KR102349235B1 - Apparatus and method for estimating state of charge for battery - Google Patents

Abstract

Disclosed are an apparatus and a method for estimating a charged state of a battery using an extended Kalman filter. The device may include: a sensing unit that measures the terminal voltage, current, and temperature of the battery and outputs a sensing signal indicating the measured terminal voltage, current, and temperature; and a control unit including at least one processor configured to execute the extended Kalman filter and operatively coupled to the sensing unit. The control unit predicts a dot product state including a state of charge and a polarization voltage of the battery, and an error covariance matrix. The controller determines a voltage measurement value, a current measurement value, and a temperature measurement value corresponding to the measured terminal voltage, the measured current, and the measured temperature based on the sensing signal. The control unit determines whether a predetermined bypass condition is satisfied by the current measurement value, the temperature measurement value, or a previously estimated state of charge. When the bypass condition is satisfied, the state of charge of the battery is estimated in the correction deactivation mode.

Description

Apparatus and method for estimating state of charge for battery

The present invention relates to an apparatus and method for estimating the state of charge of a battery, and more particularly, to an apparatus and method for estimating the state of charge of a battery using an extended kalman filter (EKF). .

Recently, as the demand for portable electronic products such as laptops, video cameras, and mobile phones has rapidly increased, and the development of electric vehicles, energy storage batteries, robots, satellites, etc. is in full swing, high-performance batteries that can be repeatedly charged and discharged have been developed. Research is being actively conducted.

Currently commercialized batteries include nickel-cadmium batteries, nickel-hydrogen batteries, nickel-zinc batteries, and lithium batteries. It is in the spotlight because of its low energy density and high energy density.

One of the important parameters in using and managing a battery is the state of charge. The state of charge is a factor representing the relative ratio of the current remaining capacity to the full charge capacity representing the maximum amount of charge that can be stored in the battery, and may be expressed as 0 to 1 or 0% to 100%.

A conventional representative technique used for estimating the state of charge of a battery is a current integration method. According to the current integration method, the current flowing through the battery is measured using a current sensor, and the state of charge of the battery is estimated based on the current integration value accumulated over time. However, due to a measurement error of the current sensor, a difference occurs between the state of charge estimated using the current integration method and the actual state of charge, and the difference increases as time passes.

In order to compensate for the above disadvantages of the current integration method, a technique for estimating the state of charge of a battery using an extended Kalman filter has been disclosed. The extended Kalman filter corrects the result of the current integration according to the difference between the predicted value and the measured value of a parameter (eg, terminal voltage) measurable from the battery, thereby providing higher accuracy than the current integration method under normal circumstances.

However, what should be noted is the nonlinearity of the terminal voltage of the battery. Specifically, the terminal voltage of the battery has a characteristic that non-linearly changes according to the current, temperature, and/or state of charge of the battery, and a system model (eg, equivalent circuit model) to accurately simulate this non-linearity It is never easy to establish. Therefore, when the nonlinearity of the battery becomes very large, the difference between the predicted value and the measured value also increases, and as a result, the estimation accuracy of the state of charge estimated using the extended Kalman filter may be lower than the state of charge estimated using only the current integration method. .

The present invention has been devised to solve the above problems, and by selectively setting the Kalman gain of the extended Kalman filter to 0 according to the state of charge, current or temperature of the battery, to more accurately estimate the state of charge of the battery. An object of the present invention is to provide an apparatus and method.

Other objects and advantages of the present invention may be understood by the following description, and will become more clearly understood by the examples of the present invention. In addition, it will be readily apparent that the objects and advantages of the present invention can be realized by means and combinations thereof indicated in the claims.

Various embodiments of the present invention for achieving the above object are as follows.

An apparatus for estimating a layered state of a battery using an extended Kalman filter according to an aspect of the present invention measures the terminal voltage, current, and temperature of the battery, and receives a sensing signal representing the measured terminal voltage, current and temperature a sensing unit to output; and a control unit including at least one processor configured to execute the extended Kalman filter and operatively coupled to the sensing unit. The control unit predicts a dot product state including a state of charge and a polarization voltage of the battery, and an error covariance matrix. The controller determines a voltage measurement value, a current measurement value, and a temperature measurement value corresponding to the measured terminal voltage, the measured current, and the measured temperature based on the sensing signal. The control unit determines whether a predetermined bypass condition is satisfied by the current measurement value, the temperature measurement value, or a previously estimated state of charge. When the bypass condition is satisfied, the state of charge of the battery is estimated in the correction deactivation mode. In the correction deactivation mode, the controller sets the Kalman gain to 0, and in response to the Kalman gain being set to 0, estimates the state of charge and the polarization voltage from the predicted inner product state, respectively.

When the bypass condition is not satisfied, the controller may estimate the state of charge of the battery in a correction activation mode. In the correction activation mode, the controller calculates the Kalman gain based on the predicted error covariance matrix. The control unit calculates a voltage prediction value representing the predicted terminal voltage of the battery based on the predicted internal product state and the current measurement value. The control unit estimates the charging state and the polarization voltage from a result of correcting the predicted inner product state based on the voltage prediction value, the voltage measurement value, and the Kalman gain, respectively. The controller corrects the predicted error covariance matrix based on the Kalman gain.

The control unit, when the current measurement value is within a predetermined current range, when the temperature measurement value is within a predetermined temperature range, or when the previously estimated state of charge is within a predetermined charge state range, the bypass condition is satisfied can be judged as

In the calibration deactivation mode, in response to the Kalman gain being set to 0, the controller may initialize the predicted error covariance matrix to be the same as an initial error covariance matrix. The initial error covariance matrix may be a predetermined matrix or a most recently corrected error covariance matrix.

When the bypass condition is satisfied, the controller may determine whether the time count is less than a predetermined threshold count. When the time count is less than the threshold count, the controller may estimate the state of charge of the battery in the correction deactivation mode.

The controller may increase the time count by a predetermined value when the state of charge of the battery is estimated in the correction deactivation mode. When the state of charge of the battery is estimated in the correction activation mode, the time count may be initialized to be equal to a predetermined initial value.

The apparatus may further include a communication unit configured to transmit a notification signal indicating the estimated state of charge to an external device according to a predetermined communication rule.

A battery pack according to another aspect of the present invention includes the above device.

In a method for estimating a charged state of a battery using an extended Kalman filter according to another aspect of the present invention, a control unit including at least one processor configured to execute the extended Kalman filter comprises: a state of charge and polarization of the battery predicting a dot product state including a voltage and an error covariance matrix; The control unit, based on the sensing signal output by the sensing unit electrically connected to the battery, the voltage measurement value, the current measurement value and the temperature corresponding to the terminal voltage, current and temperature of the battery measured by the sensing unit determining a measurement value; determining, by the control unit, whether a predetermined bypass condition is satisfied by the current measurement value, the temperature measurement value, or a previously estimated state of charge; and estimating, by the controller, the state of charge of the battery in a correction deactivation mode when the bypass condition is satisfied. The estimating of the state of charge of the battery in the correction deactivation mode may include: setting, by the controller, a Kalman gain to 0; and estimating, by the controller, the state of charge and the polarization voltage from the predicted inner product state, respectively, in response to the Kalman gain being set to zero.

In the step of estimating the state of charge of the battery in the correction deactivation mode, the controller initializes the predicted error covariance matrix to be the same as the initial error covariance matrix in response to the Kalman gain being set to 0. may include more. The initial error covariance matrix may be a predetermined matrix or a most recently corrected error covariance matrix.

According to at least one of the embodiments of the present invention, when there is a sign that the nonlinearity of the battery temporarily becomes excessively large according to the state of charge, current or temperature of the battery, the state of charge of the battery is estimated based only on the current integration. By selectively setting the Kalman gain of the Kalman filter to 0, it is possible to more accurately estimate the state of charge of the battery.

Further, according to at least one of the embodiments of the present invention, in response to the Kalman gain being set to zero, by initializing the error covariance matrix of the extended Kalman filter to be equal to a predetermined matrix, excessive increase of the components of the error covariance matrix can be suppressed.

In addition, according to at least one of the embodiments of the present invention, when the time count corresponding to the time in which the Kalman gain is continuously maintained at 0 becomes greater than or equal to the threshold count, the Kalman gain is not set to 0, so that the current integration Accordingly, it is possible to limit the charge state error.

Effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description of the claims.

The following drawings attached to this specification illustrate preferred embodiments of the present invention, and serve to further understand the technical spirit of the present invention together with the detailed description of the present invention to be described later, so that the present invention is a matter described in such drawings should not be construed as being limited only to
1 is a diagram exemplarily showing the configuration of a battery pack including an apparatus for estimating a state of charge of a battery according to an embodiment of the present invention.
FIG. 2 is a diagram showing an exemplary equivalent circuit model of the battery shown in FIG. 1 .
3 is a graph showing an exemplary SOC-OCV profile for use in determining the open circuit voltage of the voltage source of FIG. 2 during operation of the extended Kalman filter.
4 is a graph exemplarily showing a correlation between a current of a battery and an error increase rate.
5 is a flowchart of a method for estimating a state of charge of a battery according to an embodiment of the present invention.
6 is a flowchart of a method for estimating a state of charge of a battery according to another embodiment of the present invention.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. Prior to this, the terms or words used in the present specification and claims should not be construed as being limited to conventional or dictionary meanings, and the inventor should properly understand the concept of the term in order to best describe his invention. Based on the principle that it can be defined, it should be interpreted as meaning and concept consistent with the technical idea of the present invention.

Accordingly, the embodiments described in this specification and the configurations shown in the drawings are only the most preferred embodiment of the present invention and do not represent all the technical spirit of the present invention, so at the time of the present application, various It should be understood that there may be equivalents and variations.

In addition, in the description of the present invention, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the present invention, the detailed description thereof will be omitted.

Terms including an ordinal number such as 1st, 2nd, etc. are used for the purpose of distinguishing any one of various components from the others, and are not used to limit the components by such terms.

Throughout the specification, when a part "includes" a certain element, it means that other elements may be further included, rather than excluding other elements, unless specifically stated otherwise. In addition, a term such as <control unit> described in the specification means a unit that processes at least one function or operation, which may be implemented as hardware or software, or a combination of hardware and software.

In addition, throughout the specification, when a part is "connected" with another part, it is not only "directly connected" but also "indirectly connected" with another element interposed therebetween. include

1 is a diagram exemplarily showing the configuration of a battery pack 10 including an apparatus 100 for estimating a state of charge of a battery 20 according to an embodiment of the present invention.

Referring to FIG. 1 , a battery pack 10 includes a battery 20 and a device 100 (hereinafter, referred to as a 'device') for estimating a state of charge of the battery 20 . The battery 20 may be, for example, a lithium ion battery, a lithium polymer battery, a nickel cadmium battery, a nickel hydride battery, or a nickel zinc battery. Of course, the type of the battery 20 is not limited to those listed above, and is not particularly limited as long as it can be repeatedly charged and discharged.

The device 100 includes a sensing unit 110 and a control unit 120 . The device 100 may optionally further include at least one of the memory unit 130 and the communication unit 140 .

The sensing unit 110 is operatively coupled to the control unit 120 . That is, the sensing unit 110 may be electrically connected to the control unit 120 to transmit an electrical signal to the control unit 120 or to receive an electrical signal from the control unit 120 .

The sensing unit 110, the terminal voltage corresponding to the potential difference between the positive terminal and the negative terminal of the battery 20, the current of the battery 20, and the battery every predetermined period (that is, whenever the time step is increased by 1), The temperature of 20 may be measured, and signals indicating the measured terminal voltage, the measured current, and the measured temperature may be output to the controller 120 .

The sensing unit 110 includes a voltage sensor 111 . The voltage sensor 111 is electrically connected to a positive terminal and a negative terminal of the battery 20 , and is configured to measure a terminal voltage of the battery 20 .

The sensing unit 110 includes a current sensor 112 . The current sensor 112 is installed on the charge/discharge current path of the battery 20 and is configured to measure the current of the battery 20 . The current sensor measures a charging current flowing through the battery 20 when the battery 20 is being charged and a discharging current flowing through the battery 20 when the battery 20 is being discharged.

The sensing unit 110 includes a temperature sensor 113 . The temperature sensor 113 is directly attached to the outside of the battery 20 or disposed close within a predetermined distance from the battery 20 to measure the temperature of the battery 20 .

The sensing unit 110 outputs a sensing signal indicating the measured terminal voltage, current, and temperature to the control unit 120 .

The memory unit 130 is not particularly limited in its type as long as it is a storage medium capable of recording and erasing information. As an example, the memory unit 130 may be a RAM, a ROM, a register, a hard disk, an optical recording medium, or a magnetic recording medium. The memory unit 130 may be communicatively connected to the control unit 120 through, for example, a data bus, so that it can be accessed by the control unit 120 . The memory unit 130 may store, update, and/or erase a program including various control logic executed by the control unit 120 and/or data generated when the control logic is executed. The memory unit 130 may be logically divided into two or more. The memory unit 130 may be built into the control unit 120 .

The communication unit 140 includes a communication interface that supports bidirectional communication between the control unit 120 and an external device. The communication unit 140 may transmit a signal output from the control unit 120 to an external device according to a predetermined communication rule. The communication unit 140 may transmit a command from an external device to the control unit 120 . The external device may be, for example, a controller of a vehicle in which the battery pack 10 is installed.

The control unit 120, in hardware, ASICs (application specific integrated circuits), DSPs (digital signal processors), DSPDs (digital signal processing devices), PLDs (programmable logic devices), FPGAs (field programmable gate arrays), microprocessors (microprocessors) and may be implemented using at least one of electrical units for performing other functions.

The control unit 120 is operatively coupled to the sensing unit 110 . The control unit 120 may be additionally operatively coupled to the memory unit 130 and the communication unit 140 . The controller 120 includes at least one processor configured to execute the extended Kalman filter.

The control unit 120, based on the sensing signal from the sensing unit 110, whenever the time step is increased by 1, the voltage measurement value corresponding to the terminal voltage of the battery 20, the current of the battery 20 A corresponding current measurement and a temperature measurement corresponding to the temperature of the battery 20 are determined. When the battery 20 is charged, the current measurement value may have a positive value, and when the battery 20 is discharged, the current measurement value may have a negative value.

The control unit 120 is further configured to estimate the state of charge of the battery 20 using the extended Kalman filter based on the voltage measurement value, the current measurement value, and the temperature measurement value.

From now on, the extended Kalman filter will be described in more detail. The extended Kalman filter statistically calculates at least one state variable representing the internal states of the system in consideration of externally measurable parameters and system disturbance for a dynamic system having nonlinear characteristics. It is an estimating algorithm. Since the terminal voltage of the battery 20 has a characteristic that non-linearly changes according to the current, temperature, and/or state of charge of the battery 20 , the battery 20 is subjected to at least one that can be estimated through the extended Kalman filter. It can be treated as a dynamic system with an inner state of .

The extended Kalman filter sequentially executes a process for time update and a process for measurement update whenever the time step is increased by 1. Through time update, state variables and parameters related to the inner product state of the battery 20 are predicted based on previous data. Through the measurement update, parameters predicted through the time update are corrected based on the measurement data. As a result of the measurement update, estimation for each state variable representing the inner product state of the battery 20 is completed.

Hereinafter, for convenience of description, it is assumed that time step k refers to a current time step and time step k-1 refers to a previous time step. In this case, the symbol k means the number of steps, and the controller 120 may increase k by 1 whenever a predetermined time Δt elapses.

First, Equations 1 to 3 related to the time update process will be described. Equation 1 below is for predicting each state variable included in the internal state of the battery 20 treated as a dynamic system, based on previous data, and may be referred to as a 'state equation'. .

<Formula 1>

x _k,- = f(x _k-1,+ , u _k-1 ) = A _k-1 x _k-1,+ + B _k-1 u _k-1

Each symbol in Equation 1 is described as follows. x _k-1,+ is the estimated dot product state of time step k-1, u _k-1 is the measured input of time step k-1, x _k,- is the predicted dot product state of time step k, f() is a function to obtain x _{k,- from x k-1,+} and u _k-1. Meanwhile, in preparation for an initialization event of the extended Kalman filter in which the step number k of the time step k becomes 0, _{data representing the initial inner product state x 0} may be stored in advance in the memory unit 130 .

Equation 2 below is for predicting the error covariance of the extended Kalman filter.

<Formula 2>

P _k,- = A _k-1 P _k-1,+ A _k-1 ^T +B _k-1 QB _k-1 ^T

Each symbol in Equation 2 will be described as follows. P _k-1,+ is the estimated error covariance matrix of time step k-1, Q is the process noise covariance matrix, and P _k,- is the predicted error covariance matrix of time step k. The superscript ^T is the transposition matrix operator. _{A k-1} and B _k-1 of Equations 1 and 2 may be expressed by Equation 3 below.

<Equation 3>

That is, A _k-1 is a Jacobian matrix that is the result of partial differentiation of function f() _{with respect to x k-1,+ , and B k-1} is the result of partial differentiation of function f() with respect _{to u k-1.} The resulting Jacobian matrix.

Next, Equations 4 to 8 related to the process for measurement update will be described. Equation 4 below is for estimating an output variable, and may be referred to as an 'output equation'.

<Formula 4>

y _k,- = h(x _k,- , u _k ) = C _k x _k,- + D _k u _k

Each symbol in Equation 4 is described as follows. x _k,- is the predicted dot product state _{of time step k, u k} is the measured input of time step k, y _k,- is the predicted output variable of time step k, h is y from _{x k,-} and u _k This is a function to get _k,-. Alternatively, y _k,- may _{depend only on x k,-} , in which case y _k,- = h(x _k,- ) = C _k x _k,- and D _k is a given value (eg, 1) can be fixed.

Equation 5 below is for calculating the Kalman gain of the extended Kalman filter.

<Formula 5>

K _k = P _k,- C _k ^T (C _k P _k,- C _k ^T +D _k RD _k ^T ) ^-1

Each symbol in Equation 4 is described as follows. P _k,- is the predicted error covariance matrix of time step k, R is the measurement noise covariance matrix, K _k is the Kalman gain of time step k, the superscript ^T is the transpose matrix operator, and the superscript ^-1 is the inverse matrix operator. _{C k} and D _k of Equations 4 and 5 may be expressed by Equation 6 below.

<Formula 6>

That is, C _k is a Jacobian matrix that is a result of partial differentiation of function h() _{with respect to x k,- , and D k} is a Jacobian matrix that is a result of partial differentiation of function h() _{with respect to u k .}

Equation 7 below is for correcting _{x k,- from Equation 1.}

<Formula 7>

x _k,+ = x _k,- + K _K (y _k - y _k,- )

In Equation 7, y _k is the measured output variable _{of time step k, and x k,+} is the estimated dot product state of time step k. That is, x _k,+ is a result of correcting _{x k,-} by adding the value obtained by multiplying the difference between the measured output variable and the estimated output variable of the time step k by the Kalman gain to _{x k,-.} Each state variable contained in x _k,- is the estimated state variable of time step k.

Equation 8 below is for correcting the error covariance predicted by Equation 2.

<Equation 8>

P _k,+ = (I - K _K C _k )P _k,-

In Equation 8, I is a unit matrix.

The controller 120 may execute Equations 1 to 8 described above every time the time step increases by 1 in a predetermined order.

The state equation (refer to Equation 1) and the output equation (refer to Equation 4) of the extended Kalman filter are based on an equivalent circuit model of the battery 20, which will be described later with reference to FIG. 2 . That is, Equations 1 to 8 are related to the equivalent circuit model of the battery 20 .

2 is a diagram showing an exemplary equivalent circuit model 30 of the battery 20 shown in FIG. 1, and FIG. 3 is for use in determining the open circuit voltage of the voltage source of FIG. 2 during operation of the extended Kalman filter. A graph showing an exemplary SOC-OCV profile.

Referring to FIG. 2 , the equivalent circuit model 30 includes a voltage source 31 , a series resistor 32 , and an RC circuit 33 .

The voltage source 31 represents an open circuit voltage (OCV) of the battery 20 depending on the state of charge of the battery 20 . That is, the open-circuit voltage of the battery 20 may be uniquely determined when the state of charge is determined. In the memory unit 130 , the SOC-OCV profile as shown in FIG. 3 defining the correlation between the open voltage and the state of charge of the battery 20 is previously stored in the form of a lookup table or the like. As can be seen from FIG. 3 , as the state of charge of the battery 20 increases, the open-circuit voltage of the battery 20 increases non-linearly. The controller 120 may determine the open-circuit voltage related to the state of charge with reference to the SOC-OCV profile stored in the memory unit 130 . Of course, the correlation between the open voltage of the battery 20 and the state of charge may vary depending on the temperature. In this case, the memory unit 130 may store SOC-OCV profiles for a plurality of temperature values in advance, The controller 120 may select any one of the SOC-OCV profiles based on the temperature measurement value, and then determine the open circuit voltage related to the state of charge of the battery 20 with reference to the selected SOC-OCV profile.

The DC resistance 32 simulates an instantaneous change in the terminal voltage caused by the current flowing through the battery 20 . The voltage applied to the DC resistor 32 may be referred to as an 'IR voltage'. _{The resistance value R 0} of the DC resistor 32 may be a predetermined value. Alternatively, whenever the time step is increased by 1, the controller 120 may determine _{R 0} based on at least one of a state of charge and a temperature of the battery 20 . To this end, data defining a correlation between a charging state or temperature and R _{0 may be previously stored in the memory unit 130 .}

The RC circuit 33 mimics the transient change of the polarization voltage reflected in the terminal voltage of the battery 20, and includes a resistor 41 and a capacitor 42 connected in parallel with each other.

Equation 9 below is for time-updating the state of charge of the battery 20 according to the current integration method. The state of charge of the battery 20 is one state variable representing the state of the internal product of the battery 20 .

<Formula 9>

SOC _k,- = SOC _k-1,+ + (Δt/FCC)i _k-1

Each symbol in Equation 9 will be described as follows. SOC _k-1,+ is the estimated state of charge at time step k-1, i _k-1 is the current measurement at time step k-1, Δt is the interval between two adjacent time steps, FCC is the full charge of the battery 20 The full capacity, SOC _k,- is the predicted state of charge at time step k.

Equation 10 below is for predicting the polarization voltage by the RC circuit 33 of the equivalent circuit model 30 . The polarization voltage of the battery 20 is a state variable representing the internal product state of the battery 20 independently of the state of charge.

<Equation 10>

Each symbol in Equation 10 will be described as follows. V1 _k-1,+ is the estimated polarization voltage of time step k-1, V1 _k,- is the predicted polarization voltage of time step k, R ₁ is the resistance value of resistor 41, C ₁ is capacitor 42 is the capacitance of Each of R ₁ and C ₁ may be a predetermined value. Alternatively, whenever the time step is increased by 1, the controller 120 may determine each of _{R 1} and C ₁ based on at least one of a state of charge and a temperature of the battery 20 . To this end, data defining a correlation between a charging state or temperature and R ₁ and C ₁ may be stored in advance in the memory unit 130 .

The following Equation 11 can be derived from the combination of Equations 9 and 10.

<Formula 11>

Since each of the charging state and the polarization voltage of the battery 20 cannot be measured from the outside, it can be said that they are state variables constituting the internal product state of the battery 20 . In addition, the current of the battery 20 may be said to be an input causing a change in the internal state of the battery 20 . Therefore, if the state of charge and the polarization voltage are two state variables to be estimated from the battery 20, Equation 11 can be used as the state equation of Equation 1, in this case x _k-1,+ = [SOC _{k -1,+} V1 _k-1,+ ] ^T , x _k,- = [SOC _k,- V1 _k,- ] ^T , i _k-1 = u _k-1 , and A _k-1 and B in Equation 3 _k-1 is the same as Equation 12 below.

<Equation 12>

Meanwhile, the following Equation 13 is for predicting the terminal voltage of the battery 20 from the equivalent circuit model 30 .

<Equation 13>

V _k,- = OCV _k,- + V1 _k,- + R ₀ i _k = OCV(SOC _k,- ) + V1 _k,- + R ₀ i _k

In Equation 13, OCV( ) is a function for obtaining OCV _{k,- from SOC k,- using the aforementioned SOC-OCV profile, and V k,-} is a voltage predicted value representing the predicted terminal voltage of time step k . For example, referring to FIG. 3 , when SOC _k,- = A, OCV(SOC _k,- )=B. Equation 13 may be used as the output equation of Equation 4, and in this case, Equation 13 may be expressed as Equation 14 below.

<Equation 14>

V _k,- = C _k [SOC _k,- V1 _k,- ] ^T +D _k i _k = C _k [SOC _k,- V1 _k,- ] ^T + R ₀ i _k

The controller 120 may correct the result of Equation 11 by using Equation 15 below.

<Formula 15>

SOC _k,+ = SOC _k,- + K _k (y _k - V _k,- )

Equation 15 is derived from Equation 7. The terminal voltage of the battery 20 changes according to the internal state of the battery 20 and can be measured from the outside. _{Accordingly, a voltage measurement value VB k} indicating the terminal voltage of the battery 20 measured by the sensing unit 110 at time step k may be used as _{y k} in Equation 15.

4 is a graph exemplarily showing a correlation between a current of the battery 20 and an error increase rate.

In order to obtain a graph showing the correlation between the current and the error increase rate of the battery 20 as shown in FIG. 4 , that is, the error increase rate for each current, the test battery 20 designed to have the same specifications as the battery 20 A charge/discharge test should be preceded. The charge/discharge test is for observing a change in the state of charge of the test battery 20 while precisely changing the current and temperature of the test battery 20 according to a predetermined rule for a predetermined time.

While the charge/discharge test is in progress, the error between the actual state of charge of the test battery 20 and the state of charge estimated using Equations 1 to 8 (see Equations 9 to 15) described above is repeatedly calculated every predetermined time Δt and calculates the error increase rate for each time indicating _{the amount of change (ΔSOC error} ) of the error per unit time (eg, 0.1 sec). Then, by grouping the error increase rate by time for each of the current, temperature, and actual charging state used in the charge/discharge test through a predetermined algorithm, the error increase rate for each current, the error increase rate for each temperature, and the error increase rate for each charging state are obtained can

Referring to FIG. 4 , when the error increase rate for each current is equal to or greater than the predetermined threshold increase rate TH, it means that the accuracy of the state of charge estimated using the extended Kalman filter is significantly reduced.

Accordingly, by analyzing the result of the charging/discharging test for the test battery 20 , the current range in which the error increase rate for each current is equal to or greater than the critical increase rate TH may be determined. For example, when a graph as shown in FIG. 3 is obtained, 0 to C and D to E may be determined as the current range. Similarly, by analyzing the results of the charge/discharge test for the test battery 20, the temperature range in which the error increase rate for each temperature is equal to or greater than the critical increase rate (TH) and the error increase rate for each state of charge becomes the critical increase rate (TH) or greater The state of charge range may also be determined.

When the error increase rate for each current, the error increase rate for each temperature, or the error increase rate for each state of charge is equal to or greater than the threshold increase rate TH, the difference between the predicted voltage value and the voltage measurement value may be excessively large. The difference is scaled by the Kalman gain and used to correct the inner product state. When the difference is too large, the accuracy of the state of charge estimated by correcting the result of current integration according to the difference is obtained from only the result of current integration. The accuracy of the estimated state of charge may be rather low. Therefore, by selectively setting the Kalman gain to 0 according to the current measurement value, the temperature measurement value, or the previously estimated state of charge, estimating the state of charge of the battery 20 by relying only on the result of current integration using Equation 11 is can be good

Data representing the bypass condition including the determined current range, temperature range, or charge state range may be stored in advance in the memory unit 130 .

5 is a flowchart of a method for estimating the state of charge of the battery 20 according to an embodiment of the present invention. The method of FIG. 5 may be repeatedly executed every predetermined time Δt while charging and discharging of the battery 20 is in progress.

Referring to FIG. 5 , in step S500 , the control unit 120 , the previous current measurement value i _k-1 , the previous temperature measurement value T _{k-1 ,} and the previous internal product state x _{k-1 from the memory unit 130 . , +} is obtained. The previous current measurement value i _k-1 indicates the direction and magnitude of the current of the battery 20 measured by the sensing unit 110 at time step k-1. The temperature measurement value T _k-1 represents the temperature of the battery 20 measured by the sensing unit 110 at time step k-1. Said previous dot product state x _k-1,+ comprises an estimated state of charge SOC _k-1,+ of time step k-1 and an estimated polarization voltage V1 _k-1,+ .

In step S510 , the controller 120 predicts the dot product state and the error covariance matrix of the battery 20 . Specifically, the control unit 120 time-updates _{the previous inner product state x k-1,+} using Equation 11 to obtain the predicted inner product state x _{k,- including SOC k,-} and V1 _k,- Calculate. SOC _k,- is the predicted state of charge _{at time step k, and V1 k,-} is the predicted polarization voltage at time step k. In addition, the control unit 120 calculates P _k,- by timely _{updating P k-1,+ using Equation 2 .} P _k,- is the predicted error covariance matrix of time step k. _{SOC k,-} , V1 _k,- and P _k,- obtained as a result of the time update executed in step S520 are stored in the memory unit 130 by the control unit 120 .

In step S520 , the controller 120 determines a voltage measurement value V _k , a current measurement value i _{k ,} and a temperature measurement value T _k based on the sensing signal from the sensing unit 110 . The voltage measurement V _k , the current measurement value i _k and the temperature measurement value T _k correspond to the respectively measured terminal voltage, current and temperature at time step k. The controller 120 may _{store the voltage measurement value V k} , the current measurement value i _{k ,} and the temperature measurement value T _k in the memory unit 130 .

In step S530 , the controller 120 determines whether a predetermined bypass condition is satisfied by _{the current measured value i k} , the temperature measured value T _k or the predicted state of charge SOC _{k,− .} For example, the control part 120, the current measured value _k i in this case is within a predetermined current range (e. G., Less than 20A, more than 125A 160A), said temperature measurement value T _k is a predetermined temperature range (e.g., -15 Fig. less than or equal to 45 degrees), or when the predicted state of charge SOC _k,- is within a predetermined range of state of charge (eg, less than or equal to 0.08, greater than or equal to 0.93), it is determined that the bypass condition is satisfied. The fact that the bypass condition is satisfied may mean that the nonlinearity of the terminal voltage of the battery 20 may become excessively high due to _{the current measured value i k} , the temperature measured value T _{k ,} or the state of charge SOC _k,- have. When the result of step S530 is “NO”, the controller 120 estimates the state of charge of the battery 20 in the correction activation mode. When the result of step S530 is “YES”, the control unit 120 estimates the state of charge of the battery 20 in the correction deactivation mode. Steps S540, S542, S544, and S546 to be described below are steps that can be executed by the controller 120 in the correction activation mode, and steps S550, S552, and S554 to be described later are steps that can be executed by the controller 120 in the correction inactive mode admit.

In step S540 , the control unit 120 calculates a _{Kalman gain K k .} In this case, the controller 120 may calculate _{K k} based on the _{predicted error covariance matrix P k,- using Equation 5 .}

In step S542 , the controller 120 calculates _{a voltage predicted value V k,-} representing the predicted terminal voltage of the battery 20 . In this case, the controller 120 may calculate the voltage predicted value V _{k,- based on the predicted internal product state x k ,-} and the current measured value I _{k using Equation 13 .}

In step S544 , the controller 120 corrects the predicted inner product state x _{k,- based on the Kalman gain K k} , the voltage measurement value V _{k ,} and the voltage predicted value V _{k,- .} That is, the controller 120 may calculate the estimated inner product state x _k,+ . In this case, the control unit 120 calculates _{SOC k,+} by correcting _{SOC k,-} that is the result of the time update by Equation 1 (see Equation 11) using Equation 7 (see Equation 13-15). do. SOC _k,+ represents the state of charge of the battery 20 which is one of the components of the estimated inner product state x _{k,+ of time step k.} The difference between SOC _k,+ and SOC _k,- corresponds to the product between _{the difference between V k} and V _k,- of time step k and the Kalman gain K _k.

In step S546, the control unit 120 corrects the _{predicted error covariance matrix P k,-.} In this case, the controller 120 may calculate _{P k,+} by correcting _{P k,-} that is the result of the time update according to Equation 2 using Equation 8. P _k,+ is the estimated error covariance matrix of time step k. K _k , x _k,+ and P _k,+ are stored in the memory unit 130 by the control unit 120 .

In step S550, the control unit 120 sets the Kalman gain K _k to 0. Accordingly, unlike step S540, calculation of the _{Kalman gain K k using Equation 5 is omitted.}

In step S552, the control unit 120 estimates the state of charge and the polarization voltage from the predicted inner product state x _{k,- in response to the Kalman gain K k being set to 0, respectively.} That is, the control unit 120 sets the estimated inner product state x _k,+ of the time step k equal to the predicted inner product state x _{k,- of the time step k.} That is, x _k,+ = x _k,- . This is because, as K _k is set to 0 in step S550, the second term on the right side of Equation 7 (refer to Equation 15), K _k (y _k - V _k,- ), also becomes 0. In other words, since the voltage measurement V _k cannot influence the correction of the inner product state x _{k,- , x k,+} is determined solely by x _k-1,+ and i _k-1 .

In step S554 , the controller 120 initializes the predicted error covariance matrix P _k,- to be equal to the initial error covariance matrix P _{0 in response to the Kalman gain K k being set to 0 .} The initial covariance matrix P ₀ may be predetermined. Alternatively, the initial covariance matrix P ₀ may be the most recently corrected error covariance matrix P _m,+ (symbol m is an integer less than k). Referring to Equations 2 and 8, when the Kalman gain K _k is continuously set to 0, there is a fear that matrix components of _{P k,+} may excessively increase as the number of steps k increases. Accordingly, by executing step S554, it is possible to suppress an excessive increase in matrix components of _{P k,+ .}

In step S560 , the control unit 120 outputs _{data representing the estimated state of charge SOC k,+} to the communication unit 140 . Accordingly, a notification signal indicating the estimated state of charge SOC _k,+ of the time step k may be transmitted to an external device (eg, a vehicle-side controller) through the communication unit 140 .

6 is a flowchart of a method for estimating the state of charge of the battery 20 according to another embodiment of the present invention. The method of FIG. 6 may be repeatedly executed every predetermined time Δt while charging and discharging of the battery 20 is in progress.

Referring to FIG. 6 , in step S600 , the control unit 120 , the previous current measurement value i _k-1 , the previous temperature measurement value T _{k-1 ,} and the previous internal product state x _{k-1 from the memory unit 130 . , +} is obtained. The previous current measurement value i _k-1 indicates the direction and magnitude of the current of the battery 20 measured by the sensing unit 110 at time step k-1. The temperature measurement value T _k-1 represents the temperature of the battery 20 measured by the sensing unit 110 at time step k-1. Said previous dot product state x _k-1,+ comprises an estimated state of charge SOC _k-1,+ of time step k-1 and an estimated polarization voltage V1 _k-1,+ .

In step S610 , the controller 120 predicts the dot product state and the error covariance matrix of the battery 20 . Specifically, the control unit 120 time-updates _{the previous inner product state x k-1,+} using Equation 11 to obtain the predicted inner product state x _{k,- including SOC k,-} and V1 _k,- Calculate. SOC _k,- is the predicted state of charge _{at time step k, and V1 k,-} is the predicted polarization voltage at time step k. In addition, the control unit 120 calculates P _k,- by timely _{updating P k-1,+ using Equation 2 .} P _k,- is the predicted error covariance matrix of time step k. _{SOC k,-} , V1 _k,- and P _k,- obtained as a result of the time update executed in step S620 are stored in the memory unit 130 by the control unit 120 .

In step S620 , the controller 120 determines a voltage measurement value V _k , a current measurement value i _{k ,} and a temperature measurement value T _k based on the sensing signal from the sensing unit 110 . The voltage measurement V _k , the current measurement value i _k and the temperature measurement value T _k correspond to the respectively measured terminal voltage, current and temperature at time step k. The controller 120 may _{store the voltage measurement value V k} , the current measurement value i _{k ,} and the temperature measurement value T _k in the memory unit 130 .

In step S630 , the controller 120 determines whether a predetermined bypass condition is satisfied by _{the current measured value i k} , the temperature measured value T _{k ,} or the predicted state of charge SOC _{k,− .} For example, the control part 120, the current measured value _k i in this case is within a predetermined current range (e. G., Less than 20A, more than 125A 160A), said temperature measurement value T _k is a predetermined temperature range (e.g., -15 Fig. less than or equal to 45 degrees), or when the predicted state of charge SOC _k,- is within a predetermined range of state of charge (eg, less than or equal to 0.08, greater than or equal to 0.93), it is determined that the bypass condition is satisfied. The fact that the bypass condition is satisfied may mean that the nonlinearity of the terminal voltage of the battery 20 may become excessively high due to _{the current measured value i k} , the temperature measured value T _{k ,} or the state of charge SOC _k,- have. If the result of step S630 is &quot;YES&quot;, step S632 is executed.

In step S632, the control unit 120 determines whether the time count is less than a threshold count. When the result of step S630 is “NO” or when the result of step S632 is “NO”, the controller 120 estimates the state of charge of the battery 20 in the correction activation mode. When both the result of step S630 and the result of step S632 are "YES", the control unit 120 estimates the state of charge of the battery 20 in the correction deactivation mode. Steps S640, S642, S644, S646, and S648 to be described later are steps that can be executed by the controller 120 in the correction activation mode, and steps S650, S652, S654, and S656 to be described later are performed by the controller 120 in the correction inactivation mode. These are steps that can be implemented.

That the time count is equal to or greater than the threshold count means that the state of charge is estimated based only on the current integration (see Equation 11) over a long period of time (eg, 10 minutes or more), so the state of charge error according to the current integration is calculated using the equivalent circuit model (30) ) means that it is highly likely to be larger than the state of charge error caused by the inaccuracy of ). Therefore, if the result of step S632 is “NO” even if the result of step S630 is “YES”, then by estimating the state of charge of the battery 20 in the correction activation mode instead of the correction deactivation mode, the charging estimated based on current integration only A phenomenon in which the state error increases excessively can be suppressed.

In step S640 , the control unit 120 calculates a _{Kalman gain K k .} In this case, the controller 120 may calculate _{K k} based on the _{predicted error covariance matrix P k,- using Equation 5 .}

In step S642 , the controller 120 calculates _{a voltage predicted value V k,-} representing the predicted terminal voltage of the battery 20 . In this case, the controller 120 may calculate the voltage predicted value V _{k,- based on the predicted internal product state x k ,-} and the current measured value I _{k using Equation 13 .}

In step S644 , the controller 120 corrects the predicted inner product state x _{k,- based on the Kalman gain K k} , the voltage measurement value V _{k ,} and the voltage predicted value V _{k,- .} That is, the controller 120 may calculate the estimated inner product state x _k,+ . In this case, the control unit 120 calculates _{SOC k,+} by correcting _{SOC k,-} that is the result of the time update by Equation 1 (see Equation 11) using Equation 7 (see Equation 13-15). do. SOC _k,+ represents the state of charge of the battery 20 which is one of the components of the estimated inner product state x _{k,+ of time step k.} The difference between SOC _k,+ and SOC _k,- corresponds to the product between _{the difference between V k} and V _k,- of time step k and the Kalman gain K _k.

In step S646, the control unit 120 corrects the _{predicted error covariance matrix P k,-.} In this case, the controller 120 may calculate _{P k,+} by correcting _{P k,-} that is the result of the time update according to Equation 2 using Equation 8. P _k,+ is the estimated error covariance matrix of time step k. K _k , x _k,+ and P _k,+ are stored in the memory unit 130 by the control unit 120 .

In step S648, the control unit 120 initializes the time count to be equal to a predetermined initial value.

In step S650, the control unit 120 sets the Kalman gain K _k to 0. Accordingly, unlike step S640, calculation of the _{Kalman gain K k using Equation 5 is omitted.}

In step S652, the controller 120 estimates the state of charge and the polarization voltage from the predicted inner product state x _{k,- in response to the Kalman gain K k being set to 0, respectively.} That is, the control unit 120 sets the estimated inner product state x _k,+ of the time step k equal to the predicted inner product state x _{k,- of the time step k.} That is, x _k,+ = x _k,- . This is because, as K _{k is set to 0 in step S650, the second term K k} (y _k - V _k,- ) of the right side of Equation 7 (see Equation 15) also becomes 0. In other words, since the voltage measurement V _k cannot influence the correction of the inner product state x _{k,- , x k,+} is determined solely by x _k-1,+ and i _k-1 .

In step S654 , the controller 120 initializes the predicted error covariance matrix P _k,- to be equal to the initial error covariance matrix P _{0 in response to the Kalman gain K k being set to 0 .} The initial covariance matrix P ₀ may be predetermined. Alternatively, the initial covariance matrix P ₀ may be the most recently corrected error covariance matrix P _m,+ (symbol m is an integer less than k). Referring to Equations 2 and 8, when the Kalman gain K _k is continuously set to 0, there is a fear that matrix components of _{P k,+} may excessively increase as the number of steps k increases. Therefore, by executing step S654, it is possible to suppress an excessive increase in matrix components of _{P k,+ .}

In step S646, the control unit 120 increments the time count by a predetermined value.

In step S660 , the control unit 120 outputs _{data indicating the estimated state of charge SOC k,+} to the communication unit 140 . Accordingly, a notification signal indicating the estimated state of charge SOC _k,+ of the time step k may be transmitted to an external device (eg, a vehicle-side controller) through the communication unit 140 .

The embodiment of the present invention described above is not implemented only through the apparatus 100 and the method, and may be implemented through a program for realizing a function corresponding to the configuration of the embodiment of the present invention or a recording medium in which the program is recorded. And, such an implementation can be easily implemented by an expert in the technical field to which the present invention belongs from the description of the above-described embodiment.

In the above, although the present invention has been described with reference to limited embodiments and drawings, the present invention is not limited thereto and will be described below with the technical idea of the present invention by those of ordinary skill in the art to which the present invention pertains. Of course, various modifications and variations are possible within the scope of equivalents of the claims.

In addition, since the present invention described above is capable of various substitutions, modifications and changes within the scope that does not depart from the technical spirit of the present invention for those of ordinary skill in the art to which the present invention pertains, the above-described embodiments and attachments It is not limited by the illustrated drawings, and all or part of each embodiment may be selectively combined and configured so that various modifications may be made.

10: battery pack
20: battery
100: battery state of charge estimation device
110: sensing unit
120: control unit
130: memory unit
140: communication department

Claims (12)

An apparatus for estimating a state of charge of a battery using an extended Kalman filter, the apparatus comprising:
a sensing unit that measures the terminal voltage, current, and temperature of the battery and outputs a sensing signal indicating the measured terminal voltage, current, and temperature; and
a control unit including at least one processor configured to execute the extended Kalman filter and operatively coupled to the sensing unit;
The control unit, for each time step that increases by 1 in a predetermined time unit,
Predicting an inner product state and an error covariance matrix of the battery, wherein the predicted inner product state includes predicted values of each of a state of charge and a polarization voltage of the battery,
determining a voltage measurement value, a current measurement value, and a temperature measurement value corresponding to the measured terminal voltage, the measured current, and the measured temperature based on the sensing signal;
determine whether a predetermined bypass condition is satisfied by the current measurement value, the temperature measurement value, or the estimated value of the state of charge determined in a previous time step;
When the bypass condition is satisfied and the time count is less than a predetermined threshold count, in a correction deactivation mode, the Kalman gain of the extended Kalman filter is set to zero, and in response to the Kalman gain being set to zero, the predicted dot product The apparatus for estimating a state of charge of a battery, wherein the estimated values of the state of charge and the polarization voltage are determined to be the same as the predicted values of each of the state of charge and the polarization voltage of the battery included in the state, and the time count is increased by a predetermined value.
According to claim 1,
The control unit is
If the bypass condition is not satisfied, determining an estimate of the state of charge of the battery in a correction activation mode,
In the calibration activation mode, the Kalman gain is calculated based on the predicted error covariance matrix, and a voltage prediction value representing the predicted terminal voltage of the battery is calculated based on the predicted dot product state and the current measurement value And, each of the state of charge and the polarization voltage from the result of correcting the predicted values of the state of charge and the polarization voltage of the battery included in the predicted inner product state based on the voltage predicted value, the voltage measured value, and the Kalman gain determining an estimated value and correcting the predicted error covariance matrix based on the Kalman gain.
According to claim 1,
The control unit is
When the current measurement value is within a predetermined current range, when the temperature measurement value is within a predetermined temperature range, or when the estimated value of the state of charge determined at the previous time step is within a predetermined state of charge range, the bypass condition is satisfied A device for estimating a state of charge of a battery that determines that it is.
According to claim 1,
The control unit is
In the calibration deactivation mode, in response to the Kalman gain being set to 0, the predicted error covariance matrix is initialized to be equal to an initial error covariance matrix.
5. The method of claim 4,
The initial error covariance matrix is
An apparatus for estimating a state of charge of a battery, which is a predetermined matrix or a most recently corrected error covariance matrix.
delete 3. The method of claim 2,
The control unit is
When the estimated value of the state of charge of the battery is determined in the correction activation mode, the time count is initialized to be equal to a predetermined initial value.
According to claim 1,
a communication unit configured to transmit a notification signal indicating the estimated state of charge to an external device according to a predetermined communication rule;
Further comprising, a battery state of charge estimation device.
The apparatus for estimating the state of charge of the battery according to any one of claims 1 to 5, 7 and 8;
A battery pack comprising a.
A method for estimating a state of charge of a battery using an extended Kalman filter, the method comprising:
Predicting, by a control unit including at least one processor configured to execute the extended Kalman filter, an inner product state and an error covariance matrix including a state of charge and a polarization voltage of the battery, wherein the predicted inner product state of the battery including a predicted value of each of the state of charge and the polarization voltage;
The control unit, based on the sensing signal output by the sensing unit electrically connected to the battery, the voltage measurement value, the current measurement value and the temperature corresponding to the terminal voltage, current and temperature of the battery measured by the sensing unit determining a measurement value;
determining, by the control unit, whether a predetermined bypass condition is satisfied by the current measured value, the temperature measured value, or the estimated value of the state of charge determined in a previous time step; and
estimating, by the control unit, the state of charge of the battery in a correction deactivation mode when the bypass condition is satisfied and the time count is less than a predetermined threshold count;
The step of estimating the state of charge of the battery in the correction deactivation mode,
setting the Kalman gain to zero; and
in response to the Kalman gain being set to zero, determining an estimated value of each of the state of charge and the polarization voltage to be the same as the predicted values of each of the state of charge and the polarization voltage of the battery included in the predicted inner product state; and
and increasing the time count by a predetermined value.
11. The method of claim 10,
The step of estimating the state of charge of the battery in the calibration deactivation mode,
initializing, by the controller, the predicted error covariance matrix to be the same as an initial error covariance matrix in response to the Kalman gain being set to 0;
Further comprising, a method of estimating a battery state of charge.
12. The method of claim 11,
The initial error covariance matrix is
A method for estimating the state of charge of a battery, which is a predetermined matrix or a most recently corrected error covariance matrix.

Images