low pass filter;

Calculating a parameter used in the equivalent circuit model by applying the filtered current data and voltage data to the equivalent circuit model and an adaptive digital filter; and

calculating the OCV using the parameter.

3. The method of claim 2, wherein the low pass filter is a third order low pass filter.

4. The method of claim 2, wherein the equivalent circuit model is expressed by an electric circuit using a resistance parameter R, a current parameter I, a capacitor parameter C, a terminal voltage parameter V and VOC parameter Vo.

5. The method of claim 2, wherein a value of the parameter used in the equivalent circuit model is renewed by the adaptive digital filter.

6. The method of claim 1, wherein the judging of the current state of the battery for the desired period of time, and setting of the SOC of the battery using at least one of the SOCv and the SOCi comprises:

setting up the SOCv to the SOC of the battery, if the battery is in a low current state for the desired period of time; and

setting up the SOCv to the SOC of the battery, if the battery is not in a low current state for the desired period of time.

7. The method of claim 6, wherein the desired period of time is 20˜60 seconds, and the low current criteria is 2A.

8. The method of claim 1, wherein the calculating of the OCV using the equivalent circuit model in which the current data, the voltage data and the battery are simply expressed by the electric circuit comprises

calculating the OCV using an integration model of the equivalent circuit model in which the current data, the voltage data and the battery are simply expressed by the electric circuit.

9. An apparatus for measuring SOC of a battery, comprising:

a battery information obtaining part that measures current, voltage and temperature of the battery and obtains current data, voltage data and temperature data;

a current accumulating part that calculates SOCi by accumulating the current data;

a OCV calculating part that calculates OCV using an equivalent circuit model in which the current data, the voltage data and the battery are simply expressed by an electric circuit;

a SOCv estimating part that estimates SOCv using the temperature data and the OCV; and

a SOC setting part that judges a current state of the battery for a desired period of time, and sets up the SOC of the battery using at least one of the SOCv and the SOCi.

10. The apparatus of claim 9, further comprising a low pass filtering part that filters the current data and the voltage data using a low pass filter,

wherein the OCV calculating part applies the current data and the voltage data filtered by the low pass filtering part to the equivalent circuit model and an adaptive digital filter, calculates a parameter used in the equivalent circuit model and then calculates the OCV using the parameter.

11. The apparatus of claim 10, wherein the low pass filter is a third order low pass filter.

12. The apparatus of claim 10, wherein the equivalent circuit model is expressed by an electric circuit using a resistance parameter R, a current parameter I, a capacitor parameter C, a terminal voltage parameter V and VOC parameter Vo.

13. The apparatus of claim 10, wherein the adaptive digital filter renews a value of the parameter using in the equivalent circuit model, continuously.

14. The apparatus of claim 9, wherein the SOC setting part sets up the SOCv to the SOC of the battery, if the battery is in a low current state for the desired period of time, and also sets up the SOCi to the SOC of the battery in other case.

15. The apparatus of claim 14, wherein the desired period of time is 2060 seconds, and the low current criteria is 2A.

16. The apparatus of claim 9, wherein the OCV calculating part calculates the OCV through an integration model of the equivalent circuit model.

17. A computer-readable recording medium which records a program for executing the method of claim 1.