JP6376091B2 - Battery power prediction device - Google Patents

Description

  The present invention uses at least one of the output power and input power of the secondary battery as a prediction target value, and predicts the prediction target value after the lapse of a specified time from the current time based on the battery model of the secondary battery. The present invention relates to a battery power prediction apparatus.

  As this type of device, for example, as can be seen in Patent Document 1 below, a predetermined model from the present is based on a battery model (hereinafter referred to as a 1RC equivalent circuit model) composed of one DC resistor and an RC equivalent circuit. What predicts the maximum input / output power that can be sustained for a long time is known. This device will be described by taking the case of discharging a secondary battery as an example. First, each parameter of the 1RC equivalent circuit model is collectively estimated using an applicable digital filter. Then, based on the estimated parameters, the inter-terminal voltage of the secondary battery after discharging for a predetermined time is estimated, and the maximum output current when the estimated inter-terminal voltage becomes the minimum possible voltage is calculated. Then, the maximum output possible power is predicted as a multiplication value of the maximum output current and the minimum possible voltage.

Japanese Patent No. 4788307

  In the 1RC equivalent circuit model, since the time constant is small, the electrolyte resistance, the charge transfer resistance, and the electric double capacitance are simulated by one DC resistance, and the diffusion resistance is expressed by a parallel circuit of the resistor R and the capacitor C. It is said. Here, this battery model is not configured to express the current-voltage non-linear characteristics of the secondary battery. This is because the non-linear characteristic of current-voltage of the secondary battery becomes more dominant as the secondary battery becomes colder, and the non-linear characteristic cannot be ignored particularly in the region of 0 ° C. or lower. Therefore, in the apparatus described in Patent Document 1 using the 1RC equivalent circuit model, there is a risk that the prediction accuracy of the maximum input / output power can be reduced when the secondary battery is at a low temperature.

  The main object of the present invention is to provide a battery power prediction device capable of predicting at least one of the output power and the input power of the secondary battery with high accuracy even at a low temperature of the secondary battery.

  Hereinafter, means for solving the above-described problems and the operation and effects thereof will be described.

  In the present invention, at least one of the output power and the input power of the secondary battery (20a) is set as a prediction target value, and the prediction target value after elapse of a specified time from the current time based on the battery model of the secondary battery. The battery model includes a DC resistance model representing the DC resistance (Rs) of the secondary battery and a model representing the reaction resistance of the secondary battery, wherein the Butler-Volmer equation is used. A reaction resistance model including a reaction resistance parameter (β) that is derived and correlated with an exchange current density, and an RC equivalent circuit model including a parallel connection body of a resistor and a capacitor, and represents a diffusion resistance of the secondary battery. An equation including a circuit constant of the RC equivalent circuit model, and based on a discrete equation for calculating a current polarization voltage of the diffusion resistor model, A first polarization voltage calculating means for calculating a polarization voltage, and an equation including the circuit constant, wherein the current polarization voltage decreases with time and the remaining voltage is the polarization voltage remaining after the specified time has elapsed. A residual voltage predicting means for predicting the residual voltage by inputting the current polarization voltage in a continuous time formula for predicting the future polarization voltage of the diffusion resistance model newly generated after the lapse of the specified time from the present time. A second polarization voltage calculating means for predicting the future polarization voltage based on a predicted continuous time formula; a total value of the residual voltage and the future polarization voltage; a potential difference in the DC resistance model; and the reaction resistance And a power prediction unit that predicts the prediction target value after the lapse of the specified time from the current time based on a potential difference in the model.

  The internal resistance of the secondary battery is roughly divided into a direct current resistance, a reaction resistance, and a diffusion resistance. For this reason, in the said invention, the battery model is comprised from the direct current | flow resistance model, the reaction resistance model, and the diffusion resistance model.

  Here, when the secondary battery is at a low temperature, the non-linear characteristic of current-voltage due to the reaction resistance becomes dominant. Therefore, in the above invention, the reaction resistance model is a model derived from the Butler-Volmer equation in electrochemistry and expressing the non-linear characteristics of the secondary battery. Specifically, this model includes a parameter corresponding to a Butler-Volmer exchange current density and a reaction resistance parameter correlated with the temperature of the secondary battery. Since the reaction resistance parameter depends on the temperature of the secondary battery, by including the reaction resistance model including the reaction resistance parameter in the battery model, for example, the current at a low temperature that cannot be expressed by the technique described in Patent Document 1 above. -The non-linear characteristic of voltage can be expressed accurately. In the above invention, the potential difference in the reaction resistance model is used for prediction of the prediction target value after the lapse of a specified time from the present time, so that the prediction target value can be predicted with high accuracy even at low temperatures.

  Furthermore, in the above invention, the remaining voltage is calculated based on the continuous time formula using the current polarization voltage as an input. Further, the future polarization voltage is calculated based on the continuous time formula. According to the calculation method based on the continuous time formula, it is possible to greatly reduce the calculation load of each of the remaining voltage and the future polarization voltage used for prediction of the prediction target value, and consequently to reduce the calculation load when predicting the prediction target value. it can.

The lineblock diagram of the battery pack concerning one embodiment. The block diagram which shows the process of battery ECU. The figure which shows a battery model. The block diagram which shows the process of a direct-current resistance potential difference calculation part. The block diagram which shows the process of the reaction resistance potential difference calculation part. The figure which shows the relationship between a reaction resistance parameter and battery temperature. The figure which shows the temperature dependence of the electric current-voltage characteristic in reaction resistance. The block diagram which shows the process of a polarization voltage calculation part. The time chart which shows the voltage estimation result concerning related technology. The time chart which shows the voltage estimation result concerning one Embodiment. The time chart which shows the outline | summary of an electric power prediction method. The figure which shows the outline | summary of the electric power prediction method. The flowchart which shows the procedure of an allowable output power prediction process. The figure which shows an example of an electric current estimation aspect. The figure which shows an example of an electric current estimation aspect. The flowchart which shows the procedure of an allowable output power prediction process.

  DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, an embodiment in which a battery power prediction apparatus according to the invention is embodied will be described with reference to the drawings. In the present embodiment, the above-described device is applied to, for example, a vehicle including a rotating electrical machine (motor generator) as an in-vehicle main machine, or a vehicle that uses an in-vehicle auxiliary battery such as an idling stop system.

  As shown in FIG. 1, the battery pack 10 includes an assembled battery 20 and a battery ECU 30. The assembled battery 20 is composed of a series connection body of a plurality of battery cells 20a, and exchanges power with a motor generator or the like (not shown). The battery cell 20a is a secondary battery, and a lithium ion secondary battery is used in this embodiment.

  The battery pack 10 includes a voltage sensor 21, a temperature sensor 22, and a current sensor 23. The voltage sensor 21 is voltage detection means for detecting the voltage between the terminals of each battery cell 20a. The temperature sensor 22 is temperature detection means for detecting the temperature of the assembled battery 20 (each battery cell 20a). The current sensor 23 is current detection means for detecting a charge / discharge current flowing through the assembled battery 20 (each battery cell 20a).

  The battery ECU 30 is configured as a computer including a CPU, a memory 31 (storage device), an I / O (not shown), and the like. The CPU includes a single cell calculation unit 32 and a power prediction unit 33 corresponding to each of the plurality of battery cells 20a. In the present embodiment, one processing cycle of the CPU (single cell calculation unit 32, power prediction unit 33) is set to 0.1 second. Detection values of the voltage sensor 21, the temperature sensor 22, and the current sensor 23 are input to the battery ECU 30.

  The battery ECU 30 performs a process of predicting each of the allowable output power and the allowable input power of the assembled battery 20. Hereinafter, this prediction process will be described in the order of the single cell calculation unit 32 and the power prediction unit 33.

<1. Processing of Single Cell Operation Unit 32>
With reference to FIG. 2, an outline of processing performed by each single cell arithmetic unit 32 (CPU) will be described. The single cell calculation unit 32 includes a DC resistance potential difference calculation unit 34, a reaction resistance potential difference calculation unit 35, a polarization voltage calculation unit 36, an SOC conversion unit 37, and the like.

  Each calculation part 34-36 estimates each parameter of the battery cell 20a shown in FIG. Here, FIG. 3 shows a battery model expressing internal impedance and the like. In the present embodiment, the battery model is represented as a series connection body of a DC resistance model, a reaction resistance model, and a diffusion resistance model. In FIG. 3, “Rs” represents a direct current resistance representing the current resistance of the solution or the electrode, and “Vs” represents a potential difference in the direct current resistance Rs (hereinafter referred to as a direct current resistance potential difference). “ΔV” indicates a potential difference in reaction resistance (hereinafter, reaction resistance potential difference) representing an electrode interface reaction between the positive electrode and the negative electrode. “Rw1, Rw2, Rw4, Rw4” indicates a resistance component term in diffusion resistance representing ion diffusion in the active material or solution, and “Cw1, Cw2, Cw3, Cw4” indicates that the resistance changes with time. , “Vw1, Vw2, Vw3, Vw4” indicates the polarization voltage at each diffused resistor. In the present embodiment, the diffusion resistor has a configuration in which a plurality of parallel connection bodies of resistance components and capacitance components (four in this embodiment) are connected in series. An equivalent circuit in which a resistance component and a capacitance component are connected in parallel is called a Foster-type equivalent circuit.

  In the present embodiment, the reaction resistance model shown in FIG. 3 is represented only by DC resistance for convenience, and the time constant in the model is ignored. This is because in this embodiment, one processing cycle of the single cell calculation unit 32 (CPU) is set sufficiently longer than the time constant in the reaction resistance.

  First, the DC resistance potential difference calculation unit 34 will be described with reference to FIG.

  In the DC resistance potential difference calculation unit 34, the Rs calculation unit 34 a calculates the DC resistance Rs based on the battery temperature Ts detected by the temperature sensor 22. The reason why the battery temperature Ts is used for calculating the DC resistance Rs is that the DC resistance Rs depends on the temperature of the battery cell 20a. In the present embodiment, the DC resistance Rs is calculated using an Rs map in which the DC resistance Rs and the battery temperature Ts are related in advance. In the present embodiment, the Rs map is adapted such that the higher the battery temperature Ts, the lower the DC resistance Rs. In the present embodiment, the Rs map is stored in the memory 31.

  The first multiplication unit 34b calculates a DC resistance potential difference Vs as a multiplication value of the DC resistance Rs calculated by the Rs calculation unit 34a and the current detected by the current sensor 23 (hereinafter, detected current Is). Incidentally, in this embodiment, the discharge current of the battery cell 20a is represented by a negative value, and the charging current is represented by a positive value. For this reason, in this embodiment, the DC resistance potential difference Vs becomes a negative value when the battery cell 20a is discharged, and becomes a positive value when the battery cell 20a is charged.

  Next, the reaction resistance potential difference calculation unit 35 will be described with reference to FIG.

  In the reaction resistance potential difference calculation unit 35, the β calculation unit 35a calculates a reaction resistance parameter β, which is a parameter related to the reaction resistance, based on the battery temperature Ts. The battery temperature Ts is used to calculate the reaction resistance parameter β because the reaction resistance parameter β depends on the temperature of the battery cell 20a. In the present embodiment, the reaction resistance parameter β is calculated using a β map in which the natural logarithm “lnβ” of the reaction resistance parameter β and the battery temperature Ts (absolute temperature) are related in advance. Hereinafter, the reaction resistance parameter β will be described.

  The Butler-Volmer equation in electrochemistry is represented by the following equation (eq1).

In the above equation (eq1), “i” represents the current density, “io” represents the exchange current density, “αs” represents the transfer coefficient (oxidation reaction) of the electrode reaction, and “n” represents the number of charges. , “F” indicates a Faraday constant, “η” indicates an overvoltage, “R” indicates a gas constant, and “T” indicates an absolute temperature.

  In the above equation (eq1), for the sake of simplicity, assuming that the positive and negative electrodes are equivalent (that is, the charge / discharge efficiency is the same) and “a = αs = 1−αs”, the above equation (eq1) is the following equation (eq2): Become.

Using the relationship between the hyperbolic sine function and the exponential function, the above equation (eq2) is transformed into the following equation (eq3).

When the above equation (eq3) is solved for the overvoltage η, the following equation (eq4) is obtained.

On the other hand, the relationship between the overvoltage η and the potential difference ΔV in the reaction resistance is expressed by the following equation (eq5) using the proportionality constant γ. Further, the relationship between the current density i and the current I flowing through the battery cell is expressed by the following equation (eq6) using the proportionality constant γ.

Substituting the above equations (eq5) and (eq6) into the above equation (eq4) leads to the following equation (eq7).

Here, the above equation (eq7) is arranged as the following equation (eq8).

In the above equation (eq8), “β” represents the reaction resistance parameter, “α” represents a constant, “γ” represents a compatible constant, and the charge / discharge current I flowing through the battery cell and the reaction resistance potential difference ΔV are expressed as follows. It shows that it can be related by the reaction resistance parameter β. As can be seen from the above equation (eq8), the reaction resistance parameter β derived from the Butler-Volmer equation is an inverse hyperbola in an inverse hyperbolic sine function with the current flowing through the battery cell as an independent variable and the reaction resistance potential difference ΔV as a dependent variable. This is a coefficient that determines the relationship between the sine function and the reaction resistance potential difference ΔV.

  Here, since the exchange current density io depends on the battery temperature T, the reaction resistance parameter β also depends on the battery temperature T. For this reason, in the present embodiment, according to the Arrhenius plot, a β map in which the natural logarithm of the reaction resistance parameter β is adapted to be the inverse of the battery temperature Ts or a linear expression with respect to the battery temperature Ts is stored in the memory 31. Yes. Here, FIG. 6 illustrates a β map in which the natural logarithm of the reaction resistance parameter β is adapted in the form of a linear expression with respect to the reciprocal of the battery temperature Ts.

  The reaction resistance parameter β calculated by the β calculator 35a is input to the potential difference calculator 35b. The potential difference calculator 35b receives the reaction resistance parameter β and the detected current Is as input, and calculates the reaction resistance potential difference ΔV based on the above equation (eq8). As shown in FIG. 7, the above equation (eq8) is an equation in which the reaction resistance potential difference ΔV becomes nonlinear with respect to the current when the temperature becomes low. When the temperature of the battery cell is low, the reaction resistance potential difference ΔV that accurately represents the current-voltage nonlinear characteristic can be calculated by using the reaction resistance parameter β.

  Next, the polarization voltage calculation unit 36 will be described with reference to FIG. The polarization voltage calculator 36 calculates a parameter related to the diffusion resistance. Hereinafter, after describing parameters related to the diffusion resistance, a method for calculating the polarization voltage will be described.

  Based on the diffusion equation in electrochemistry, the Warburg impedance Z relating to the diffusion resistance is derived. Here, the impedance Z is expressed by the following equation (eq9).

In the above equation (eq9), the numerator is based on the Nernst equation indicating that the electromotive force is proportional to the natural logarithm of the surface concentration of the substance. In the molecule, “Co” indicates an average concentration, “ΔC” indicates a concentration change with respect to the average concentration Co, and “x” indicates a position from the electrode. The denominator is based on Fick's first law indicating that the amount of a substance passing through a unit area per unit time is proportional to the concentration gradient. In the denominator, “D” indicates a diffusion coefficient.

  By arranging the above equation (eq9), the following equation (eq10) is derived.

Here, it is assumed that when an AC voltage is applied to the electrodes, if the voltage changes with a sine wave, the concentration also changes with a sine wave. At this time, the density change ΔC is expressed by the following equation (eq11) using the imaginary number j and the angular velocity ω.

In the above equation (eq11), | Δv | and | ΔC | indicate complex amplitudes. Here, Fick's second law is expressed by the following equation (eq12).

The left side of the above equation (eq12) is “jωΔC” because it is the time differential value of the above equation (eq11). For this reason, the general solution of the above equation (eq12) is expressed by the following equation (eq13) using the constants k1 and k2.

Here, “L” is defined as the diffusion length. When the condition that the density change ΔC is 0 when “x = L” is imposed, the above equation (eq13) becomes the following equation (eq14).

Therefore, the concentration change ΔC is expressed by the following equation (eq15).

When the above equation (eq15) is partially differentiated by x, the following equation (eq16) is obtained.

When the above equations (eq15) and (eq16) are substituted into the above equation (eq10), a part of the above equation (eq10) is expressed as the following equation (eq17).

Therefore, the Warburg impedance Z of the above equation (eq10) can be expressed by the following equation (eq18) using the Laplace operator s (= j × ω).

In the present embodiment, “Rd” is referred to as a first parameter, and “τd” is referred to as a second parameter. Here, the diffusion coefficient D is expressed by the following equation (eq19).

In the above equation (eq19), “Do” represents a temperature independent constant, and “E” represents activation energy. When the above equation (eq19) is used, the first and second parameters Rd, τd are represented by the following equation (eq20).

The above equation (eq20) indicates that the first and second parameters Rd, τd depend on the battery temperature T. Taking the natural logarithm of the above equation (eq20), the following equation (eq21) is derived.

In this embodiment, according to the Arrhenius plot, an Rd map and τd map in which the natural logarithm of the first and second parameters Rd and τd is adapted to be a reciprocal of the battery temperature T or a linear expression with respect to the battery temperature Ts are stored in the memory. 31 is stored.

  In the polarization voltage calculation unit 36, the Rd calculation unit 36a calculates the first parameter Rd using the Rd map and the battery temperature Ts described above. The τd calculation unit 36b calculates the second parameter τd using the τd map and the battery temperature Ts described above.

  Based on the first parameter Rd, the resistance component calculation unit 36c calculates the resistance values Rw1 to Rw4 of the first to fourth resistances constituting the diffusion resistance by the following equation (eq22).

In the above formula (eq22), “m” represents a positive integer (in this embodiment, m = 1, 2, 3, 4). The capacitance component calculation unit 36d calculates the capacitances Cw1 to Cw4 of the first to fourth capacitors constituting the diffusion resistance based on the first and second parameters Rd and τd by the following equation (eq23).

The resistance value Rwm of the m-th resistor and the capacitance Cwm of the m-th capacitor can be calculated by the above equations (eq22) and (eq23) because an equivalent circuit that matches the Warburg impedance represented by the above equation (eq18) and the Warburg impedance This is based on the result of investigation by the inventors of the literature on the constant law of the equivalent circuit in the series equivalent to the above. In addition, as said literature, there exists "Modeling Ni-mH battery using Cauer and Foster structures. E. Kuhn et al. JOUNAL of Power Sources 158 (2006)", for example.

  The first voltage calculation unit 36e is based on the detection current Is, the first resistance value Rw1 calculated by the resistance component calculation unit 36c, and the first capacitance Cw1 calculated by the capacitance component calculation unit 36d. A first polarization voltage Vw1 is calculated. In the present embodiment, the first polarization voltage Vw1 is sequentially calculated based on an expression obtained by discretizing a transfer function representing a parallel circuit of a resistor and a capacitor constituting the diffusion resistance model. Specifically, in the present embodiment, the first polarization voltage Vw1, the detection current Is calculated in the previous processing cycle, and the current (current processing cycle) detection current Is are input, and the following equation (eq24) Is used to calculate the first polarization voltage Vw1.

In the above equation (eq24), “ΔT” indicates one processing cycle. Similarly to the first voltage calculation unit 36e, the second, third, and fourth voltage calculation units 36f, 36g, and 36h also have resistance values Rw2, Rw3, and Rw4, capacitances Cw2, Cw3, and Cw4, and a detection current Is. Based on the above, the second, third, and fourth polarization voltages Vw2, Vw3, and Vw4 are calculated.

  Returning to the description of FIG. 2, the first addition unit 38 calculates the total polarization voltage Vw0 as an addition value of the polarization voltages Vw1 to Vw4 calculated by the voltage calculation units 36e to 36h. The second addition unit 39 calculates an addition value of the DC resistance potential difference Vs, the reaction resistance potential difference ΔV, and the total polarization voltage Vw0. The open end voltage calculation unit 40 calculates the open end voltage OCV of the battery cell 20a by subtracting the output value of the second addition unit 39 from the inter-terminal voltage CCV of the battery cell 20a detected by the voltage sensor 21.

  The SOC conversion unit 37 calculates the charging rate (SOC) of the battery cell 20a based on the open end voltage OCV calculated by the second addition unit 40. In the present embodiment, the SOC is calculated using an SOC map in which the open-circuit voltages OCV and SOC are related in advance. In the present embodiment, the SOC map is stored in the memory 31.

  Subsequently, the voltage estimation accuracy of the battery cell 20a based on the above-described battery model will be described in comparison with related technology. The related technology is a battery model including a series connection body of one DC resistance and 1RC equivalent circuit. FIG. 9 shows the estimation result of the inter-terminal voltage of the battery cell according to the related art, and FIG. 10 shows the estimation result of the inter-terminal voltage of the battery cell according to this embodiment. Here, in the figure, the estimated value corresponds to a value obtained by adding the open-ended voltage to the output value of the second addition value 39. This open end voltage is calculated based on the SOC calculated by the SOC conversion unit 37, for example. 9 and 10 show test results at a low temperature (for example, −15 ° C.) at which nonlinearity appears in the current-voltage characteristics.

  As shown in FIG. 9, in the related art, a deviation occurs between the measured voltage value and the estimated voltage value depending on the magnitude of the charge / discharge current of the battery cell 20a. On the other hand, in the present embodiment, as shown in FIG. 10, the deviation between the actually measured voltage value and the estimated voltage value is greatly reduced. This is because the reaction resistance model includes a nonlinear expression with respect to the current flowing through the battery cell, thereby accurately expressing the nonlinear characteristic. Therefore, according to the battery model including the reaction resistance model according to the present embodiment, the prediction accuracy of the allowable output power and the allowable input power can be improved.

<2. Processing of Power Prediction Unit 33>
Subsequently, a prediction process of the allowable output power Wout and the allowable input power Win performed by the power prediction unit 33 will be described.

  First, the prediction process of the allowable output power Wout will be described. This process is a process for predicting, as the allowable output power Wout, the maximum power that the assembled battery 20 can continuously discharge over the period from the current time until the discharge side specified time Tdff elapses. In the present embodiment, the discharge side specified time Tdff is set to 10 seconds. In the allowable output power prediction process, for example, the output performance of the assembled battery that supplies power to the motor generator as the in-vehicle main engine at the time of acceleration of the vehicle is sufficient, or power is supplied to the starter for engine restart in the idling stop system. Is used to predict whether the output performance of the auxiliary battery supplying the battery is sufficient.

  The outline of the prediction process will be described with reference to FIGS. 11 and 12. In this process, as shown in FIG. 11, first, it is assumed that a predetermined discharge current Ip is supplied to the battery cell 20a for each battery cell 20a over a period from the current time until the discharge-side specified time Tdff elapses. In this case, the inter-terminal voltage Vff of the battery cell 20a at the timing when the discharge side specified time Tdff elapses from the present is predicted. At this time, if the minimum value of the inter-terminal voltage Vff of each battery cell 20a matches the allowable lower limit voltage VLlimit of the battery cell 20a, the predicted sum of the respective inter-terminal voltages Vff of each battery cell 20a and a predetermined value Can be predicted as the allowable output power Wout. Here, as shown in FIG. 12, the total remaining voltage Vwr is used to predict the inter-terminal voltage Vff. The total residual voltage Vwr is a polarization voltage remaining at a timing when the current polarization voltage in the diffusion resistance decreases with time and the discharge side specified time Tdff elapses from now. That is, when the current charge is stored in the m-th capacitor constituting the diffusion resistance model, the influence of the current polarization voltage on the diffusion resistance remains even after the discharge-side specified time Tdff has elapsed. For this reason, by calculating the total residual voltage Vwr, it is possible to take into account the effect of the amount of charge accumulated in the m-th capacitor, and to improve the prediction accuracy of the allowable output power Wout.

  FIG. 13 shows a procedure for predicting the allowable output power Wout according to the present embodiment. This process is repeatedly executed by the power prediction unit 33 (CPU), for example, at a predetermined processing cycle.

  In this series of processes, first, in step S10, for each battery cell 20a, the total remaining voltage at the timing when the discharge-side specified time Tdff elapses from the current time based on the continuous time expression expressed by the following expression (eq25). Predict Vwr.

In the above equation (eq25), the current first, second, third and fourth polarization voltages Vw1, Vw2, Vw3 and Vw4 are the first, second, third and fourth voltage calculators 36f, 36g and 36h. It is calculated by. In this step, the remaining voltage is predicted for each of the polarization voltages Vw1 to Vw4, and the total remaining voltage Vwr is calculated as the total value of the predicted remaining voltages. The total remaining voltage Vwr decreases with time, as can be seen from the above equation (eq25).

  In subsequent step S11, the virtual operating current Ip of the battery cell 20a is set to the allowable upper limit discharge current Idlimit (<0). In the subsequent step S12, when it is assumed that the virtual operating current Ip (allowable upper limit discharge current Idlimit) flows through the period from the present to the discharge side specified time Tdff, the discharge from the present is performed for each battery cell 20a. The inter-terminal voltage Vff of the battery cell 20a at the timing when the side specified time Tdff elapses is predicted based on the following equation (eq26).

In the above equation (eq26), the DC resistance potential difference Vs is the virtual operating current Ip in place of the detected current Is in the DC resistance potential difference calculation unit 34 of FIG. Use the value calculated in. As the reaction resistance potential difference ΔV, a value calculated by replacing the current input to the potential difference calculation unit 35b with the virtual operation current Ip instead of the detection current Is in the reaction resistance potential difference calculation unit 35 of FIG. The total residual voltage Vwr uses the value calculated in step S10.

  Further, in the above equation (eq26), “Vws” is a future newly generated at the timing when the discharge side specified time Tdff elapses from the present when it is assumed that no charge is accumulated in each capacitor Cwm constituting the diffusion resistance model. The polarization voltage (hereinafter, the future polarization voltage) is shown. The future polarization voltage Vws is predicted based on the continuous time expression expressed by the following expression (eq27).

Diffusion of timing when the total remaining voltage Vwr predicted based on the above equation (eq26) and the future polarization voltage Vws predicted based on the above equation (eq27) elapses from the current discharge side specified time Tdff. Represents the potential difference in resistance.

  Further, in the above equation (eq26), ΔOCV is the amount of change in the open-circuit voltage OCV when it is assumed that the virtual operating current Ip flows through the battery cell 20a over the period from the current time until the discharge side specified time Tdff elapses ( Hereinafter, an open-circuit voltage change amount) is shown. The open circuit voltage change amount ΔOCV may be calculated based on the current open circuit voltage OCV calculated by the open circuit voltage calculator 40, for example. Specifically, for example, when it is assumed that the virtual operating current Ip flows for a period from the current time until the discharge side specified time Tdff elapses, for each battery cell 20a, the discharge side specified time Tdff from the current time. The amount of change in SOC in the period until it elapses is calculated. The change in the SOC is a negative value during discharge. Then, the change in the SOC is added to the current SOC calculated by the SOC conversion unit 37, and the added value is converted into an open-circuit voltage using an SOC map. Then, the open circuit voltage change amount ΔOCV is calculated by subtracting the current open circuit voltage OCV from the converted open circuit voltage.

  Incidentally, the ratio of the open-circuit voltage change amount ΔOCV to the inter-terminal voltage Vff calculated in step S12 is small because the discharge-side specified time Tdff is set to 10 seconds in the present embodiment. For this reason, the term of the open circuit voltage change amount ΔOCV may be removed from the right side of the above equation (eq26).

  In subsequent step S13, the minimum value Vmin of the inter-terminal voltage Vff of each battery cell 20a calculated in step S12 is selected, and whether or not the selected minimum value Vmin is less than the allowable lower limit voltage VLlimit of the battery cell 20a. to decide. Here, FIG. 14 shows an example of a situation in which a negative determination is made in this step. In FIG. 14, for convenience, the number of battery cells is three.

  If a negative determination is made in step S13, it is determined that all the inter-terminal voltages Vff of each battery cell 20a are equal to or lower than the allowable lower limit voltage VLlimit, and the process proceeds to step S14. In step S14, the product of the sum of the inter-terminal voltages Vff predicted in step S12 and the allowable upper limit discharge current Idlimit is predicted as the allowable output power Wout after the discharge side specified time Tdff has elapsed from the present. .

  On the other hand, if an affirmative determination is made in step S13, the process proceeds to step S15, in which the discharge current having the minimum value Vmin of the voltage between the terminals of each battery cell 20a after the lapse of the discharge side specified time Tdff from the present as the allowable lower limit voltage VLlimit Is calculated as the maximum discharge current. The magnitude of the maximum discharge current is a value equal to or less than the magnitude of the allowable upper limit discharge current Idlimit, and a predetermined search method is used to calculate the maximum discharge current. In this embodiment, a bisection method is used as the predetermined search method. The maximum discharge current is the maximum value of the discharge current that satisfies both the current limit range and the voltage limit range. Then, the searched maximum discharge current is set as the virtual operating current Ip. In the present embodiment, the search range by the bisection method is a range from 0 to the allowable upper limit discharge current Idlimit. Here, FIG. 15 shows an example of the virtual operating current Ip set when the battery cell 20a is discharged. In FIG. 15, for convenience, the number of battery cells is two.

  In the subsequent step S16, for each battery cell 20a, it is assumed that the discharge current of the battery cell 20a is the virtual operating current Ip (maximum discharge current) over the period from the current time until the discharge side specified time Tdff elapses. The inter-terminal voltage Vff of the battery cell 20a at the timing when the discharge side specified time Tdff elapses from the present is predicted. In subsequent step S14, the product of the sum of the respective inter-terminal voltages Vff predicted in step S16 and the maximum discharge current searched in step S15 is obtained after the discharge side specified time Tdff has elapsed from the present. Is estimated as the allowable output power Wout.

  Next, a process for predicting the allowable input power Win will be described. This process is a process for predicting the maximum power that can be continuously charged in the assembled battery 20 in the period from the present to the timing when the charging side specified time Tcff elapses. In the present embodiment, the charging side specified time Tcff is basically set to one processing cycle (0.1 second) of the power prediction unit 33 (CPU).

  FIG. 16 shows a procedure for predicting the allowable input power Win. This process is repeatedly executed by the power prediction unit 33 (CPU), for example, at a predetermined processing cycle.

  In this series of processing, first, in step S20, the total remaining voltage Vwr at the timing when the charge side specified time Tcff elapses from the present is predicted. Here, the total remaining voltage Vwr may be calculated by changing the discharge side specified time Tdff in the above equation (eq25) to the charge side specified time Tcff.

  In the following step S21, the virtual operating current Ip of the battery cell 20a is set to the allowable upper limit charging current Ilimit (> 0). In the subsequent step S22, when it is assumed that the virtual operating current Ip (allowable upper limit charging current Ilimit) is passed over the period from the current time until the charging side specified time Tcff elapses, charging is performed for each battery cell 20a from the present time. The inter-terminal voltage Vff of the battery cell 20a at the timing when the side specified time Tcff elapses is predicted. Here, the terminal-side voltage Vff may be predicted by changing the discharge-side specified time Tdff in the equations (eq26) and (eq27) to the charge-side specified time Tcff.

  In subsequent step S23, the maximum value Vmax of the inter-terminal voltage Vff of each battery cell 20a predicted in step S22 is selected, and whether or not the selected maximum value Vmax exceeds the allowable upper limit voltage VHlimit of the battery cell 20a. to decide. Here, FIG. 14 shows an example of a situation in which a negative determination is made in this step.

  If a negative determination is made in step S23, the process proceeds to step S24, and the product of the sum of the inter-terminal voltages Vff of each battery cell 20a predicted in step S22 and the allowable upper limit charging current Ilimit is determined from the current charge side specified time. Predicted as allowable input power Win after Tcff has elapsed.

  On the other hand, when an affirmative determination is made in step S23, the process proceeds to step S25, and first, the charging side specified time Tcff is changed from 0.1 seconds to 10 seconds, which is the same as the discharging side specified time Tdf. Then, the charging current having the maximum value Vmax of the inter-terminal voltage of each battery cell 20a at the timing when the charging side specified time Tcff elapses from the present as the allowable upper limit voltage VHlimit is calculated as the maximum charging current. The magnitude of the maximum charging current is a value less than or equal to the magnitude of the allowable upper limit charging current Ilimit, and a predetermined search method is used to calculate the maximum charging current. In this embodiment, a bisection method is used. The maximum charging current is the maximum value of the charging current that satisfies both the current limiting range and the voltage limiting range. Then, the searched maximum charging current is set as the virtual operating current Ip. In the present embodiment, the search range by the bisection method is a range from 0 to the allowable upper limit charging current Ilimit. Here, FIG. 15 shows an example of the virtual operating current Ip set at the time of charging.

  In subsequent step S26, for each of the battery cells 20a, when it is assumed that the charging current of the battery cell 20a is the virtual operating current Ip (maximum charging current) over the period from the current charging time Tcff has elapsed. The inter-terminal voltage Vff of the battery cell 20a at the timing when the charging side specified time Tcff elapses from the present is predicted. In subsequent step S24, a charging side specified time Tcff elapses from the present value by multiplying the sum of the inter-terminal voltages Vff of each battery cell 20a predicted in step S26 by the maximum charging current searched in step S25. Predicted as allowable input power Win at the timing.

  Here, the reason why the length of the charging-side specified time Tcff is changed according to the determination result in the previous step S23 will be described below in comparison with the prediction of the allowable output power Wout during discharging. First, the discharge time of the assembled battery 20 will be described. Here, the diagram shown in FIG. 14 in which the operating point of the battery cell 20a is determined by the voltage between terminals and the charge / discharge current is referred to as an IV diagram. As shown in FIG. 14, when the discharge current used for predicting the allowable output power Wout is the allowable upper limit discharge current Idlimit, the voltage between the terminals of the battery cell 20a decreases every time and discharges. Electric power is reduced. For this reason, even if the allowable output power Wout is calculated from the operating point after the discharge side specified time Tdf has elapsed, and the discharge side specified time Tdiff continues to be discharged at the allowable output power Wout calculated from the current operating point, the discharge current remains within the allowable upper limit. The discharge current Idlimit is not exceeded.

  Then, the time of charge of the assembled battery 20 is demonstrated. As shown in FIG. 14, when the charging current used for predicting the allowable input power Win is the allowable upper limit charging current Ilimit, when the charging is continued at the allowable upper limit charging current Ilimit, the respective voltages of the battery cells 20a increase. As a result, the charging power value increases. That is, since the charging power after the charging side specified time Tcff is larger than the charging power at the time when the allowable upper limit charging current Ilimit is reached, the operation is allowed when operated with the allowable charging power Win obtained from the operating point after the charging side specified time Tcff has elapsed. After passing the operating point exceeding the upper limit charging current Ilimit, the operating point corresponding to the predicted allowable input power Win can be shifted. This is because at the time of charging, the future polarization voltage Vws increases toward the convergence value “Ip × Rwm” every moment, so that the voltage between the terminals of the battery cell 20a increases every moment. In particular, when the charging side specified time Tcff becomes longer, there is a greater concern that the charging current exceeds the allowable upper limit charging current Ilimit. For this reason, when the charging current used for prediction of the allowable input power Win is the allowable upper limit charging current Ilimit, the charging-side specified time Tcff is set to 0.1 seconds, which is one processing cycle.

  On the other hand, as shown in FIG. 15, when the charging current used for prediction of the allowable input power Win is equal to or lower than the allowable upper limit charging current Ilimit, the charging current used for prediction and the allowable upper limit charging current Ilimit tend to be separated in FIG. It is in. Therefore, in the IV diagram, for each of the battery cells 20a, the charging current is set to the allowable upper limit charging current Ilimit during the transition from the current operating point to the operating point corresponding to the predicted allowable input power Win. There is little concern to exceed. Therefore, when the charging current used for prediction of the allowable input power Win is equal to or less than the allowable upper limit charging current Ilimit, the charging side specified time Tcff should be set short in order to avoid a situation where the charging current exceeds the allowable upper limit charging current Ilimit. I don't need it. Therefore, the charging side specified time Tcff is changed from 0.1 second to 10 seconds.

  As described above, in this embodiment, the battery model is configured by the reaction resistance model using the reaction resistance parameter β in addition to the DC resistance model and the diffusion resistance model. By including the reaction resistance model including the reaction resistance parameter in the battery model, it is possible to accurately represent the current-voltage non-linear characteristics at low temperatures that could not be expressed by the technique described in Patent Document 1. For this reason, according to this embodiment, the allowable input / output powers Win and Wout of the assembled battery 20 can be predicted with high accuracy even at low temperatures.

  Further, in the present embodiment, the total remaining voltage Vwr is calculated based on the continuous time formula expressed by the above formula (eq25), and the future polarization voltage Vws is calculated based on the continuous time formula expressed by the above formula (eq27). Predicted. According to the calculation method based on the continuous time formula, it is possible to calculate each of the total remaining voltage Vwr and the future polarization voltage Vws used for prediction of the allowable input / output powers Win and Wout by one calculation. For this reason, the calculation load in predicting the allowable input / output powers Win and Wout can also be suitably reduced.

(Other embodiments)
The above embodiment may be modified as follows.

  The predetermined search method is not limited to the bisection method, and other search methods such as the golden section method may be used.

  The discharge side specified time Tdff is not limited to 10 seconds, and may be set to several seconds to several tens of seconds, for example, 2 seconds, 5 seconds, 30 seconds, and the like. The same applies to the charge-side specified time Tcff after being changed in step S25 of FIG.

  -The diffusion resistance model is not limited to an RC equivalent circuit model in which four parallel connection bodies of resistors and capacitors are connected in series, but is an RC equivalent circuit model in which a plurality of the parallel connection bodies (other than four) are connected. May be. The diffusion resistance model is not limited to an RC equivalent circuit in which a plurality of parallel connection bodies of resistors and capacitors are connected, and the parallel connection body may be a single RC equivalent circuit for simplification.

  The secondary battery is not limited to a lithium ion secondary battery, but may be another secondary battery such as a nickel metal hydride battery.

  In each of the above embodiments, the battery temperature used for each process is not limited to the detection value of the temperature sensor 22, and may be a battery temperature estimated by some method.

  -The application object of this invention is not restricted to a vehicle.

  20a ... battery cell, 30 ... battery ECU.

Claims (8)

A battery that uses at least one of the outputable power and the inputable power of the secondary battery (20a) as a prediction target value, and predicts the prediction target value after a lapse of a specified time from the current time based on the battery model of the secondary battery. In the power prediction device,
The battery model is
A DC resistance model representing the DC resistance (Rs) of the secondary battery;
A model representing a reaction resistance of the secondary battery, the reaction resistance model including a reaction resistance parameter (β) derived from a Butler-Volmer equation and correlated with an exchange current density;
An RC equivalent circuit model including a parallel connection body of a resistor and a capacitor, and a diffusion resistance model representing a diffusion resistance of the secondary battery;
A formula including circuit constants of the RC equivalent circuit model, the first polarization voltage calculating means for calculating the current polarization voltage based on a discrete expression for calculating the current polarization voltage of the diffusion resistance model;
An equation including the circuit constant, wherein the current polarization voltage decreases to a continuous time equation for predicting a residual voltage, which is the polarization voltage remaining after the lapse of the specified time from the present, and the current polarization voltage. A residual voltage predicting means for predicting the residual voltage by inputting
Second polarization voltage calculation means for predicting the future polarization voltage based on a continuous time formula for predicting the future polarization voltage of the diffusion resistance model newly generated after the lapse of the specified time from the present time;
Based on the total value of the residual voltage and the future polarization voltage, the potential difference in the DC resistance model, and the potential difference in the reaction resistance model, power prediction means for predicting the prediction target value after the lapse of the specified time from the present A battery power prediction apparatus comprising: The reaction resistance parameters are defined as constants α and γ, current flowing in the secondary battery as I, temperature of the secondary battery as T, and potential difference in the reaction resistance as ΔV.
The battery power prediction apparatus according to claim 1, wherein the battery power prediction apparatus satisfies the following conditions. The secondary battery is each of a plurality of battery cells connected in series,
The power prediction means includes
Assuming that the specified time is the discharge-side specified time, and the discharge current of the secondary battery is set to the allowable upper limit discharge current over the period from the current time until the discharge-side specified time elapses for each of the secondary batteries. First discharge time prediction means for predicting the voltage between the terminals of the secondary battery after the discharge side specified time has elapsed from the present when
When the minimum value of the voltage between the terminals of the secondary battery predicted by the first discharge time prediction unit is equal to or higher than the allowable lower limit voltage, the secondary battery predicted by the first discharge time prediction unit First output power predicting means for predicting the output possible power after elapse of the discharge side specified time from the current time based on the sum of the voltages between the terminals and the allowable upper limit discharge current;
When the minimum value of the inter-terminal voltage of the secondary battery predicted by the first discharge time prediction unit is less than the allowable lower limit voltage, each of the secondary batteries after the lapse of the discharge side specified time from the present time A discharge current search for searching for a maximum discharge current that is a discharge current of the secondary battery having a minimum value of the inter-terminal voltage of the battery as the allowable lower limit voltage and that is equal to or lower than the allowable upper limit discharge current according to a predetermined search method. Means,
For each of the secondary batteries, the discharge side specified time elapses from the present when it is assumed that the discharge current of the secondary battery is set to the maximum discharge current over a period from the current time until the discharge side specified time elapses. Second discharge time prediction means for predicting the voltage across the terminals of the secondary battery later;
Based on the sum of the inter-terminal voltages of the secondary battery predicted by the second discharge time prediction means and the maximum discharge current, the output possible power after the discharge side specified time has elapsed is predicted The battery power prediction apparatus according to claim 1, further comprising: a second output power prediction unit that performs the operation.   The battery power prediction device according to claim 3, wherein the discharge current search means searches for the maximum discharge current using a range from 0 to the allowable upper limit discharge current as a search range. The secondary battery is each of a plurality of battery cells connected in series,
The power prediction means includes
The specified time is defined as one processing cycle of the battery power prediction device, and the charging current of the secondary battery is set to the allowable upper limit charging current over the period from the present to the passage of the one processing cycle for each of the secondary batteries. First charging time prediction means for predicting the voltage across the terminals of the secondary battery after the one processing cycle from the present when it is assumed that
When the maximum value of the inter-terminal voltage of the secondary battery predicted by the first charging time prediction unit is equal to or lower than the allowable upper limit voltage, the secondary battery predicted by the first charging time prediction unit First input power predicting means for predicting the input possible power after the one processing cycle from the present based on the sum of the voltages between the terminals and the allowable upper limit charging current;
When the maximum value of the voltage between the terminals of the secondary battery predicted by the first charging time prediction unit exceeds the allowable upper limit voltage, the specified time is set to a predetermined time longer than the one processing cycle, The charging current of the secondary battery having the maximum value of the voltage between the terminals of the secondary battery after the lapse of the predetermined time as the allowable upper limit voltage, wherein the charging current is equal to or lower than the allowable upper limit charging current. Charging current search means for searching for current according to a predetermined search method;
For each of the secondary batteries, the 2nd battery after the predetermined time elapses from the present when it is assumed that the charging current of the secondary battery is set to the maximum charging current over a period from the present to the elapse of the predetermined time. A second charging time prediction means for predicting a voltage between terminals of the secondary battery;
Based on the sum of the voltages across the terminals of the secondary battery predicted by the second charging time prediction means and the maximum charging current, the input possible power after the predetermined time has elapsed is predicted from the present time. The battery power prediction apparatus according to any one of claims 1 to 4, further comprising a two-input power prediction unit.   6. The battery power prediction apparatus according to claim 5, wherein the charging current search means searches for the maximum charging current using a range from 0 to the allowable upper limit charging current as a search range. The RC equivalent circuit model is a model configured by N (N is an integer of 1 or more) the parallel connection bodies,
In the diffusion resistance model, the resistance value of the resistor of the m-th (m is a positive integer) constituting the RC equivalent circuit model is defined as Rwm, and the capacitance of the capacitor is defined as Cwm.
The residual voltage prediction means defines the residual voltage Vwr by defining the specified time as Tff and the current polarization voltage as Vwm.
Based on
The second polarization voltage calculating means defines the future polarization voltage Vws as Ip when a current flowing through the secondary battery over a period from the present to the lapse of the specified time is defined as Ip.
The battery power prediction apparatus according to claim 1, which predicts based on For each of the secondary batteries, further comprising OCV change prediction means for predicting the amount of change in the open-circuit voltage of the secondary battery in the period from the present to the lapse of the specified time,
The battery power prediction apparatus according to any one of claims 1 to 7, wherein the power prediction unit predicts the prediction target value after the lapse of the specified time from the present time by further using a change amount of the open-circuit voltage. .