JP2018151176A - Estimation device, estimation method, and estimation program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an estimation device, an estimation method, and an estimation program with which it is possible to improve the accuracy of estimating the SOC value of a secondary battery.SOLUTION: The estimation device pertaining to an embodiment comprises an observation value acquisition unit, an arithmetic unit, and a switching unit. The arithmetic unit estimates, on the basis of an observation value acquired by the observation value acquisition unit, an SOC value that indicates the state of charge of a secondary battery by estimation computation using a Kalman filter for an equivalent circuit model of the secondary battery. When the effect of an open-circuit voltage of the secondary battery on SOC value estimation computation is relatively large, the switching unit causes an SOC value calculated by estimation computation, where a maximum error priority parameter is used as a parameter of the Kalman filter, to be outputted from the arithmetic unit, and when the effect of the open-circuit voltage of the secondary battery is relatively small, causes an SOC value calculated by estimation computation, where an average error priority parameter is used as a parameter of the Kalman filter, to be outputted from the arithmetic unit.SELECTED DRAWING: Figure 5

Description

  Embodiments disclosed herein relate to an estimation apparatus, an estimation method, and an estimation program.

  Conventionally, for example, an electric vehicle such as a hybrid vehicle or an electric vehicle has been equipped with a power source that supplies electric power to a motor that is a power source. A chargeable / dischargeable secondary battery that stores regenerative power is used.

  In general, an electric vehicle is equipped with an estimation device that estimates an SOC (State Of Charge) value so that electric power from the secondary battery can be stably supplied to the motor. The SOC value is a value indicating a charging state of the secondary battery, and is, for example, a charging rate of the secondary battery.

  As a method for estimating the SOC value of the secondary battery, a method for estimating the SOC value by applying a Kalman filter to an equivalent circuit model in which the secondary battery is regarded as an electric circuit is known. (For example, refer to Patent Document 1).

Japanese Patent Laid-Open No. 2006-065828

  In the estimation method described above, the SOC value is estimated using the SOC-OCV characteristic of the secondary battery from the OCV (Open Circuit Voltage) estimated value indicating the open circuit voltage of the secondary battery, but a large current flows. In such a case, the estimation error of the SOC value from the OCV value becomes large. That is, when the influence of the OCV on the estimation calculation of the SOC value is relatively large as in the case where a large current flows, the estimation error of the SOC value is large and the estimation accuracy of the SOC value is lowered.

  One aspect of the embodiments has been made in view of the above, and an object thereof is to provide an estimation device, an estimation method, and an estimation program that can improve the estimation accuracy of the SOC value.

  An estimation apparatus according to an aspect of the embodiment includes an observation value acquisition unit, a calculation unit, and a switching unit. The observed value acquisition unit acquires observed values of the voltage and current of the secondary battery. The arithmetic unit is based on the observation value acquired by the observation value acquisition unit, and an SOC value indicating the state of charge of the secondary battery by an estimation calculation applying a Kalman filter to the equivalent circuit model of the secondary battery Is estimated. When the influence of the open circuit voltage of the secondary battery on the estimation calculation of the SOC value is relatively large, the switching unit sets a maximum error priority parameter that prioritizes suppression of the maximum error of the estimation calculation. When the first SOC value, which is the SOC value calculated by the estimation calculation as a parameter of the Kalman filter, is output from the calculation unit and the influence of the open circuit voltage is relatively small, the estimation is performed. A second SOC value that is the SOC value calculated by the estimation calculation using an average error priority parameter giving priority to suppression of an average error of the calculation as a parameter of the Kalman filter is output from the calculation unit.

  According to one aspect of the embodiment, it is possible to provide an estimation device, an estimation method, and an estimation program that can improve the estimation accuracy of the SOC value.

FIG. 1 is a diagram illustrating a configuration example of a vehicle-mounted system including an estimation device according to an embodiment. FIG. 2 is a flowchart showing a flow of setting parameters of the Kalman filter in the estimation apparatus. FIG. 3 is an explanatory diagram illustrating an example of an equivalent circuit model of a secondary battery. FIG. 4 is a diagram illustrating an example of SOC-OCV characteristics of the secondary battery. FIG. 5 is a functional block diagram of the charging state estimation unit. FIG. 6 is a diagram showing a temporal transition of the estimation error of the SOC value. FIG. 7 is a flowchart illustrating an example of a processing procedure executed by the charge state estimation unit. FIG. 8 is a flowchart illustrating another example of the processing procedure executed by the charge state estimation unit.

  Hereinafter, embodiments of an estimation device, an estimation method, and an estimation program disclosed in the present application will be described in detail with reference to the accompanying drawings. In addition, this invention is not limited by embodiment shown below. In the following, a vehicle-mounted system will be described as an application example of a power supply system including an estimation device, but the present invention is not limited to this example. For example, a power supply system including an estimation device can be applied in addition to the on-vehicle system, such as a power generation system that converts natural energy such as sunlight or wind power into electric power.

[1. Configuration of on-board system]
FIG. 1 is a diagram illustrating a configuration example of a vehicle-mounted system including an estimation device according to an embodiment. The vehicle mounting system 100 shown in FIG. 1 is mounted on a vehicle such as a hybrid vehicle (HEV: Hybrid Electric Vehicle) and an electric vehicle (EV: Electric Vehicle) (not shown).

  The on-vehicle system 100 includes a power supply system 1, a vehicle control device 2, a motor 3, and a power conversion unit 4. The vehicle control device 2 starts the power supply system 1, connects the power supply system 1 and the power conversion unit 4, and drives the power conversion unit 4 to control the motor 3.

  The power conversion unit 4 can convert DC power from the power supply system 1 into AC power and output it to the motor 3, and converts AC power that is regenerative power of the motor 3 into DC power to convert the power into the power system 1. Can be supplied to.

  The vehicle control device 2 can control charging / discharging of the secondary battery 5 based on an SOC (State Of Charge) value indicating a state of a secondary battery 5 described later notified from the power supply system 1. For example, the vehicle control device 2 can increase the SOC value by supplying regenerative power of the motor 3 to the power supply system 1 when the SOC value decreases. The SOC value is the charging rate (%) of the secondary battery 5, but may be the remaining capacity (Ah) of the secondary battery 5.

  The power supply system 1 includes a secondary battery 5, a battery monitoring system 6, and a relay 7. The secondary battery 5 is, for example, a lithium ion secondary battery or a nickel hydride secondary battery. The secondary battery 5 is, for example, an assembled battery, and is configured by connecting a plurality of battery blocks in series. Each battery block includes a plurality of battery cells connected in series. The secondary battery 5 may be configured by connecting a plurality of battery blocks in parallel.

  The battery monitoring system 6 brings the power supply system 1 into an operating state when an ignition switch (not shown) provided in the vehicle is turned on. When the power supply system 1 is in an operating state, the secondary battery 5 is connected to the power conversion unit 4 and the SOC value estimation process of the secondary battery 5 is repeatedly performed. The battery monitoring system 6 notifies the vehicle control device 2 of the estimated SOC value every time the SOC value is estimated.

  Further, the battery monitoring system 6 puts the power supply system 1 in a stopped state when the ignition switch is turned off. When the power supply system 1 is in the stopped state, the secondary battery 5 is disconnected from the power conversion unit 4 and the SOC value estimation process of the secondary battery 5 is stopped.

  The battery monitoring system 6 has a block monitoring unit (not shown) provided in each of a plurality of battery blocks constituting the secondary battery 5, and each block monitoring unit has an abnormality of a corresponding battery block, etc. Can be detected.

  The battery monitoring system 6 includes a battery state monitoring unit 10, a voltage sensor 11 that detects an instantaneous value of the voltage u across the secondary battery 5 (hereinafter referred to as a voltage observation value u), and a current flowing through the secondary battery 5. a current sensor 12 that detects an instantaneous value of i (hereinafter referred to as a current observation value i).

The battery state monitoring unit 10 includes a switch control unit 20 and a charge state estimation unit 21 (an example of an estimation device). The switch control unit 20 controls ON / OFF of the relay 7 based on an ignition signal _SIGN indicating the state of the ignition switch.

For example, when the ignition signal _SIGN indicates ON of the ignition switch, the switch control unit 20 turns on the relay 7 to connect the secondary battery 5 to the power conversion unit 4, and the vehicle-mounted system 100 uses the secondary battery 5. Make it available for use. Further, when the ignition signal _SIGN indicates that the ignition switch is OFF, the switch control unit 20 turns off the relay 7 to disconnect the secondary battery 5 from the power conversion unit 4, and the vehicle-mounted system 100 detects the secondary battery 5. Stop using.

  The switch control unit 20 can also control ON / OFF of the relay 7 in response to a request from the vehicle control device 2. Further, when the switch control unit 20 detects an abnormality in the secondary battery 5 in a state where the relay 7 is ON, the switch control unit 20 can also turn off the relay 7 to stop the power supply system 1.

  When the power supply system 1 is in an operating state, the charging state estimation unit 21 performs an estimation calculation for estimating the SOC value of the secondary battery 5, and stops the SOC value estimation calculation when the power supply system 1 is in a stopped state. .

  Specifically, the charging state estimation unit 21 acquires the voltage observation value u and the current observation value i from the voltage sensor 11 and the current sensor 12. The charging state estimation unit 21 calculates an SOC value indicating the charging state of the secondary battery 5 by an estimation calculation in which a Kalman filter is applied to the equivalent circuit model of the secondary battery 5 based on the voltage observation value u and the current observation value i. presume. The charge state estimation unit 21 outputs the SOC value, which is the result of the estimation calculation, to the vehicle control device 2 every time the estimation calculation is performed.

  When the current i flowing through the secondary battery 5 is large, the influence of the SOC value on the OCV (Open Circuit Voltage) value, which is the open circuit voltage of the secondary battery 5, increases, and the estimation error of the SOC value may increase. is there. In addition, during a certain period of time after a large current i flows through the secondary battery 5, the SOC value is greatly affected by the SOC value, and the SOC value estimation error may increase.

  Therefore, the charging state estimation unit 21 according to the present embodiment estimates the SOC value by changing the parameter of the Kalman filter according to the state of the influence of the OCV value on the SOC value estimation calculation (hereinafter referred to as the OCV influence). Suppresses an increase in error. The OCV value is the value of the open circuit voltage of the secondary battery 5, in other words, the voltage value of the secondary battery 5 in the no-load state. The OCV value is a value estimated from the current observation value i and the voltage observation value u.

  FIG. 2 is a flowchart showing a flow of setting the parameters of the Kalman filter in the charge state estimation unit 21. As shown in FIG. 2, the charge state estimation unit 21 determines whether or not the OCV influence is relatively large (step S10). The charging state estimation unit 21 is, for example, when the current observation value i is greater than or equal to the threshold value ith and when the current observation value i is greater than or equal to the threshold value ith and less than the threshold value ith until a certain time elapses. Can be assumed to be a state in which the OCV influence is relatively large.

  When it is determined that the OCV influence is relatively large (step S10: Yes), the charging state estimation unit 21 sets the Kalman filter parameter as the maximum error priority parameter (step S11). The maximum error priority parameter is a parameter that prioritizes suppression of the maximum error in the SOC value estimation calculation. By using the maximum error priority parameter for the estimation calculation, for example, when the current i flowing through the secondary battery 5 is large, the error of the SOC value can be suppressed, the convergence can be improved, and the estimation accuracy of the SOC value is improved. Can be made.

  On the other hand, when it is determined that the OCV influence is not relatively large (step S10: No), the charging state estimation unit 21 sets the Kalman filter parameter as an average error priority parameter (step S12). The average error priority parameter is a parameter that prioritizes suppression of the average error in the SOC value estimation calculation. By using the average error priority parameter for the estimation calculation, for example, when the current i flowing through the secondary battery 5 is small, the average error can be suppressed.

  When the current i flowing through the secondary battery 5 is small, the maximum error in the estimation calculation of the SOC value is smaller than when the current i flowing through the secondary battery 5 is large. Therefore, it is desirable to suppress the average error rather than the maximum error. The average error is suppressed by using the average error priority parameter. Thereby, when the current i flowing through the secondary battery 5 is small, the estimation accuracy of the SOC value can be improved.

  In the above description, the parameters of the Kalman filter are changed according to the magnitude of the current observation value i, but the present invention is not limited to this example. For example, when the state of the OCV influence is known in advance, the charging state estimation unit 21 uses the maximum error priority parameter during a period in which the OCV influence is assumed to be relatively large, and the state where the OCV influence is relatively small. An average error priority parameter can also be used in the assumed period. That is, the charging state estimation unit 21 can switch parameters using time instead of current as a trigger.

  In addition, the charging state estimation unit 21 can perform the SOC value estimation calculation using the maximum error priority parameter and the SOC value estimation calculation using the average error priority parameter in parallel. In this case, the charging state estimation unit 21 outputs the SOC value obtained by the estimation calculation using the maximum error priority parameter when the OCV influence is relatively large. Further, when the OCV influence is relatively small, the charging state estimation unit 21 can output the SOC value obtained by the estimation calculation using the average error priority parameter.

[2. Charge state estimation unit 21]
The charge state estimation unit 21 performs an estimation calculation for estimating the SOC value by applying a Kalman filter to an equivalent circuit model obtained by modeling the secondary battery 5.

  FIG. 3 is an explanatory diagram illustrating an example of an equivalent circuit model of the secondary battery 5. As shown in FIG. 3, the equivalent circuit model 30 of the secondary battery 5 includes a power supply 31, a resistance element 32, a first RC circuit 33, and a second RC circuit 34.

  The power supply 31 is an ideal power supply with no internal impedance, and the voltage value of the power supply 31 is the voltage value of the secondary battery 5 in the no-load state, which is the OCV value described above. In the example shown in FIG. 3, the OCV value at the time of charging is represented as OCVa, and the OCV value at the time of discharging is represented as OCVb.

  The resistance element 32 is a resistance element having a resistance value R <b> 0, and the voltage u <b> 0 of the resistance element 32 increases in direct proportion to the change in the current i flowing through the secondary battery 5. The first RC circuit 33 is a parallel circuit of a resistance element having a resistance value R1 and a capacitor having a capacitance C1, and the voltage u1 of the first RC circuit 33 is caused by a change in the current i flowing through the secondary battery 5. On the other hand, it changes transiently.

  The second RC circuit 34 is a parallel circuit of a resistance element having a resistance value R2 and a capacitor having a capacitance C2, and the voltage u2 of the second RC circuit 34 varies with a change in the current i flowing through the secondary battery 5. On the other hand, it changes transiently.

The both-ends voltage u of the equivalent circuit model 30 shown in FIG. 3 can be expressed by the sum of the OCV value, the voltage u0, the voltage u1, and the voltage u2, as shown in the following formula (1).

The voltage u0 can be expressed by the following formula (2) from the current i flowing through the secondary battery 5 and the resistance value R0. The differential equation between the voltage u1 of the first RC circuit 33 and the voltage u2 of the second RC circuit 34 can be expressed as shown in the following formulas (3) and (4).

The charge state estimation unit 21 estimates the SOC value by the estimation calculation using the equivalent circuit model 30 described above. The following formulas (5) and (6) show a system model representing the state of the secondary battery 5. Equation (5) is a state equation indicating the state of the secondary battery 5, and Equation (6) is an observation equation relating to the observed value of the secondary battery 5. In the following formulas (5) and (6), “k” represents a discretized time index (k-th time; hereinafter, sometimes referred to as time k).

In the above formulas (5) and (6), “x _k ” is a state vector indicating the state of the secondary battery 5 at time k, and as shown in the above formula (7), the SOC value and voltage at time k. u1 and voltage u2. Further, “i _k ” is the above-described current observation value i that is a control input from the outside, and “y _k ” is the voltage u across the secondary battery 5 at time k as shown in the above equation (8). It is. “A”, “B”, “C”, and “D” are, for example, matrices obtained from (1) to (4) above.

“V _k ” is system noise, and “w _k ” is observation noise. The system noise v _k and the observation noise w _k are noises that follow a multidimensional normal distribution. System noise is also called process noise.

  FIG. 4 is a diagram illustrating an example of the SOC-OCV characteristic of the secondary battery 5. In FIG. 4, the horizontal axis represents SOC (%), and the vertical axis represents OCV (V). The charging state estimation unit 21 temporarily estimates the SOC value based on the OCV value estimated using the system model and the SOC-OCV characteristic shown in FIG. The charging state estimation unit 21 corrects the provisionally estimated SOC value with a Kalman filter based on an observation error that is a difference between the estimated value of the both-end voltage u estimated using the system model and the voltage observation value u, and corrects it. The SOC value is actually estimated by using the SOC value as the estimated value of the SOC.

  Hereinafter, the SOC estimation calculation performed by the charge state estimation unit 21 will be described in more detail, including processing when the use of the secondary battery 5 is resumed after the use of the secondary battery 5 is stopped. In the following description, the discretized time index “k” is written in parentheses instead of subscripts.

[2.1. Configuration of charge state estimation unit 21]
FIG. 5 is a functional block diagram of the charge state estimation unit 21. As illustrated in FIG. 5, the charging state estimation unit 21 includes a storage unit 40, an A / D conversion unit 41 (an example of an observation value acquisition unit), a calculation unit 42, and a switching unit 43. The calculation unit 42 includes a battery state estimation unit 50, a gain setting unit 51, and a filter unit 52. The gain setting unit 51 and the filter unit 52 function as a Kalman filter.

  The charge state estimation unit 21 is a microcomputer including, for example, a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), a flash memory, an A / D conversion unit, an input / output port, and the like. The CPU of the microcomputer functions as the calculation unit 42 and the switching unit 43 by reading and executing a program stored in the ROM or flash memory. The storage unit 40 is configured by a RAM or the like.

  Note that all or part of the calculation unit 42 and the switching unit 43 can be configured by hardware such as an application specific integrated circuit (ASIC) or a field programmable gate array (FPGA). In addition, the above-described program can be read from a recording disk (for example, CD or DVD) by a reading unit (not shown) and stored in a ROM or flash memory.

  The storage unit 40 stores various parameters of the equivalent circuit model 30 used by the charging state estimation unit 21, various parameters of the Kalman filter, information indicating SOC-OCV characteristics, and the like. In addition, the storage unit 40 stores an SOC value that is the latest result of the estimation calculation performed by the charge state estimation unit 21.

  The A / D converter 41 repeatedly digitally converts the voltage observation value u output from the voltage sensor 11 every predetermined time, and outputs the result as the voltage observation value u (k). The A / D conversion unit 41 repeatedly digitally converts the current observation value i output from the current sensor 12 every predetermined time, and outputs the result as the current observation value i (k).

  The battery state estimation unit 50 acquires the voltage observation value u (k) and the current observation value i (k) from the A / D conversion unit 41. Further, the battery state estimation unit 50 acquires the post-condition estimation value x ^ (k−1) and various parameters of the equivalent circuit model 30 from the storage unit 40.

The a posteriori state estimated value x ^ (k-1) is a state vector of the secondary battery 5 that is estimated by the charge state estimating unit 21 at time k1-1. Here, state estimation value x (k), which is a state vector of secondary battery 5 at time k, is expressed by the following equation (9), and includes SOC value SOC (k), voltage u1 (k), and voltage u2 ( k).

The battery state estimation unit 50 determines the prior state estimated value x ^-(k) based on the posterior state estimated value x ^ (k-1), the voltage observed value u (k), and the current observed value i (k). Is calculated. For example, the battery state estimation unit 50 can obtain the prior state estimated value x ^-(k) by performing the calculation of the following equation (10) corresponding to the above equation (5). Further, the current observation value i (k) is an integrated value of the current i flowing through the secondary battery 5 in a predetermined period as shown in the following formula (11), and the predetermined period is, for example, 1 second.

  In the above equation (10), f (x ^ (k-1), i (k)) is the post-state state estimated value x ^ (k-1), the current observation value i (k), and the prior state estimated value. It is a function indicating the relationship with x ^-(k). In the above formula (10), the current observation value i (k) is used, but the current observation value i (k−1) may be used.

Further, the battery state estimation unit 50 estimates the voltage estimated value u which is an estimated value of the both-ends voltage u of the secondary battery 5 based on the observed current value i (k) and the prior state estimated value x ^-(k). ^ (K) is calculated. For example, the battery state estimation unit 50 can obtain the estimated voltage value u ^ (k) by performing the calculation of the following formula (12) corresponding to the above formula (6). The battery state estimation unit 50 outputs the prior state estimation value x ^-(k) to the filter unit 52.

Further, the battery state estimation unit 50 calculates a difference u˜ (k) between the voltage observation value u (k) and the voltage estimation value u ^ (k). For example, the battery state estimation unit 50 can obtain the difference u˜ (k) by performing the calculation of the following equation (13). The battery state estimation unit 50 outputs the difference u˜ (k) to the gain setting unit 51.

The gain setting unit 51 acquires the difference u˜ (k) from the battery state estimation unit 50. Calculating the gain setting unit 51, the difference u ~ and (k), and the posterior error covariance matrix P (k-1), and the observation noise sigma _w, based on the system noise sigma _v, Kalman gain G (k) is To do. Note that the observation noise Σ _w is the covariance of the observation noise w _k described above, and the system noise Σ _v is the covariance of the system noise v _k described above.

The gain setting unit 51 obtains the prior error covariance matrix P- (k) at time k by, for example, the calculation of the following equation (14). In the following formula (14), “A” can be, for example, the Jacobian of f (x ^ (k−1), i (k)) in the above formula (10). For example, the gain setting unit 51 can obtain the Jacobian “A” by the calculation of the following equation (15) based on the a posteriori state estimated value x ^ (k−1) at the time k−1.

Next, the gain setting unit 51, based on a pre-error covariance matrix P- (k) and system noise sigma _v at time k, to calculate the Kalman gain G (k). For example, the gain setting unit 51 can obtain the Kalman gain G (k) by, for example, the calculation of the following equation (16).

Further, the gain setting unit 51 calculates the posterior error covariance matrix P (k) at time k based on the prior state estimated value x ^-(k), the Kalman gain G (k), and the difference u ~ (k). Calculate. For example, the gain setting unit 51 can obtain the posterior error covariance matrix P (k) by the calculation of the following equation (17).

The filter unit 52 calculates the posterior state estimated value x ^ (k) based on the prior state estimated value x ^-(k) and the Kalman gain G (k). Specifically, the filter unit 52 corrects the prior state estimated value x ^-(k) based on the Kalman gain G (k) and the difference u ~ (k). ) Is calculated. For example, the filter unit 52 can obtain the correction values x to (k) by the calculation of the following equation (18).

The filter unit 52 calculates the posterior state estimated value x ^ (k) based on the prior state estimated value x ^-(k) and the corrected values x ~ (k). For example, the filter unit 52 can obtain the posterior state estimated value x ^ (k) by the calculation of the following equation (19).

Then, the filter unit 52 extracts the posterior state estimated value SOC ^ (k) of the SOC included in the posterior state estimated value x ^ (k). For example, the filter unit 52 can obtain the posterior state estimated value SOC ^ (k) by the calculation of the following equation (20).

  The filter unit 52 stores the extracted posterior state estimated value x ^ (k) in the storage unit 40 and outputs the calculated posterior state estimated value SOC ^ (k) to the vehicle control device 2. Note that the calculation method of the posterior state estimated value SOC ^ (k) is not limited to the above-described example. The calculation method of the posterior state estimated value SOC ^ (k) only needs to be able to calculate the posterior state estimated value SOC ^ (k) by an estimation calculation in which a Kalman filter is applied to the equivalent circuit model. For example, the charging state estimation unit 21 can also calculate the posterior state estimated value SOC ^ (k) by a known estimation process.

[2.2. Parameter setting process]
As described above, the charging state estimation unit 21 includes the storage unit 40 and the switching unit 43. The switching unit 43 includes a determination unit 44 and a parameter setting unit 45.

The storage unit 40 stores the above-described maximum error priority parameter and average error priority parameter. The maximum error priority parameters as parameters include the first system noise sigma _v1 and the first observation noise sigma _w1. Similarly, the average error priority parameter includes a second system noise Σ _v2 and a second observation noise Σ _w2 as parameters.

  In addition, the storage unit 40 has information on a period T1 (hereinafter referred to as a first period T1) in which the OCV influence, which is an influence of the OCV value, on the SOC value estimation calculation is relatively large, and the OCV influence is relatively Information of a small period T2 (hereinafter referred to as the second period T2) is stored.

The information of the first period T1 includes information such as time t1 to time t2 and time t4 to time t5, and the information of the second period T2 includes information such as time t2 to time t3 and time t5 to time t6. Time etc. t1 to t6, the elapsed time from the start of the estimation calculation of the SOC value in the state indicating an ON from the state where the ignition signal S _IGN indicates OFF of the ignition switch, available rechargeable battery 5 It is the elapsed time after connecting to the power converter 4 so that Information such as the periods T1 and T2 and the times t1 to t6 can be determined by the type, characteristics, and the like of the secondary battery 5. For example, if it is known that a predetermined time from the start of the estimation calculation is a period with a large error, the period is set in advance as the first period T1, and the period thereafter is set as the second period T2. Can be set in advance.

  The determination unit 44 determines the state of the OCV influence in the mode selected from the first mode and the second mode, and the OCV influence is relatively large or the OCV influence is relatively small. It is determined whether it is. The first mode is a mode for determining the state of the OCV influence based on the current observation value i (t), and the second mode is the first period T1 and the second period stored in the storage unit 40. In this mode, the state of the OCV influence is determined based on the information of T2.

  First, the first mode will be described. As described above, since the OCV influence increases as the current observation value i (t) increases, the determination unit 44 determines the state of the OCV influence based on the current observation value i (t).

  For example, when the current observation value i (t) is greater than or equal to the threshold value ith and the current observation value i (t) is greater than or equal to the threshold value ith and less than the threshold value ith, the determination unit 44 determines that the current observation value i (t) is within the predetermined time Tth. Determines that the OCV influence is relatively large. On the other hand, when the current observation value i (t) is less than the threshold value ith except for the case where the current observation value i (t) is less than the threshold value ith and within the predetermined time Tth, the OCV influence is relatively small. It is determined that

  Next, the second mode will be described. The determination unit 44 determines whether the current time t is included in the first period T1 based on the information of the first period T1 and the information of the second period T2 stored in the storage unit 40. Is included. For example, it is assumed that the period from time t1 to time t2 is the first period T1, and the period from time t2 to time t3 is the second period T2.

  In this case, the determination unit 44 determines that the current time t is included in the first period T1 if the current time t is within the period from time t1 to time t2 (t1 ≦ t <t2). Further, the determination unit 44 determines that the current time t is included in the second period T2 if the current time t is within the period from time t2 to time t3 (t2 ≦ t <t3). When the current time t is included in the first period T1, the OCV influence is relatively large. When the current time t is included in the second period T2, the OCV influence is relatively small.

  Note that only the information of the first period T1 can be stored in the storage unit 40. In this case, if the current time t is within the first period T1, the determination unit 44 determines that the current time t is included in the first period T1, and the current time t must be within the first period T1. For example, it is determined that the current time t is included in the second period T2.

  The parameter setting unit 45 causes the calculation unit 42 to output the estimated value of the SOC calculated by the estimation calculation using the parameters of the Kalman filter according to the determination result of the determination unit 44. The calculation unit 42 performs an SOC value estimation calculation using the maximum error priority parameter (hereinafter referred to as a first estimation calculation) and an SOC value estimation calculation using the average error priority parameter (hereinafter referred to as a second estimation calculation). Can be switched or performed in parallel.

  First, switching processing between the first estimation calculation and the second estimation calculation will be described. When the determination unit 44 determines that the OCV influence is relatively large, the parameter setting unit 45 sets the Kalman filter parameter as the maximum error priority parameter. Further, when the determination unit 44 determines that the OCV influence is relatively small, the parameter setting unit 45 sets the Kalman filter parameter as the average error priority parameter.

For example, when the determination unit 44 determines that the OCV influence is relatively large, the parameter setting unit 45 sets the first system noise Σ _v1 and the first observation noise Σ _w1 in the calculation unit 42. Calculation unit 42 performs calculation of the equation (14) the first observation noise sigma _w1 as observation noise sigma _w, performs calculation of the equation (16) the first system noise sigma _v1 as system noise sigma _v . Thereby, the error of the SOC value can be suppressed and the convergence can be improved, and the estimation accuracy of the SOC value can be improved.

The parameter setting unit 45 sets the second system noise Σ _v2 and the second observation noise Σ _w2 in the calculation unit 42 when the determination unit 44 determines that the OCV influence is relatively small. Calculation unit 42 performs calculation of the equation (14) the second observation noise sigma _w2 as observation noise sigma _w, performs calculation of the equation (16) the second system noise sigma _v2 as system noise sigma _v . Thereby, the average error of the SOC value can be suppressed, and the estimation accuracy of the SOC value can be improved.

  FIG. 6 is a diagram showing a temporal transition of the estimation error of the SOC value. In FIG. 6, the horizontal axis indicates the time since the start of the estimation calculation, and the vertical axis indicates the SOC value estimation error. As shown in FIG. 6, when the maximum error priority parameter is used, the variation range of the SOC value estimation error becomes small, and when the average error priority parameter is used, the average value of the SOC value estimation error becomes small. .

  Next, selection of the output of the SOC value calculated by the parallel processing of the first estimation calculation and the second estimation calculation will be described. First, the parameter setting unit 45 sets both the maximum error priority parameter and the average error priority parameter in the calculation unit 42. Then, based on the determination result of the state of the OCV influence by the determination unit 44, the parameter setting unit 45 selects either the SOC value calculated by the first estimation calculation or the SOC value calculated by the second estimation calculation. An output set value indicating whether to output is set in the calculation unit 42.

  The parameter setting unit 45 sets an output setting value indicating that the SOC value calculated by the first estimation calculation is output to the calculation unit 42 when the determination unit 44 determines that the OCV influence is relatively large. . Also, the parameter setting unit 45 outputs an output setting value indicating that the SOC value calculated by the second estimation calculation is output to the calculation unit 42 when the determination unit 44 determines that the OCV influence is relatively small. Set.

  The calculation unit 42 repeatedly performs the first estimation calculation and the second estimation calculation in parallel, and calculates the SOC value calculated by the first estimation calculation and the second estimation calculation based on the output set value. One of the set SOC values is output to the vehicle control device 2.

  For example, when the output setting value indicates that the SOC value calculated by the first estimation calculation is output, the calculation unit 42 outputs the SOC value calculated by the first estimation calculation to the vehicle control device 2. Further, when the output setting value indicates that the SOC value calculated by the second estimation calculation is output, the calculation unit 42 outputs the SOC value calculated by the second estimation calculation to the vehicle control device 2.

  As described above, the charging state estimation unit 21 can calculate by parallel processing of the first estimation calculation and the second estimation calculation, and can select and output one of the calculated SOC values. . Therefore, compared with the case where the parameter of the Kalman filter is switched, the SOC value estimated by the estimation calculation using the parameter corresponding to the state of the OCV influence can be output quickly.

  The parameters of the Kalman filter set by the parameter setting unit 45 are fixed values or initial values. When the parameter of the Kalman filter set by the parameter setting unit 45 is a fixed value, the calculation unit 42 performs the estimation calculation of the SOC value by continuously using the parameter set by the parameter setting unit 45 as the parameter of the Kalman filter. .

  When the parameter of the Kalman filter set by the parameter setting unit 45 is an initial value, the calculation unit 42 uses the parameter set by the parameter setting unit 45 as an initial value when the estimation calculation is started, and performs the initial SOC. Perform value estimation. Thereafter, the calculation unit 42 can perform the second and subsequent SOC value estimation calculations while repeating the parameter update.

Calculation unit 42, for example, can be based on the calculated values or observed values such as correction values x ~ (k), and updates the system noise sigma _v used for estimating calculation of the second and subsequent SOC value. The arithmetic unit 42 can be based on the calculated values or observed values, such as voltage observations u (t) and state estimate x ^, and updates the observation noise sigma _w.

The arithmetic unit 42, the other to a fixed value one of the observation noise sigma _w and system noise sigma _v as an initial value, it is also possible to perform estimation calculation of the SOC value. In the above-described example, the observation noise Σ _w and the system noise Σ _v are described as examples of the parameters set in the calculation unit 42 by the switching unit 43. However, the parameters set in the calculation unit 42 are the observation noises. but it is not limited to Σ _w and system noise Σ _v. Parameter set to the operation unit 42 by the switching unit 43 may be a parameter other than the observation noise sigma _w and system noise sigma _v.

[3. Processing by the charge state estimation unit 21]
Next, a flow of processing executed by the charging state estimation unit 21 which is an example of an estimation device will be described using a flowchart.

  First, switching processing between the first estimation calculation and the second estimation calculation will be described. FIG. 7 is a flowchart illustrating an example of a processing procedure executed by the charge state estimation unit 21, and is a process that is repeatedly executed.

  As illustrated in FIG. 7, the switching unit 43 of the charging state estimation unit 21 determines whether or not the charging state estimation unit 21 is set to the first mode (step S20). When it is determined that the first mode is set (step S20: Yes), the switching unit 43 determines whether or not the current observation value i (k) is equal to or greater than the threshold value isth (step S21). When the switching unit 43 determines that the current observation value i (k) is greater than or equal to the threshold value ith (step S21: Yes), the switching unit 43 sets the maximum error priority parameter in the calculation unit 42 (step S23).

  When the switching unit 43 determines that the current observation value i (k) is not greater than or equal to the threshold value ith (step S21: No), the switching unit 43 is constant from the timing when the current observation value i (k) is greater than or equal to the threshold value ith and less than the threshold value ith. It is determined whether or not time Tth has elapsed (step S22).

  When the switching unit 43 determines in step S22 that the predetermined time Tth has elapsed from the timing when it becomes less than the threshold value isth (step S22: Yes), the switching unit 43 sets the average error priority parameter in the calculation unit 42 (step S25). In addition, when the switching unit 43 determines that the predetermined time Tth has not elapsed since the timing when the current observation value i (k) is equal to or greater than the threshold value ith and less than the threshold value ith (step S22: No), the maximum error priority parameter is set. It sets to the calculating part 42 (step S23).

  Further, when the switching unit 43 determines in step S20 that the charging state estimation unit 21 is not set to the first mode (step S20: No), whether or not the current time t is the first period T1. Is determined (step S24).

  If the switching unit 43 determines that the current time t is not the first period T1 (step S24: No), the switching unit 43 sets the average error priority parameter in the calculation unit 42 (step S25). Further, when the switching unit 43 determines that the time t is the first period T1 (step S24: Yes), the switching unit 43 sets the maximum error priority parameter in the calculation unit 42 (step S26).

  When the processing of steps S23, S25, and S26 by the switching unit 43 is completed, the calculation unit 42 of the charging state estimation unit 21 performs the SOC value estimation calculation using the parameters set by the switching unit 43 ( Step S27). Calculation unit 42 outputs the SOC value, which is the result of the estimation calculation in step S27, to vehicle control device 2 (step S28).

  Next, selection of the output of the SOC value calculated by the parallel processing of the first estimation calculation and the second estimation calculation will be described. FIG. 8 is a flowchart showing another example of a processing procedure executed by the charge state estimation unit 21, and is a process that is repeatedly executed.

  Note that the processing of steps S31 and S33, which will be described later, is executed only for the first time in the SOC value estimation calculation that is repeatedly performed, for example, at the time of the first estimation calculation when the ignition is turned on. The Further, the processes of steps S35 to S37 and S40 are the same as the processes of steps S20 to S22 and S24 shown in FIG.

  As shown in FIG. 8, the switching unit 43 of the charging state estimation unit 21 sets a maximum error priority parameter as an initial parameter for the calculation unit 42 (step S31), and the calculation unit 42 uses the maximum error priority parameter. Then, the SOC value estimation calculation is executed (step S32). Further, the switching unit 43 sets an average error priority parameter as an initial parameter for the calculation unit 42 (step S33), and the calculation unit 42 performs an SOC value estimation calculation using the average error priority parameter (step S33). S34).

  When the switching unit 43 determines that the current observation value i (k) is greater than or equal to the threshold value isth (step S35: Yes), or when the current observation value i (k) is greater than or equal to the threshold value ith and less than the threshold value ith. If it is determined that the predetermined time Tth has not elapsed (step S37: No), the SOC value estimated by the calculation unit 42 using the maximum error priority parameter is output from the calculation unit 42 (step S38). Further, when the switching unit 43 determines that the current time t is the first period T1 (step S40: Yes), the switching unit 43 outputs the SOC value estimated by the calculation unit 42 using the maximum error priority parameter from the calculation unit 42. (Step S38).

  In addition, the switching unit 43 determines that a certain time Tth has elapsed from the timing when the current observation value i (k) is greater than or equal to the threshold value ith and less than the threshold value ith (step S37: Yes), or the current time t is the first time When it is determined that it is not the period T1 of 1 (step S40: No), the SOC value estimated and calculated by the calculation unit 42 using the average error priority parameter is output from the calculation unit 42 (step S39).

  As described above, the charging state estimation unit (an example of the estimation device) according to the present embodiment includes the A / D conversion unit 41 (an example of the observation value acquisition unit), the calculation unit 42, and the switching unit 43. The A / D conversion unit 41 acquires the observed voltage value u (k) and the observed current value i (k) (an example of the observed value of the voltage and current of the secondary battery). The calculation unit 42 performs the charge state of the secondary battery 5 by an estimation calculation in which a Kalman filter is applied to the equivalent circuit model 30 of the secondary battery 5 based on the observed voltage value u (k) and the observed current value i (k). Is estimated. When the OCV influence (an example of the influence of the open circuit voltage of the secondary battery 5 on the SOC value estimation calculation) is relatively large, the switching unit 43 gives priority to the maximum error suppression of the estimation calculation. The first SOC value calculated by the estimation calculation using the priority parameter as the parameter of the Kalman filter is output from the calculation unit 42, and priority is given to the suppression of the average error of the estimation calculation when the OCV influence is relatively small. The second SOC value calculated by the estimation calculation using the average error priority parameter as the parameter of the Kalman filter is output from the calculation unit 42. As a result, the SOC value estimation calculation can be performed with the parameter corresponding to the state of the influence of the OCV, so that the estimation accuracy of the SOC value can be improved.

  In addition, when the current observation value i (k) is the threshold value ith, or when the current observation value i (k) is greater than or equal to the threshold value ith and less than the threshold value ith, a certain time Tth (an example of a set time) has elapsed. If not, the OCV influence is relatively large. The switching unit 43 determines whether the current observation value i (k) is greater than or equal to the threshold value isth, or whether a certain time Tth has elapsed from the timing when the current observation value i (k) is greater than or equal to the threshold value ith and less than the threshold value isth. One of the first SOC value and the second SOC value is output from the computing unit 42 based on whether or not it is either. Thereby, when the current i flowing through the secondary battery 5 is large, the error of the SOC value can be suppressed, the convergence can be improved, and the estimation accuracy of the SOC value can be improved. Further, when the current i flowing through the secondary battery 5 is small, the maximum error of the SOC value estimation calculation is smaller than when the current i flowing through the secondary battery 5 is large, and the average error can be reduced by using the average error priority parameter. Since it can suppress, the estimation precision of SOC value can be improved.

  In addition, the switching unit 43 displays the elapsed time after connecting the secondary battery 5 to the power conversion unit 4 so that the secondary battery 5 can be used by the power conversion unit 4 and the motor 3 (an example of another device). Based on this, either one of the first SOC value and the second SOC value is output from the computing unit 42. Thereby, the estimation accuracy of the SOC value can be improved without using the current observation value i (k).

  Further, the switching unit 43 sets the parameter according to the OCV influence state among the maximum error priority parameter and the average error priority parameter as the Kalman filter parameter in the calculation unit 42, and the calculation unit 42 is set by the switching unit 43. Estimate with parameters. Thereby, the estimation accuracy of the SOC value can be improved while reducing the processing load as compared with the case where the estimation calculation is performed with the maximum error priority parameter and the average error priority parameter, respectively.

  In addition, the calculation unit 42 performs in parallel an estimation calculation using the Kalman filter parameter as the maximum error priority parameter and an estimation calculation using the Kalman filter parameter as the average error priority parameter, and performs the first calculation according to the state of the OCV influence. One of the SOC value and the second SOC value is output. Thereby, compared with the case where the parameter of a Kalman filter is switched, the SOC value estimated by the estimation calculation using the parameter according to the state of the influence of OCV can be output rapidly.

The parameters of the Kalman filter include observation noise Σ _w (an example of a parameter related to noise of an observation value) and system noise Σ _v (an example of a parameter related to noise of an equivalent circuit model). Thereby, the estimation accuracy of the SOC value can be improved.

  Further effects and modifications can be easily derived by those skilled in the art. Thus, the broader aspects of the present invention are not limited to the specific details and representative embodiments shown and described above. Accordingly, various modifications can be made without departing from the spirit or scope of the general inventive concept as defined by the appended claims and their equivalents.

DESCRIPTION OF SYMBOLS 1 Power supply system 2 Vehicle control apparatus 3 Motor (an example of another apparatus)
4 Power converter (an example of another device)
DESCRIPTION OF SYMBOLS 5 Secondary battery 6 Battery monitoring system 7 Relay 10 Battery state monitoring part 11 Voltage sensor 12 Current sensor 20 Switch control part 21 Charging state estimation part (an example of an estimation apparatus)
30 equivalent circuit model 40 storage unit 41 A / D conversion unit (an example of an observation value acquisition unit)
42 arithmetic unit 43 switching unit 44 determination unit 45 parameter setting unit 50 battery state estimation unit 51 gain setting unit 52 filter unit 100 vehicle mounting system

Claims (8)

An observation value acquisition unit for acquiring observation values of the voltage and current of the secondary battery;
An arithmetic unit that estimates an SOC value indicating a state of charge of the secondary battery based on the observation value acquired by the observation value acquisition unit, by an estimation calculation in which a Kalman filter is applied to the equivalent circuit model of the secondary battery. When,
When the influence of the open circuit voltage of the secondary battery on the estimation calculation of the SOC value is relatively large, a maximum error priority parameter giving priority to suppression of the maximum error of the estimation calculation is a parameter of the Kalman filter. When the first SOC value, which is the SOC value calculated by the estimation calculation, is output from the calculation unit and the influence of the open circuit voltage is relatively small, the average error of the estimation calculation is A switching unit that outputs a second SOC value, which is the SOC value calculated by the estimation calculation using an average error priority parameter that prioritizes suppression as a parameter of the Kalman filter, from the calculation unit. Estimating device. When the observed value of the current is greater than or equal to a threshold value, or when the set time has not elapsed since the timing when the observed value of the current has decreased from the threshold value to less than the threshold value, the influence is relatively large. Yes,
The switching unit is
Based on whether the observed value of the current is greater than or equal to the threshold or whether a set time has not elapsed since the time when the observed value of the current is greater than or equal to the threshold and less than the threshold. The estimation apparatus according to claim 1, wherein any one of the first SOC value and the second SOC value is output from the calculation unit. The switching unit is
One of the first SOC value and the second SOC value is calculated based on an elapsed time since the secondary battery is connected to the other device so that the secondary battery can be used by the other device. The estimation apparatus according to claim 1, wherein the estimation device outputs the signal from a unit. The switching unit is
Of the maximum error priority parameter and the average error priority parameter, a parameter according to the state of influence is set as the Kalman filter parameter in the arithmetic unit,
The computing unit is
The estimation apparatus according to claim 1, wherein the estimation calculation is performed using a parameter set by the switching unit. The computing unit is
The estimation calculation with the Kalman filter parameter as the maximum error priority parameter and the estimation calculation with the Kalman filter parameter as the average error priority parameter are performed in parallel.
The estimation apparatus according to claim 1, wherein one of the first SOC value and the second SOC value is output according to the state of influence. The Kalman filter parameters are:
The estimation apparatus according to claim 1, further comprising a parameter related to noise of the observed value and a parameter related to noise of the equivalent circuit model. Obtaining observation values of voltage and current of the secondary battery;
Estimating an SOC value indicating a state of charge of the secondary battery by an estimation calculation in which a Kalman filter is applied to the equivalent circuit model of the secondary battery based on the observed value;
When the influence of the open circuit voltage of the secondary battery on the estimation calculation of the SOC value is relatively large, a maximum error priority parameter giving priority to suppression of the maximum error of the estimation calculation is a parameter of the Kalman filter. When the SOC value calculated by the estimation calculation is output and the influence of the open circuit voltage is relatively small, an average error priority parameter giving priority to suppression of the average error of the estimation calculation is set as the Kalman filter. A step of outputting the SOC value calculated by the estimation calculation as a parameter of the estimation method. A procedure for obtaining observations of voltage and current of the secondary battery;
A procedure for estimating an SOC value indicating a state of charge of the secondary battery by an estimation calculation in which a Kalman filter is applied to the equivalent circuit model of the secondary battery based on the observed value;
When the influence of the open circuit voltage of the secondary battery on the estimation calculation of the SOC value is relatively large, a maximum error priority parameter giving priority to suppression of the maximum error of the estimation calculation is a parameter of the Kalman filter. When the SOC value calculated by the estimation calculation is output and the influence of the open circuit voltage is relatively small, an average error priority parameter giving priority to suppression of the average error of the estimation calculation is set as the Kalman filter. And a procedure for outputting the SOC value calculated by the estimation calculation using the parameter as a parameter.

Images