JP5694088B2 - Secondary battery deterioration management system - Google Patents

Description

  The present invention relates to a secondary battery deterioration management system, and more particularly to secondary battery deterioration management based on changes in parameter values in a battery model capable of estimating the internal state of the secondary battery.

  A power supply system is used in which power is supplied to a load by a rechargeable secondary battery, and the secondary battery can be charged even during operation of the load as necessary. Typically, an electric vehicle such as a hybrid vehicle or an electric vehicle equipped with an electric motor driven by a secondary battery as a driving force source is equipped with such a power supply system.

  In the power supply system of an electric vehicle, the stored power of the secondary battery is used as the drive power of the electric motor as the driving force source, and the electric power generated when the electric motor regenerates and the generator that generates power with the rotation of the engine. The secondary battery is charged by generated power or the like. Therefore, in charge / discharge management of the secondary battery, it is required to accurately estimate a state of charge (SOC) with respect to a fully charged state. That is, it is necessary to accurately estimate the charging rate of the secondary battery even during charging / discharging or immediately after charging / discharging to limit excessive charging / discharging of the secondary battery. In addition, since the battery parameters (internal resistance, etc.) of the secondary battery gradually change (deteriorate) with use, it is necessary to accurately estimate the state of the secondary battery corresponding to such aging deterioration. Is required.

  For example, Patent Document 1 (Japanese Patent Laid-Open No. 2008-241246), Patent Document 2 (Japanese Patent Laid-Open No. 2010-60384), and Patent Document 3 (Japanese Patent Laid-Open No. 2010-60406) describe the inside of an active material of a secondary battery. Based on the internal state including the lithium concentration distribution inside the battery by the battery model including the diffusion equation model for estimating the concentration distribution of the reaction participating substance (typically lithium in the lithium ion secondary battery) It is described that the SOC of the secondary battery is estimated.

  Further, Patent Documents 1 to 3 describe that parameter identification is performed online during use of the secondary battery with respect to DC resistance and diffusion coefficient (diffusion resistance), which are parameters in the battery model formula. . Then, it is described that the deterioration of the secondary battery is evaluated by obtaining the change rate (parameter change rate) from the initial state value for the current parameter value obtained by parameter identification.

  Patent Document 4 (Japanese Patent Application Laid-Open No. 2011-15520) discloses a parameter change when estimating a current state of a secondary battery mounted on a vehicle using a battery model according to Patent Documents 1 to 3. Learning to rate is described. In particular, it is described that the parameter change rate of the direct current resistance and the parameter change rate of the diffusion coefficient similar to those of Patent Documents 1 to 3 are learned online using a map classified by battery temperature and SOC. Thereby, it is possible to estimate the parameter change rate with high accuracy in correspondence with the behavior of the internal resistance changing with the battery temperature and the SOC. As a result, the estimation error of the state of the secondary battery including the SOC can be reduced by suppressing the error of the parameter value in the battery model formula.

  Furthermore, in Patent Document 4, an in-vehicle secondary battery that can be charged by a power source external to the vehicle is subjected to an actual charge / discharge test (hereinafter referred to as “IV test”) according to a predetermined test pattern during external charging by the power source. The current direct current resistance and the diffusion coefficient of the secondary battery are calculated by executing (also referred to as). Furthermore, by using the battery model formula that reflects the DC resistance and diffusion coefficient calculated at this opportunity, the IV test is performed on the software in a simulated manner, thereby increasing the increase rate of the internal resistance of the secondary battery. The calculation is described.

  Patent Document 5 (Japanese Patent Laid-Open No. 2007-5304) introduces a variable forgetting factor by determining an optimum value of a time weighting factor when estimating a model parameter related to a secondary battery by regression analysis. Have been described. This makes it possible to estimate the parameters so as to have an early influence on recently obtained data.

JP 2008-241246 A JP 2010-60384 A JP 2010-60406 A JP 2011-15520 A JP 2007-5304 A

  In Patent Document 4, an IV test (hereinafter, referred to as “IV test”) virtually executed on software using a battery model that reflects a DC resistance and a diffusion coefficient obtained by an actual IV test at the time of external charging. The internal resistance can be evaluated by “virtual IV test”.

  However, since the DC resistance and diffusion coefficient reflected in the battery model formula in the virtual IV test are limited to those obtained in the external charging, the secondary battery used in a wide SOC range and temperature range. However, satisfactory deterioration management cannot always be performed.

  Further, in Patent Document 4, when the secondary battery is used, the parameter change rate for the diffusion coefficient and the DC resistance can be obtained for each predetermined learning region defined by the combination of the SOC region and the battery temperature region. It is known that the internal resistance of a secondary battery is determined by the diffusion coefficient indicating the degree of diffusion of reaction-related substances typified by lithium inside the battery and the DC resistance when current is generated due to an electrochemical reaction. Yes.

  However, in the internal resistance at the time of actual charge / discharge (hereinafter also referred to as “total internal resistance”), the influences of the diffusion resistance and DC resistance occupying the total internal resistance differ depending on the SOC and the battery temperature. For this reason, even if the parameter change rate of the DC resistance and the parameter change rate of the diffusion coefficient are obtained for each learning region, an appropriate deterioration evaluation of the secondary battery is performed based directly on these parameter change rates. It is difficult to do. On the other hand, performing a charge / discharge test (IV test) on an actual machine in order to obtain an internal resistance (that is, total internal resistance) corresponding to various SOC ranges and battery temperature ranges is costly. In addition, when charging is stopped during external charging and the charge / discharge test is performed for each SOC, there is a problem that the charging time becomes long.

  The present invention has been made to solve such problems, and the object of the present invention is to provide a DC resistance learned for each learning region according to a combination of the SOC and battery temperature of the secondary battery, and Based on the parameter change rate of the diffusion coefficient, the deterioration evaluation of the secondary battery is appropriately executed.

  In one aspect of the present invention, a deterioration management system for a secondary battery includes a detection unit, a battery state estimation unit, a first storage unit, a parameter change rate learning unit, a second storage unit, and a deterioration evaluation unit. With. The detection means detects the battery voltage, battery current, and battery temperature of the secondary battery. The battery state estimation means sequentially estimates the state quantity according to the battery model for estimating the state quantity of the secondary battery including at least the SOC using data including at least one of the battery voltage and the battery current. The first storage means stores in advance an initial state value for each predetermined learning region defined by a combination of the SOC and the battery temperature for a predetermined parameter in the parameter group used in the battery model. The parameter change rate learning means is a parameter change rate that is a ratio of a current parameter value to an initial state value of a predetermined parameter by parameter identification based on the data detected by the detection means and the battery model during use of the secondary battery. Are learned sequentially. The second storage means stores the parameter change rate learned value calculated by the parameter change rate learning means for each learning region. The degradation evaluation means is a result of a virtual test performed for each learning region to obtain a behavior of the secondary battery when the secondary battery performs at least one of charging and discharging according to a predetermined test pattern using a battery model. Based on this, the deterioration degree of the secondary battery is evaluated.

Preferably, the deterioration evaluation unit is configured to perform two tests based on a virtual test result using a battery model in which a parameter value calculated based on the parameter change rate learning value stored in the second storage unit is substituted for a predetermined parameter. More preferably, the deterioration evaluation unit substitutes, for each learning region, a parameter value calculated based on the parameter change rate learning value stored in the second storage unit for a predetermined parameter. Means for executing a virtual test using a battery model; means for calculating an internal resistance value of a secondary battery in a predetermined test pattern based on a result of the virtual test for each learning area; and for each learning area The reference value of the internal resistance value set in advance to the internal resistance value for each learning region calculated based on the result of the virtual test Means for calculating an index value indicating the degree of deterioration.

  Further preferably, the deterioration evaluation means includes a limit value relating to a parameter change rate of the predetermined parameter obtained based on a result of the virtual test for each learning region, and a parameter change rate learning value stored in the second storage means. The deterioration degree of the secondary battery is evaluated based on the comparison for each learning region.

  More preferably, the predetermined parameter includes a plurality of parameters. For each learning region, the deterioration evaluation means fixes the values of the other parameters except for the first parameter among the plurality of parameters to the initial state values, while changing the first parameter to change the battery model. First means corresponding to the means for executing the used virtual test a plurality of times and the internal resistance value of the secondary battery in the predetermined test pattern is an allowable maximum value based on the results of the plurality of virtual tests. Means for obtaining the parameter change rate limit value of the parameter as a limit value for each learning region, the parameter change rate learned value stored in the second storage means for the first parameter, the parameter change rate limit value, And a means for calculating an index value indicating the degree of deterioration of the secondary battery based on the comparison for each learning region.

  More preferably, the deterioration evaluation means includes a third storage means for storing a parameter change rate limit value for each learning region obtained in advance by a plurality of virtual tests before using the secondary battery. In the calculation of the index value, the parameter change rate limit value for each learning area is read from the third storage means.

  Alternatively, preferably, the deterioration evaluation means indicates that the internal resistance value is the highest among the means for calculating the comparison evaluation value indicating the result of the comparison for each learning region and the comparison evaluation value for each learning region. Index value calculating means for calculating the index value according to the maximum value.

  More preferably, the parameter change rate learning means updates the parameter change rate learning value of the learning region at that time each time a predetermined learning condition is satisfied. The degradation management system further includes fourth storage means for storing the number of times the parameter change rate learning means has updated the parameter change rate learning value for each learning region. The deterioration evaluation means includes a means for calculating a comparative evaluation value indicating a comparison result for each learning area, and a learning area selected based on the number of learning times for each learning area stored in the fourth storage means. Means for calculating an index value according to the comparative evaluation value.

  More preferably, the parameter change rate learning means updates the parameter change rate learning value of the learning area at that time each time a predetermined learning condition is satisfied. The degradation management system further includes fourth storage means for storing the number of times the parameter change rate learning means has updated the parameter change rate learning value for each learning region. The deterioration evaluation means includes means for calculating a comparison evaluation value indicating the result of comparison for each learning area, and learning for each learning area stored in the fourth storage means for the comparison evaluation value for each learning area. Means for calculating an index value according to an average value weighted by the number of times.

  More preferably, the parameter change rate learning means further includes means for reducing the number of learning times for each learning area stored in the fourth storage means as time elapses.

  Preferably, the battery model is an open-circuit voltage that is a function of a concentration of a reaction-participating substance on the surface of the active material inside the secondary battery, and a voltage equation expressed by a voltage change amount due to a charge / discharge current inside the secondary battery; And a diffusion equation that defines the distribution of the concentration of the substances involved in the reaction inside the active material. The predetermined parameter includes a parameter indicating DC resistance in the voltage equation.

  Preferably, the battery model is an open-circuit voltage that is a function of a concentration of a reaction-participating substance on the surface of the active material inside the secondary battery, and a voltage equation expressed by a voltage change amount due to a charge / discharge current inside the secondary battery; And a diffusion equation that defines the distribution of the concentration of the substances involved in the reaction inside the active material. The predetermined parameter includes a diffusion parameter representing the diffusion rate of the reaction participating substance in the diffusion equation.

  According to the present invention, the deterioration evaluation of the secondary battery is easily and appropriately performed based on the parameter change rate of the DC resistance and the diffusion coefficient learned for each learning region according to the combination of the SOC of the secondary battery and the battery temperature. Can be executed.

1 is a block diagram showing a schematic configuration of a power supply system using a secondary battery as a power source to which a secondary battery deterioration management system according to an embodiment of the present invention is applied. It is a functional block diagram for demonstrating the control structure by ECU shown in FIG. It is a conceptual diagram explaining the outline of the internal structure of the secondary battery expressed with a battery model. It is a conceptual diagram which shows the structural example of the map which shows the change characteristic of the open circuit voltage with respect to the change of local SOC. It is a conceptual diagram explaining the map which shows the relationship between the average lithium density | concentration in an active material model, and a charging rate. It is a conceptual diagram explaining the structure of the initial state value map of DC resistance. It is a conceptual diagram explaining the structure of the initial state value map of a spreading | diffusion coefficient. It is a flowchart explaining the control processing procedure of the state estimation and parameter change rate learning of a secondary battery using the battery model formula by embodiment of this invention. It is a conceptual diagram explaining the structure of the parameter change rate map of DC resistance. It is a conceptual diagram explaining the structure of the parameter change rate map of a diffusion coefficient. It is a conceptual diagram explaining the variation of the update process of the map value (learning value) of parameter change rate. 3 is a schematic functional block diagram for explaining the function of a deterioration evaluation unit according to Embodiment 1. FIG. It is a wave form diagram of the battery current in the charge / discharge test (IV) test for evaluating the total internal resistance of a secondary battery. It is a wave form diagram of a battery voltage in an IV test. It is a flowchart for demonstrating the control processing with respect to each learning area | region by a deterioration evaluation part. It is a flowchart explaining the control processing procedure of a virtual IV test. FIG. 10 is a schematic functional block diagram for explaining a function of a deterioration evaluation unit according to a modification of the first embodiment. 6 is a flowchart for illustrating a control process for each learning region in a secondary battery deterioration management system according to a modification of the first embodiment. 6 is a conceptual diagram illustrating a configuration of a learning count map of DC resistance in a secondary battery deterioration management system according to Embodiment 2. FIG. 6 is a conceptual diagram illustrating the configuration of a learning coefficient map of diffusion coefficients in a secondary battery deterioration management system according to Embodiment 2. FIG. 6 is a flowchart illustrating a control process related to learning count management in a secondary battery deterioration management system according to a second embodiment. 5 is a flowchart illustrating a control process by a deterioration evaluation index calculation unit in a deterioration management system for a secondary battery according to a second embodiment. 10 is a flowchart illustrating a control process by a deterioration evaluation index calculation unit in a deterioration management system for a secondary battery according to a modification of the second embodiment. 10 is a flowchart illustrating a control process related to the management of the number of learnings in the secondary battery deterioration management system according to the third embodiment. FIG. 10 is a schematic functional block diagram for explaining a function of a deterioration evaluation unit according to a fourth embodiment. It is a flowchart explaining the control processing for calculating | requiring the limit value of the parameter change rate of DC resistance by a virtual IV test. FIG. 10 is a schematic functional block diagram for explaining a function of a deterioration evaluation unit according to a modification of the fourth embodiment. It is a flowchart explaining the control processing for calculating | requiring the limit value of the parameter change rate of a diffused resistance by a virtual IV test.

  Embodiments of the present invention will be described below in detail with reference to the drawings. In the following, the same or corresponding parts in the drawings are denoted by the same reference numerals, and the description thereof will not be repeated in principle.

(overall structure)
FIG. 1 is a block diagram showing a schematic configuration of a power supply system using a secondary battery as a power source to which a secondary battery deterioration management system according to an embodiment of the present invention is applied.

Referring to FIG. 1, secondary battery 10 supplies driving power for load 50. The load 50 is constituted by, for example, a traveling motor mounted on an electric vehicle, a hybrid vehicle, or the like. Furthermore, the load 50 charges the secondary battery 10 with the regenerative power of the electric motor. The secondary battery 10 is typically composed of a lithium ion battery.

  The secondary battery 10 is provided with a current sensor 20 for measuring battery current, a voltage sensor 30 for measuring battery voltage, and a temperature sensor 40 for measuring battery temperature. Hereinafter, the measured value by the current sensor 20 is denoted as battery current Ib, the measured value by the voltage sensor 30 is denoted by battery current Vb, and the measured value by the temperature sensor 40 is denoted by battery temperature Tb.

  The battery current Ib, the battery voltage Vb, and the battery temperature Tb measured by the sensors 20 to 40 are sent to an electronic control unit (ECU) 100. The battery current Ib is defined as a positive value (Ib> 0) when the secondary battery 10 is discharged, and a negative value (Ib <0) when charged.

  The ECU 100 includes a microprocessor, a memory, an A / D converter, a D / A converter, and the like (not shown). By executing a predetermined program stored in advance in the memory, the ECU 100 uses a predetermined signal using data / input from a sensor or the like. An arithmetic process is executed, and an output signal / data based on the arithmetic process result is generated. In the present embodiment, the ECU 100 performs a secondary operation according to a battery model to be described later based on battery data (which collectively represents Ib, Vb, and Tb) detected by the current sensor 20, the voltage sensor 30, and the temperature sensor 40. Battery information represented by the charging rate (SOC) is generated by dynamically estimating the internal state of the battery 10.

  In particular, the ECU 100 actually operates when the load 50 is operated by the secondary battery 10, that is, when the load 50 is driven by the power supplied from the secondary battery 10 or when the secondary battery 10 is charged by the regenerative power from the load 50. It is possible to perform parameter estimation in the battery model formula based on the battery data during the load operation.

  The battery information obtained by the ECU 100 is sent to the load control device 60. The load control device 60 generates a control command for controlling the driving state of the load 50 based on the battery information. For example, when the SOC of the secondary battery 10 decreases, a control command that limits the power used by the load 50 is generated. Conversely, when the SOC of the secondary battery 10 is high, a control command that suppresses the generation of regenerative power by the load 50 is generated.

  FIG. 2 is a functional block diagram for illustrating a control configuration by ECU 100 shown in FIG. FIG. 2 particularly shows a configuration relating to internal state estimation and deterioration management of the secondary battery 10.

  2, ECU 100 includes a data collection unit 110, a battery state estimation unit 120, an initial value storage unit 130, a change rate storage unit 140, a parameter change rate learning unit 150, and a deterioration evaluation unit 200. including.

  The data collection unit 110 collects data used for the battery model for estimating the state of the secondary battery 10. Specifically, the data collection unit 110 obtains the detected value of the battery voltage Vb, the detected value of the battery current Ib, and the detected value of the battery temperature Tb of the secondary battery 10 from the voltage sensor 30, the current sensor 20, and the temperature sensor 40. Get each.

  The battery state estimation unit 120 includes a battery model 125 for dynamically estimating the internal state of the secondary battery 10. The battery state estimation unit 120 uses the data (for example, the battery voltage Vb and the battery temperature Tb) acquired by the data collection unit 110 and uses the internal state (behavior) of the secondary battery 10 according to the battery model formula that configures the battery model 125. ). Furthermore, the battery state estimation part 120 estimates the charging rate (SOC) of the secondary battery 10 based on the estimation result.

  The initial value storage unit 130 stores initial state values of parameters used for the battery model 125.

  When the secondary battery 10 is used, the parameter change rate learning unit 150 calculates a parameter change rate indicated by the ratio of the current parameter value to the initial state value for the parameter whose initial state value is stored in the initial value storage unit 130. presume. That is, the parameter change rate is estimated online using the data acquired by the data collection unit 110.

  The parameter change rate learning unit 150 sequentially calculates a parameter change rate learning value based on the parameter change rate estimation result when online. The change rate storage unit 140 stores the parameter change rate learned value calculated by the parameter change rate learning unit 150. In other words, the parameter change rate learning value stored in the change rate storage unit 140 is sequentially updated while online. The parameter change rate estimation and its learning will be described in detail later.

  The degradation evaluation unit 200 evaluates the total internal resistance (hereinafter, also referred to as “IV resistance”) in the IV test in which the secondary battery 10 is charged and / or discharged according to a predetermined test pattern, whereby the secondary battery 10 The degradation index value Pdg is calculated. As will be described in detail later, the degradation evaluation unit 200 actually charges and discharges the secondary battery 10 under various conditions by virtually executing the IV test using the battery model 125. Rather, it is configured to evaluate IV resistance.

(Explanation of battery model type)
Next, an example of a battery model used for state estimation of the secondary battery 10 will be described. The battery model described below is constructed by including a non-linear model so that the internal behavior can be dynamically estimated in consideration of the electrochemical reaction inside the secondary battery. Although the kind of secondary battery is not limited, the following battery model will be described assuming that a lithium ion battery is applied as the secondary battery 10.

  FIG. 3 is a conceptual diagram illustrating an outline of the internal configuration of the secondary battery expressed by a battery model.

  Referring to FIG. 3, secondary battery 10 includes a negative electrode 12, a separator 14, and a positive electrode 15. The separator 14 is configured by impregnating an electrolytic solution into a resin provided between the negative electrode 12 and the positive electrode 15.

Each of the negative electrode 12 and the positive electrode 15 is composed of an aggregate of spherical active materials 18. When the secondary battery 10 is discharged, a chemical reaction that releases lithium ions Li ⁺ and electrons e ⁻ is performed on the interface of the active material 18 of the negative electrode 12. On the other hand, a chemical reaction that absorbs lithium ions Li ⁺ and electrons e ⁻ is performed on the interface of the active material 18 of the positive electrode 15. When the secondary battery 10 is charged, a reaction opposite to the above reaction is performed with respect to the emission and absorption of the electron e ⁻ .

The negative electrode 12 is provided with a current collector 13 that absorbs electrons e ⁻ , and the positive electrode 15 is provided with a current collector 16 that emits electrons e ⁻ . The negative current collector 13 is typically made of copper, and the positive current collector 16 is typically made of aluminum. The current collector 13 is provided with a negative terminal, and the current collector 16 is provided with a positive terminal. By the exchange of lithium ions Li ⁺ through the separator 14, the secondary battery 10 is charged / discharged to generate a charging current or a discharging current.

  That is, the charge / discharge state inside the secondary battery varies depending on the lithium concentration distribution in the active material 18 of the electrodes (the negative electrode 12 and the positive electrode 15). This lithium corresponds to a reaction participating substance in the lithium ion battery.

Pure electric resistance (pure resistance) Rd against the movement of electrons e ⁻ at the negative electrode 12 and the positive electrode 15, and charge transfer resistance (reaction resistance) Rr that acts equivalently as an electric resistance when a reaction current is generated at the active material interface, Is equivalent to the DC resistance when the secondary battery 10 is viewed macroscopically. The macroscopic DC resistance, also designated DC resistance R _a in the following. Further, the diffusion of lithium Li in the active material 18 is governed by the diffusion coefficient D _s . That is, the diffusion coefficient D _s is a parameter indicating the magnitude of the diffusion resistance.

  Subsequently, an example of the battery model 125 will be described. In the battery model described here, a model that ignores the influence is constructed in consideration of the small influence of the electric double layer capacitor at normal temperature. Furthermore, the battery model is defined as a model per unit electrode plate area of the electrode. By using a model per unit electrode plate area of the electrode, the model can be generalized to the design capacity.

First, the battery voltage V output is a voltage of the secondary battery 10, using a battery temperature T, a battery current I, an open circuit voltage (OCV) U and, above secondary battery 10 as a whole macroscopic DC resistance R _a The following equation (1) (voltage equation) is established. Here, the battery current I represents a current value per unit electrode plate area. That is, when the battery current flowing through the positive and negative terminals (current value measurable by an ammeter) is Ib and the double-sided electrode plate area of the battery is S, the battery current I is defined as I = Ib / S. Hereinafter, “current” and “current estimation value” described in the battery model refer to the current per unit electrode plate area unless otherwise specified.

θ ₁ and θ ₂ represent local SOC on the surface of the positive electrode active material and local SOC on the surface of the negative electrode active material, respectively. The open circuit voltage OCV is expressed as a potential difference between the positive electrode open potential U ₁ and the negative electrode open potential U ₂ .

As shown in FIG. 4, positive electrode open potential U ₁ and negative electrode open potential U ₂ have characteristics that change depending on local SOC θ ₁ and local SOC θ ₂ , respectively. Therefore, in the initial state of the secondary battery 10, by measuring the relationship between the local SOC θ ₁ and the positive electrode open potential U ₁ and the relationship between the local SOC θ ₂ and the negative electrode open potential U ₂ , the local SOC θ ₁ it is possible to create a positive open-circuit potential U ₁ (θ ₁₎ variation characteristic and characteristic map that prestores change characteristics of negative electrode open-circuit potential U ₂ (θ ₂₎ with respect to changes in the local SOC [theta] ₂ of the relative change.

Also, DC resistance _{R a} is local SOC (θ _1), having a characteristic that varies according to changes in the local SOC (θ ₂₎ and battery temperature. That is, the DC resistance _Ra is shown as a function of SOC (more specifically, local SOC (θ ₁ , θ ₂ )) and battery temperature T. Thus, based on the measured experimental results in the initial state of the secondary battery 10, it is possible to pre-create the initial state value map 131 of DC resistance R _a as shown in FIG.

Referring to FIG. 6, in initial state value map 131, an SOC range obtained by dividing 0 (%) to 100 (%) into a plurality, and a temperature range obtained by dividing −Ty (° C.) to Tx (° C.) into a plurality of each region defined by the combination of the initial state value of the DC resistance R _a is stored. Since these plural regions coincide with the learning category of the parameter change rate described later, they are also referred to as “learning regions”. Hereinafter, as an example, in the present embodiment, the range of local SOC (θ ₁ or θ ₂ ) is used as the SOC range of each map.

Referring to FIG. 3 again, as described above, in the spherical active material models of the negative electrode 12 and the positive electrode 15, the local SOC θ _i (i = 1, 2) on the active material surface (interface with the electrolyte) is It is defined by the following equation (2). As in the case of the local SOC θ _i , in the following description, the subscript represented by i is defined as a positive electrode when 1 and a negative electrode when 2.

In the formula (2), c _{se, i} is the average lithium concentration at the active material interface, and c _{s, i, max} is the critical lithium concentration in the active material.

In the active material handled by the spherical model, the lithium concentration c _{s, i} has a distribution in the radial direction. That is, the lithium concentration distribution in the active material assumed to be spherical is defined by the diffusion equation of the polar coordinate system shown in the following equation (3).

In the formula (3), D _{s, i} is a diffusion coefficient of lithium in the active material. The diffusion coefficient D _{s, i} has a characteristic that varies depending on the SOC and the battery temperature. Hereinafter, when D _{s, i} (i = 1, 2) is generically referred to, it is also simply referred to as a diffusion coefficient D _s .

For the diffusion coefficient D _s, as with DC resistance R _a of the above, based on the measured experimental results in the initial state of the secondary battery 10, it is possible to pre-create the initial state value map 132 as shown in FIG. 7 .

Referring to FIG. 7, in the initial state value map 132, the initial state value of the diffusion coefficient D _s is stored for each learning region divided in common with the initial state value map 131.

  The initial state storage unit 130 of FIG. 2 is configured by the initial state value map 131 shown in FIG. 6 and the initial state value map 132 shown in FIG. The initial value storage unit 130 corresponds to a “first storage unit”.

  The boundary conditions of the diffusion equation (3) are set as the following expressions (4) and (5).

  Equation (4) indicates that the concentration gradient at the center of the active material is zero. In the formula (5), the change in lithium concentration at the electrolyte interface of the active material means that it changes as lithium enters and leaves from the active material surface.

In equation (5), r _{s, i} represents the active material radius, ε _{s, i} represents the volume fraction of the active material, and as _{, i} represents the active material surface area per unit electrode volume. These values are determined from electrode specifications, that is, design values. F is a Faraday constant.

Furthermore, j ^Li in the formula (5) is the amount of lithium produced per unit volume and time, and for the sake of simplicity, assuming that the reaction is uniform in the electrode thickness direction, the electrode thickness L _i and the unit electrode Using the battery current I per plate area, the following equation (6) is used.

  By inputting the battery current I or the battery voltage V and solving these equations (1) to (6) simultaneously, the internal state of the secondary battery 10 is estimated while calculating the estimated voltage value or estimated current value. Thus, the charging rate can be estimated.

  By using this battery model, for example, it is possible to estimate the charging rate of the secondary battery with the battery voltage V as an input. When the battery voltage V is input, the charging rate is calculated using a map showing the relationship between the average lithium concentration in the active material model and the charging rate as shown in FIG. That is, the current SOC can be estimated according to the map of FIG. 5 from the average lithium concentration in the active material model calculated by the battery model formula.

  Here, a control processing procedure for estimating the state of the secondary battery using the battery model described above will be described with reference to FIG. The process shown in FIG. 8 is executed by the ECU 100 at every predetermined calculation cycle.

  FIG. 8 shows the control processing procedures for both state estimation and parameter change rate learning of the secondary battery. The control processing in steps S100 to S190 in the first half is performed by the battery state estimation unit 120 shown in FIG. The state estimation of the secondary battery 10 is realized.

  Referring to FIG. 8, ECU 100 measures battery voltage Vb by voltage sensor 30 in step S100. In step S110, ECU 100 measures battery temperature Tb by temperature sensor 40. The battery voltage Vb and the battery temperature Tb in the current calculation cycle are used as the battery voltage V and the battery temperature T in the battery model formula, respectively. ECU 100 (data collection unit 110) also collects battery current Ib obtained by measurement.

In step S120, ECU 100 calculates local SOC θ _i (θ ₁ and θ ₂ ) on the active material surface based on lithium concentration distribution c _{se, i} in the previous calculation cycle, according to equation (2).

Further, ECU 100, in step S130, as shown in FIG. 4, calculated from the characteristic map of open-circuit potential U _i (theta _i) for the local SOC [theta] _i, open-circuit potential U _i a (U ₁ and U _2), As the potential difference between the calculated open circuit potentials U ₁ and U ₂ , an open circuit voltage estimated value U # is calculated.

Further, ECU 100, in step S140, the local SOC [theta] _i and calculated in step S130, based on the measured battery temperature T, obtaining the DC resistance R _a. For example, the DC resistance _Ra is determined according to the initial state value map 131.

ECU100 further by step S150, the battery voltage V (V = Vb), calculated opened by using the voltage estimate U♯ and DC resistance R _a, the estimated value of the battery current based on the following equation (7) I _te is calculated.

Next, in step S160, ECU 100 substitutes battery current estimated value I _te into battery current I in equation (6) to calculate lithium generation amount j ^Li per unit volume / time. By using the lithium production amount j ^Li per unit volume / time as the boundary condition of the equation (5) and solving the diffusion equation of the equation (3), the lithium concentration distribution in the active material of each of the positive and negative electrodes is determined. . Note that the diffusion coefficient D _{s, i} in the expression (3) is also obtained based on the SOC (local SOC) and the battery temperature. For example, the diffusion coefficient D _{s, i} is determined according to the initial state value map 132.

When solving the diffusion equation (3), the ECU 100 uses the diffusion equation discretized according to the position and time, and uses the diffusion concentration equation c _{s, i, k} (t + Δt) inside the active material (where Δt is discrete) A time step (corresponding to an operation execution cycle) is indicated, and k represents a discrete position number discretized in the radial direction (step S170). Since the method of discretizing the diffusion equation by position and time is well known, detailed description will not be repeated here.

Next, in step S180, ECU 100 calculates an average lithium concentration c _save inside the active material according to the following equation (8). However, in the formula (8), N is the number of divisions when the spherical active material is discretized in the radial direction.

In step S190, ECU 100 uses a prestored map showing the relationship between the average lithium concentration c _save in the active material and the charging rate (SOC) of secondary battery 10 as shown in FIG. To calculate the SOC.

  In this way, ECU 100 (battery state estimation unit 120) determines secondary battery 10 charging rate (SOC), open-circuit voltage estimated value U #, and unit pole from battery voltage Vb and battery temperature Tb measured by the sensor. An estimated value of the battery current per plate area can be calculated. Moreover, the estimated value of the current flowing through the entire battery can be calculated by multiplying the estimated current value per unit electrode plate area by the double-sided electrode plate area of the battery, from the above-described definition formula of the battery current I.

In the above battery model formula, a corresponding separate spherical active material model was set for each of the negative electrode 12 and the positive electrode 15. However, in order to reduce the calculation load of the ECU, a single spherical model having the averaged characteristics of the negative electrode 12 and the positive electrode 15 may be used as an active material model common to the positive and negative electrodes. When a single spherical model is used, the local SOCs θ ₁ and θ ₂ are common values.

(Update parameter values)
Next, parameters in the battery model will be described.

Some of the plurality of parameters of the battery model formula change due to deterioration of the battery accompanying the use of the secondary battery. For example, the DC resistance R _a is changed gradually due to the deterioration of the battery. When the difference between the direct current resistance _Ra (initial state value) in the initial state (typically when new) and the actual direct current resistance _Ra (current parameter value) is large, the SOC estimation error is It tends to occur.

  Similarly, due to the deterioration of the battery, the diffusion rate of the reaction-participating substance in the active material decreases (that is, the diffusion coefficient decreases), and as a result, the so-called diffusion resistance increases. The increase in the diffusion resistance has a great influence on the battery performance and the current-voltage characteristics, particularly in the case where charging / discharging with a large current is continued. Therefore, in an electric vehicle (hybrid vehicle, electric vehicle, etc.) that charges or discharges a battery with a large current, it is preferable to estimate a change in diffusion resistance, that is, a change in diffusion coefficient in the active material.

Thus, in this embodiment, the DC resistance R _a and the diffusion coefficient D _s in the formula (3) in the formula (1), by sequentially estimating the parameters rate of change in the same manner as Patent Documents 1 to 4, the parameters Update the value.

First, a description will be given parameter change rate of the DC resistance R _a. For DC resistance R _a, the parameter rate of change gr from the initial state value R _an, is defined by the following formula (9).

gr = R _a / R _an (9)
Parameter rate of change learning unit 150, the parameter change rate gr of DC resistance R _a, estimated using the recursive least-square method with forgetting factor to be described below. First, the sequential least square method with a forgetting factor will be described.

  According to the sequential least square method, in the system represented by the linear regression model expressed by the following equation (10), the parameter Θ in the equation (8) is the time update equation expressed by the equations (11) to (13). , (14), and (15) are estimated by sequentially calculating according to the initial conditions. In each equation, the estimated value of the parameter Θ is indicated by Θ #.

In the equations (11) and (13), λ is a forgetting factor, and usually λ <1.0. Further, P is a covariance matrix, and an initial value P (0) of the equation (15) is a matrix obtained by multiplying a diagonal element of the unit matrix I by a constant γ, and γ is usually a large value of about 10 ² to 10 ^3. Use the value. The initial value Θ # ₀ of the parameter Θ # is usually a zero vector.

  Using such a sequential least square method with a forgetting factor, the DC resistance change rate gr is estimated as follows.

The DC resistance R _a of the secondary battery that has changed over time (deteriorated) from the new state can be expressed as R _a = gr · R _an by the definition of the equation (9), and this is substituted into the equation (1). When rewritten in the form of equation (10), equation (16) is obtained as a linear regression model equation based on the battery model equation. In the equation (16), the local SOCs θ ₁ and θ ₂ are comprehensively expressed as “θ”.

When the secondary battery 10 is in use (on-line), the value estimated in the charging rate estimation process is used as the open circuit voltage U (θ) on the left side of the equation (16), and V is the measured battery voltage. Y can be calculated by using Vb. For the right side of the equation (16), the initial state value R _an of the DC resistance is obtained by referring to the initial state value map 131 using the battery temperature Tb and the local SOCs θ ₁ and θ ₂ as arguments. As the battery current I, Z can be calculated by substituting the current value per unit plate area calculated from the current battery current (measured value) Ib.

Such using computed Y and Z in (11) to (15) of the forgetting factor with recursive least square method, as the estimated parameter theta, can be sequentially estimating parameters change rate gr of DC resistance R _a It becomes. It is also possible to apply a least square method of another method such as a collective least square method.

Next, estimation of the parameter change rate of the diffusion coefficient D _s will be described.
Also for the diffusion coefficient D _s , the parameter change rate gd of the diffusion coefficient is defined as the change rate with respect to the initial state parameter value (D _sn ) according to the following equation (17).

gd = D _s / D _sn (17)
ECU 100 (data collection unit 110) repeatedly obtains data on battery voltage Vb, battery current Ib, and battery temperature Tb at a predetermined period in a time range in which the influence of the diffusion resistance is greatly expressed in the battery voltage.

Further, ECU 100 uses the battery data in that range to obtain an estimated value I _te of the battery current using a battery model when the parameter change rate is a certain candidate value, and evaluates an error from the actual battery current Ib. An evaluation function is calculated. Further, for example, by using the well-known GSM method (Golden section method), the parameter change rate that minimizes the evaluation function can be searched by executing the above process a predetermined number of times while switching the parameter change rate. it can.

  The GSM method is a kind of bisection method and has a feature that an optimum value satisfying the allowable error can be obtained by a known search function by determining the search range and the allowable error. The lithium diffusion coefficient in the active material of a lithium ion battery after a certain use condition and period of use can be determined in advance by a deterioration test or the like, and to what extent the diffusion coefficient changes compared to the initial state to the maximum Can be predicted in advance.

  Therefore, by setting the range of the change rate that can change to the maximum as the search range, there is an advantage that the calculation time required for estimating the diffusion coefficient change rate can be predicted in advance. This is suitable for application to a secondary battery mounted on a hybrid vehicle or an electric vehicle. Since details of the GSM method are known, detailed description will not be repeated here. Note that the estimation of these parameter change rates is described in detail in Patent Documents 1 to 3.

  In the present embodiment, the parameter change rate is learned for each predetermined learning region defined by the combination of the SOC (local SOC) of the secondary battery 10 and the battery temperature Tb, as in Patent Document 4. .

Figure 9 is for storing the parameter rate of change in DC resistance R _a, it is a conceptual diagram illustrating a schematic configuration of a change rate map 141.

  Referring to FIG. 9, in change rate map 141, learning value grl of parameter change rate gr defined by equation (9) is stored for each learning region set in common with initial state value maps 131 and 132. The The initial value of the parameter change rate learning value grl is 1.0 in each learning region. Then, when a predetermined learning condition is satisfied, parameter change rate learning unit 150 determines a map value based on the estimated value of parameter change rate gr in a learning region corresponding to the SOC (local SOC) and battery temperature at that time. (Parameter change rate learned value grl) is updated.

  For example, a new learning value gl1 is calculated according to the following equation (18) based on the map value gl0 before update and the estimated value g of the parameter change rate when the learning condition is satisfied. Note that the change in the learning value is smoothed by the coefficient α (0 <α <1) in the equation (18).

gl1 = α · g + (1−α) · gl0 (18)
FIG. 10 is a conceptual diagram illustrating a schematic configuration of the change rate map 142 for storing the parameter change rate learned value gdl of the diffusion coefficient D _s .

  Referring to FIG. 10, in change rate map 142, learning of parameter change rate gd defined by equation (17) is performed for each learning region set in common with initial state value maps 131 and 132 and change rate map 141. The value gdl is stored. The initial value of the parameter change rate learning value gdl is 1.0 in each learning region. When a predetermined learning condition is established, parameter change rate learning unit 150 determines a map value (parameter) based on the estimated value of parameter change rate gd in a learning region corresponding to the SOC (local SOC) and battery temperature at that time. The change rate learned value gdl) is updated. The learning value can be updated according to the above equation (18).

  The change rate storage unit 140 of FIG. 2 is configured by the change rate map 141 shown in FIG. 8 and the change rate map 142 shown in FIG. The change rate maps 141 and 142 correspond to “second storage means”.

  FIG. 11 is a conceptual diagram illustrating a variation regarding the update of the learning value in the change rate map.

  Referring to FIG. 11, it is assumed that a learning condition for the parameter change rate is satisfied in learning area AR1. At this time, at least the map value of the learning area AR1 is updated according to the above equation (18).

  Furthermore, the map value may be updated for at least a part of the learning areas AR2 to AR8 adjacent to the learning area AR1. At this time, it is preferable to set the coefficient α in the equation (18) to a small value as compared with the update of the map value in the learning area AR1.

  In this way, when the learning condition is satisfied, the learning result can be reflected even in a similar SOC region and / or battery temperature region.

  Next, a control processing procedure for parameter change rate learning in the secondary battery evaluation system according to the embodiment of the present invention will be described.

  Referring to FIG. 8 again, ECU 100 executes the parameter change rates gr and gd described above in parallel with steps S100 to S190. Then, ECU 100 determines in step S200 whether a learning condition for the parameter change rate is satisfied.

DC resistance R _a is preferably estimated in a region where the battery voltage to the battery current changes linearly. Therefore, when the charging / discharging conditions of the secondary battery 10 are satisfied, the learning conditions can be set so that step S200 is determined as YES for learning of the parameter change rate gr.

The diffusion coefficient D _s indicates the degree of diffusion of lithium in the active material after the once generated battery current becomes zero. Therefore, after the secondary battery 10 is charged / discharged, the diffusion coefficient is calculated based on the change in the battery voltage after the battery current becomes zero. Therefore, the learning condition can be set so that step S200 is determined as YES for learning of the parameter change rate gd when the battery current = 0 period continues to some extent.

Thus, the parameter change rate gr of DC resistance R _a, the parameter change rate gd of the diffusion coefficient D _s, it is preferred to learn independently. That, also the processing of steps S200, S210, are performed separately for each of the DC resistance R _a and the diffusion coefficient D _s.

  When the learning condition is satisfied (when YES is determined in S200), ECU 100 updates the map value stored in change rate map 141 or 142 for the parameter change rate for which the learning condition is satisfied in step S210. Thereby, at least the parameter change rate learning value in the learning region corresponding to the SOC (local SOC) when the learning condition is satisfied and the battery temperature is updated.

  On the other hand, ECU 100 skips the process of step S210 when the learning condition is not satisfied (NO in S200). Therefore, the map value of the parameter change rate stored in the change rate map 141 or 142 is maintained.

As described above, in the secondary battery deterioration evaluation system according to the present embodiment, when the secondary battery 10 is used, the predetermined parameters (typically, the DC resistance _Ra and the diffusion coefficient D _s ) in the battery model are: It is assumed that the change (parameter change rate) with respect to the initial state value is estimated online. Then, based on the estimated parameter change rate, the change rate maps 141 and 142 shown in FIGS. 9 and 10 are used for each predetermined learning category defined by a combination of SOC (local SOC) and battery temperature. The parameter change rate is learned.

(Embodiment 1)
Next, the deterioration evaluation index in the secondary battery deterioration evaluation system according to the first embodiment will be described in detail.

  FIG. 12 is a schematic functional block diagram for explaining the function of deterioration evaluation unit 200 (FIG. 2) according to the first embodiment.

  The deterioration evaluation unit 200 does not directly use the parameter change rates (learned values) stored in the change rate maps 141 and 142, but based on a virtual IV test using the battery model 125, the learning region is determined. The IV resistance is calculated independently as a section (that is, for each learning region).

13 and 14 are waveform diagrams for explaining the IV test.
FIG. 13 is a typical current waveform diagram in the IV test. Referring to FIG. 13, before time ta, the battery current is 0 and the secondary battery is in a relaxed state. A constant battery current is output from the battery during a period from time ta to time tb. The battery current at this time is, for example, 1C, and the period from time ta to time tb is, for example, 10 seconds. 1C is a current value when charging or discharging the entire capacity of the battery in one hour.

  The battery current is zero during the period from time tb to time tc. This period is a predetermined time as a time for relaxing the battery, for example.

  During the period from time tc to time td, the secondary battery is charged with a constant battery current. The battery current at this time is, for example, 1C, and the period from time tc to time td is, for example, 10 seconds. During the period from time td to time te, the battery current becomes zero. This period is, for example, a predetermined time as a time for relaxing the secondary battery.

  In the IV test, at least one of discharging and charging is performed based on a predetermined pattern. For example, by measuring the behavior of the battery voltage when the secondary battery 10 is charged / discharged according to the charge / discharge pattern of FIG.

  FIG. 14 is a waveform diagram of the battery voltage during the IV test. FIG. 14 shows a waveform diagram of a period corresponding to the discharge of the charge / discharge shown in FIG.

  Referring to FIG. 14, at times ta to tb, a current is output from the secondary battery (battery current = I0). While the battery is output from the secondary battery, the battery voltage decreases. At time tb, the output of current is stopped. Thereby, the battery voltage is recovered.

  From the voltage behavior (voltage waveform) at this time, the voltage change ΔV during the IV test is obtained. Therefore, the total internal resistance (IV resistance) during the IV test can be obtained by dividing ΔV by I0. Similarly, the IV resistance can be obtained from the voltage behavior even during charging (time tc to td) in the IV test.

  In the IV test, only one of charging and discharging may be performed, and both charging and discharging may be performed. When both charging and discharging are performed, the IV resistance value according to the IV test is obtained from the maximum value of the IV resistance during charging and the IV resistance during charging, or the average value of both.

Referring again to FIG. 12 stored in the change rate map 141, which is sequentially updated at the time of use of the secondary battery 10, online learning values of the parameters change rate gr of DC resistance R _a (grl) is, for each learning area Has been. Similarly, the change rate map 142 stores an online learning value (gdl) of the parameter change rate gd of the diffusion coefficient D _s for each learning region.

  In the secondary battery deterioration evaluation system according to the present embodiment, the IV test is executed in a simulated manner using the battery model, not the actual IV test. Hereinafter, the above-described IV test on software without actual charge / discharge is also referred to as a “virtual IV test”.

The ECU 100 (degradation evaluation unit 200) calculates the IV resistance by executing the virtual IV test using the battery model for each learning region as described above. At this time, the value of the diffusion coefficient D _s in DC resistance R _a and the formula in the formula of the battery model (1) (3) is online learning values of the parameters rate of change grl, is calculated based on gdl. That is, the DC resistance _Ra is determined according to the product of the map value (R _an ) of the initial state value map 131 and the map value of the change rate map 141 in each learning region (R _a = R _an · grl). Similarly, the diffusion coefficient D _s is determined according to the product of the map value (D _sn ) of the initial state value map 132 and the map value of the change rate map 142 (D _s = D _sn · gdl) in each learning region. .

  In this way, the IV resistance value RIV by the virtual IV test can be obtained based on the parameter change rate learned online for each learning region when the secondary battery 10 is used.

  The IV resistance value RIV obtained by the virtual IV test is stored in the storage map 201 for each learning area. That is, the storage map 201 is configured to store a map value (a test value in the virtual IV test) for each learning area common to the change rate maps 141 and 142.

  ECU 100 (degradation evaluation unit 200) further includes a reference value map 205 for storing a reference value of IV resistance. The reference value map 205 stores an initial value RIVn of IV resistance as a map value for each learning region common to the change rate maps 141 and 142. The initial value RIVn in each learning region can be obtained in advance based on, for example, a virtual IV test when the parameter change rate is set to gr = gd = 1.0. Alternatively, the initial value RIVn can be obtained by executing an IV test using an actual machine in various SOC regions and temperature regions in advance.

  The arithmetic unit 202, for each learning region, IV resistance indicating a comparison result between the IV resistance value RIV (virtual test value) stored in the storage map and the reference value (initial value RIVn) stored in the reference value map 205. The change rate Kr is calculated. The IV resistance change rate Kr corresponds to a “comparison evaluation value”.

  In the example of FIG. 12, Kr = RIV / RIVn. That is, as the deterioration progresses and the internal resistance increases, the IV resistance change rate Kr increases.

  The IV resistance change rate Kr is calculated for each learning region and stored in the storage map 210. That is, the storage map 210 is configured to store the IV resistance change rate Kr as a map value for each learning region common to the change rate maps 141 and 142.

  FIG. 15 shows a flowchart for explaining the control process for each learning area described above by the deterioration evaluation unit 200.

  Referring to FIG. 15, ECU 100 determines a learning region in which the virtual IV test is executed in step S300. In step S200, one of all the learning areas is sequentially selected. Thereby, the target learning area is determined.

  In step S310, ECU 100 reads online change values grl and gdl of parameter change rates gr and gd from change rate maps 141 and 142 for the target learning region determined in step S300.

Further, ECU 100, in step S320, parameter change rate learning value read out in step S310 grl, gdl and initial value R _an,, from D _sn, DC resistance R _a and in the battery model in a virtual I-V test The diffusion coefficient D _s is determined.

  In step S330, ECU 100 executes a virtual IV test using the battery model. Hereinafter, in the virtual IV test, the battery current according to the current waveform in FIG. 14 is input to the battery model, and the voltage waveform (battery voltage) in FIG. 14 is obtained as the output of the battery model.

  FIG. 16 shows a flowchart for explaining the control processing procedure in step S330 (virtual IV test) in more detail.

Referring to FIG. 16, ECU 100, when the virtual I-V test is started, the step S510, the DC resistance R _a and the diffusion coefficient D _s of formula (3) of the formula of the battery model (1), step The value determined in S320 is read. Further, ECU 100 manages time data t in the virtual IV test in step S520. Specifically, t = 0 is set at the start of the virtual IV test. On the other hand, when the virtual IV test is executed, the time data t is incremented by Δt each time the following steps S530 to S560 are executed.

  In step S530, ECU 100 inputs battery current I in time data t determined in step S520. The battery current I is determined based on a current waveform in a predetermined IV test shown in FIG.

  In step S540, ECU 100 determines the boundary condition of the diffusion equation by substituting battery current I determined in step S530 into equation (6) and reflecting it in equation (5). Further, in step S542, ECU 100 calculates a lithium concentration distribution according to equation (3) in accordance with the boundary condition determined in step S540.

  In step S546, ECU 100 calculates a local SOC according to equation (2) according to the lithium concentration distribution calculated in step S542. In step S548, ECU 100 calculates the term of open-circuit voltage OCV in equation (1) based on the local SOC calculated in step S546 by the same processing as step S130 in FIG.

Further, ECU 100, in step S550, OCV calculated by the step S548, the battery current I and determined in the step S530, using a DC resistance R _a determined in step S510, the battery voltage V in accordance with equation (1) (Simulation value) is calculated. ECU 100 stores battery current I and battery voltage V at current time data t.

  Further, in step S570, ECU 100 determines whether or not current time data t has reached predetermined value tfin corresponding to the end of the virtual IV test. The predetermined value tfin corresponds to, for example, time tb in FIG.

  Until time data t reaches tfin (NO determination in step S570), ECU 100 repeatedly executes the processes of steps S530 to S570 while increasing time data t by Δt in step S520. When the time data t reaches tfin (YES in S570), the ECU 100 proceeds to step S580 and ends the IV virtual test of the target learning area.

  By the series of processes shown in FIG. 16, it is understood that the behavior of the battery voltage V can be simulated for the target learning region by using the battery current I according to the current waveform of FIG.

  Referring to FIG. 15 again, ECU 100 calculates IV resistance value RIV in the learning region based on the calculation result of the battery model in the virtual IV test in step S340. For example, the IV resistance value RIV can be calculated according to ΔV / I0 by obtaining the voltage change amount ΔV in FIG. 14 from the behavior (voltage waveform) of the battery voltage V obtained by the virtual IV test.

  In step S350, ECU 100 acquires initial value RIVn of the IV resistance of the learning region from reference value map 205. Further, ECU 100 calculates IV resistance change rate Kr (Kr = RIV / RIVn) in the learning region and stores it in storage map 210 in step S360.

  Through the series of processes shown in FIG. 15, the IV resistance is obtained based on the result of the virtual IV test in each learning region, and the comparison result with the reference value of IV resistance (the initial value RIVn in FIG. 15) is calculated. can do.

  Referring to FIG. 12 again, the deterioration index calculation unit 250 completes the virtual IV test and the calculation of the IV resistance change rate Kr based on the test result in all the learning regions, and stores all the data stored in the storage map 210. The maximum value of the IV resistance change rate Kr in the learning region is output as the deterioration index value Pdg. In addition, about the virtual IV test for every learning area | region, a process can also be deform | transformed so that execution in a part of learning area | region may be abbreviate | omitted previously.

  By adopting such a configuration, in the secondary battery deterioration evaluation system according to the first embodiment, the virtual IV test for each learning region when the parameter change rate is learned online during use of the secondary battery. By evaluating the total internal resistance (IV resistance) based on the results of the above, it is possible to manage the deterioration of the secondary battery.

Therefore, the rate of change of the DC resistance _Ra and the diffusion coefficient D _s is estimated with high accuracy for each learning region, and the influence of the DC resistance _Ra and the diffusion coefficient D _{s on} the total internal resistance depending on the SOC and the battery temperature. Therefore, it is possible to appropriately manage and evaluate the increase in the internal resistance of the secondary battery.

(Modification of Embodiment 1)
FIG. 17 is a block diagram illustrating the function of deterioration evaluation unit 200 according to a modification of the first embodiment.

  Compared with FIG. 17 in FIG. 12, in the modification according to the first embodiment, the reference value map 205 stores not the initial IV resistance RIVn but the allowable maximum value RIVmax for each learning region. Different from Form 1. The allowable maximum value RIVmax of the IV resistance in each learning region can be determined in advance based on the required output of the load 50 corresponding to the SOC and battery temperature regions.

  On the other hand, the calculation of the total internal resistance (IV resistance) in each learning region is the same as that in the first embodiment, and therefore detailed description will not be repeated.

  The calculation unit 202 calculates the comparison result between the IV resistance value RIV (storage map) obtained in the virtual IV test and the reference value (allowable maximum value RIVmax) stored in the reference value map 205 for each learning region. An IV resistance deterioration rate Kr # shown is calculated. In the example of FIG. 17, Kr # = RIV / RIVmax is indicated. That is, the IV resistance deterioration rate Kr # increases as the deterioration progresses and the internal resistance increases. In FIG. 17, the IV resistance deterioration rate Kr # corresponds to the “comparison evaluation value”.

  FIG. 18 is a flowchart for illustrating control processing for each learning region according to the modification of the first embodiment.

  FIG. 18 is different from FIG. 15 in that the modification of the first embodiment executes steps S350 # and S360 # in place of steps S350 and S360 shown in FIG. Since the processes in steps S300 to S340 for calculating the IV resistance value are the same as those in the first embodiment (FIG. 15), detailed description will not be repeated.

  In step S350 #, ECU 100 obtains allowable maximum value RIVmax of IV resistance in the learning region from reference value map 205. Further, ECU 100 calculates IV resistance deterioration rate Kr # (Kr # = RIV / RIVmax) in the learning region and stores it in storage map 210 in step S360 #.

  Referring to FIG. 17 again, when the calculation of the virtual IV test and the IV resistance deterioration rate Kr # based on the test result is completed in all learning regions, deterioration index calculation unit 250 is stored in storage map 210. The maximum value of the IV resistance deterioration rates Kr # in all the learning regions is output as the deterioration index value Pdg.

  As described above, according to the modification of the first embodiment, as in the first embodiment, the result of the virtual IV test for each learning region when the parameter change rate is learned online while using the secondary battery. By evaluating the total internal resistance (IV resistance) based on the above, deterioration of the secondary battery can be managed.

(Embodiment 2)
As described in the first embodiment, the secondary battery deterioration evaluation system according to the present embodiment is premised on learning the parameter change rate of the battery model when the secondary battery is used. Therefore, in Embodiment 2, the deterioration of the secondary battery is evaluated by reflecting the number of learning in each learning area in online learning.

  In the secondary battery deterioration management system according to the second embodiment, when the secondary battery 10 is used, a learning frequency map shown in FIGS. 19 and 20 is further provided to count the learning frequency in each learning region. .

Referring to FIG. 19, change rate storage section 140 shown in FIG. 2 further includes _a learning frequency map 143 related to DC resistance Ra that forms a pair with change rate map 141 shown in FIG. 9. The learning frequency map 143 stores the learning frequency Ngr of the parameter change rate gr for each learning area that is divided in common with the change rate map 141. The learning number Ngr is counted up every time the map value of the learning area is updated.

Referring to FIG. 20, the change rate storage unit 140 illustrated in FIG. 2 further includes a learning frequency map 144 regarding the diffusion coefficient D _s that forms a pair with the change rate map 142 illustrated in FIG. 10. The learning number map 144 stores the learning number Ngd of the parameter change rate gd for each learning region that is divided in common with the change rate map 142. The learning count Ngd is counted up every time the map value of the learning area is updated. The learning frequency maps 143 and 144 correspond to a “fourth storage unit”.

  FIG. 21 is a flowchart illustrating a control process related to learning count management in the secondary battery deterioration management system according to the second embodiment.

  Referring to FIG. 21, in the second embodiment, step S210 executed when the learning condition is satisfied includes steps S212 and S214.

  In step S212, ECU 100 updates the map value of change rate map 141 and / or 142 for the parameter for which the learning condition is satisfied in step S200. Thereby, the online learning value (grl and / or gdl) of the parameter change rate gr and / or gd is updated.

  Further, ECU 100 counts up the number of learnings stored in learning number maps 143 and 144 for the learning region in which the map values of change rate maps 141 and 142 are updated in step S214.

Each time a parameter change rate gr learning condition of DC resistance R _a is satisfied, the number of times of learning Ngr the corresponding learning region increases in the number of times of learning maps 143. Similarly, every time the learning condition for the parameter change rate gd of the diffusion coefficient D _s is satisfied, the learning number Ngd of the corresponding learning region in the learning number map 144 increases. In this way, it is possible to manage the number of learnings in online learning when the secondary battery 10 is used.

  As shown in FIG. 11, when the parameter change rate is learned also in the adjacent learning areas AR2 to AR8, the increase in the number of learning times in the learning areas AR2 to AR8 is larger than the increase in the learning area AR1. It may be small.

  FIG. 22 is a flowchart for illustrating processing by deterioration index calculation unit 250 (FIG. 12) in the deterioration management system for a secondary battery according to the second embodiment.

  Referring to FIG. 22, in the second embodiment, ECU 100 implements the function of deterioration index calculation unit 250 by executing the processes of steps S410 to S430.

  In step S410, ECU 100 reads learning times Ngr and Ngd for all learning areas from learning frequency maps 143 and 144. In step S420, ECU 100 determines at least one representative region based on the number of learning times Ngr, Ngd for each learning region.

  For example, the maximum number of learning times Ngr and Ngd can be determined as the representative region. In addition, for a predetermined one of the learning times Ngr and Ngd, the learning region having the maximum learning frequency can be determined as the representative region. Alternatively, a plurality of representative regions may be determined based on the number of learning times. In step S420, the representative area can be determined based on any predetermined rule. Basically, the representative area is determined so as to exclude the learning area where the learning frequency is small.

  In step S430, ECU 100 reads from the storage map 210 the IV resistance data (a general term for the IV resistance change rate Kr and the IV resistance deterioration rate Kr #) of at least one representative region determined in step S420. Then, a deterioration index value Pdg is set based on the read IV resistance data. For example, the degradation index value Pdg can be set according to the maximum value among the IV resistance data of at least one representative region.

  According to the secondary battery degradation evaluation system according to the second embodiment, it is possible to evaluate the IV resistance of the secondary battery by eliminating the data in the learning area where the number of times of learning was small during use of the secondary battery 10. . Accordingly, it is possible to prevent the deterioration state of the internal resistance of the secondary battery from being erroneously evaluated based on the learning result in the learning region where the number of learning times is small. Further, in the learning region where the IV resistance is not evaluated and the learning frequency is small, it is possible to omit the execution of the virtual IV test.

(Modification of Embodiment 2)
In the modification of the second embodiment, another example of calculating the deterioration index value reflecting the number of learning in each learning area will be described. That is, in the modification of the second embodiment, the function of the deterioration index calculation unit 250 is different from that in the second embodiment. Since other parts are the same as those in the second embodiment, detailed description will not be repeated.

  FIG. 23 is a flowchart for illustrating control processing by deterioration index calculation unit 250 (FIG. 12) in the secondary battery deterioration evaluation system according to the modification of the second embodiment.

  Referring to FIG. 23, ECU 100 reads learning times Ngr and Ngd for each learning area from learning number maps 143 and 144 in step S410 similar to FIG. Further, in step S440, ECU 100 calculates a weighted average value of IV resistance data in all learning regions, in which the number of learnings in each learning region is weighted.

  For example, the weighted average value IVave of the IV resistance data is expressed by the following equation (19) using the IV resistance data K (map value of the storage map 210) and the learning count Ngr (map value of the learning count map 143) for each learning region. Can be asked according to. In Equation (19), “Σ” represents the sum total over the entire learning region.

IVave = Σ (K · Ngr) / Σ (Ngr) (19)
In step S450, ECU 100 determines deterioration index value Pdg according to the weighted average value IVave of the IV resistance data obtained in step S440.

  According to the secondary battery deterioration evaluation system according to the modification of the second embodiment, it is possible to evaluate the IV resistance through all the learning regions in consideration of the learning frequency for each learning region. Therefore, it is possible to evaluate the deterioration of the internal resistance of the secondary battery by appropriately considering the learning frequency.

(Embodiment 3)
In the third embodiment, a modification of the management of the number of learning times used in the second embodiment and the modification is shown. That is, in the third embodiment, the management of the number of learning (FIG. 21) is different from that in the second embodiment and its modification. Since other parts are the same as those in the second embodiment or the modification thereof, detailed description will not be repeated.

Referring to FIG. 24, in the third embodiment, ECU 110 further executes step S216.
The ECU 100 updates the online learning values grl and gdl of the parameter change rate gr and / or gd for the parameter for which the learning condition is satisfied in step S200 in step S212 similar to FIG. Furthermore, the learning frequency (Ngd and / or Ngd) stored in the learning frequency maps 143 and 144 is counted up for the learning region in which the map values of the change rate maps 141 and 142 are updated by the same step S214 as in FIG. The

  In step S216, the ECU 100 executes a process of reducing the map value of the number of learnings as time passes. For example, every time a predetermined time elapses, the map value of each learning area of the learning frequency maps 143 and 144 is subtracted by a predetermined value. Alternatively, by managing the elapsed time from when the previous learning condition is established for each learning area, the map value of the number of learning times in the learning area may be subtracted by a predetermined value each time the predetermined time elapses.

  That is, in step S216, a reduction process with the passage of time is added to the map values of the learning number maps 143 and 144 so that the number of learnings in the past does not affect with the passage of time. As a result, even if the total number of learning times is the same, the map value in the learning region where the recent number of learning times is large can be relatively increased.

  In the third embodiment, the deterioration index value Pdgd is calculated using the number of learning times Ngr and Ngd managed according to FIG. 24 in the second embodiment or its modification.

  In this case, when the usage condition (typically, battery temperature) of the secondary battery 10 changes greatly, such as the area where the vehicle equipped with the secondary battery 10 is used has changed from a low temperature area to a high temperature area. In addition, the deterioration of the internal resistance of the secondary battery can be evaluated by appropriately reflecting the recent use conditions of the secondary battery 10.

(Embodiment 4)
In the first embodiment, a configuration has been described in which both the DC resistance _Ra and the diffusion coefficient D _s , which are parameters used in the battery model equation, change over time, and both parameters are learned online. However, depending on the type of the secondary battery, there is mainly one in which only one of the DC resistance _Ra and the diffusion coefficient D _s changes over time.

In such a secondary battery, the other of the DC resistance _Ra and the diffusion coefficient D _s does not change greatly from the initial state value even when the deterioration progresses. Therefore, considering the calculation load of the ECU 100, for only one of the DC resistance R _a and the diffusion coefficient D _s, it is preferable to online learning parameter rate of change. In the fourth embodiment, a deterioration management system for such a type of secondary battery will be described. First, the aging of the DC resistance R _a, the deterioration evaluating secondary batteries small type negligibly aged deterioration of the diffusion coefficient D _s is described.

  FIG. 25 is a schematic functional block diagram for explaining the function of the deterioration evaluation unit according to the fourth embodiment.

Referring to FIG. 25, the deterioration evaluating system for a secondary battery according to the fourth embodiment, only the parameters change rate gr of DC resistance R _a is online learning. The online learning value grl of the parameter change rate gr is stored for each learning region in the change rate map 141 as in the first embodiment. On the other hand, estimation and online learning are not executed for the parameter change rate gd of the diffusion coefficient D _s . That is, gd = 1.0 is fixed in the entire learning region. Thereby, the calculation load of ECU100 is reduced.

  Further, a limit value map 206 for storing the limit value grmax of the parameter change rate gr is provided in advance. The limit value map 206 stores a limit value grmax for each learning region common to the change rate map 141. The limit value grmax can be obtained in advance for each learning region by a virtual IV test using a battery model. The limit value map 206 corresponds to “third storage unit”.

  FIG. 26 is a flowchart illustrating a control process for obtaining a limit value of the parameter change rate of the DC resistance by a virtual IV test. The limit value calculation process described below can be executed in advance prior to the use of the secondary battery 10. Alternatively, even after the secondary battery 10 is mounted on the power supply system, the calculation process can be executed by the ECU 100 using the idle time when the load 50 is not operated.

In step S610, ECU 100 reads the SOC (local SOC) and battery temperature Tb in the learning region for calculating the parameter change rate limit value and the value of diffusion coefficient D _s used in equation (3) of the battery model. Since the parameter change rate gd is fixed at 1.0, the map value of the initial state value map 132 can be used as it is as the diffusion coefficient D _s .

In step S620, ECU 100 sets a candidate value for parameter change rate gr. Then, ECU 100, in step S630, the DC resistance R _a in the battery model, and candidate values of gr, which is set in in the step S620, the determined according the product of the initial state parameter value R _an,.

ECU100, in step S640, executes a virtual I-V tested using the battery model in this way DC resistance R _a and the diffusion coefficient D _s have been determined. The virtual IV test can be executed by processing similar to steps S520 to S580 in FIG.

  In step S650, the ECU 100 performs the learning region similar to the case where the IV resistance value based on the behavior of the battery voltage obtained by the virtual IV test in step S640 is stored in the reference value map 205 in FIG. Is compared with the maximum allowable value RIVmax.

  Further, ECU 100 proceeds to step S660 to determine whether or not the search for parameter change rate gr has been completed. A predetermined pattern can be provided in advance for the search for the parameter change rate gr.

  The ECU 100 repeatedly executes the processes in steps S630 and S640 while switching the parameter change rate gr candidate values in step S620 until the search for the parameter change rate gr is completed (when NO is determined in S660). As a result, at the end of the search for the parameter change rate gr (when YES is determined in S660), IV resistance values based on the virtual IV test are obtained for a plurality of candidate values.

  When the search for parameter change rate gr is completed (when YES is determined in S660), ECU 100 proceeds to step S670 to calculate limit value grmax of parameter change rate gr corresponding to allowable maximum value RIVmax of IV resistance.

  For example, among the plurality of candidate values of the parameter change rate gr, the limit of the parameter change rate gr is obtained by complementing between two candidate values whose IV resistance value based on the virtual IV test is closest to the allowable maximum value RIVmax. The value grmax is determined. The obtained limit value grmax of the parameter change rate gr is stored as a map value of the corresponding learning region in the limit value map 206 of FIG.

  The limit value map 206 shown in FIG. 25 can be created by executing the control process shown in FIG. 26 for each learning region.

  Referring to FIG. 25 again, in the secondary battery deterioration management system according to the fourth embodiment, deterioration evaluation unit 200 uses parameter calculation unit 202 to store parameter change rate gr stored in change rate map 142 for each learning region. The DC resistance deterioration rate Kgr indicating the comparison result between the online learning value grl and the limit value grmax stored in the limit value map 206 is calculated. That is, it is represented by Kgr = gr / grmax. As the deterioration progresses and the internal resistance increases, the DC resistance deterioration rate Kgd increases. In FIG. 25, the DC resistance deterioration rate Kgr corresponds to the “comparative evaluation value”.

  The DC resistance deterioration rate Kgr calculated by the calculation unit 202 is stored in the storage map 212 for each learning region. That is, the storage map 212 is configured to store the DC resistance deterioration rate Kgr as a map value for each learning region common to the change rate map 141.

  When the calculation of the DC resistance deterioration rate Kgr is completed in the entire learning region, the deterioration index calculation unit 250 calculates the deterioration index value Pdg of the secondary battery 10 based on the DC resistance deterioration rate Kgr stored in the storage map 212. . For example, as in the first embodiment, the maximum value of the DC resistance deterioration rates Kgr of all the learning regions stored in the storage map 212 can be output as the deterioration index value Pdg.

  Alternatively, the DC resistance deterioration rate Kgr stored in the storage map 212 is reflected in accordance with the second embodiment, the modification thereof, or the third embodiment, reflecting the learning frequency Ngr (learning frequency map 143) of the parameter change rate gr. It is also possible to calculate the deterioration index value Pdg based on the above.

Thus, according to the secondary battery deterioration management system according to the fourth embodiment, in a secondary battery of a type in which only _a specific parameter (DC resistance R _a ) in the battery model changes mainly over time, There is no need to conduct a virtual IV test after online learning of parameter change rates. This is because the limit value of the parameter change rate corresponding to the allowable maximum value RIVmax of the IV resistance can be obtained by the virtual IV test even before the secondary battery 10 is used.

  Therefore, the secondary battery deterioration management system according to the fourth embodiment realizes the same secondary battery deterioration management as in the first to third embodiments and the modifications thereof while reducing the calculation load of the ECU 100. Can do.

(Modification of Embodiment 4)
In the modification of the fourth embodiment, as opposed to the fourth embodiment, with respect to aging of the diffusion coefficient D _s, deterioration evaluation of the secondary battery of small type enough to aging of DC resistance R _a is negligible Will be described.

  FIG. 27 is a schematic functional block diagram for explaining the function of the deterioration evaluation unit according to the modification of the fourth embodiment.

With reference to FIG. 27 and FIG. 26, in the secondary battery deterioration evaluation system according to the modification of the fourth embodiment, only the parameter change rate gd of the diffusion coefficient D _s is learned online. The online learning value of the parameter change rate gd is stored for each learning region in the change rate map 142 as in the first embodiment. Meanwhile, for the parameter change rate gr of DC resistance R _a, estimation and online learning is not performed. That is, gr = 1.0 is fixed in the entire learning region. Thereby, the calculation load of ECU100 is reduced.

When the diffusion coefficient D _s decreases, the internal resistance (IV resistance) increases. Therefore, the limit value map 206 stores the limit value gdmin of the parameter change rate gd for each learning region. As in the fourth embodiment, the limit value gdmin in each learning region can be obtained in advance by a virtual IV test using a battery model.

  FIG. 28 is a flowchart for explaining a control process for obtaining a limit value of a parameter change rate of a diffusion coefficient by a virtual IV test. The processing shown in FIG. 28 can also be executed at the same opportunity as the control processing according to the flowchart shown in FIG.

ECU100, in step S610♯, parameter change rate SOC of the learning region for calculating a limit value (local SOC) and battery temperature Tb and reads the value of DC resistance R _a used in equation battery model (1). Since it is fixed to the parameter rate of change gr = 1.0, the map value of the initial state value map 131, can be directly used as a DC resistance R _a.

In step S620 #, ECU 100 sets a candidate value for parameter change rate gd. Then, ECU 100, in step S630♯, the diffusion coefficient D _s in the battery model, and candidate values gd set in step S620♯, determined according to the product of the initial state parameter value D _an,.

ECU100, in step S640♯, executes the virtual I-V tested using this battery model the DC resistance R _a and the diffusion coefficient D _s is determined so. The virtual IV test can be executed by processing similar to steps S520 to S580 in FIG.

  In step S650 #, ECU 100 executes the same IV resistance value based on the behavior of the battery voltage obtained by the virtual IV test in step S640 # as that stored in reference value map 205 in FIG. The maximum allowable value RIVmax in the learning area is compared.

  Further, ECU 100 proceeds to step S660 # to determine whether or not the search for parameter change rate gd has been completed. A predetermined pattern can be provided in advance for the search for the parameter change rate gd.

  ECU 100 repeatedly executes the processes in steps S630 # and S640 # while switching the candidate value of parameter change rate gd in step S620 # until the search for parameter change rate gd is completed (when NO is determined in S660 #). . As a result, when the search for parameter change rate gd is completed (YES in S660 #), IV resistance values based on the virtual IV test are obtained for a plurality of candidate values.

  When the search for parameter change rate gd is completed (when YES in S660 #), ECU 100 proceeds to step S670 # to limit value gdmin of parameter change rate gd corresponding to allowable maximum value RIVmax of IV resistance. Is calculated.

  For example, among the plurality of candidate values of the parameter change rate gd, the IV resistance value based on the virtual IV test is complemented between two candidate values closest to the allowable maximum value RIVmax, thereby limiting the parameter change rate gd. The value gdmin is determined. The obtained limit value grmax of the parameter change rate gr is stored as a map value of the corresponding learning region in the limit value map 206 of FIG.

  The limit value map 206 shown in FIG. 27 can be created by executing the control process shown in FIG. 28 for each learning region.

Referring to FIG. 27 again, in the deterioration management system for a secondary battery according to the modification of the fourth embodiment, deterioration evaluation unit 200 uses parameters stored in change rate map 141 for each learning region by calculation unit 202. A diffusion coefficient deterioration rate Kgd indicating a comparison result between the online learning value gdl of the change rate gd and the limit value gdmin stored in the limit value map 206 is calculated. That is, Kgd = gdmin / gd. Since the diffusion coefficient D _s decreases as the deterioration progresses, the diffusion coefficient deterioration rate Kgd increases as the deterioration progresses and the internal resistance increases. In FIG. 25, the diffusion coefficient deterioration rate Kgd corresponds to the “comparison evaluation value”.

  The diffusion coefficient deterioration rate Kgd calculated by the calculation unit 202 is stored in the storage map 212 for each learning region. That is, the storage map 212 is configured to store the diffusion coefficient deterioration rate Kgd as a map value for each learning area common to the change rate map 142.

  When the calculation of the diffusion coefficient deterioration rate Kgd is completed in the entire learning region, the deterioration index calculation unit 250 calculates the deterioration index value Pdg of the secondary battery 10 based on the diffusion coefficient deterioration rate Kgd stored in the storage map 212. . For example, as in the first embodiment, the maximum value of the diffusion coefficient deterioration rates Kgd of all the learning regions stored in the storage map 212 can be output as the deterioration index value Pdg.

  Alternatively, the diffusion coefficient deterioration rate Kgd stored in the storage map 212 reflects the learning frequency Ngd (learning frequency map 144) of the parameter change rate gd in accordance with the second embodiment, the modification thereof, or the third embodiment. It is also possible to calculate the deterioration index value Pdg based on the above.

Thus, according to the secondary battery deterioration management system according to the modification of the fourth embodiment, for a secondary battery of a type in which only the diffusion coefficient D _s in the battery model changes mainly over time. Therefore, the same deterioration management as that in the fourth embodiment can be realized.

  As described above, according to the fourth embodiment and the modification thereof, in the secondary battery of a type in which only the specific parameter in the battery model mainly changes over time, the calculation load of the ECU 100 is reduced, The deterioration of the secondary battery can be managed based on the result of the virtual IV test for each learning region.

  In the embodiment described above, the secondary battery is described as a lithium ion battery. However, the secondary battery deterioration management system according to the present invention is applicable to other secondary batteries other than lithium ion batteries. Can be applied without any particular limitation. For example, in a nickel metal hydride battery, the method of the present invention is similarly applied by calculating the concentration distribution of protons as a reaction participating substance inside the active material by a diffusion equation and defining the open-circuit voltage as a function of protons on the active material surface. It becomes possible to do. Further, with respect to other types of secondary batteries, the same deterioration management can be performed if the rate of change from the parameter value in the initial state is estimated for a predetermined parameter of the same battery model formula. .

  The embodiment disclosed this time should be considered as illustrative in all points and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims. The embodiment disclosed this time should be considered as illustrative in all points and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

  The present invention can be applied to deterioration management of secondary batteries.

10 secondary battery, 12 negative electrode, 13, 16 current collector, 14 separator, 15 positive electrode, 18 active material, 20 current sensor, 30 voltage sensor, 40 temperature sensor, 50 load, 60 load control device, 110 data collection unit, 120 Battery state estimation unit, 125 battery model, 130 initial value storage unit, 131, 132 initial state value map, 140 change rate storage unit, 141, 142 change rate map, 143, 144 learning frequency map, 150 parameter change rate learning unit, 200 Degradation evaluation unit, 201, 210 Storage map, 202 Calculation unit, 205 Reference value map, 206 Limit value map, 250 Degradation index calculation unit, AR1 to AR8 learning region, D _sn initial state value (diffusion coefficient), D _s diffusion Coefficient, I, Ib battery current, Kgr DC resistance deterioration rate, Kgd diffusion coefficient deterioration rate, Kr IV resistance resistance Deterioration rate, Kr♯ IV resistance change rate, NGD, Ngr number of times of learning, Pdg degradation index value, R _a DC resistance, R _an, initial value (DC resistance), RIV IV resistance value (virtual IV Test Results), RIvMax Allowable maximum value (IV resistance), RIVn initial value (IV resistance), T, Tb battery temperature, U ₁ positive electrode open potential, U ₂ negative electrode open potential, V, Vb battery voltage, gd parameter change rate (diffusion coefficient), gdl Parameter change rate learning value (diffusion coefficient), gdmin, grmax Parameter change rate limit value, gr parameter change rate (DC resistance).

Claims (11)

Detection means for detecting the battery voltage, battery current and battery temperature of the secondary battery;
Battery state estimation means for sequentially estimating the state quantity according to a battery model for estimating the state quantity of the secondary battery including at least SOC using data including at least one of the battery voltage and the battery current; ,
For a predetermined parameter of the parameter group used in the battery model, first storage means for storing an initial state value in advance for each predetermined learning region defined by a combination of the SOC and the battery temperature;
A parameter change rate, which is a ratio of the current parameter value to the initial state value of the predetermined parameter, is sequentially determined by parameter identification based on the data detected by the detection means and the battery model during use of the secondary battery. A parameter change rate learning means for learning;
Second storage means for storing the parameter change rate learning value calculated by the parameter change rate learning means for each learning region;
Based on the result of a virtual test that is performed for each learning region and that determines the behavior of the secondary battery using the battery model when the secondary battery performs at least one of charging and discharging according to a predetermined test pattern. Te, Bei example a deterioration evaluation means for evaluating the degree of deterioration of the secondary battery,
The deterioration evaluation means includes
Based on a result of the virtual test using the battery model in which a parameter value calculated based on the parameter change rate learning value stored in the second storage unit is substituted for the predetermined parameter, the secondary battery A secondary battery degradation management system that evaluates the degradation level of batteries. The deterioration evaluation means includes
For executing the virtual test using the battery model in which the parameter value calculated based on the parameter change rate learning value stored in the second storage unit is substituted for the predetermined parameter for each learning region Means of
Means for calculating an internal resistance value of the secondary battery in the predetermined test pattern based on the result of the virtual test for each learning region;
Based on a comparison for each learning area between a reference value of the internal resistance value preset for each learning area and the internal resistance value for each learning area calculated based on the result of the virtual test Te, wherein and means for calculating an index value indicating the degree of degradation of the secondary battery, according to claim 1 Symbol placement of the rechargeable battery deterioration management system. Detection means for detecting the battery voltage, battery current and battery temperature of the secondary battery;
Battery state estimation means for sequentially estimating the state quantity according to a battery model for estimating the state quantity of the secondary battery including at least SOC using data including at least one of the battery voltage and the battery current; ,
For a predetermined parameter of the parameter group used in the battery model, first storage means for storing an initial state value in advance for each predetermined learning region defined by a combination of the SOC and the battery temperature;
A parameter change rate, which is a ratio of the current parameter value to the initial state value of the predetermined parameter, is sequentially determined by parameter identification based on the data detected by the detection means and the battery model during use of the secondary battery. A parameter change rate learning means for learning;
Second storage means for storing the parameter change rate learning value calculated by the parameter change rate learning means for each learning region;
Based on the result of a virtual test that is performed for each learning region and that determines the behavior of the secondary battery using the battery model when the secondary battery performs at least one of charging and discharging according to a predetermined test pattern. And a deterioration evaluation means for evaluating the deterioration degree of the secondary battery,
The deterioration evaluation means includes
The learning of a limit value related to the parameter change rate of the predetermined parameter obtained based on the result of the virtual test for each learning region and the parameter change rate learned value stored in the second storage unit A secondary battery deterioration management system that evaluates a deterioration degree of the secondary battery based on a comparison for each region. The plant Teipa parameter includes a plurality of parameters,
The deterioration evaluation means includes
For each learning region, while fixing the values of the other parameters except the first parameter among the plurality of parameters to the initial state value, the battery model is changed by changing the first parameter, respectively. Means for performing the used virtual test multiple times;
Based on the results of the plurality of virtual tests, the parameter change rate limit value of the first parameter corresponding to the internal resistance value of the secondary battery in the predetermined test pattern is an allowable maximum value. Means for determining each learning area as a limit value;
Based on the comparison of the parameter change rate learned value stored in the second storage means for the first parameter and the parameter change rate limit value for each learning region, the secondary battery and means for calculating an index value indicating the degree of deterioration, the deterioration management system for a secondary battery according to claim 3 Symbol placement. The deterioration evaluation means includes
A third storage means for storing the parameter change rate limit value for each learning area obtained in advance by the plurality of virtual tests before using the secondary battery;
5. The secondary battery deterioration management system according to claim 4 , wherein, in the calculation of the index value, the parameter change rate limit value for each learning region is read from the third storage unit. The deterioration evaluation means includes
Means for calculating a comparative evaluation value indicating a result of comparison for each learning region;
Wherein one of said comparison evaluation value for each learning region, said containing and index value calculating means for calculating the index value according to the maximum value indicating that the internal resistance value is highest, according to claim 2 or 4 Secondary battery deterioration management system. The parameter change rate learning means updates the parameter change rate learning value of the learning region at that time each time a predetermined learning condition is satisfied,
The deterioration management system includes:
A fourth storage means for storing, for each learning area, the number of times the parameter change rate learning means has updated the parameter change rate learning value;
The deterioration evaluation means includes
Means for calculating a comparative evaluation value indicating a result of comparison for each learning region;
Or means for calculating the index value according to the comparative evaluation value in the learning area selected based on the number of learning times for each learning area stored in the fourth storage means. 4. A secondary battery deterioration management system according to 4. The parameter change rate learning means updates the parameter change rate learning value of the learning region at that time each time a predetermined learning condition is satisfied,
The deterioration management system includes:
A fourth storage means for storing, for each learning area, the number of times the parameter change rate learning means has updated the parameter change rate learning value;
The deterioration evaluation means includes
Means for calculating a comparative evaluation value indicating a result of comparison for each learning region;
Means for calculating the index value according to an average value weighted by the number of learning times for each learning area stored in the fourth storage means for the comparative evaluation value for each learning area. Item 5. The secondary battery deterioration management system according to Item 2 or 4 . The secondary according to claim 7 or 8 , wherein the parameter change rate learning means further includes means for reducing the number of learnings for each learning region stored in the fourth storage means as time elapses. Battery deterioration management system. The battery model is
A voltage equation represented by the voltage variation due to charge and discharge current in the interior of the open voltage is a function of the concentration of the reaction-participating material in the surface of the active material inside the secondary battery, and the secondary battery,
A diffusion equation defining a distribution of the concentration of the reaction-participating substance inside the active material,
The secondary battery deterioration management system according to any one of claims 1 to 9 , wherein the predetermined parameter includes a parameter indicating a direct current resistance in the voltage equation. The battery model is
A voltage equation represented by the voltage variation due to a charge-discharge current within the function in which the open circuit voltage of the concentration of the reaction-participating material in the surface of the active material inside the secondary battery, and the secondary battery,
A diffusion equation defining a distribution of the concentration of the reaction-participating substance inside the active material,
Wherein the predetermined parameter, the at diffusion equation including diffusion parameter representing the rate of diffusion of the reaction involved material, a secondary battery deterioration management system according to any one of claims 1-9.

Images