JP6213333B2 - Estimation program, estimation method, and estimation apparatus - Google Patents

Description

  The present invention relates to an estimation program, an estimation method, and an estimation apparatus.

  2. Description of the Related Art In recent years, in an electric vehicle, an electric motorcycle, or the like, an operation is known in which a use time loss due to a charging time is eliminated by exchanging a secondary battery for each module. Moreover, a lithium ion secondary battery may be used as a secondary battery used for an electric vehicle, an electric motorcycle, or the like. In the operation of replacing the battery module, it is necessary to manage the SoC (States of Charge) that is the charging rate of the lithium ion secondary battery in each battery module. SoC can be measured by measuring the amount of electricity inside the lithium ion secondary battery, but it is difficult to measure during actual operation. For this reason, the SoC in actual operation is estimated from the measurable terminal voltage and current. The SoC is estimated by applying a Kalman filter to an equivalent electric circuit model in which a lithium ion secondary battery is regarded as an electric circuit.

JP 2008-010420 A JP 2012-149947 A

  However, the Kalman filter corrects the SoC based on the error of the predicted terminal voltage of the lithium ion secondary battery. However, since the Kalman filter has a feature of correcting the correction capability, the error accumulates with time. Therefore, when the Kalman filter predicts a large error, the accumulated error includes a large error. For this reason, the Kalman filter excessively corrects the SoC over time, so that the estimation accuracy of the SoC is significantly deteriorated.

  In one aspect, the present invention is to provide an estimation program, an estimation method, and an estimation apparatus that can prevent deterioration in estimation accuracy of a charging rate.

  In one aspect, the estimation program calculates a charging rate and a predicted terminal voltage of the battery using a Kalman filter based on the measured current and measured terminal voltage of the battery, and the measured terminal voltage and the predicted terminal voltage. The process which calculates the difference with is executed. Further, the estimation program causes the computer to execute a process of turning off the Kalman gain for a predetermined step when the amount of change in the measured current exceeds a predetermined value. The estimation program causes the computer to execute a process of setting the component corresponding to the charging rate to a different value in the matrix representing the predicted noise of the Kalman filter when the difference exceeds a threshold value after the predetermined step has elapsed.

  It is possible to prevent the deterioration of the charging rate estimation accuracy.

FIG. 1 is a block diagram illustrating an example of the configuration of the estimation apparatus according to the embodiment. FIG. 2 is an explanatory diagram illustrating an example of an equivalent electric circuit model of a lithium ion secondary battery. FIG. 3 is a graph showing an example of the OCV characteristic curve. FIG. 4 is an explanatory diagram illustrating an example of the SoC estimation process using the Kalman filter. FIG. 5 is an explanatory diagram illustrating an example of switching the predicted noise Σv. FIG. 6 is a flowchart illustrating an example of processing performed by the estimation apparatus according to the embodiment. FIG. 7 is a flowchart illustrating an example of the Kalman filter Σv switching process. FIG. 8 is an explanatory diagram illustrating an example of a computer that executes an estimation program.

  Hereinafter, embodiments of an estimation program, an estimation method, and an estimation apparatus disclosed in the present application will be described in detail based on the drawings. The disclosed technology is not limited by the present embodiment. Further, the following embodiments may be appropriately combined within a consistent range.

  FIG. 1 is a block diagram illustrating an example of the configuration of the estimation apparatus according to the embodiment. The estimation apparatus 100 illustrated in FIG. 1 includes a measurement unit 101, an output unit 102, a storage unit 120, and a control unit 130. The estimation apparatus 100 is incorporated in, for example, a lithium ion secondary battery module. Moreover, the estimation apparatus 100 may include a control unit that controls charging / discharging of the cell. Note that the estimation device 100 may be mounted as a function of a control device such as an electric vehicle or an electric motorcycle to control the lithium ion secondary battery module.

  Measurement unit 101 measures the current and terminal voltage of the lithium ion secondary battery. The measurement part 101 measures an electric current and a terminal voltage, for example with an ammeter and a voltmeter, respectively. The measurement unit 101 outputs the measured current and terminal voltage to the control unit 130 as a measurement current and a measurement terminal voltage, respectively. In addition, the measurement unit 101 acquires the operation mode of the lithium ion secondary battery and outputs it to the control unit 130.

  When the charging rate information is input from the control unit 130, the output unit 102 outputs the charging rate information to, for example, another device. Here, the other device is, for example, a vehicle control device if it is an electric vehicle, an electric motorcycle, or the like, and is a display unit that displays a screen if it is a portable information terminal such as a smartphone. When the charging rate information is input, the other device displays the charging rate on a meter or a display unit of the vehicle so that the user can visually recognize the charging rate information.

  The storage unit 120 is realized by, for example, a semiconductor memory device such as a RAM (Random Access Memory) or a flash memory, or a storage device such as a hard disk or an optical disk. The storage unit 120 includes a condition storage unit 121. In addition, the storage unit 120 stores information used for processing in the control unit 130.

  The condition storage unit 121 stores an OCV (Open Circuit Voltage) characteristic curve, various parameters of a Kalman filter, each parameter of an equivalent electric circuit model, a predetermined value of a change amount of a measured current, a threshold value a, a threshold value b, and the like. The OCV characteristic curve is a graph showing the OCV-SoC characteristic of the lithium ion secondary battery, and details will be described later. Examples of the various parameters of the Kalman filter include Σv indicating prediction noise, Σw indicating measurement noise, and the like. Σv indicating the predicted noise stores at least two kinds of values used in normal time and values used when the amount of change in the measured current exceeds a predetermined value. In addition, the condition storage unit 121 uses, for example, a value corresponding to the reliability of the equivalent electric circuit model obtained from the current change pattern, that is, the predetermined value of the change amount of the measurement current, as Σv indicating the prediction noise. , At least two types or more may be stored. The condition storage unit 121 may store values corresponding to these reliability levels, for example, as a list.

Examples of each parameter of the equivalent electric circuit model include R ₀ , R ₁ , R ₂ , C ₁ , and C ₂ described later. Each parameter of the equivalent electric circuit model is set in advance so as to match an actual lithium ion secondary battery. The predetermined value of the change amount of the measurement current indicates a predetermined change amount when the measurement current increases in the same direction. The predetermined value of the change amount of the measurement current becomes a trigger for executing a process of switching the component corresponding to the SoC in the matrix representing the predicted noise of the Kalman filter when the change amount of the measurement current exceeds the predetermined value. It is a threshold value.

  The threshold value a is a threshold value for resetting the P matrix when the component corresponding to the SoC of the Kalman filter covariance matrix P (hereinafter also simply referred to as the P matrix) exceeds the threshold value a, that is, for setting the initial value. It is. The threshold value a indicates a SoC shift value that may be shifted in one step, and is represented by the following formula (1). Here, c is the average slope of the intermediate region of the OCV characteristic curve, and for example, the slope when SoC is 50% can be used. I indicates the maximum current assumed. sca indicates the charge capacity of the lithium ion secondary battery. That is, the threshold value a is obtained by converting the allowable change amount of SoC into the terminal voltage.

  The threshold value b is a threshold value for a difference between the measured terminal voltage and the predicted terminal voltage predicted by the Kalman filter, and is a threshold value for switching a component corresponding to the SoC in the matrix representing the predicted noise of the Kalman filter. That is, the threshold value b is a threshold value for switching the component corresponding to the SoC in the matrix representing the prediction noise of the Kalman filter when the value of the difference exceeds the threshold value b.

  The control unit 130 is realized, for example, by executing a program stored in an internal storage device using a RAM as a work area by a CPU (Central Processing Unit), an MPU (Micro Processing Unit), or the like. Further, the control unit 130 may be realized by an integrated circuit such as an application specific integrated circuit (ASIC) or a field programmable gate array (FPGA). The control unit 130 includes a calculation unit 131 and a setting unit 132, and realizes or executes information processing functions and operations described below. Note that the internal configuration of the control unit 130 is not limited to the configuration illustrated in FIG. 1, and may be another configuration as long as the information processing described below is performed.

When the measurement current and the measurement terminal voltage are input from the measurement unit 101, the calculation unit 131 calculates the SoC (charge rate) and the predicted terminal voltage using a Kalman filter. In addition, the calculation unit 131 calculates the difference y ^* (k) between the measurement terminal voltage and the predicted terminal voltage as an error. Further, the calculation unit 131 outputs the Kalman filter covariance matrix P (k) to the setting unit 132 for each step. Further, the calculation unit 131 receives the covariance matrix P (k) for which the processing in the setting unit 132 has been completed from the setting unit 132 for each step. The calculation unit 131 receives the switching information of the predicted noise Σv from the setting unit 132.

Here, the SoC estimation process using the Kalman filter will be described with reference to FIGS. 2 and 3. FIG. 2 is an explanatory diagram illustrating an example of an equivalent electric circuit model of a lithium ion secondary battery. As shown in FIG. 2, the equivalent electric circuit model of the lithium ion secondary battery has a current source V _{OCV_DC} (SoC) and a V _{OCV_CC} (SoC) representing a potential difference OCV that changes in accordance with a change in SoC. Here, the current sources V _{OCV_DC} (SoC) and V _{OCV_CC} (SoC) represent the potential difference OCV during discharging and the potential difference OCV during charging, respectively. Also, the equivalent electric circuit model shown in FIG. 2 includes a resistor R ₀ that represents a change in voltage v ₀ generated with respect to a current change, and an RC circuit that represents a change in transient voltages v ₁ and v ₂ with respect to the current change. (R ₁ C ₁ and R ₂ C ₂ ). That is, the equivalent electric circuit model is used to make the internal resistance of the lithium ion secondary battery a function that depends on the measurement current i. Note that the accuracy of the internal resistance greatly affects the SoC estimation accuracy using the Kalman filter.

The terminal voltage v of the equivalent electric circuit model shown in FIG. 2 is expressed by the sum of the potential difference OCV, the voltage v ₀ , the voltage v _1, and the voltage v ₂ . That is, the terminal voltage v can be expressed by the following formula (2). The calculation unit 131 estimates the SoC using the Kalman filter and using the fact that the OCV changes according to the SoC.

  FIG. 3 is a graph showing an example of the OCV characteristic curve. The OCV characteristic curve shown in FIG. 3 is stored in the condition storage unit 121. The calculation unit 131 refers to the OCV characteristic curve stored in the condition storage unit 121 when performing SoC estimation using a Kalman filter.

The calculation unit 131 uses the equivalent electric circuit model to calculate the current v ₁ , v ₂ , and OCV based on v ₁ , v ₂ , and OCV one step before and the input measurement current i, Predict the predicted terminal voltage. In addition, the calculation unit 131 calculates SoC from the OCV using the OCV characteristic curve. Here, the corresponding equations are shown in the following equations (3) to (5). Here, Formula (3) shows the current state estimated value. Formula (4) shows the terminal voltage based on the OCV characteristic. Formula (5) shows the state estimated value one step before.

Next, the calculation unit 131 calculates the difference between the actually measured measurement terminal voltage and the predicted terminal voltage. The calculation unit 131 considers noise assumed for v ₁ , v ₂ , and SoC and the terminal voltage, and causes v ₁ , v ₂ , and SoC that cause an error between the measurement terminal voltage and the predicted terminal voltage. To estimate the error. The calculating unit 131 corrects the estimated v ₁ , v ₂ , and SoC errors in addition to the predicted v ₁ , v ₂ , and SoC, and thereby outputs the correct estimated SoC value as the charging rate information. Output to.

は、ステップｋ−１のカルマンフィルタの状態推定値を示し、以下、１ステップ前の状態推定値という。
［文字２］

は、ステップｋのカルマンフィルタの状態推定値を示し、以下、状態推定値という。
［文字３］

は、ステップｋの測定端子電圧を示し、以下、測定端子電圧という。
［文字４］

は、ステップｋのカルマンフィルタの状態推定値の逆行列を示し、以下、状態推定値の逆行列という。
［文字５］

は、ステップｋのカルマンフィルタの状態推定値の修正値を示し、以下、修正値という。Ｇ（ｋ）は、ステップｋのカルマンゲインを示す。Ａは、ヤコビアンを示す。Ｐ（ｋ）は、ステップｋの誤差の共分散行列、つまり推定値の精度を示す。Σｖは、予測ノイズを示す。Σｗは、測定ノイズを示す。 Next, details of the SoC estimation process using the Kalman filter will be described with reference to FIG. FIG. 4 is an explanatory diagram illustrating an example of the SoC estimation process using the Kalman filter. First, characters used in the description of FIG. 4 will be described. k indicates the number of steps of the Kalman filter.
[Character 1]

Indicates a state estimated value of the Kalman filter in step k-1, and is hereinafter referred to as a state estimated value one step before.
[Character 2]

Indicates the estimated state value of the Kalman filter in step k, and is hereinafter referred to as the estimated state value.
[Character 3]

Indicates the measurement terminal voltage of step k, and hereinafter referred to as the measurement terminal voltage.
[Character 4]

Indicates an inverse matrix of the state estimation value of the Kalman filter in step k, and is hereinafter referred to as an inverse matrix of the state estimation value.
[Character 5]

Indicates a correction value of the estimated state value of the Kalman filter in step k, and is hereinafter referred to as a correction value. G (k) represents the Kalman gain of step k. A indicates Jacobian. P (k) represents the error covariance matrix of step k, that is, the accuracy of the estimated value. Σv indicates prediction noise. Σw represents measurement noise.

  Here, the covariance matrix P (k) represents a ratio of whether the cause of the prediction error is due to an internal resistance error or an SoC error. The estimation capability of the covariance matrix P (k) deteriorates due to an increase in the ratio to the SoC that is inherently difficult to increase due to an error caused by an error in the equivalent electric circuit model.

  Since the Kalman filter is a feedback system process, in the description of the SoC estimation process in FIG. 4, the process will be described from “S” for convenience. In this embodiment, the calculation unit 131 executes processes corresponding to steps S11 to S16 and steps S20 to S22 of the following Kalman filter, and the setting unit 132 executes processes corresponding to steps S17 to S19. . Therefore, in the following description, it is assumed that the control unit 130 executes a Kalman filter process.

  When the measurement current i (k) is input, the control unit 130 uses the following equation (6) to calculate the state estimation value based on the state estimation value one step before and the measurement current i (k). An inverse matrix, that is, a predicted value is calculated (step S11). Here, the measurement current i (k) shows a case where the measurement current i (k) is input every second, and has the relationship of the following formula (7).

When the measurement terminal voltage is input, the control unit 130 uses the following equation (8) to measure the measurement terminal voltage based on the measurement terminal voltage, the inverse matrix of the state estimation value, and the measurement current i (k). And the difference y ^* (k) between the predicted terminal voltage y (k) and the predicted terminal voltage y (k) (step S12). Moreover, Formula (8) can also be expressed as the following Formula (9) using the predicted terminal voltage y (k). Here, the measurement terminal voltage has the relationship of the following formula (10).

  Next, the control unit 130 calculates the Jacobian A using the following equation (11) based on the state estimated value one step before (Step S13).

Based on the Jacobian A, the one-step previous covariance matrix P (k−1), and the prediction noise Σv, the control unit 130 uses the following equation (12) to inverse the covariance matrix P ⁻ ( k) is calculated (step S14). Here, when executing step S14, the control unit 130 calculates the inverse matrix P ⁻ (k) of the covariance matrix using the prediction noise Σv based on the result determined in step S19 described later.

Based on the inverse matrix P ⁻ (k) of the covariance matrix and the measurement noise Σw, the control unit 130 calculates the Kalman gain G (k) using the following equation (13) (step S15).

The control unit 130 calculates the covariance matrix P (k) using the following equation (14) based on the Kalman gain G (k) and the inverse matrix P ⁻ (k) of the covariance matrix (step) S16).

The control unit 130 determines whether or not the component corresponding to the SoC of the covariance matrix P (k) is larger than the threshold value a (step S17). Here, in the example of FIG. 4, the covariance matrix P (k) is a 3 × 3 matrix because three elements of v ₁ , v ₂ , and SoC are used in the equivalent electric circuit model. The corresponding component is the (3, 3) component. In the equivalent electric circuit model, for example, when four elements such as v ₁ , v ₂ , v ₃ , and SoC are used, the component corresponding to the SoC is the (4, 4) component, so the SoC The component corresponding to is appropriately changed depending on the equivalent electric circuit model to be used.

  When the component corresponding to the SoC of the covariance matrix P (k) is larger than the threshold value a (Yes at Step S17), the control unit 130 resets the covariance matrix P (k) and returns to Step S14. That is, the control unit 130 sets an initial value in the covariance matrix P (k) and returns to step S14 (step S18). When the component corresponding to the SoC of the covariance matrix P (k) is equal to or less than the threshold value a (No at Step S17), the control unit 130 returns to Step S14 without resetting the covariance matrix P (k). The control unit 130 repeats steps S14 to S18 for each step.

  Moreover, the control part 130 performs the process which switches prediction noise (SIGMA) v, when the variation | change_quantity of the measurement electric current i (k) exceeds predetermined value (step S19). The control unit 130 determines whether or not a predetermined step has elapsed since the amount of change in the measured current i (k) exceeds a predetermined value. The control unit 130 turns off the Kalman gain G (k) of the Kalman filter until a predetermined step elapses after the change amount of the measurement current i (k) exceeds a predetermined value, and the measurement terminal voltage and the prediction terminal after the predetermined step elapses. The predicted noise Σv is switched based on the difference from the voltage. Details of the process of switching the predicted noise Σv will be described later. That is, when the amount of change in the measured current i (k) exceeds a predetermined value, the control unit 130 corresponds to SoC in the matrix of the predicted noise Σv according to the difference between the measured terminal voltage and the predicted terminal voltage. Set the components to different values.

Here, an example of the prediction noise Σv is shown in the following equation (15). In Equation (15), l, m, and n represent parameters corresponding to v ₁ , v ₂ , and SoC, respectively.

  As the switching of the predicted noise Σv, the control unit 130 reduces the (3, 3) component, which is a component corresponding to SoC, to 1/10, for example, in the example of Expression (15). That is, the control unit 130 sets a value that narrows the SoC correction range. The control unit 130 executes the process of step S14 using the switched prediction noise Σv. Further, the control unit 130 does not switch the predicted noise Σv when the change amount of the measured current i (k) is equal to or less than a predetermined value. Further, the control unit 130 has been switched to a value such that the previous predicted noise Σv has a narrow SoC correction range, and the amount of change in the measured current i (k) decreases in the measured current i (k). If the amount exceeds the predetermined value, the initial value is switched. That is, the control unit 130 switches the prediction noise Σv in accordance with a prediction error that occurs due to a large error in the equivalent electric circuit model.

Next, the control unit 130 uses the following equation (16) based on the difference y ^* (k) calculated in step S12 and the Kalman gain G (k) calculated in step S15, A correction value for correcting the estimated value is calculated (step S20). When the Kalman gain G (k) is turned off in step S19, the control unit 130 sets G (k) = 0 and calculates a correction value for correcting the state estimation value. That is, the correction value is zero.

  Based on the inverse matrix of the state estimation value calculated in step S11 and the correction value calculated in step S20, the control unit 130 calculates the state estimation value using the following equation (17) (step S21). ). Here, the state estimated value can also be expressed by the following equation (18). The state estimated value is an inverse matrix of the state estimated value when the Kalman gain G (k) is turned off in step S19. In other words, the state estimated value is a predicted value by a coulomb counter.

  Based on the state estimated value, the control unit 130 calculates SoC using the following equation (19) (step S22).

  Thus, the control part 130 can calculate SoC every second by repeating the process of step S11-S22 for every step as a SoC estimation process, for example.

  Returning to FIG. 1, as an initial setting, the setting unit 132 first refers to the condition storage unit 121 by the control unit 130, the threshold value a, the threshold value b, the prediction noise Σv, and the covariance matrix P (k). The initial value of is set. The setting unit 132 receives the measurement current from the measurement unit 101, that is, the measurement current i (k) of the step. The setting unit 132 receives the covariance matrix P (k) from the calculation unit 131 for each step.

  When the covariance matrix P (k) is input from the calculation unit 131, the setting unit 132 determines whether the component corresponding to the SoC of the covariance matrix P (k) is greater than the threshold value a. The setting unit 132 resets the covariance matrix P (k) when the component corresponding to the SoC of the covariance matrix P (k) is larger than the threshold value a. That is, the setting unit 132 resets the covariance matrix P (k) enlarged due to the accumulation of errors. On the other hand, the setting unit 132 does not reset the covariance matrix P (k) when the component corresponding to the SoC of the covariance matrix P (k) is equal to or less than the threshold value a. When the determination of the covariance matrix P (k) is completed, the setting unit 132 outputs the initialized covariance matrix P (k) or the input covariance matrix P (k) to the calculation unit 131.

  In addition, when the measurement current i (k) is input from the measurement unit 101 as the prediction noise Σv switching process, the setting unit 132 determines whether or not the amount of change in the measurement current i (k) exceeds a predetermined value. To do. The setting unit 132 generates switching information that does not switch the predicted noise Σv when the change amount of the measured current i (k) is equal to or less than a predetermined value. When the change amount of the measurement current i (k) exceeds a predetermined value, the setting unit 132 determines whether a predetermined step has elapsed after the change amount of the measurement current i (k) exceeds the predetermined value. The setting unit 132 turns off the Kalman gain G (k) of the Kalman filter, that is, G (k) = 0, when a predetermined step has not elapsed since the change amount of the measured current i (k) exceeds the predetermined value. To do. For example, the predetermined step may be 4 to 9 steps. For example, if one step is set to 1 second, the setting unit 132 turns off the Kalman gain G (k) for 4 to 9 seconds. The predetermined value of the change amount of the measurement current i (k) can be set to, for example, 5 to 10 amperes.

  When a predetermined step has elapsed since the amount of change in the measured current i (k) exceeds a predetermined value, the setting unit 132 determines whether the next step after the predetermined step has elapsed, that is, the predetermined step + 1 step. judge. The setting unit 132 generates switching information that does not switch the prediction noise Σv, assuming that the prediction noise Σv is the same as the previous step, when the predetermined step is not the first step, that is, after the second step and the second step.

The setting unit 132 determines whether or not the difference y ^* (k) between the measurement terminal voltage and the predicted terminal voltage y (k) is larger than the threshold value b when the predetermined step + 1 step. When the difference y ^* (k) is larger than the threshold value b, the setting unit 132 determines whether or not the previous prediction noise Σv has been switched to a value Σv ₂ that makes the SoC correction width narrow. Setting unit 132, when the previous prediction noise [sigma] v has not been switched to a value [sigma] v _2, such as compensation range SoC is narrowed refers to the condition storage unit 121, switches the prediction noise [sigma] v to the prediction noise [sigma] v ₂ Generate switching information.

When the previous predicted noise Σv has been switched to a value Σv ₂ that reduces the SoC correction range, the setting unit 132 determines whether or not the measurement current i (k) has changed in a decreasing direction. To do. The setting unit 132 generates switching information for switching the prediction noise Σv to the prediction noise Σv ₁ that is the initial value when the measurement current i (k) changes in a decreasing direction. The setting unit 132 generates switching information that does not switch the predicted noise Σv when the measurement current i (k) does not change in the decreasing direction. When the difference y ^* (k) is equal to or less than the threshold value b, the setting unit 132 generates switching information for switching the prediction noise Σv to the prediction noise Σv ₁ that is an initial value. The setting unit 132 outputs the generated switching information to the calculation unit 131.

Here, an example of switching the predicted noise Σv will be described with reference to FIG. FIG. 5 is an explanatory diagram illustrating an example of switching the predicted noise Σv. As shown in the example of FIG. 5, in the state where the measured current i (k) is indicated by the current value 51, the value of the initial value Σv ₁ is used for the predicted noise Σv of the Kalman filter, assuming that the equivalent electric circuit model is reliable. However, in the Kalman filter, when the measured current i (k) increases in the same direction, for example, from discharge to discharge or from charge to charge, an unexpected voltage change is likely to occur. The error is greatly estimated, and the SoC value is deteriorated.

In the SoC estimation using the Kalman filter, for example, when the measured current i (k) changes from the current value 51 to the current value 52, an unexpected voltage change is likely to occur. Therefore, in the state of the current value 52, the SoC error is large. Become. For this reason, when the change occurs, until the next change of the measured current i (k), for example, when the current value 52 changes to the current value 51 again, among the matrix of the predicted noise Σv of the Kalman filter, The component corresponding to SoC is set to a value that narrows the SoC correction range. A value that narrows the SoC correction range can be, for example, 1/10 of the initial value. That is, the setting unit 132 switches the prediction noise Σv (covariance matrix Σv) to the prediction noise Σv ₂ according to the reliability of the equivalent electric circuit model. As a result, the control unit 130 can narrow the SoC error range, that is, the blur width, and thus can prevent the estimated SoC value from deteriorating.

  Next, the operation of the estimation apparatus 100 according to the embodiment will be described.

  FIG. 6 is a flowchart illustrating an example of processing performed by the estimation apparatus according to the embodiment. The control unit 130 of the estimation apparatus 100 reads various parameters from the condition storage unit 121 and sets them as initial values in the calculation unit 131 and the setting unit 132. For example, the control unit 130 refers to the condition storage unit 121, sets the predetermined value of the change amount of the measured current i (k), the threshold value a, and the threshold value b in the setting unit 132, and calculates the predicted noise Σv. It is set to 131 (step S101).

  The estimation apparatus 100 starts estimating the SoC of the connected lithium ion secondary battery. The measurement unit 101 outputs the measured current and terminal voltage to the control unit 130 as a measurement current and a measurement terminal voltage, respectively. In addition, the measurement unit 101 acquires the operation mode of the lithium ion secondary battery and outputs it to the control unit 130.

When the measurement current and the measurement terminal voltage are input from the measurement unit 101, the calculation unit 131 calculates the SoC (charge rate) and the predicted terminal voltage using a Kalman filter. In addition, the calculation unit 131 calculates the difference y ^* (k) between the measurement terminal voltage and the predicted terminal voltage as an error. Further, the calculation unit 131 calculates the Kalman filter covariance matrix P (k) for each step and outputs the calculation result to the setting unit 132 (step S102).

  When the covariance matrix P (k) is input from the calculation unit 131, the setting unit 132 determines whether the component corresponding to the SoC of the covariance matrix P (k) is larger than the threshold value a (step S103). . When the component corresponding to the SoC of the covariance matrix P (k) is larger than the threshold value a (Yes at Step S103), the setting unit 132 sets an initial value to the covariance matrix P (k) and resets it (Step S103: Yes). Step S104). The setting unit 132 does not reset the covariance matrix P (k) when the component corresponding to the SoC of the covariance matrix P (k) is equal to or less than the threshold value a (No at Step S103). When the determination of the covariance matrix P (k) is completed, the setting unit 132 outputs the initialized covariance matrix P (k) or the input covariance matrix P (k) to the calculation unit 131.

  The setting unit 132 performs switching processing of the predicted noise Σv of the Kalman filter (step S105). FIG. 7 is a flowchart illustrating an example of the Kalman filter Σv switching process. When measurement current i (k) is input from measurement unit 101, setting unit 132 determines whether or not the amount of change in measurement current i (k) exceeds a predetermined value (step S151). When the amount of change in the measured current i (k) is equal to or smaller than the predetermined value (step S151: No), the setting unit 132 generates switching information that does not switch the predicted noise Σv.

  When the change amount of the measurement current i (k) exceeds a predetermined value (step S151: Yes), the setting unit 132 determines whether a predetermined step has elapsed after the change amount of the measurement current i (k) exceeds the predetermined value. It is determined whether or not (step S152). The setting unit 132 turns off the Kalman gain G (k) of the Kalman filter, that is, if the predetermined amount has not elapsed since the change amount of the measured current i (k) exceeds the predetermined value (No at Step S152), that is, G (K) = 0 is set (step S153).

  When a predetermined step has elapsed since the amount of change in the measured current i (k) exceeds a predetermined value (step S152: affirmative), the setting unit 132 sets the next step after the predetermined step has elapsed, that is, the predetermined step + 1 step. It is determined whether or not (step S154). If the setting unit 132 is not the predetermined step + 1 step, that is, if it is the predetermined step + the second step or later (No in step S154), the setting information 132 is the same as the previous step, and the switching information that does not switch the prediction noise Σv. Is generated (step S155).

When it is the predetermined step + 1 step (step S154: affirmative), the setting unit 132 determines whether or not the difference y ^* (k) between the measurement terminal voltage and the predicted terminal voltage y (k) is larger than the threshold value b. Determination is made (step S156). Setting unit 132, when the difference ^y * (k) is greater than the threshold b (step S156: Yes), whether the previous prediction noise [sigma] v is already switched to the value [sigma] v _2, such as compensation range SoC narrows It is determined whether or not (step S157). When the previous predicted noise Σv has not been switched to a value Σv ₂ that makes the SoC correction width narrow (No at Step S157), the setting unit 132 refers to the condition storage unit 121 to determine the predicted noise Σv. Switching information for switching to the predicted noise Σv ₂ is generated (step S158).

When the previous predicted noise Σv has been switched to the predicted noise Σv ₂ (step S157: affirmative), the setting unit 132 determines whether or not the measured current i (k) has changed in a decreasing direction (step S157: Yes). Step S159). Setting unit 132, if the measured current i (k) is changed in the direction to become smaller (step S159: Yes), generates the switching information for switching the predicted noise [sigma] v to the prediction noise [sigma] v ₁ is the initial value (step S160 ). The setting unit 132 generates switching information that does not switch the predicted noise Σv when the measured current i (k) does not change in the decreasing direction (No at Step S159).

Setting unit 132, when the difference ^y * (k) is the threshold value b or less (Step S156: No), it generates the switching information for switching the predicted noise [sigma] v to the prediction noise [sigma] v ₁ is the initial value (step S160). The setting unit 132 outputs the generated switching information to the calculation unit 131.

  Returning to the description of FIG. 6, when the setting process of the covariance matrix P (k) and the switching process of the prediction noise Σv are completed, the calculating unit 131 calculates the SoC state estimated value based on the Kalman filter ( Step S106). Note that the calculation unit 131 may estimate the SoC using only the coulomb counter without using the Kalman filter calculation result according to the operation mode of the lithium ion secondary battery.

  The calculation unit 131 calculates the SoC from the calculated state estimated value according to the equation (19) (step S107). The calculation unit 131 outputs the calculated SoC to the output unit 102 (step S108). The calculation unit 131 increments k in order to advance the Kalman filter to the next step (step S109), and returns to the process of step S102. The estimating apparatus 100 repeats steps S102 to S109. This prevents degradation in estimation accuracy due to accumulation of SoC errors over time, and prevents degradation in estimation accuracy based on an increase in SoC errors due to an increase in current. In addition, when the current is constant for a predetermined time or longer, the prediction noise is automatically switched according to the reliability of the Kalman filter, so the estimation accuracy can be obtained without preparing a value corresponding to the reliability of the current change pattern in advance. Can be prevented.

  As described above, the estimation apparatus 100 calculates the charging rate and the predicted terminal voltage of the battery using the Kalman filter based on the measured current and the measured terminal voltage of the battery, and calculates the difference between the measured terminal voltage and the predicted terminal voltage. . Further, the estimation device 100 turns off the Kalman gain for a predetermined step when the amount of change in the measured current exceeds a predetermined value. When the difference exceeds the threshold value after a predetermined step has elapsed, the estimation device 100 sets the component corresponding to the charging rate in the matrix representing the prediction noise of the Kalman filter to a different value. As a result, it is possible to prevent deterioration in the estimation accuracy of the charging rate. That is, the estimation apparatus 100 prevents deterioration in estimation accuracy based on an increase in SoC error due to an increase in current.

  Further, when the measurement current increases in the same direction and exceeds a predetermined value as the amount of change in the measurement current, the estimation apparatus 100 turns off the Kalman gain for a predetermined step. When the difference exceeds the threshold value after a predetermined step has elapsed, the estimation device 100 sets the component corresponding to the charging rate in the matrix representing the prediction noise of the Kalman filter to a different value. As a result, it is possible to prevent deterioration in estimation accuracy based on an increase in SoC error due to an increase in current.

  Moreover, the estimation apparatus 100 sets the charging rate correction width to a value that is narrowed as a different value. As a result, the fluctuation range of the SoC correction of the Kalman filter can be narrowed, and the SoC correction according to the reliability of the equivalent electric circuit model can be performed.

  In the above embodiment, the covariance matrix P (k) is reset to the initial value when the component corresponding to the SoC of the P matrix (covariance matrix P (k)) of the Kalman filter exceeds the threshold value a. It is not limited to this. The estimation apparatus 100 may reset the covariance matrix P (k) to an initial value after elapse of a predetermined time at a predetermined interval, that is, a constant interval, for example. The predetermined interval can be set to 500 seconds, for example. As a result, it is possible to prevent more errors from being accumulated than the predetermined time interval.

  Further, in the above embodiment, the processing when the component corresponding to the charging rate of the P matrix of the Kalman filter exceeds a predetermined value and the processing when the change amount of the measured current exceeds the predetermined value are performed together. It is not limited to this. The estimation apparatus 100 may perform only processing when the component corresponding to the charging rate of the P matrix of the Kalman filter exceeds a predetermined value, or only processing when the amount of change in the measured current exceeds a predetermined value. . Thereby, the SoC correction process according to the cost can be performed.

  In the above embodiment, a lithium ion secondary battery is used as an example of the secondary battery, but the present invention is not limited to this. The estimation device 100 can estimate the SoC for other types of batteries as long as the OCV characteristic curve of the target battery and the equivalent electric circuit model are obtained. Other batteries can be applied to, for example, lithium ion polymer secondary batteries, calcium ion secondary batteries, nanowire batteries, nickel metal hydride secondary batteries, lead storage batteries, and the like. Moreover, as a lithium ion secondary battery, it can apply to a lithium cobalt oxide battery, a lithium iron phosphate battery, etc. Thereby, SoC of various batteries can be estimated.

In the above embodiment, _two RC circuits of R ₁ C ₁ and R ₂ C ₂ are used as an RC circuit representing a transient voltage change with respect to a current change as an equivalent electric circuit model. However, it is not limited to this. The estimation apparatus 100 may use, for example, three or more equivalent electric circuit models as the equivalent electric circuit model. As a result, the accuracy of the equivalent electric circuit model is improved, so that the SoC estimation accuracy by the Kalman filter is improved.

  In addition, each component of each part illustrated does not necessarily need to be physically configured as illustrated. In other words, the specific form of distribution / integration of each part is not limited to the one shown in the figure, and all or a part thereof may be functionally or physically distributed / integrated in arbitrary units according to various loads and usage conditions. Can be configured. For example, the calculation unit 131 and the setting unit 132 may be integrated, and the calculation unit 131 may reset the covariance matrix P (k) and switch the prediction noise Σv.

  Furthermore, various processing functions performed in each unit may be executed entirely or arbitrarily on a CPU (or a microcomputer such as an MPU or MCU (Micro Controller Unit)). In addition, various processing functions may be executed in whole or in any part on a program that is analyzed and executed by a CPU (or a microcomputer such as an MPU or MCU) or hardware based on wired logic. It goes without saying that it is good.

  By the way, the various processes described in the above embodiments can be realized by executing a program prepared in advance by a computer. Therefore, in the following, an example of a computer that executes a program having the same function as in the above embodiment will be described. FIG. 8 is an explanatory diagram illustrating an example of a computer that executes an estimation program.

  As shown in FIG. 8, the computer 200 includes a CPU 201 that executes various arithmetic processes, a battery information interface 202 that receives information from the battery, and a current measurement interface 203 that receives information on measured current and voltage. Further, the computer 200 has an external interface 204 used for exchange of various information connected to other devices, operation of the computer 200 performed by providing a UI (User Interface) to other devices, and the like. The external interface 204 is a general-purpose interface that can be used for writing a program or the like from another device to a flash memory 206 described later. The computer 200 also includes a RAM 205 that temporarily stores various types of information and a flash memory 206. Each component 201 to 206 is connected to the bus 207. The computer 200 is, for example, a control device for an electric vehicle or an electric motorcycle, a smartphone, a notebook personal computer, or the like.

  The flash memory 206 stores an estimation program having the same function as each processing unit of the calculation unit 131 and the setting unit 132 illustrated in FIG. The flash memory 206 stores a condition storage unit 121. The flash memory 206 stores various data for realizing the estimation program. The battery information interface 202 has a function similar to the function of acquiring the operation mode of the lithium ion secondary battery of the measurement unit 101 shown in FIG. The current measurement interface 203 has the same function as the ammeter and voltmeter of the measurement unit 101 shown in FIG. The external interface 204 has the same function as that of the output unit 102 shown in FIG. Further, the external interface 204 may be connected to, for example, a data logger so that the charge / discharge state of the lithium ion secondary battery can be recorded.

  The CPU 201 reads out each program stored in the flash memory 206, develops it in the RAM 205, and executes it to perform various processes. In addition, these programs can cause the computer 200 to function as the calculation unit 131 and the setting unit 132 illustrated in FIG.

  Note that the above estimation program is not necessarily stored in the flash memory 206. For example, the computer 200 may read and execute a program stored in a storage medium readable by the computer 200. The storage medium readable by the computer 200 corresponds to, for example, a portable recording medium such as a CD-ROM, a DVD disk, a USB (Universal Serial Bus) memory, a semiconductor memory such as a flash memory, a hard disk drive, and the like. Alternatively, the estimation program may be stored in a device connected to a public line, the Internet, a LAN (Local Area Network), etc., and the computer 200 may read and execute the estimation program therefrom. The computer 200 may be, for example, a module incorporated in a lithium ion secondary battery, a so-called embedded microcomputer using an MPU, or the like.

DESCRIPTION OF SYMBOLS 100 Estimation apparatus 101 Measuring part 102 Output part 120 Storage part 121 Condition storage part 130 Control part 131 Calculation part 132 Setting part

Claims (5)

On the computer,
Based on the measured current and measured terminal voltage of the battery, the charging rate and predicted terminal voltage of the battery are calculated using a Kalman filter, and the difference between the measured terminal voltage and the predicted terminal voltage is calculated,
When the amount of change in the measured current exceeds a predetermined value, the Kalman gain is turned off for a predetermined step, and when the difference exceeds a threshold value after the predetermined step has elapsed, the matrix representing the predicted noise of the Kalman filter, An estimation program that executes processing for setting a component corresponding to a charging rate to a different value.   In the setting process, when the measurement current increases in the same direction and exceeds the predetermined value as a change amount of the measurement current, the Kalman gain is turned off during a predetermined step, and the difference is calculated after the predetermined step has elapsed. 2. The estimation program according to claim 1, wherein when the value exceeds a threshold, a component corresponding to the charging rate is set to a different value in a matrix representing the prediction noise of the Kalman filter.   The estimation program according to claim 1 or 2, wherein the setting process sets the different value as a value that narrows the correction range of the charging rate. Computer
Based on the measured current and measured terminal voltage of the battery, the charging rate and predicted terminal voltage of the battery are calculated using a Kalman filter, and the difference between the measured terminal voltage and the predicted terminal voltage is calculated,
When the amount of change in the measured current exceeds a predetermined value, the Kalman gain is turned off for a predetermined step, and when the difference exceeds a threshold value after the predetermined step has elapsed, the matrix representing the predicted noise of the Kalman filter, An estimation method characterized by executing processing for setting a component corresponding to a charging rate to a different value. Based on the measured current of the battery and the measured terminal voltage, the Kalman filter is used to calculate the charging rate and the predicted terminal voltage of the battery, and a calculation unit that calculates the difference between the measured terminal voltage and the predicted terminal voltage;
When the amount of change in the measured current exceeds a predetermined value, the Kalman gain is turned off for a predetermined step, and when the difference exceeds a threshold value after the predetermined step has elapsed, the matrix representing the predicted noise of the Kalman filter, An estimation apparatus comprising: a setting unit that sets a component corresponding to a charging rate to a different value.

Images