DEVICE FOR ASSESSING BATTERY STATUS
The battery status estimating device improves SOC and SOH estimation accuracy by combining current integration and OCV methods with feedback-based recursive updates, addressing errors in existing technologies.
Patent Information
- Application Number
- DE112016000834
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2015-02-19
- Filing Date
- 2016-01-22
- Publication Date
- 2024-11-07
- Estimated Expiration
- 2036-01-22
AI Technical Summary
Existing methods for estimating the state of charge (SOC) and state of health (SOH) of secondary batteries are prone to errors due to influences from offset currents, equivalent circuit parameters, and integration errors, leading to inaccurate estimates over time.
A battery status estimating device that combines current integration and open circuit voltage (OCV) estimation methods, using feedback information to recursively update SOC and equivalent circuit parameters, thereby correcting estimation errors.
The device provides more accurate estimation of battery status by eliminating the influence of offset currents and equivalent circuit parameter errors, enabling real-time detection of battery degradation and precise SOC estimation.
Smart Images

Figure 00000000_0000_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present invention relates to a battery status estimating device configured to estimate an internal status of the battery, and more particularly to a battery status estimating device configured to estimate an internal status of a secondary battery, such as the state of charge and the state of health. TECHNICAL BACKGROUND
[0002] A technology for accurately estimating an internal status of a secondary battery, such as a state of charge (SOC) and a state of health (SOH), is important for electric vehicles, railway vehicles, installed power storage systems, and other systems that efficiently use a secondary battery.
[0003] A well-known prior art example for SOC estimation is a current integration method for estimating the current SOC using an initial SOC value and an integrated value of the measured current. Another well-known example is an open-circuit voltage (OCV) estimation method for estimating the current SOC by estimating the battery's OCV based on a circuit model equivalent to the battery and using an OCV-SOC curve.
[0004] The current integration method and the OCV estimation method have different characteristics. Specifically, the current integration method can accurately track short-term changes in SOC, but is sensitive to the influence of errors in parameters such as the initial electricity quantity, SOH, and offset current. In particular, the current offset error is integrated, resulting in the SOC estimation accuracy deteriorating over time.
[0005] The OCV estimation method, on the other hand, primarily uses a measured voltage to estimate SOC. Therefore, unlike the current integration method, the error of each parameter is not accumulated. However, it is known that the OCV estimation method is highly sensitive to the influence of errors in the equivalent circuit parameters and the measured voltage. Therefore, the accuracy of the short-term estimation is poor, for example, due to the occurrence of an outlier in the SOC estimate.
[0006] Under these circumstances, numerous SOC estimation methods have been proposed, which feature a suitable use of the current integration method in combination with the OCV estimation method to achieve the advantages of both methods while compensating their disadvantages.
[0007] A specific example of an SOC estimation method is a technology for calculating a final SOC estimation value by weighting and combining an SOC estimation value estimated by the current integration method and an SOC estimation value estimated by the OCV estimation method depending on the battery consumption (see, for example, Patent Literature 1).
[0008] Furthermore, as described above, technology for accurate SOH estimation is also important. An accurate SOH estimation enables the timing of secondary battery replacement to be determined appropriately and leads to an improvement in the accuracy of SOC estimation.
[0009] A specific example of an SOH estimation method is a technology for calculating an SOH estimation value (i.e., a battery capacity estimation value) that is not sensitive to the current integration error by using SOC deviation amounts of the current integration method and the OCV estimation method in a period in which a charge / discharge current exceeds a predetermined threshold (see, for example, Patent Literature 2).
[0010] In addition, reference is made to Patent Literature 3, which describes a nonlinear adaptive observer approach for estimating the battery state of charge. Patent Literature 4 describes a method for estimating the battery state of charge using a recursive estimator based on current integration and an OCV model. STATE OF THE TECHNICAL PATENT LITERATURE PTL 1: JP 4 583 765 B2 PTL 2: JP 5 419 832 B2 PTL 3: DE 10 2012 210 883 A1 PTL 4: JP 2013- 238 402 A NON-PATENT LITERATURE
[0011] NPL 1: Shuichi Adachi, “Basics of system identification”, Tokyo, Denki University Press, September 2009, pp. 170-177. SUMMARY OF THE INVENTION TECHNICAL PROBLEM
[0012] However, the prior art suffers from the following problems. In the prior art disclosed in Patent Literature 1, the weighting when combining the two SOCs is determined based on the magnitude of a moving average current. This results in the current offset error significantly affecting the final SOC estimate when the current varies greatly over a period of time.
[0013] Furthermore, even if the moving average value varies appropriately, the errors due to SOC and current offset for the SOC estimate measured by the current integration method cannot be completely eliminated as long as an approach of weighting and combining two SOC estimates is applied.
[0014] In the prior art disclosed in Patent Literature 2, the SOH estimation value is highly sensitive to the influences of current integration errors and equivalent circuit parameters when the current value does not exceed the threshold value for a sufficient time during the integration period. Therefore, the influence of a current integration error cannot be completely eliminated unless a method is adopted to improve the SOH estimation accuracy by reducing the current integration error.
[0015] The present invention has been conceived to solve the above-mentioned problems and has an object to provide a battery status estimating device capable of estimating the internal status of a battery more accurately than the prior art. SOLUTION TO THE PROBLEM
[0016] The object underlying the invention is achieved by a device for estimating a battery status having the features of independent patent claim 1 and independent patent claim 7, respectively. Advantageous developments of the device for estimating a battery status according to the invention are specified in dependent claims 1 to 6 and 8, respectively. ADVANTAGEOUS EFFECTS OF THE INVENTION
[0017] According to the present invention, it is possible to provide a battery status estimation device capable of estimating the internal status of a battery more accurately than the prior art. To this end, the battery status estimation device is configured, in a basic configuration in which the current integration method and the OCV estimation method are used in combination, to estimate the SOC of the secondary battery using the new SOC estimation method that includes the use of feedback information based on the estimation result. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a configuration drawing of a battery status estimating apparatus according to a first embodiment of the present invention. Fig. 2 is a configuration drawing of an SOC estimation unit according to a first embodiment of the present invention. Fig. 3 is a circuit diagram illustrating a configuration example of an equivalent circuit model of a secondary battery for use with an OCV estimation unit according to the first embodiment of the present invention. Fig. 4 is a configuration drawing of an OCV estimation unit according to the first embodiment of the present invention. Fig. 5 is a flowchart illustrating a series of processes to be executed by the battery status estimating apparatus according to the first embodiment of the present invention. Fig. 6 is a flowchart illustrating a series of processes to be executed by the SOC estimation unit according to the first embodiment of the present invention. Fig. 7 is a flowchart illustrating a series of processes to be executed by the OCV estimation unit according to the first embodiment of the present invention. Fig. 8 is a configuration drawing illustrating a battery status estimating apparatus according to a second embodiment of the present invention. Fig. 9 is a flowchart illustrating a series of processes to be executed by the battery status estimating apparatus according to the second embodiment of the present invention. Fig. 10 is a configuration drawing of a battery status estimating apparatus according to a third embodiment of the present invention. Fig. 11 is a configuration drawing of a feedback SOC estimation unit according to the third embodiment of the present invention. Fig. 12 is a flowchart illustrating a series of processes to be executed by the battery status estimating apparatus according to the third embodiment of the present invention. Fig. 13 is a flowchart illustrating a series of processes to be executed by the feedback SOC estimation unit according to the third embodiment of the present invention. DESCRIPTION OF THE EMBODIMENTS
[0018] A battery status estimation device according to preferred embodiments of the present invention will be described below with reference to the accompanying drawings. In the drawings, identical or corresponding components are denoted by the same reference numerals. Redundant descriptions are omitted. The battery status estimation device according to the present invention is configured to estimate an internal status of a secondary battery; in particular, to estimate the internal status of the secondary battery during operation in, for example, electric vehicles, railway vehicles, installed energy storage systems, and other systems. First embodiment
[0019] Fig. 1 is a configuration drawing of a battery status estimation device 100 according to a first embodiment of the present invention. Fig. 1 also shows a secondary battery 101 that can be connected to the battery status estimation device 100. Assuming that the secondary battery 101 is a lithium-ion battery, the battery status estimation device 100 will be described below. The configuration of the secondary battery 101 generally includes chargeable / dischargeable batteries. The secondary battery 101 can also be, for example, a lead-acid storage battery, a nickel-hydrogen battery, or an electrochemical double-layer capacitor.
[0020] In Fig. 1, the battery status estimation device 100 includes a current detection unit 102, a voltage detection unit 103, an offset current subtraction unit 104, an SOC estimation unit 105, an OCV estimation unit 106, a SOC-OCV conversion unit 107, an OCV-SOC conversion unit 108, an SOC subtraction unit 109, an OCV subtraction unit 110, a first recursive estimation unit 111, and a second recursive estimation unit 112.
[0021] The current detection unit 102 is configured to perform current detection processing. Specifically, the current detection unit 102 detects a charge / discharge current of the secondary battery 101 as a detected current I and outputs the detected current I to the offset current subtraction unit 104 and the SOC estimation unit 105.
[0022] The voltage detection unit 103 is configured to perform voltage detection processing. Specifically, the voltage detection unit 103 detects a voltage between terminals during charging / discharging of the secondary battery 101 as a detected voltage V and outputs the detected voltage V to the OCV estimation unit 106.
[0023] The offset current subtraction unit 104 is configured to perform offset current subtraction processing. Specifically, the offset current subtraction unit 104 subtracts an offset current I off input from the first recursive estimation unit 111 from the detected current I input from the current detection unit 102, and outputs the determined value as a corrected current I' to the OCV estimation unit 106 and the second recursive estimation unit 112.
[0024] The SOC estimation unit 105 is configured for SOC estimation processing using a current integration method. Specifically, the SOC estimation unit 105 uses the current integration method to estimate a first state of charge SOC1. Specifically, the SOC estimation unit 105 calculates the first state of charge SOC1, an elapsed time t, and k and an integrated electricity quantity Qc based on the detected current I input from the current detection unit 102, a health state SOH of the secondary battery 101, the offset current I off and an initial electricity quantity error ΔQ C0 the secondary battery 101, which are input from the first recursive estimation unit 111.
[0025] The parameters required to calculate the SOC using the current integration method are referred to as "SOC estimation parameters." In this description, an example is given in which the SOC estimation parameters are the state of health SOH, the offset current I off and the initial electricity quantity error ΔQ C0 are.
[0026] The SOC estimation unit 105 outputs the calculated first state of charge SOC1 to the SOC-OCV conversion unit 107 and the SOC subtraction unit 109, and outputs the calculated elapsed time and the integrated electricity quantity Q C to the first recursive estimation unit 111.
[0027] Furthermore, the SOC of the secondary battery 101 estimated by the battery status estimation device 100 is the first state of charge SOC1 output by the SOC estimation unit 105. In other words, the first state of charge SOC1 is output by the battery status estimation device 100 as a final result of estimating the SOC of the secondary battery 101.
[0028] The following describes the rules for calculating the first state of charge SOC1 by the SOC estimation unit using formulas.
[0029] When the SOC of the secondary battery 101 is estimated using the current integration method, which is the conventional way, an SOC estimated value SOCc can be represented by the expression (1) or (2), where SOCc represents the estimated value of the SOC and Qc represents the estimated value of the integrated electricity amount. SOCC(k+1)=SOCC(k)+tsFCCI(k) {QC(k+1)=QC(k)+tsI(k)SOCC(k)=1FCCQC(k)
[0030] In the above expressions, t S represents a sampling period, k represents a sampling time, and FCC represents a full charge capacity (FCC). The FCC can be represented by expression (3) using an initial full charge capacity FCC0 and the health state SOH. FCC = FCC0×SOH
[0031] In the following, the actual value of the electricity quantity of the secondary battery 101 is represented by Q*. The representation of Q* in terms of the integrated electricity quantity Qc will be discussed.
[0032] The detected current I includes the offset current I off , which is a constant, as the detection error of the current detection unit 102. Furthermore, the initial value of the integrated electricity quantity Qc includes the initial electricity quantity error ΔQ C0(=Qc (0) - Q* (0)), measured as a function of Q*. Therefore, it can be assumed that Q* can be determined by calculating the offset current I off corresponding electricity quantity and the initial electricity quantity error ΔQ C0 be subtracted from the integrated amount of electricity Qc. This means that expression (4) is satisfied. Q*(k)=QC(k)−tkIoff−ΔQC0
[0033] In expression (4) t k = kt s fulfilled, where t k represents the elapsed time from an initial time corresponding to k=0. Specifically, the elapsed time t k a time that has passed since the start of the estimation of the SOC of the secondary battery 101; more precisely, it represents a time that has passed since the start of the execution of a functional sequence of a later-described Fig. 5 shown flowchart, has passed.
[0034] In order to correct the second equation according to expression (2) using expressions (3) and (4), the first state of charge SOC1 is further represented by the following expression. SOC1(k)=1FCC0⋅SOH(QC(k)−tkIoff−ΔQC0)
[0035] The SOC estimation unit 105 calculates the first state of charge SOC1 based on the expression (5). As can be seen from the expression (5), the SOC estimation unit 105 can calculate an accurate SOC whose error is immediately eliminated when the SOC estimation unit 105 uses the precisely estimated SOC estimation parameters (namely, the state of health SOH, the offset current I off and the initial electricity quantity error ΔQ C0 ).
[0036] Next, a specific configuration example of the SOC estimation unit 105 will be described with reference to Fig. 2 described. Fig. 2 is a configuration drawing of the SOC estimation unit 105 according to the first embodiment of the present invention. Fig. 2, the SOC estimation unit 105 includes a coefficient multiplier 201, an integrator 202, a value storage unit 203, an integrator 204, a value storage unit 205, a multiplier 206, a subtractor 207, a value storage unit 208, a subtractor 209, a coefficient multiplier 210, and a coefficient multiplier 211.
[0037] The coefficient multiplier 201 is configured for sampling period multiplication processing. Specifically, the coefficient multiplier 201 multiplies the detected current I input by the current detection unit 102 by the sampling period t. s and outputs the determined value to the integrator 202.
[0038] The integrator 202 is configured for integrated electricity quantity calculation processing. Specifically, the integrator 202 adds an output value of the integrator 202 at a previous time to an output value of the coefficient multiplier 201 to calculate the integrated electricity quantity Qc and outputs the calculated integrated electricity quantity Qc. In short, the integrator 202 calculates the first equation from Expression (2). When the SOC estimation is started, the first state of charge SOC1 obtained at the end of the previous SOC estimation can be multiplied by the full charge capacity FCC to be used as the value of Qc (0).
[0039] In other cases, if a long time has passed since the end of the previous SOC estimation, the overvoltage of the secondary battery 101 is assumed to be sufficiently low. Therefore, the voltage V detected by the voltage detection unit 103 can be assumed to be the open-circuit voltage of the secondary battery 101. Therefore, it is possible to use the OCV-SOC conversion unit 108 described below to convert the open-circuit voltage to the SOC and use the value obtained by multiplying the converted SOC by the full charge capacity FCC as the value for Q. C (0) to use.
[0040] The value storage unit 203 is configured to process the stored sampling period output. Specifically, the value storage unit 203 stores the sampling period t s and returns the stored sampling period t s to the integrator 204.
[0041] The integrator 204 is configured to process the calculation of the elapsed time. Specifically, the integrator integrates the output value of the value storage unit 203 (namely, the sampling period t s ) to calculate the elapsed time t k and returns the calculated elapsed time t k The integrator 204 sets the initial value of t k to 0 before starting to integrate the output value of the value storage unit 203.
[0042] The value storage unit 205 is configured to process the stored current offset output. Specifically, the value storage unit 205 stores the offset current I off and outputs the stored offset current I off to the multiplier 206.
[0043] The multiplier 206 is configured to process the multiplication of elapsed time. Specifically, the multiplier 206 multiplies the output value of the integrator 204 (namely, the elapsed time t k ) with the output value of the value storage unit 205 (namely the offset current I off ) and outputs the determined value to the subtractor 207.
[0044] Subtractor 207 is configured for offset subtraction processing. Specifically, subtractor 207 subtracts the output value of multiplier 206 from the output value of integrator 202 and outputs the resulting value to subtractor 209.
[0045] The value storage unit 208 is configured to process the stored initial electricity quantity error output. Specifically, the value storage unit 208 stores the initial electricity quantity error ΔQ C0 and returns the stored initial electricity quantity error ΔQC0 to the subtractor 209.
[0046] The subtractor 209 is configured for initial electricity quantity error subtraction processing. Specifically, the subtractor 209 subtracts the output value of the value storage unit 208 from the output value of the subtractor 207 and outputs the obtained value to the coefficient multiplier 210.
[0047] The coefficient multiplier 210 is configured to process the initial FCC inverse multiplication. Specifically, the coefficient multiplier 210 multiplies the output value of the subtractor 209 by the inverse of the initial full load capacity FCC0 and outputs the determined value to the coefficient multiplier 211. The inverse of the initial full load capacity FCC0 is hereinafter referred to as the "reciprocal FCC0 -1 " designated.
[0048] The coefficient multiplier 211 is configured for SOH inverse multiplication processing. Specifically, the coefficient multiplier 211 multiplies the output value of the coefficient multiplier 210 by the inverse of the health state SOH and outputs the determined value as the first charge state SOC1. The inverse of the health state SOH is hereinafter referred to as the "reciprocal SOH." -1 " designated.
[0049] Respective values stored in the value storage unit 205 and the value storage unit 208, as well as the state of health SOH used by the coefficient multiplier 211 to calculate the first state of charge SOC1, are repeatedly updated by using the output by the first recursive estimation unit 111 described later.
[0050] As described in Fig. 1, the OCV estimation unit 106 is configured to perform OCV estimation processing using an OCV estimation method. Specifically, the OCV estimation unit 106 uses the OCV estimation method to estimate the second open-circuit voltage OCV2 of the secondary battery 101. Specifically, the OCV estimation unit 106 calculates the second open-circuit voltage OCV2 and a voltage drop amount v of the secondary battery 101 based on the corrected current I' input from the offset current subtraction unit 104, the detected voltage V input from the voltage detection unit 103, and equivalent circuit parameters input from the second recursive estimation unit 112.
[0051] The OCV estimation unit 106 outputs the calculated second open-circuit voltage OCV2 to the OCV-SOC conversion unit 108 and the OCV subtraction unit 110, and outputs the calculated voltage drop amount v to the second recursive estimation unit 112.
[0052] In the following, rules for calculating the second open-circuit voltage OCV2 by the OCV estimation unit 106 by using formulas with reference to Fig. 3. described. Fig. 3 is a circuit diagram illustrating an example configuration of an equivalent circuit model of the secondary battery 101 that can be applied to the OCV estimation unit 106 according to the first embodiment of the present invention.
[0053] The second open-circuit voltage OCV2 is determined by subtracting an overvoltage η of the secondary battery 101 from the detected voltage V corresponding to the voltage between the terminals of the secondary battery 101. That is, expression (6) is satisfied. OCV2(k)=V(k)−η(k)
[0054] It is also possible to approximately calculate the overvoltage η by using the equivalent circuit model of the secondary battery 101.
[0055] In Fig. In Figure 3, R0 represents a DC resistance, which takes into account a solution resistance and a charge-transfer resistance with a small time constant. Furthermore, R1 represents a diffusion resistance, and C1 represents an electrochemical double-layer capacitance. In this case, R0, R1, and C1 are the equivalent circuit parameters. Furthermore, the voltage drop v is a voltage drop measured for the parallel connection region of R1 and C1.
[0056] Based on the Fig. 3, the overvoltage η of the secondary battery 101 is further calculated with the voltage drop amount v as the state variable of the equivalent circuit according to expression (7). {v(k+1)=e−tsR1C1v(k)+R1(1−e−tsR1C1)I'(k)η(k)=R0I'(l)+v(k)
[0057] In the first embodiment, the Fig. 3 is assumed. The configuration of the equivalent circuit model is not limited to the Fig. 3. For example, diffusion can be represented by a multi-stage RC circuit or the charge transfer resistance can be separated from the DC resistance if the sampling period t s is short enough. In this way, it is possible to apply various configuration forms of the equivalent circuit model of the secondary battery 101.
[0058] In this description, a case is further illustrated in which the state quantity of the equivalent circuit is the voltage drop amount v, but a charge amount q of a capacitor C1 can be assumed as the state quantity of the equivalent circuit instead of the voltage drop amount v.
[0059] In the following, a specific configuration example of an OCV estimation unit 106 will be described with reference to Fig. 4 described. Fig. 4 is a configuration drawing of the OCV estimation unit 106 according to the first embodiment of the present invention. Fig. 4, the OCV estimation unit 106 includes an overvoltage calculation unit 401 and a subtractor 402.
[0060] The overvoltage calculation unit 401 is configured for overvoltage calculation processing. Specifically, the overvoltage calculation unit 401 calculates the voltage drop amount v and the overvoltage η according to Expression (7) based on the corrected current I' input by the offset current subtraction unit 104. Furthermore, the overvoltage calculation unit 401 outputs the calculated voltage drop amount v to the second recursive estimation unit 112 and outputs the calculated overvoltage η to the subtractor 402.
[0061] The subtractor 402 is configured for overvoltage subtraction processing. Specifically, the subtractor 402 subtracts the overvoltage η from the detected voltage V input by the voltage detection unit 103 and outputs the determined value as the second open-circuit voltage OCV2.
[0062] The equivalent circuit parameters used by the overvoltage calculation unit 401 to calculate the voltage drop amount v and the overvoltage η according to Expression (7) are recursively updated using the output of the second recursive estimation unit 112 described later.
[0063] As explained with reference to the description of the Fig. 1, the SOC-OCV conversion unit 107 is configured to perform SOC-OCV conversion processing. Specifically, the SOC-OCV conversion unit 107 converts the first state of charge SOC1 input by the SOC estimation unit 105 into a first open-circuit voltage OCV1 of the secondary battery 101 and outputs the first open-circuit voltage OCV1 obtained by the conversion to the OCV subtraction unit 110.
[0064] The relationship between the SOC and the open-circuit voltage depends little on the temperature and degree of degradation of the secondary battery 101. Therefore, the SOC-OCV conversion unit 107 converts the open-circuit voltage to the SOC based on the previously acquired measurement data. Specifically, the SOC-OCV conversion unit 107 generates, for example, a linear interpolation function or an approximate curve based on a limited number of measurement data and uses the function or curve to convert the open-circuit voltage to the SOC.
[0065] The OCV-SOC conversion unit 108 is configured for OCV-SOC conversion processing. Specifically, the OCV-SOC conversion unit 108 converts the second open-circuit voltage OCV2 input from the OCV estimation unit 106 into a second state of charge SOC2 of the secondary battery 101 and outputs the converted second state of charge SOC2 to the SOC subtraction unit 109.
[0066] Furthermore, the OCV-SOC conversion unit 108 may use a similar function as that of the SOC-OCV conversion unit 107 to convert the second open circuit voltage OCV2 input from the OCV estimation unit 106 into the second state of charge SOC2.
[0067] The SOC subtraction unit 109 is configured for SOC error calculation processing. Specifically, the SOC subtraction unit 109 subtracts the first state of charge SOC1 input from the SOC estimation unit 105 from the second state of charge SOC2 input from the OCV-SOC conversion unit 108, and outputs the obtained value to the first recursive estimation unit 111 as an SOC error ε1.
[0068] The OCV subtraction unit 110 is configured for OCV error calculation processing. Specifically, the OCV subtraction unit 110 subtracts the first open-circuit voltage OCV1 input from the SOC-OCV conversion unit 107 from the second open-circuit voltage OCV2 input from the OCV estimation unit 106, and outputs the obtained value to the second recursive estimation unit 112 as an open-circuit voltage error ε2.
[0069] The first recursive estimation unit 111 is configured for recursive estimation processing for the current integration correction, recursively updating estimated values of the SOC estimation parameters. Specifically, the first recursive estimation unit 111 uses a recursive estimation method based on the elapsed time t. k and the integrated electricity quantity Qc inputted from the SOC estimation unit 105 and that inputted by the SOC subtraction unit 109, and the SOC error ε1, to thereby calculate the estimated values of the SOC estimation parameters.
[0070] In addition, the first recursive estimation unit 111 outputs the calculated SOC estimation parameters to the SOC estimation unit 105 and outputs the offset current I alone in addition to the calculated SOC estimation parameters. off to the offset current subtraction unit 104.
[0071] The recursive estimation method used by the first recursive estimation unit 111 may be, for example, a recursive least squares (RLS) method (see Non-Patent Literature 1).
[0072] The recursive estimation method is not limited to RLS and can be, for example, a recursive total least squares (RTLS) method, a recursive partial least squares (RPLS) method, or a Kalman filter.
[0073] The update formula for RLS is formed according to the following expression (8). θ1(k)=θ1(k−1)+P1(k−1)ϕ1(k)λ1+ϕ1T(k)P1(k−1)ϕ1(k)∈1(k)P1(k)= 1λ1{P1(k−1)−P1(k−1)ϕ1(k)ϕ1T(k)P1(k−1)λ1+ϕ1T(k)P1(k−1)ϕ1(k)}
[0074] In the expression (8), where N represents the number of independent variables, θ1, φ1, P1, ε1, and λ1 each represent a vector of independent variables of dimension N×1, an N×1 vector, an N×N covariance matrix, an error signal, and a forgetting coefficient.
[0075] For example, φ1 and θ1 can be determined by the following expression (9). ϕ1(k)=[QC(k) / FCC0 −tk / FCC −1]Tθ1(k)=[1 / SOH(k) Ioff(k) / SOH(k) ΔQC0(k) / (FCC0⋅SOH(k))]T
[0076] Based on the expression (5), the representation of SOC1(k) = φ1 T (k)θ1(k-1) is obtained. Therefore, it can be understood that the SOC error ε1, which is an output value of the SOC subtraction unit 109, has a second state of charge SOC2 as its teaching signal and serves as the prediction error of RLS.
[0077] θ1(k), determined from the first equation of expression (8), is used to solve the second equation according to expression (9). This results in the estimated values of the SOC estimation parameters, namely SOH(k), I off (k) and ΔQ C0 (k), is determined.
[0078] The initial value θ1(0) of the independent variable vector can be determined by using preliminary information about SOH(0), I off (0) and ΔQ C0 (0). If no preliminary information is known, for example, SOH(0) = 1, I off (0) = 0 and ΔQ C0 (0) = 0.
[0079] In this description, an example case is shown in which the SOC estimation parameters are the health state SOH, the offset current I off and the initial electricity quantity error ΔQ C0 However, the SOC estimation parameters are not limited to these values.
[0080] Specifically, this description illustrates a case where the health state SOH is given as an error factor expressed according to Expression (5) to estimate the health state. However, instead of the health state SOH, the full charge capacity FCC can be given as an error factor to estimate the full charge capacity FCC.
[0081] In addition, this description describes an example case in which the initial electricity quantity error ΔQ C0 is given as an error factor of the current integration method to calculate the initial electricity quantity error ΔQ C0 However, instead of an initial electricity quantity error ΔQ C0 also an initial amount of electricity Q C0 as an error factor of the current integration method to determine the initial amount of electricity Q C0 to appreciate.
[0082] In addition, this description describes an example case in which the initial electricity quantity error ΔQ C0 is given as the error factor of the current integration method to calculate the initial electricity quantity error ΔQ C0 Instead of the initial electricity quantity error ΔQ C0 However, an initial state of charge SOC ini as an error factor of the current integration method to determine the initial state of charge SOC ini to appreciate.
[0083] This description also shows an example case in which the initial electricity quantity error ΔQ C0 is given as an error factor of the current integration method to calculate the initial electricity quantity error ΔQ C0 to be estimated. An initial SOC error ΔSOC ini can still be considered as an error factor of the current integration method instead of the initial electricity quantity error ΔQ C0be given to determine the initial SOC error ΔSOC ini to appreciate
[0084] To give a specific example: If the full charge capacity FCC is estimated instead of the state of health SOH and the initial state of charge SOC ini instead of the initial electricity quantity error ΔQ C0 is estimated, the first state of charge SOC1 is calculated according to expression (10), which replaces expression (5). SOC1(k)=1FCC(QC(k)−tkIoff)+SOCini
[0085] If the initial electricity quantity error ΔQ C0 is sufficiently small, only the health status SOH and the offset current I off be estimated without the initial electricity quantity error ΔQ C0 to appreciate.
[0086] In this way, it is possible to estimate the first state of charge SOC1 as long as the SOC estimation parameters include at least the state of health SOH and the offset current I offfrom the parameters health status SOH, full charge capacity FCC, offset current I off , the initial amount of electricity Q C0 , the initial electricity quantity error ΔQ C0 , the initial state of charge SOC ini and the initial SOC error ΔSOC ini include.
[0087] In other cases, it is also possible to estimate the first state of charge SOC1 as long as the SOC estimation parameters include at least the full charge capacity FCC and the offset current I off under the parameters health status SOH, full charge capacity FCC, offset current I off , the initial amount of electricity Q C0 , the initial electricity quantity error ΔQ C0 , the initial state of charge SOC ini and the initial SOC error ΔSOC ini include.
[0088] The second recursive estimation unit 112 is configured for recursive estimation processing for equivalent circuit correction, recursively updating estimated values of the equivalent circuit parameters. Specifically, the second recursive estimation unit 112 uses a recursive estimation method based on the corrected current I' input from the offset current subtraction unit 104, the voltage drop amount v input from the OCV estimation unit 106, and the open-circuit voltage error ε2 input from the OCV subtraction unit 110, to thereby calculate the estimated values of the equivalent circuit parameters. Furthermore, the second recursive estimation unit 112 outputs the calculated equivalent circuit parameters to the OCV estimation unit 106.
[0089] The recursive estimation method to be used by the second recursive estimation unit 112 is, for example, RLS, similar to the case of the first recursive estimation unit 111. In this case, the update formula for RLS is constructed as shown in the following expression (11). θ2(k)=θ2(k−1)+P2(k−1)ϕ1(k)λ2+ϕ2⊤(k)P2(k−1)ϕ2(k)ε2(k)P2(k)= 1λ2{P2(k−1)−P2(k−1)ϕ2(k)ϕ2⊤(k)P2(k−1)λ2+ϕ2⊤(k)P2(k−1)ϕ2(k)}
[0090] In the expression (11), where N represents the number of independent variables, θ2, φ2, P2, ε2, and λ2 each represent an N×1 vector of independent variables, an N×1 vector, an N×N covariance matrix, an error signal, and a forgetting coefficient.
[0091] Furthermore, φ2 and θ2 can be determined, for example, by the following expression (12). ϕ2(k)=[I'(k)v(k)I'(k−1)]⊤θ2(k)=[R0(k)e−tsR1(k)C1(k)R1(k)(1−e−tsR1(k)C1(k))]⊤
[0092] In this case, based on expressions (6) and (7), expression (13) is satisfied. ε2(k)=OCV2(k)−OCV1(k)=V(k)−OCV1(k)−η(k)=V(k)−OCV1(k)−R0(k−1)I'(k) −e−tsR1(k−1)C1(k−1)v(k−1)−R1(k−1)(1−e−tsR1(k−1)C1(k−1))I'(k−1)=V(k)−OCV1(k)−ϕ2⊤(k)θ2(k−1)
[0093] As is clear from expression (13), (V-OCV1) is the learning signal of the open circuit voltage error ε2, which is an output value of the OCV subtraction unit 110, and serves as a prediction error for RLS.
[0094] The θ2(k) obtained from the first equation in expression (11) is used to solve the second equation in expression (12). From this, the estimated values of the equivalent circuit parameters, namely R0(k), R1(k), and C1(k), are obtained.
[0095] The initial value θ2(0) of the independent variable vector can be determined, for example, by holding data depending on the battery temperature for each parameter as prior information about R0(0), R1(0) and C1(0) and inputting the battery temperature of the secondary battery 101.
[0096] As can be seen from expression (12), the second recursive estimation unit 112 is configured to estimate all equivalent circuit parameters. However, the second recursive estimation unit 112 may also be configured to estimate only R0 and R1 of the equivalent circuit parameters by fixing a part of the parameters, for example, a time constant τ (= C1R1) of the equivalent circuit.
[0097] In this case, for example, associated data for each temperature or degree of degradation of the secondary battery 101 is used to estimate the value of the time constant τ. Furthermore, the following expression (14) is used instead of expression (7). {q(k+1)=e−tsτq(k)+τ(1−e−tsτ)I'(k)η(k)=R0I'(k)+R1τq(k)
[0098] When using expression (14) in this way, the following expression (15) is used instead of expression (12). In this case, the charge quantity q of the capacitor C1 is the state quantity of the equivalent circuit instead of the voltage drop amount v. ϕ2(k)=[I'(k)q(k) / τ]⊤θ2(k)=[R0(k)R1(k)]⊤
[0099] In the following, with reference to the flow charts of the Fig. 5 to Fig. 7 describes a series of processes for estimating the SOC of the secondary battery 101, which are executed by the battery status estimating apparatus 100 according to the first embodiment. Fig. 5 is a flowchart showing a series of processes to be executed by the battery status estimating apparatus 100 according to the first embodiment of the present invention.
[0100] Fig. 6 is a flowchart illustrating a series of processes to be executed by the SOC estimation unit 105 according to the first embodiment of the present invention. Fig. 7 is a flowchart showing a series of processes to be executed by the OCV estimation unit 106 according to the first embodiment of the present invention.
[0101] A series of steps S102 to S112 of arithmetic processing carried out in Fig. 5 corresponds to the arithmetic processing of a period or time period performed by the battery status estimation device 100. This arithmetic processing is performed with a period corresponding to the sampling period t s repeated.
[0102] Furthermore, a number of Fig. 6 shown steps S201 to S211 of an arithmetic processing which is carried out in the Fig. 5 is to be executed. In addition, a series of Fig. 7 shown steps S401 and S402 of arithmetic processing of the arithmetic processing shown in the Fig. 5 is executed.
[0103] In the flowcharts of the Fig. 5 to Fig. 7, the number of each step corresponds to the respective component of the device 100 for estimating of the battery status. In other words: As described above, in the flowcharts of the Fig. 5 to Fig. 7 each component of the battery status estimating device 100 executes a step having the same number as the component.
[0104] As in Fig. 5, the battery status estimating device 100 executes the series of steps S102 to S112 of the arithmetic processing in the sampling period t s out of.
[0105] Furthermore, in step S105, the battery status estimating device 100 executes the series of steps S201 to S211 of the arithmetic processing shown in Fig. 6. In addition, the battery status estimating device 100 executes the series of steps S401 and S402 of the Fig. 7 shows the arithmetic processing.
[0106] With regard to the steps of the flowcharts of the Fig. 5 to Fig. 7, the order of execution by the battery status estimating device 100 is not limited to that shown in the Fig. 5 to Fig. 7. It can be changed as long as the dependency relationship between the steps is maintained.
[0107] As described above, the battery status estimation device according to the first embodiment, in the first configuration, includes: the first recursive estimation unit configured to estimate and output the SOC estimation parameters; the SOC estimation unit configured to calculate and output the first state of charge, elapsed time, and integrated electricity amount measured since the start of estimation of the SOC of the secondary battery based on the detected current input from the current detection unit and the SOC estimation parameters input from the first recursive estimation unit;the offset current subtraction unit configured to output, as the corrected current, a value obtained by subtracting the offset current input from the first recursive estimation unit from the detected current input from the current detection unit; the OCV estimation unit configured to calculate and output the second open-circuit voltage based on the corrected current input from the offset current subtraction unit and the detected voltage input from the voltage detection unit; the OCV-SOC conversion unit configured to convert the second open-circuit voltage input from the OCV estimation unit into the second state of charge for output;and the SOC subtraction unit configured to output, as the SOC error, a value obtained by subtracting the first state of charge input by the SOC estimation unit from the second state of charge input by the OCV-SOC conversion unit;
[0108] In the first configuration, the first recursive estimation unit is configured to recursively estimate and update the SOC estimation parameters based on the elapsed time and integrated electricity quantity input by the SOC estimation unit, and the SOC error input by the SOC subtraction unit. Furthermore, the SOC estimation unit is configured to output, as the SOC of the secondary battery, the first state of charge calculated using the SOC estimation parameters updated by the first recursive estimation unit.
[0109] In this way, the problem of SOC estimation is reduced to the problem of estimating the parameters of the current integration method (namely, offset current, SOH, and initial electricity quantity error). Each parameter is estimated reliably and accurately. As a result, the SOC estimated by the current integration method also assumes reliable and accurate values.
[0110] Therefore, based on the SOH and the resistance values of the equivalent circuit, it is possible to detect the degree of degradation of the secondary battery in real time, and also to perform accurate SOC estimation by explicitly considering the offset current, the SOH and the initial electricity quantity error, which are error factors of the current integration method, and directly eliminating their influence.
[0111] In addition to the components of the first configuration, the battery status estimation device in the second configuration further includes: the SOC-OCV conversion unit configured to convert the first state of charge input from the SOC estimation unit into the first open-circuit voltage for output; the OCV subtraction unit configured to output, as the open-circuit voltage error, a value obtained by subtracting the first open-circuit voltage input from the SOC-OCV conversion unit from the second open-circuit voltage input from the OCV estimation unit; and the second recursive estimation unit configured to estimate and output the equivalent circuit parameters.
[0112] In the second configuration, the OCV estimation unit is configured to calculate and output the second open-circuit voltage and the state quantity of the equivalent circuit based on the corrected current input from the offset current subtraction unit, the detected voltage input from the voltage detection unit, and the equivalent circuit parameters input from the second recursive estimation unit.
[0113] Furthermore, the second recursive estimation unit is configured to recursively estimate and update the equivalent circuit parameters based on the corrected current input by the offset current subtraction unit, the state variable input by the OCV estimation unit, and the open-circuit voltage error input by the OCV subtraction unit. Furthermore, the OCV estimation unit is configured to calculate the second open-circuit voltage and the state variable using the equivalent circuit parameters updated by the second recursive estimation unit.
[0114] In this way, not only the parameters of the current integration method are recursively estimated, but also the parameters of the OCV estimation method (namely, equivalent circuit parameters). Therefore, it is possible to obtain a value through estimation that adapts to the variation in an equivalent circuit parameter that depends on the temperature and degradation of the secondary battery. This further improves the accuracy of the estimation of the parameters of the current integration method.
[0115] In summary, the first embodiment is configured to recursively estimate and update the SOC estimation parameters and the equivalent circuit parameters based on the estimation results obtained by the SOC estimation unit and the OCV estimation unit. Each updated parameter is used as feedback information to correct the estimation results. This makes it possible to implement a battery status estimation device capable of estimating the internal status of a battery more accurately than the prior art. Second embodiment
[0116] The battery status estimation device 100 according to a second embodiment of the present invention has a different configuration from the first embodiment. In the second embodiment, descriptions of the same aspects as in the first embodiment are omitted. Aspects different from those of the first embodiment will be mainly described.
[0117] Fig. Fig. 8 is a configuration drawing of the battery status estimation device 100 according to the second embodiment of the present invention. Fig. 8, the battery status estimation device 100 includes the current detection unit 102, the voltage detection unit 103, the offset current subtraction unit 104, the SOC estimation unit 105, the OCV estimation unit 106, the SOC-OCV conversion unit 107, the OCV-SOC conversion unit 108, the SOC subtraction unit 109, the OCV subtraction unit 110, a feedback addition unit 501, a high-frequency component removal unit 502, a coefficient multiplication unit 503, a first recursive estimation unit 504, and the second recursive estimation unit 112.
[0118] The SOC estimation unit 105 is configured to calculate the first state of charge SOC1, the elapsed time t k and the integrated amount of electricity Q Cbased on the detected current I input from the current detection unit 102 and the SOC estimation parameters input from the first recursive estimation unit 504. Furthermore, the SOC estimation unit 105 is configured to output the calculated first state of charge SOC1 to the feedback addition unit 501 and the calculated elapsed time t k and output the integrated electricity quantity Qc to the first recursive estimation unit 504.
[0119] The feedback addition unit 501 is configured for correction value addition processing. Specifically, the feedback addition unit 501 adds an SOC correction value L input from the coefficient multiplication unit 503 to the first SOC1 input from the SOC estimation unit 105, and outputs the obtained value as the third SOC3 to the first recursive estimation unit 504, the SOC-OCV conversion unit 107, and the SOC subtraction unit 109.
[0120] In the second embodiment, the SOC of the secondary battery 101 estimated by the battery status estimation device 100 is the third state of charge SOC3 output by the feedback addition unit 501. In other words, the third state of charge SOC3 is output as the final result of the estimation of the SOC of the secondary battery 101 by the battery status estimation device 100.
[0121] The SOC-OCV conversion unit 107 is configured to convert the third state of charge SOC3 input from the feedback addition unit 501 into the first open circuit voltage OCV1 and output the converted first open circuit voltage OCV1 to the OCV subtraction unit 110.
[0122] The SOC subtraction unit 109 is configured to subtract the third state of charge SOC3 input from the feedback addition unit 501 from the second state of charge SOC2 input from the OCV-SOC conversion unit 108 and output the obtained value as an SOC error u to the high-frequency component removal unit 502.
[0123] The high-frequency component removal unit 502 is configured for high-frequency component removal processing. Specifically, the high-frequency component removal unit 502 removes high-frequency components from the SOC error u input from the SOC subtraction unit 109 and outputs the SOC error y obtained through the removal to the coefficient multiplication unit 503.
[0124] As the high-frequency component removing unit 502, for example, an exponential moving average filter or an integrator with a forgetting coefficient can be used.
[0125] As the low-pass filter, for example, a filter having a filter characteristic shown in the following expression (16) can be used. y(k)=w1y(k−1)+w2u(k)
[0126] In expression (16), w1 and w2 are forgetting coefficients that are specified in advance. The conditions 0 <w1 < 1 und w2 = 1 - w1 resultieren in einem exponentiell gleitenden Durchschnittsfilter. Die Bedingung 0 < w1 = w2 < 1 resultiert in einem Integrator mit einem Vergessenskoeffizienten. Die Bedingung w1 = w2= 1 resultiert in einem allgemeinen Integrator. Der Vergessenskoeffizient ist so beschaffen, dass, wenn w1 größer wird, es unwahrscheinlicher wird, dass u in der Vergangenheit vergessen wird.
[0127] The SOC error u is sensitive to the influence of a voltage measurement error of the first state of charge (SOC1), as well as the influence of an equivalent circuit model error and an equivalent circuit parameter estimation error. The former error is an error with an average of 0, while the latter errors depend on the current value and the measurement accuracy. These errors appear as outliers, especially when a steeply rising current with a large amplitude flows. Therefore, the forgetting coefficient is set to smooth out the former influence and remove the latter influences.
[0128] For example, if T is set such that w1 = exp(-t s / T) as the forgetting coefficient, it is advantageous to set the value of T to a value of at least 10×t s or more (i.e. T is ten times the sampling period t sor more), so that the error of the OCV estimation method can be sufficiently reduced.
[0129] The coefficient multiplication unit 503 is configured for gain multiplication processing. Specifically, the coefficient multiplication unit 503 multiplies the SOC error y input from the high-frequency component removal unit 502 by a gain K and outputs the obtained value to the feedback addition unit 501 as the SOC correction value L.
[0130] The first recursive estimation unit 504 is configured to perform recursive estimation processing for current integration correction. Specifically, the first recursive estimation unit 504 calculates the estimated values of the SOC estimation parameters based on the elapsed time t. k and the integrated amount of electricity Q C, which are input from the SOC estimation unit 105, and the third state of charge SOC3, which is input from the feedback addition unit 501. Furthermore, the first recursive estimation unit 504 outputs the calculated SOC estimation parameters to the SOC estimation unit 105 and outputs only the offset current I from the calculated SOC estimation parameters. off to the offset current subtraction unit 104.
[0131] The offset current subtraction unit 104 subtracts the offset current I off input from the first recursive estimation unit 504 from the detected current I input from the current detection unit 102, and outputs the determined value as the corrected current I' to the OCV estimation unit 106 and the second recursive estimation unit 112.
[0132] The first recursive estimation unit 504 can construct RLS with the third charge state SOC3 as a learning signal by ε1(k) = SOC3(k) - φ1 T(k)θ1(k-1). Similar to the first embodiment, the second state of charge SOC2 can be used as a learning signal by inputting the second state of charge SOC2 instead of the third state of charge SOC3 to the first recursive estimation unit 504.
[0133] The following describes in more detail the third state of charge SOC3 output from the feedback addition unit 501. The battery status estimation device 100 according to the second embodiment calculates the third state of charge SOC3 according to the following expression (17). SOC3(k)=SOC1(k)+L(k−1)=1FCC0⋅SOH(k−1)(Qc(k)−tkIoff(k−1)−ΔQC0(k−1))+L(k−1)
[0134] With expression (17), it is possible to accurately determine the SOC as long as the first recursive estimation unit 504 correctly estimates the SOC estimation parameters, even if the SOC correction value L serving as a feedback correction term is omitted.
[0135] In reality, however, the first recursive estimation unit 504 does not always determine correct estimates. For example, it is problematic that if an initial value of the estimated value differs from a real value, and a significant amount of time is required for the initial value to converge to the real value, an accurate estimate of the SOC cannot be determined within this period.
[0136] The advantage of expression (17) is that even in such cases, the first state of charge SOC1 is corrected by the SOC correction value L. This allows the third state of charge SOC3 to quickly converge to a value close to the real SOC.
[0137] In addition, the value of the SOC correction value L, which serves as a feedback signal, is determined based on the error between the third state of charge (SOC3) and the second state of charge (SOC2). This error is calculated using the OCV estimation method, which is hardly affected by the current offset error, the SOH error, or the initial electricity quantity error. Therefore, the third state of charge (SOC3) follows the second state of charge (SOC2) in the medium and long term.
[0138] Furthermore, the value of the SOC correction value L is determined after passing through the high-frequency component removal unit 502. Therefore, the third state of charge SOC3 follows the change in SOC of the first state of charge SOC1 on short time scales.
[0139] The second state of charge (SOC2) is a value calculated based on the OCV estimation method, which does not involve error integration and is therefore accurate in the medium and long term. On the other hand, the first state of charge (SOC1) is calculated directly using the current integration method and can therefore accurately track short-term changes in SOC.
[0140] Finally, the third state of charge SOC3 can be described as an accurate measurement that combines the two advantages of the current integration method and the OCV measurement method.
[0141] Next, a series of processes for determining the SOC of the secondary battery 101 performed by the battery status estimating apparatus 100 according to the second embodiment will be described with reference to FIG. Fig. 9 is executed. Fig. 9 is a flowchart showing a series of processes executed by the battery status estimating apparatus 100 according to the second embodiment of the present invention.
[0142] A number of in Fig. The arithmetic processing steps S102 to S112 and S501 to S504 shown in Figure 9 correspond to arithmetic processing of one period performed by the battery status estimation device 100 according to the second embodiment. This arithmetic processing is performed with a period equal to the length of the sampling period t. s repeated.
[0143] In the flowchart according to Fig. 9, the number of each step corresponds to a corresponding component of the battery status estimating device 100 according to the second embodiment. In other words, as described above, in the flowchart according to Fig. 9 each component of the battery status estimating device 100 performs a step having the same number as the component.
[0144] As in Fig. 9, the battery status estimating device 100 performs the series of steps S102 to S112 and S501 to S504 of the arithmetic processing in the sampling period t s out of.
[0145] For each step in the flowchart according to Fig. 9, the order in which the battery status estimating device 100 executes these is not limited to the one shown in Fig. 9. It can be changed as long as the dependency relationship between the steps is maintained.
[0146] In the second embodiment, an example is illustrated in which the first recursive estimation unit 504 uses the third state of charge SOC3 input from the feedback addition unit 501 to recursively estimate and update the SOC estimation parameters. However, as described above, the second state of charge SOC2 can be used as a learning signal by inputting the second state of charge SOC2 instead of the third state of charge to the first recursive estimation unit 504.
[0147] In other words, as a modification example, the first recursive estimation unit 504 may be configured to obtain the input of the second state of charge SOC2 from the OCV-SOC estimation unit 108 and use the input second state of charge SOC2 to recursively estimate and update the SOC estimation parameters.
[0148] Furthermore, in the second embodiment, a case where the SOC of the secondary battery 101 estimated by the battery status estimation device 100 is the third state of charge SOC3 output from the feedback addition unit 501 is exemplified. However, the SOC of the secondary battery 101 may also be the first state of charge SOC1 output from the first recursive estimation unit 504.
[0149] As described above, the battery status estimation device according to the second embodiment, in the first configuration, includes: the first recursive estimation unit configured to estimate and output the SOC estimation parameters; the SOC estimation unit configured to calculate and output the first state of charge, as well as the elapsed time and the integrated amount of electricity measured since the start of the estimation of the SOC of the secondary battery based on the detected current input from the current detection unit and the SOC estimation parameters input from the first recursive estimation unit;the offset current subtraction unit configured to output, as the corrected current, a value obtained by subtracting the offset current input from the first recursive estimation unit from the detected current input from the current detection unit; the OCV estimation unit configured to calculate and output the second open-circuit voltage based on the corrected current input from the offset current subtraction unit and the detected voltage input from the voltage detection unit; the OCV-SOC conversion unit configured to convert the second open-circuit voltage input from the OCV estimation unit into the second state of charge for output;the feedback addition unit configured to output, as the third state of charge, a value obtained by adding an input SOC correction value to the first state of charge input from the SOC estimation unit; the SOC subtraction unit configured to output, as the SOC error, a value obtained by subtracting the third state of charge input from the feedback addition unit from the second state of charge input from the OCV-SOC conversion unit; the high-frequency component removal unit configured to output a value obtained by removing high-frequency components of the SOC error input from the SOC subtraction unit;and the coefficient multiplication unit configured to output a value as the SOC correction value obtained by multiplying the output value of the high-frequency component removal unit by a gain;
[0150] In the first configuration, the first recursive estimation unit is configured to recursively estimate and update the SOC estimation parameters based on the elapsed time and integrated electricity quantity input by the SOC estimation unit, and the third state of charge input by the feedback addition unit or the third state of charge input by the SOC estimation unit. Furthermore, the SOC estimation unit is configured to calculate the first state of charge, the elapsed time, and the integrated electricity quantity using the SOC estimation parameters updated by the first recursive estimation unit.
[0151] In this way, the battery status estimation device has a feedback system and can therefore accurately estimate the SOC by using the third state of charge without waiting for the SOC estimation parameters of the current integration method to converge to a value close to the real value.
[0152] As a result, a difference between the second state of charge estimated based on the OCV estimation method and the first state of charge estimated based on the current integration method is selected and passed through the high-frequency component removal unit to remove outlier values caused by high-frequency components due to the voltage measurement error and the equivalent circuit model error included in the second state of charge.
[0153] As a result, the OCV estimation method obtains a reliable error between the first SOC and the second SOC in the low and medium frequency range. The error value is then multiplied by a gain and fed back to correct the first SOC, ultimately outputting the third SOC. The third SOC can follow the short-term SOC change of the current integration method, which is reliable for short runtimes, and can simultaneously follow the medium and long-term SOC change of the OCV estimation method, which is reliable for medium and long runtimes.
[0154] Therefore, it is possible to accurately estimate the SOC by using feedback to eliminate the errors of the SOC estimation performed using the current integration method.
[0155] Furthermore, similar to the first embodiment, parameters of the current integration method are recursively estimated. Therefore, because each estimation parameter converges to the real value, the estimation accuracy for the first SOC is improved to exceed that for the third SOC. Therefore, it is possible to estimate the SOC with high precision and obtain an accurate SOH estimate in real time, eliminating the influence of the current offset error.
[0156] In addition to the components of the first configuration, the battery status estimation device in the second configuration further includes: the SOC-OCV conversion unit configured to convert the third state of charge input from the feedback addition unit into the first open-circuit voltage for output; the OCV subtraction unit configured to output, as the open-circuit voltage error, a value obtained by subtracting the first open-circuit voltage input from the SOC-OCV conversion unit from the second open-circuit voltage input from the OCV estimation unit; and the second recursive estimation unit configured to estimate and output the equivalent circuit parameters.
[0157] In the second configuration, the OCV estimation unit is configured to calculate and output the second open-circuit voltage and the state quantity of the equivalent circuit based on the corrected current input from the offset current subtraction unit, the detected voltage input from the voltage detection unit, and the equivalent circuit parameters input from the second recursive estimation unit.
[0158] Furthermore, the second recursive estimation unit is configured to recursively estimate and update the equivalent circuit parameters based on the corrected current input by the offset current subtraction unit, the state variable input by the OCV estimation unit, and the open-circuit voltage error input by the OCV subtraction unit. Furthermore, the OCV estimation unit is configured to calculate the second open-circuit voltage and the state variable using the equivalent circuit parameters updated by the second recursive estimation unit.
[0159] In this way, not only the parameters of the current integration method are recursively estimated, but also the parameters of the OCV estimation method (namely, equivalent circuit parameters). Therefore, it is possible to obtain a value that adapts to variations in an equivalent circuit parameter that depend on the temperature and degradation of the secondary battery. Therefore, the accuracy of the estimation of the current integration method parameters and the SOC is further improved.
[0160] In summary, the second embodiment is configured to recursively estimate and update the SOC correction value based on the estimation results obtained by the SOC estimation unit and the OCV estimation unit. Each updated SOC correction value is used as feedback information to correct the estimation result. This makes it possible to implement a battery status estimation device capable of estimating the internal status of a battery more accurately than the prior art.
[0161] Moreover, similar to the first embodiment, it is possible to further improve the estimation accuracy by further incorporating the configuration in which the SOC estimation parameters and the equivalent circuit parameters are recursively estimated and updated based on the estimation results obtained by the SOC estimation unit and the OCV estimation unit, and each updated parameter is used as feedback information to correct the estimation results. Third embodiment
[0162] The battery status estimation device 100 according to a third embodiment of the present invention has a configuration different from those of the first and second embodiments. In the third embodiment, the description of aspects similar to the first and second embodiments is omitted. Aspects different from those of the first and second embodiments will be mainly described.
[0163] Fig. 10 is a configuration drawing of the battery status estimation device 100 according to the third embodiment of the present invention. Fig. 10, the battery status estimation device 100 includes the current detection unit 102, the voltage detection unit 103, a feedback SOC estimation unit 601, the OCV estimation unit 106, the SOC-OCV conversion unit 107, the OCV-SOC conversion unit 108, the SOC subtraction unit 109, the OCV subtraction unit 110, the high-frequency component removal unit 502, the coefficient multiplication unit 503, and the second recursive estimation unit 112.
[0164] The feedback SOC estimation unit 601 is configured to calculate the first state of charge SOC1 based on the detected current I input from the current detection unit 102 and the SOC correction value L input from the coefficient multiplication unit 503, and output the first state of charge SOC1 to the SOC-OCV conversion unit 107 and the SOC subtraction unit 109.
[0165] The feedback SOC estimation unit 601 adds the SOC correction value L to the value obtained by the general current integration method represented by Expression (1) to estimate the SOC. In other words, the feedback SOC estimation unit 601 calculates the first state of charge SOC1 according to Expression (18) below. SOC1(k)=SOC1(k−1)+tsFCCI(k−1)+L(k−1)
[0166] As can be seen from the above expression, in the third embodiment, the error model of the current integration method, as represented in expression (5), is not adopted. The SOC calculated by the current integration method is only corrected with the SOC correction value L. The full charge capacity FCC, the offset current I off and the initial SOC error ΔSOC ini, which are all significant error factors in the current integration method, are all low-frequency errors. Therefore, it is possible to accurately estimate SOC using feedback correction alone, without directly estimating these low-frequency errors.
[0167] The following describes a specific configuration of the feedback SOC estimation unit with reference to Fig. 11 described. Fig. 11 is a configuration drawing of the feedback SOC estimation unit 601 according to the third embodiment of the present invention.
[0168] In Fig. 11, the feedback SOC estimation unit 601 comprises a coefficient multiplier 701, a coefficient multiplier 702 and a summing station 703.
[0169] The coefficient multiplier 701 is configured to perform sampling period multiplication processing. Specifically, the coefficient multiplier 701 multiplies the detected current I input from the voltage detection unit 102 by the sampling period t. s and outputs the determined value to the coefficient multiplier 702.
[0170] Coefficient multiplier 702 is configured to perform reciprocal FCC multiplication processing. Specifically, coefficient multiplier 702 multiplies the value output by coefficient multiplier 701 by the inverse of the full charge capacity FCC and outputs the determined value to summing station 703. If an accurate value of the full charge capacity FCC is unknown, the value of the initial full charge capacity FCC0 or other values, for example, can be used.
[0171] Summing station 703 is configured to perform summation processing. Specifically, summing station 703 adds the output of coefficient multiplier 702 to the SOC correction value L and the output of summing station 703 determined in the previous sampling period, and outputs the thus determined value as the first state of charge SOC1. The first state of charge SOC1 determined at the end of the previous SOC estimation can be used as the value for SOC1(0).
[0172] In other cases, if a long period of time has passed since the end of the last SOC estimation, the detected voltage V, which is determined by the voltage detection unit 103 can be considered as the open-circuit voltage of the secondary battery 101. The value obtained by converting the open-circuit voltage into the SOC by the OCV-SOC conversion unit 108 can be used as the value for SOC1(0).
[0173] With a feedback SOC estimation unit 601 configured in this way, the feedback SOC estimation unit 601 can perform the calculation according to Expression (18).
[0174] In the following, with reference to the flowcharts in Fig. 12 and Fig. 13 describes a series of processes for estimating the SOC of the secondary battery 101, which is executed by the battery status estimating apparatus 100 according to the third embodiment. Fig. 12 is a flowchart illustrating a series of processes performed by the device 100 for estimating the battery status according to the third embodiment of the present invention. Fig. 13 is a flowchart showing a series of processes executed by the feedback SOC estimation unit 601 according to the third embodiment of the present invention.
[0175] A series of steps S102 and S103, step S101, steps S106 to S110, step S112 and steps S502 and S503 of a Fig. 12 corresponds to arithmetic processing in a period or time span executed by the battery status estimation device 100. This arithmetic processing is performed with a period equal to the sampling period t s repeated.
[0176] Furthermore, the sequence of steps S701 to S703 corresponds to a Fig. 13 shown arithmetic processing of an arithmetic processing which is in the Fig. 12 shown step S601 is executed.
[0177] In the flowcharts according to Fig. 12 and Fig. 13, the number of each step corresponds to a corresponding component of the battery status estimation device 100 according to the third embodiment. In other words, as described above, in the flowcharts according to Fig. 12 and Fig. 13 each component of the battery status estimating device 100 executes a step having the same number as that of the component.
[0178] As in Fig. 12, the battery status estimating device 100 performs a series of steps S102 and S103, step S101, steps S106 to S110, step S112, and steps S502 and S503 of arithmetic processing in the sampling period t s out of.
[0179] In step S601, the battery status estimating device 100 also performs a series of steps S701 to S703 of the Fig. 13 shows the arithmetic processing.
[0180] Regarding each step of the flowcharts of the Fig. 12 and Fig. 13, the order of execution by the battery status estimating device 100 is not limited to the Fig. 12 and Fig. 13 is restricted and can be changed as long as the dependency relationship between the steps is maintained.
[0181] As described above, the battery status estimation device according to the third embodiment, in the first configuration, includes: the feedback SOC estimation unit configured to calculate the first state of charge based on the detected current input from the current detection unit and the SOC correction value for correcting the SOC of the secondary battery for output; the OCV estimation unit configured to calculate the second open-circuit voltage based on the detected current input from the current detection unit and the detected voltage input from the voltage detection unit for output; the OCV-SOC conversion unit configured to convert the second open-circuit voltage input from the OCV estimation unit into the second state of charge for output;the SOC subtraction unit configured to output, as the SOC error, a value obtained by subtracting the first state of charge input from the feedback SOC estimation unit from the second state of charge input from the OCV-SOC conversion unit; the high-frequency component removal unit configured to output a value obtained by removing high-frequency components of the SOC error input from the SOC subtraction unit; and the coefficient multiplication unit configured to output, as the SOC correction value, a value obtained by multiplying the output value of the high-frequency component removal unit by the gain.
[0182] In this way, the battery status estimation device has a feedback system and can accurately estimate the SOC by correcting influences of errors in the FCC, the offset current, and the initial SOC, which are error factors of the current integration method.
[0183] As a result, the low and medium frequency components of the error between the first state of charge based on the current integration method and the second state of charge estimated based on the OCV estimation method are fed back to calculate the first state of charge, so that the obtained SOC estimation value follows the state of charge SOC2 of the OCV estimation method, which is reliable in the low and medium frequency range, and the SOC estimation value of the current integration method, which is reliable in the high frequency range.
[0184] By using the high-frequency component removal unit in the feedback system, it is possible to remove high-frequency components from the error between SOC2 and SOC1, which occur due to the voltage measurement error and the equivalent circuit model errors of the OCV estimation method.
[0185] Despite the simple system, which lacks the recursive estimation unit to implement the current integration method, the feedback correction effect allows the SOC to converge to a value close to the real value in a short time, even if the current integration method has the initial SOC error. Furthermore, the influences of the current offset error and the FCC error are reduced, thus achieving a reliable and accurate calculation of the SOC estimate.
[0186] Furthermore, unlike the second embodiment, in the third embodiment, the SOC correction value L is not used to correct the first state of charge SOC1 and calculate the third state of charge SOC3, but to directly calculate the first state of charge SOC1 as shown in Expression (18). Thus, it is possible to prevent a phenomenon in which the SOC1, the error between the SOC2 and the SOC1, and the SOC correction value L increase or decrease unconditionally due to the influence of the current offset error.
[0187] In addition to the components of the first configuration, the battery status estimation device in the second configuration further includes: the OCV subtraction unit configured to output, as the open-circuit voltage error, a value obtained by subtracting the first open-circuit voltage input from the SOC-OCV estimation unit from the second open-circuit voltage input from the OCV estimation unit; and the second recursive estimation unit configured to estimate and output the equivalent circuit parameters.
[0188] In the second configuration, the OCV estimation unit is configured to calculate and output the second open-circuit voltage and the state quantity of the equivalent circuit based on the detected current input from the current detection unit, the detected voltage input from the voltage detection unit, and the equivalent circuit parameters input from the second recursive estimation unit.
[0189] Furthermore, the second recursive estimation unit is configured to recursively estimate and update the equivalent circuit parameters based on the detected current input by the current detection unit, the state variable input by the OCV estimation unit, and the open-circuit voltage error input by the OCV subtraction unit. Furthermore, the OCV estimation unit is configured to calculate the second open-circuit voltage and the state variable using the equivalent circuit parameters updated by the second recursive estimation unit.
[0190] In this way, the parameters of the OCV estimation method (namely, equivalent circuit parameters) are estimated recursively. Therefore, it is possible to obtain a value by estimation that adapts to the variation of an equivalent circuit parameter depending on the temperature and degradation of the secondary battery. This means that the accuracy of the SOC estimation is improved by the OCV estimation method. and the accuracy of the estimation of the SOC calculated by the feedback SOC estimation unit with respect to the SOC is also improved.
[0191] In summary, in the third embodiment, the characteristics of the OCV estimation method in which the OCV estimation method does not involve accumulation of errors and has a relatively small error with respect to the real value of the SOC based on an average time are used as feedback information to correct the error of the current integration method.
[0192] Thus, it is possible to implement a battery status estimation device capable of estimating the internal status of the battery more accurately than in the prior art, although the battery status estimation device has a simple system that does not include components for estimating the error factor of the current integration method, such as the first recursive estimation unit in the first and second embodiments.
[0193] Furthermore, similar to the first and second embodiments, it is possible to further improve the estimation accuracy by continuing to use the configuration in which the equivalent circuit parameters are recursively estimated and updated based on the estimation results obtained by the feedback SOC estimation unit, and each updated parameter is used as feedback information to correct the estimation result.
Claims
[1] A battery status estimating device (100) configured to estimate at least a state of charge of a secondary battery (101) as an estimated state of charge, the battery status estimating device (100) comprising: a current detection unit (102) configured to detect either a charging current or a discharging current of the secondary battery (101) as a detected current; a voltage detection unit (103) configured to detect a voltage between terminals of the secondary battery (101) as a detected voltage; a first recursive estimation unit (111) configured to estimate state of charge estimation parameters including a state of health of the secondary battery (101) and an offset current of the current detection unit (102); an SOC estimation unit (105) configured to calculate at least a first state of charge based on the detected current output by the current detection unit (102) and on the state of charge estimation parameters output by the first recursive estimation unit (111); an offset current subtraction unit (104) configured to output, as a corrected current, a value obtained by subtracting the offset current output from the first recursive estimation unit (111) from the detected current output from the current detection unit (102); a SOC-OCV conversion unit (107) configured to convert the first state of charge output output by the SOC estimation unit (105) into a first open circuit voltage; a second recursive estimation unit (112) configured Estimate OCV estimation parameters that include equivalent circuit parameters; an OCV estimation unit (106) configured to calculate a second open-circuit voltage and a state quantity of an equivalent circuit corresponding to the OCV estimation parameters based on the corrected current output by the offset current subtraction unit (104), the detected voltage output by the voltage detection unit (103), and the OCV estimation parameters output by the second recursive estimation unit (112); an OCV-SOC conversion unit (108) configured to convert the second open circuit voltage output from the OCV estimation unit (106) into a second state of charge; an OCV subtraction unit (110) configured to output a value as the open-circuit voltage error, which is determined by subtracting the first open-circuit voltage output by the SOC-OCV conversion unit (107) from the second open-circuit voltage output by the OCV estimation unit (106); and an SOC subtraction unit (109) configured to output a value as a state of charge error determined by subtracting the first state of charge output by the SOC estimation unit (105) from the second state of charge output by the OCV-SOC conversion unit (108), wherein the first recursive estimation unit (111) is configured to recursively estimate and update the state of charge estimation parameters based at least on the state of charge error output by the SOC subtraction unit, wherein the second recursive estimation unit (112) is configured to recursively estimate and update the OCV estimation parameters based at least on the corrected current output by the offset current subtraction unit (104) and on the open circuit voltage error output by the OCV subtraction unit (110), and wherein the battery status estimating device (100) is configured to determine the first state of charge as the estimated state of charge. [2] The battery status estimating device according to claim 1, wherein the second recursive estimation unit (112) is configured to recursively estimate and update the OCV estimation parameters based on the corrected current output from the offset current subtraction unit (104), the state quantity output from the OCV estimation unit (106), and the open circuit voltage error output from the OCV subtraction unit (110). [3] Apparatus for estimating a battery status according to claim 2, wherein the first recursive estimation unit (111) is configured to estimate the state of charge estimation parameters including the state of health or a full charge capacity of the secondary battery (101) and the offset current, wherein the SOC estimation unit (105) is configured to calculate the time elapsed since the start of estimating the state of charge of the secondary battery, an integrated amount of electricity measured since the start of estimating the state of charge of the secondary battery, and the first state of charge based on the detected current output by the current detection unit (102) and the state of charge estimation parameters output by the first recursive estimation unit (111), and wherein the first recursive estimation unit (111) is configured to recursively estimate and update the state of charge estimation parameters based on the state of charge error output from the SOC subtraction unit (109) and on the elapsed time and the integrated electricity amount output from the SOC estimation unit (105). [4] A battery status estimating apparatus according to claim 1, further comprising: a feedback addition unit (501) configured to correct the first state of charge output from the SOC estimation unit (105) to output the corrected first state of charge as a third state of charge; a high-frequency component removal unit (502) configured to output a value obtained by removing high-frequency components from the state of charge error output by the SOC subtraction unit (109); and a coefficient multiplication unit (504) configured to output a value as a state of charge correction value obtained by multiplying an output value of the high-frequency component removal unit (502) by a gain, wherein the feedback addition unit (501) is configured to output, as the third state of charge, a value determined by adding the first state of charge output by the SOC estimation unit (105) and the state of charge correction value output by the coefficient multiplication unit (503); wherein the SOC-OCV conversion unit (107) is configured to convert the third state of charge into the first open circuit voltage instead of the first state of charge, wherein the SOC subtraction unit (109) is configured to output as the state of charge error a value obtained by subtracting the third state of charge from the second state of charge instead of the first state of charge, wherein the first recursive estimation unit (504) is configured to recursively estimate and update the state of charge estimation parameters based at least on the third state of charge or the second state of charge serving as a substitute for the state of charge error output by the SOC subtraction unit (109), and wherein the battery status estimating device (100) is configured to determine the first charge level or the third charge level as the estimated charge level. [5] A battery status estimating device according to claim 4, further comprising: an offset current subtraction unit (104) configured to output, as a corrected current, a value obtained by subtracting the offset current output by the first recursive estimation unit (504) from the detected current output by the current detection unit (102); wherein the OCV estimation unit (106) is configured to calculate the second open-circuit voltage and a state quantity of an equivalent circuit corresponding to the OCV estimation parameters based on the corrected current output from the offset current subtraction unit (104), the detected voltage output from the voltage detection unit (103), and the OCV estimation parameters output from the second recursive estimation unit (112); wherein the second recursive estimation unit (112) is configured to recursively estimate and update the OCV estimation parameters based on the corrected current output by the offset current subtraction unit (104), on the state variable output by the OCV estimation unit (106), and on the open circuit voltage error output by the OCV subtraction unit (110), and wherein the battery status estimating device (100) is configured to determine the first charge level or the third charge level as the estimated charge level. [6] Apparatus for estimating a battery status according to claim 5, wherein the first recursive estimation unit (504) is configured to estimate the state of charge estimation parameters including the state of health or a full charge capacity of the secondary battery (101) and the offset current, wherein the SOC estimation unit (105) is configured to calculate the time elapsed since the start of estimating the state of charge of the secondary battery, an integrated amount of electricity measured since the start of estimating the state of charge of the secondary battery, and the first state of charge based on the detected current output by the current detection unit (102) and the state of charge estimation parameters output by the first recursive estimation unit (504), and wherein the first recursive estimation unit (504) is configured to recursively estimate and update the state of charge estimation parameters based on the third state of charge or the second state of charge serving as a substitute for the state of charge error output from the SOC subtraction unit (109) and on the elapsed time and the integrated amount of electricity output from the SOC estimation unit (105). [7] A battery status estimating device (100) configured to estimate at least a state of charge of a secondary battery (101) as an estimated state of charge, the battery status estimating device (100) comprising: a current detection unit (102) configured to detect either a charging current or a discharging current of the secondary battery (101) as a detected current; a voltage detection unit (103) configured to detect a voltage between terminals of the secondary battery (101) as a detected voltage; a coefficient multiplication unit (503) configured to calculate a state of charge correction value; a feedback SOC estimation unit (601) configured to calculate a first state of charge based at least on the detected current output from the current detection unit (102) and on the state of charge correction value output from the coefficient multiplication unit (503); a SOC-OCV conversion unit (107) configured to convert the first state of charge output from the feedback SOC estimation unit (601) into a first open circuit voltage; a second recursive estimation unit (112) configured to estimate OCV estimation parameters including equivalent circuit parameters; an OCV estimation unit (106) configured to estimate at least a second open-circuit voltage based on the detected current output by the current detection unit (102), the detected voltage output by the voltage detection unit (103), and the OCV estimation parameters output by the second recursive estimation unit (112); an OCV-SOC conversion unit (108) configured to convert the second open circuit voltage output from the OCV estimation unit (106) into a second state of charge; an OCV subtraction unit (110) configured to output a value as an open circuit voltage error obtained by subtracting the first open circuit voltage output by the SOC-OCV conversion unit (107) from the second open circuit voltage output by the OCV estimation unit (106); an SOC subtraction unit (109) configured to output a value as a state of charge error determined by subtracting the first state of charge output by the feedback SOC estimation unit (601) from the second state of charge output by the OCV-SOC conversion unit (108); and a high-frequency component removal unit (502) configured to output a value obtained by removing high-frequency components from the state of charge error output from the SOC subtraction unit (109); wherein the coefficient multiplication unit (503) is configured to output, as the state of charge correction value, a value determined by multiplying an output value of the high-frequency component removal unit (502) by a gain, wherein the second recursive estimation unit (112) is configured to recursively estimate and update the OCV estimation parameters based at least on the open circuit voltage error output by the OCV subtraction unit (110), and wherein the battery status estimating device (100) is configured to determine the first state of charge as the estimated state of charge. [8] The battery status estimation device according to claim 7, wherein the OCV estimation unit (106) is configured to calculate the second open-circuit voltage and a state quantity of an equivalent circuit corresponding to the OCV estimation parameters based on the detected current output by the current detection unit (102), the detected voltage output by the voltage detection unit (103), and the OCV estimation parameters output by the second recursive estimation unit (112), wherein the second recursive estimation unit (112) is configured to recursively estimate and update the OCV estimation parameters based on the detected current output by the current detection unit (102), on the state quantity output by the OCV estimation unit (106), and on the open circuit voltage error output by the OCV subtraction unit (110), and wherein the battery status estimating device (100) is configured to determine the first state of charge as the estimated state of charge.
Citation Information
Patent Citations
Energy storage device remaining capacity calculation device
JP4583765B2
Battery capacity calculation device and battery capacity calculation method
JP5419832B2
Nonlinear adaptive observation approach for battery state-of-charge estimation
DE102012210883A1
Battery charge ratio estimation device
JP2013238402A
JP002013238402A
