JP7501384B2 - Battery Deterioration Prediction System - Google Patents

Description

The present invention relates to a battery deterioration prediction system.

Patent Document 1 discloses a system for predicting the degree of deterioration of a secondary battery. The system described in Patent Document 1 predicts the degree of deterioration of a secondary battery using a parameter that is thought to be correlated with the amount of impurities adhering to the electrode surface of the secondary battery.

JP 2013-181875 A

For example, in secondary batteries through which a large current flows, such as secondary batteries for in-vehicle use, electrode polarization is likely to become evident. In the system described in Patent Document 1, the emergence of such polarization is not taken into consideration, making it difficult to improve the accuracy of predicting deterioration of secondary batteries.

The present invention was made in consideration of these problems, and aims to provide a battery deterioration prediction system that can improve the accuracy of predicting deterioration of secondary batteries.

According to one aspect of the present invention, there is provided a secondary battery (21) including an electrode resistance acquisition unit (107) for acquiring a negative electrode resistance (R _a ) and a positive electrode resistance (R _c );
A current value acquisition unit (102) that acquires a current value (I) of the secondary battery;
an OCP acquisition unit (106) that acquires an open circuit potential (OCP _a ) of a negative electrode of the secondary battery and an open circuit potential (OCP _c ) of a positive electrode of the secondary battery;
a polarization calculation unit (108) that calculates a negative electrode polarization (ΔV _a ) and a positive electrode polarization (ΔV _c ) from the negative electrode resistance and the positive electrode resistance acquired by the electrode resistance acquisition unit and the current value of the secondary battery acquired by the current value acquisition unit;
a CCP calculation unit (109) that calculates a closed circuit potential (CCP _a ) of the negative electrode based on the open circuit potential of the negative electrode acquired by the OCP acquisition unit and the negative electrode polarization calculated by the polarization calculation unit, and calculates a closed circuit potential (CCP _c ) of the positive electrode based on the open circuit potential of the positive electrode acquired by the OCP acquisition unit and the positive electrode polarization calculated by the polarization calculation unit;
The battery deterioration prediction system (1) includes a capacity prediction unit (110) that predicts the negative electrode capacity (Q _a ), positive electrode capacity (Q _c ), and positive/negative electrode SOC deviation capacity (Q _Li ) of the secondary battery based on at least one of the negative electrode closed circuit potential and the positive electrode closed circuit potential calculated in the CCP calculation unit, and predicts the battery capacity (Q _B ) of the secondary battery based on the predicted negative electrode capacity, positive electrode capacity, and positive/negative electrode SOC deviation capacity.

The battery degradation prediction system of the above aspect includes a CCP calculation unit that calculates the closed circuit potential of the negative electrode based on the open circuit potential and negative electrode polarization of the negative electrode, and calculates the closed circuit potential of the positive electrode based on the open circuit potential and positive electrode polarization of the positive electrode. The capacity prediction unit predicts the negative electrode capacity, positive electrode capacity, and positive/negative electrode SOC deviation capacity of the secondary battery based on this closed circuit potential, and predicts the battery capacity of the secondary battery based on the predicted negative electrode capacity, positive electrode capacity, and positive/negative electrode SOC deviation capacity. In this way, by predicting the battery capacity of the secondary battery while taking into account the polarization of the positive electrode and the polarization of the negative electrode, the battery capacity can be predicted with high accuracy.

As described above, according to the above aspect, it is possible to provide a battery deterioration prediction system that can predict the degree of deterioration of a secondary battery with high accuracy.
In addition, the symbols in parentheses described in the claims and the means for solving the problems indicate a correspondence with the specific means described in the embodiments described below, and do not limit the technical scope of the present invention.

1 is a conceptual diagram of a battery deterioration prediction system and a power supply system including the same in a first embodiment. 1 is a functional block diagram showing the configuration of a battery deterioration prediction system according to a first embodiment. 4 is a flowchart for explaining a process in which the battery deterioration prediction system predicts a battery capacity Q _B in the first embodiment. FIG. 1A is a diagram showing a schematic diagram of the relationship between the open circuit voltage and the charging rate of a secondary battery before deterioration, and the relationship between the closed circuit voltage and the charging rate, in embodiment 1; and FIG. 1B is a diagram showing a schematic diagram of the relationship between the open circuit voltage and the charging rate of a secondary battery after deterioration, and the relationship between the closed circuit voltage and the charging rate, in embodiment 1. 5 is a graph showing how the first battery and the second battery deteriorate in the first embodiment. 4 is a flowchart for illustrating a process in which the battery deterioration prediction system in the first embodiment controls a current flowing through the secondary battery by using a negative electrode capacity Q _a , a positive electrode capacity Q _c , and a positive/negative electrode SOC deviation capacity Q _Li of the secondary battery. 5 is a schematic diagram showing an updated OCP characteristic and an updated OCV characteristic in the first embodiment. 5 is a schematic diagram showing an updated OCV characteristic and a differential characteristic in the first embodiment. 13 is a flowchart for explaining a flow of updating a prediction formula in the second embodiment. FIG. 11A is a conceptual diagram showing a relationship between a resistance difference obtained from a first usage history and a capacity retention rate, based on voltage behavior at the end of charging in a second embodiment; and FIG. 11B is a conceptual diagram showing a relationship between a resistance difference obtained from a second usage history and a capacity retention rate, based on voltage behavior at the end of charging in a second embodiment. FIG. 13A is a conceptual diagram showing a relationship between the voltage change and the capacity retention rate obtained from a first usage history, based on the voltage relaxation behavior after charging is suspended in variant 1; and FIG. 13B is a conceptual diagram showing a relationship between the voltage change and the capacity retention rate obtained from a second usage history. 13A and 13B are conceptual diagrams showing a relationship between the AC impedance at a specific frequency obtained from a first usage history and the capacity maintenance rate in a modified embodiment 2; and FIG. 13C and 13D are conceptual diagrams showing a relationship between the AC impedance at a specific frequency obtained from a second usage history and the capacity maintenance rate in a modified embodiment 2. 13A and 13B are conceptual diagrams showing a relationship between the AC impedance at a specific frequency obtained from a first usage history and the capacity maintenance rate in variant embodiment 3; and FIG. 13C and 13D are conceptual diagrams showing a relationship between the AC impedance at a specific frequency obtained from a second usage history and the capacity maintenance rate in variant embodiment 3. 13A, 13B, and 13C are conceptual diagrams showing a relationship between the AC impedance at a specific frequency obtained from the first usage history and the capacity maintenance rate in the fourth modified embodiment. FIG. 11 is a functional block diagram showing the configuration of a battery deterioration prediction system in a third embodiment. FIG. 13 is a functional block diagram showing the configuration of a battery deterioration prediction system in a fourth embodiment. FIG. 13 is a conceptual diagram showing a battery deterioration prediction system and a vehicle in a fifth embodiment. FIG. 13 is a functional block diagram showing the configuration of a battery deterioration prediction system in a sixth embodiment. 23 is a flowchart for explaining an estimation process in a data non-acquired period in a battery deterioration prediction system according to a sixth embodiment. 20A to 20E are conceptual diagrams showing changes in (a) ignition, (b) temperature, (c) current value, (d) voltage value, and (e) SOC during a period in which data has not been acquired in the sixth embodiment.

(Embodiment 1)
An embodiment of a battery deterioration prediction system will be described with reference to FIGS. 1 to 8. FIG.
The battery deterioration prediction system 1 of this embodiment is a system that predicts deterioration of a secondary battery 21. In this embodiment, the secondary battery 21 is mounted on a vehicle such as an electric vehicle or a hybrid vehicle and used.

As shown in FIG. 1, the secondary battery 21 is connected to an inverter 22 and a charging device 23. The inverter 22 converts the DC power supplied from the secondary battery 21 into AC power and outputs the AC power to a three-phase AC motor (not shown).

As shown in Fig. 1, the secondary battery 21 includes a plurality of battery cells 211 connected in series to each other. Each battery cell 211 is, for example, a lithium ion secondary battery. The negative electrode of the secondary battery 21 is made of a negative electrode active material capable of absorbing and releasing lithium ions, such as graphite. The positive electrode of the secondary battery 21 can be a ternary electrode containing Ni, Mn, and Co, such as LiNi1 _{/ 3Co1} /3Mn1 _{/ 3O2} . The secondary battery 21 may be configured by connecting the plurality of battery cells 211 in parallel to each other to form a cell block, and connecting a plurality of such cell blocks in series to each other.

As shown in FIG. 1, a discharge switch 24 is provided between the secondary battery 21 and the inverter 22. A charge switch 25 is provided between the secondary battery 21 and the charging device 23. When power is supplied from the secondary battery 21 to the inverter 22, the discharge switch 24 is turned on, and when the secondary battery 21 is charged, the charge switch 25 is turned on. The on/off operation of the charge switch 25 and the discharge switch 24 is controlled by a battery management unit (i.e., BMU) installed in the vehicle.

As shown in FIG. 1, a current sensor 26 that measures the current value of the secondary battery 21 is connected to the secondary battery 21. A temperature sensor 27 that measures the temperature of the secondary battery 21 is installed in the secondary battery 21. Information about the secondary battery 21 acquired by the current sensor 26 and the temperature sensor 27 is transmitted to the battery deterioration prediction system 1.

2, the battery deterioration prediction system 1 includes a temperature acquisition unit 101, a current value acquisition unit 102, a timer 103, an SOC calculation unit 104, a change amount calculation unit 105, an OCP acquisition unit 106, an electrode resistance acquisition unit 107, a polarization calculation unit 108, a CCP calculation unit 109, a capacity prediction unit 110, a learning unit 111, an OCV characteristic calculation unit 112, a differential characteristic calculation unit 113, and a current control unit 114. These are configured, for example, by a BMU mounted on the vehicle. Each of these units includes a volatile memory unit that functions as a calculation area, a non-volatile memory unit that stores various programs, and a processor. In other words, each of these units is configured to be able to execute a program that performs its own function. At least one of these units may be configured by an electronic circuit (i.e., hardware) for performing its own function.

Next, steps in which the battery deterioration prediction system 1 predicts the battery capacity Q _B will be described with reference to FIGS.

First, in step S1, the timer 103 acquires the elapsed time t from when the secondary battery 21 started to be used, the current value acquisition unit 102 acquires the current value I of the secondary battery 21, and the temperature acquisition unit 101 acquires the temperature T of the secondary battery 21.

The usage time of the secondary battery 21 measured by the timer 103 is sent to the memory unit 115 and stored. The memory unit 115 is, for example, a rewritable non-volatile memory. The battery deterioration prediction system 1 continuously predicts the battery capacity _QB , and the start time of one prediction operation is _ts , the end time is _te , and the time from the start time _ts to the end time _te is called an execution cycle. If the execution cycle is a short period such as 1 sec, the accuracy of the capacity prediction is likely to improve, but the calculation load increases. On the other hand, if the execution cycle is a long period, it is difficult to improve the accuracy of the capacity prediction, but the calculation load is reduced. Taking such a viewpoint into consideration, the execution cycle can be appropriately determined.

The current value acquisition unit 102 periodically receives information on the current value of the secondary battery 21 measured by the current sensor 26, and stores the received information on the current value of the secondary battery 21 in the memory unit 115. The current value acquisition unit 102 calculates the current value I from the distribution of the current values of the secondary battery 21 acquired during the execution cycle. The current value I can be, for example, an average value calculated from the frequency distribution of the current values of the secondary battery 21 acquired during the execution cycle. Note that, in order to reduce the calculation load, it is also possible to adopt, as the current value I, the average value of the current values of the secondary battery 21 acquired during the execution cycle.

The temperature acquisition unit 101 periodically receives information on the temperature of the secondary battery 21 measured by the temperature sensor 27, and stores the received temperature information of the secondary battery 21 in the memory unit 115. The temperature T of the secondary battery 21 is calculated from the distribution of the temperature of the secondary battery 21 acquired during the execution cycle. The temperature T can be, for example, an average value calculated from the frequency distribution of the temperature of the secondary battery 21 acquired during the execution cycle. Note that, in order to reduce the calculation load, it is also possible to adopt, as the temperature T, the average value of the temperature of the secondary battery 21 acquired during the execution cycle, etc.

In step S2, the SOC calculation unit 104 calculates an integrated value of the current value of the secondary battery 21 acquired during the execution cycle, and calculates the charging rate of the secondary battery 21 based on the integrated value. The SOC calculation unit 104 calculates the charging rate of the secondary battery 21 by a known method using the current value of the secondary battery 21 acquired by the current value acquisition unit 102. The charging rate of the secondary battery 21 indicates the charging state of the secondary battery 21, and is expressed as a percentage as the ratio of the remaining capacity to the full charge capacity of the secondary battery 21. Hereinafter, the charging rate of the secondary battery 21 is referred to as SOC. SOC is an abbreviation for State Of Charge. The SOC calculation unit 104 can calculate the SOC of the secondary battery 21 based on, for example, the integrated value of the current value of the secondary battery 21 acquired by the current value acquisition unit 102. Information on the SOC calculated by the SOC calculation unit 104 is stored in the memory unit 115.

In step S3, the change amount calculation unit 105 calculates ΔDOD. The change amount calculation unit 105 can obtain ΔDOD from the difference between the SOC at the start time _ts and the end time _te of the execution cycle. Note that DOD is an abbreviation for Depth of discharge, which indicates the depth of discharge of the secondary battery 21.

In step S4, the electrode resistance acquisition unit 107 calculates the negative electrode resistance R _a and the positive electrode resistance R _c of the secondary battery 21. The electrode resistance acquisition unit 107 calculates the negative electrode resistance R a and the positive electrode resistance R _c of the secondary battery 21 based on the temperature T of the secondary battery 21, the current value I of the secondary battery 21, the change amount ΔDOD of the SOC, and the closed circuit potential of the negative electrode or the closed circuit potential of the positive _electrode of the secondary battery 21. Here, the temperature T is the temperature T of the secondary battery 21 acquired by the temperature acquisition unit 101 in step S1. The current value I is the current value I of the secondary battery 21 calculated in step S121. The change amount ΔDOD is the change amount ΔDOD of the SOC calculated by the change amount calculation unit 105 in step S3. The closed circuit potential of the negative electrode and the closed circuit potential of the positive electrode of the secondary battery 21 are the closed circuit potentials of the negative electrode and the positive electrode of the secondary battery 21 acquired by the CCP calculation unit 109 in the previous execution cycle. Hereinafter, the closed circuit potential of the negative electrode of the secondary battery 21 will be referred to as CCP _a , and the closed circuit potential of the positive electrode of the secondary battery 21 will be referred to as CCP _c . CCP is an abbreviation for Closed Circuit Potential. The negative electrode resistance R _a can be expressed as a function of the temperature T of the secondary battery 21, CCP _a , the amount of change ΔDOD, and the charge/discharge current value I, and the positive electrode resistance R _c can be expressed as a function of the temperature T of the secondary battery 21, CCP _c , the amount of change ΔDOD, and the charge/discharge current value I. This will be described below.

First, the negative electrode resistance R _a will be described.
The negative electrode resistance R _a increases due to the formation of a coating (SEI: Solid Electrolyte Interface) on the negative electrode surface due to oxidation-reduction decomposition of the electrolyte of the secondary battery 21 and additives to the electrolyte. Since the coating is produced by the above-mentioned chemical reaction, the negative electrode resistance R _a follows the Arrhenius law. Therefore, the negative electrode resistance R _a can be expressed as a function of temperature T.

In addition, the formation of a film on the negative electrode surface is due to oxidation-reduction and follows Tafel's law. Therefore, the negative electrode resistance R _a can be expressed as a function of the negative electrode potential CCP _a .

In addition, repeated charge/discharge cycles of the secondary battery 21 cause repeated expansion and contraction of the active material of the negative electrode, which causes cracks in the surface coating, and eventually exposes the negative electrode surface through the cracks in the coating. The amount of coating increases as new coating is formed on this exposed surface, causing a further increase in the negative electrode resistance R _a . The larger ΔDOD is, the greater the degree of expansion and contraction of the active material is. Therefore, the negative electrode resistance R _a can be expressed as a function of ΔDOD.

In addition, in the negative electrode, the active material itself cracks and the diameter becomes smaller due to repeated expansion and contraction of the active material as described above. The cracking of the active material itself combines an element that reduces the negative electrode resistance R _a and an element that increases the negative electrode resistance R _a . First, due to the cracking of the active material itself, a new surface (i.e., a surface on which no coating is formed) is formed on the active material, so that the reaction area increases. Therefore, the cracking of the active material itself is a factor in the decrease of the negative electrode resistance R _a . On the other hand, when a new surface is formed on the active material, the formation of the coating is promoted on the new surface, so that the amount of coating increases and the negative electrode resistance R _a increases. In consideration of the above, the negative electrode resistance R _a can be expressed by a function of ΔDOD from the theory shown below.

The pulverization rate, which is the rate at which the active material of the negative electrode breaks, is expressed as dr/dt, where r is the particle diameter of the active material and t is time. Here, it is considered that the pulverization rate dr/dt progresses more easily as the particle diameter r of the active material increases. In other words, the pulverization rate dr/dt can be considered to be proportional to the particle diameter r of the active material. Therefore, the pulverization rate can be expressed as the following formula (1).

In the above formula (1), k is a constant, which is hereafter also referred to as the pulverization constant. When this is solved, the following formula (2) is obtained.

In the above formula (2), a is a constant.
Furthermore, since the degree of expansion and contraction of the active material increases as the ΔDOD increases, the pulverization constant is considered to be proportional to the ΔDOD.

In the above formula (3), β and γ are constants. When this is solved, the following formula (4) is obtained.

In the above formula (4), η and ζ are constants. By coupling the above formula (2) and formula (4), the following formula (5) can be derived.

Here, r ₀ is the radius of the active material at the initial stage (i.e., when t = 0), and A, B, and C are constants. As described above, the negative electrode resistance R _a increases due to the formation of a coating on the negative electrode surface, and the formation rate of the coating on the negative electrode surface has a correlation with the diameter of the active material of the negative electrode, so the negative electrode resistance R _a can be expressed by an equation including the pulverization function f (t, ΔDOD), that is, a function of ΔDOD. Note that the values in the parentheses on the right side of the formula (5) may be further corrected by adding a constant. In addition, in this embodiment, the constants A, B, and C can be corrected by the learning unit 111 described later.

In addition, the above-mentioned cracking of the surface coating of the negative electrode and the cracking of the negative electrode active material itself also depend on the charge/discharge current of the secondary battery 21. That is, the larger the charge/discharge current value, the more the current tends to flow intensively through the low resistance parts of the active material, so the degree of expansion and contraction may differ depending on the part of the active material. This makes it easier for distortion to occur in the active material, causing cracking of the surface coating of the negative electrode and cracking of the negative electrode active material itself. Therefore, cracking of the surface coating of the negative electrode and cracking of the negative electrode active material itself can be expressed as a function of the charge/discharge current value I, or a function of the C rate that is correlated with the charge/discharge current value I. Here, the 1C rate indicates the current value that fully charges or fully discharges the rated capacity of the battery in one hour in the case of constant current charge/discharge measurement.

To summarize the above, the negative electrode resistance R _a can be expressed as the following equation (7) using a function g _A (T, CCP _a ) that takes into account the formation of a coating on the surface of the active material, a function g _B (T, CCP _a , ΔDOD, I) that takes into account the cracking of the coating formed on the surface of the active material, and a function g _C (T, CCP _a , ΔDOD, I) that takes into account the cracking of the active material itself.

Based on the above theory, the negative electrode resistance R _a is expressed as a function of the temperature T of the secondary battery 21, CCP _a , the amount of change ΔDOD, and the charge/discharge current value I.

Next, the positive electrode resistance _Rc will be described.
The positive electrode resistance _Rc increases with the deterioration of the positive electrode surface. Since the positive electrode surface is deteriorated by a chemical reaction, the positive electrode resistance _Rc follows the Arrhenius law. Therefore, the positive electrode resistance _Rc can be expressed as a function of temperature T.

In addition, the deterioration of the positive electrode surface is due to the reductive decomposition of the positive electrode surface, and therefore follows Tafel's law. Therefore, the positive electrode resistance _Rc can be expressed as a function of _CCPc .

In addition, the positive electrode active material expands and contracts repeatedly due to repeated charge and discharge cycles of the secondary battery 21, causing cracks on the surface of the altered positive electrode active material, and forming a new positive electrode surface that is not altered. This new positive electrode surface eventually undergoes alteration, causing a further increase in the positive electrode resistance _Rc . The larger ΔDOD is, the greater the degree of expansion and contraction of the active material is. Therefore, the positive electrode resistance _Rc can be expressed as a function of ΔDOD.

In addition, the deterioration of the positive electrode surface is promoted by repeated expansion and contraction of the active material of the positive electrode, which causes cracks in the active material of the positive electrode to progress and the diameter of the active material to become smaller. The cracks in the active material itself combine a factor that reduces the positive electrode resistance R _c and a factor that increases the positive electrode resistance R _c . First, the cracks in the active material itself form a new surface (i.e., the surface before the deterioration) on the active material, so that the cracks in the active material itself are a factor in the decrease of the positive electrode resistance R _c . On the other hand, when a new surface is formed on the active material, the new surface will eventually deteriorate and the positive electrode resistance R _c will increase. In consideration of the above, the positive electrode resistance R _c can be expressed by an equation including the pulverization function f (t, ΔDOD) of the above equation (6) based on the same theory as the negative electrode resistance R _a , that is, a function of ΔDOD.

The cracking of the positive electrode active material itself also depends on the charge/discharge current I. That is, the larger the charge/discharge current value I, the more the current tends to flow intensively through the low resistance parts of the active material, which can cause differences in the degree of expansion and contraction depending on the part of the active material. This makes it easier for distortion to occur in the active material, causing the positive electrode active material itself to crack. Therefore, the cracking of the positive electrode active material itself can be expressed as a function of the charge/discharge current value I, or a function of the C rate, which is correlated with the charge/discharge current value I.

To summarize the above, the positive electrode resistance _Rc can be expressed as the following equation (8) using a function _hA (T, _CCPa ) that takes into account the deterioration of the surface of the active material, a function _hB (T, _CCPa ,ΔDOD,I) that takes into account the cracking of the deteriorated surface of the active material, and a function _hC (T, _CCPa ,ΔDOD,I) that takes into account the cracking of the active material itself.

Based on the above theory, the positive electrode resistance R _c is expressed as a function of the temperature T of the secondary battery 21, CCP _c , the amount of change ΔDOD, and the charge/discharge current value I.

Here, the CCP _a and CCP _c used in step S4 are those calculated by the CCP calculation unit 109 in step S7 (described later) in the implementation cycle immediately preceding the current implementation cycle. If there is no CCP _a and CCP _c obtained in the previous implementation cycle (for example, at the time of system startup, etc.), the initial CCP _{a and CCP c are} calculated as follows.

First, the initial negative electrode polarization ΔV _a is calculated from the product of the current value I acquired in step S1 and the initial value of the negative electrode resistance R _a described later, and the initial positive electrode polarization ΔV _c is calculated from the product of the current value I acquired in step S1 and the initial value of the positive electrode resistance R _c . The initial value of the negative electrode resistance R _a and the initial value of the positive electrode resistance R _c are, for example, the negative electrode resistance and the positive electrode resistance acquired in the secondary battery 21 in the initial state (for example, the state of the secondary battery 21 at the time of shipment from the factory) of the same type as the secondary battery 21 of this embodiment, and are stored in the memory unit 115 in advance. The negative electrode resistance R _a and the positive electrode resistance R _c in the initial state can be determined, for example, by an AC impedance method or IV measurement. Alternatively, the initial state negative electrode resistance R _a and the positive electrode resistance R _c can also be determined by dismantling the secondary battery 21 in the initial state, creating a half cell using the positive electrode and a half cell using the negative electrode, and measuring the resistance of each half cell.

Then, based on the initial OCP characteristic stored in the storage unit 115 and the SOC calculated in step S2, the open circuit potential of the negative electrode of the secondary battery 21 corresponding to the SOC and the open circuit potential of the positive electrode of the secondary battery 21 corresponding to the SOC are obtained. Each open circuit potential is the potential of each electrode of the secondary battery 21 when a long period of time has passed without current passing between the secondary battery 21 and the external circuit. Hereinafter, the open circuit potential of the negative electrode of the secondary battery 21 is referred to as OCP _a , and the open circuit potential of the positive electrode of the secondary battery 21 is referred to as OCP _c . OCP is an abbreviation for Open Circuit Potential. The initial OCP characteristic indicates the relationship between the SOC and OCP _a , and the relationship between the SOC and OCP _c , of the secondary battery 21 in the initial state, and is stored in advance in the storage unit 115. Next, CCP _a is obtained by calculating OCP _a + ΔV _a , and CCP _c is obtained by calculating OCP _c + ΔV _c .

As described above, when there is no CCP _a and CCP _c obtained in the previous implementation cycle (when there is no previous implementation cycle, such as at system startup), initial CCP _a and CCP _c are calculated.

In step S5, the polarization calculation unit 108 calculates the negative electrode polarization ΔV _a = I × R _a and calculates the positive electrode polarization ΔV _c = I × R _c , where I is the current value I of the secondary battery 21 acquired by the current value acquisition unit 102 in step S1, and R _a and R _c are the negative electrode resistance R _a and the positive electrode resistance R _c calculated by the electrode resistance acquisition unit 107 in step S4, respectively.

In step S6, the OCP acquisition unit 106 calculates OCP _a and OCP _c . The OCP acquisition unit 106 acquires OCP _a and OCP _c based on the SOC of the secondary battery 21 calculated by the SOC calculation unit 104 in step S2 and the updated OCP characteristics calculated in step S10 (described later) of the previous execution cycle stored in the storage unit 115. The updated OCP characteristics indicate the relationship between the SOC and OCP _a of the secondary battery 21 after deterioration and the relationship between the SOC and OCP _c . Details of how to acquire the updated OCP characteristics will be described in step S10 (described later). The information on OCP _a and OCP _c acquired by the OCP acquisition unit 106 is stored in the storage unit 115.

In step S7, the CCP calculation unit 109 calculates CCP _a and CCP _c of the secondary battery 21. The CCP calculation unit 109 acquires ΔV _a and ΔV _c calculated by the polarization calculation unit 108 in step S5, and acquires OCP _a and OCP _c acquired by the OCP acquisition unit 106 in step S6. The CCP calculation unit 109 then rewrites OCP _a to CCP _a using OCP _a + ΔV _a , and rewrites OCP _c to CCP _c using OCP _c + ΔV _c .

Here, the polarization of the secondary battery 21 becomes apparent due to deterioration. That is, due to the occurrence of polarization, the closed circuit voltage of the secondary battery 21 rises when the secondary battery 21 is charged, and the closed circuit voltage drops when the secondary battery 21 is discharged. However, as deterioration progresses, the closed circuit voltage of the secondary battery 21 rises further when the secondary battery 21 is charged, and the closed circuit voltage drops further when the secondary battery 21 is discharged. For example, FIG. 4(a) shows a schematic diagram of the relationship between the SOC and voltage of the secondary battery 21 before deterioration when it is charged, and FIG. 4(b) shows a schematic diagram of the relationship between the SOC and voltage of the secondary battery 21 after deterioration when it is charged. In FIG. 4(a) and FIG. 4(b), the open circuit voltage is represented by a solid line, and the closed circuit voltage is represented by a dashed line. It is assumed that the voltage scale on the vertical axis is the same in FIG. 4(a) and FIG. 4(b). In the following, the open circuit voltage is referred to as OCV, and the closed circuit voltage is referred to as CCV. OCV stands for Open Circuit Voltage, and CCV stands for Closed Circuit Voltage.

4(a) and 4(b), it can be seen that the polarization ΔV of the secondary battery after deterioration is larger than that before deterioration. In view of this, the battery deterioration prediction system 1 of this embodiment rewrites the OCP in predicting deterioration of the secondary battery into a CCP that takes into account the polarization ΔV, and predicts the battery capacity _QB using this CCP.

2 and 3, in step S8, the capacity prediction unit 110 calculates the negative electrode capacity _Qa , the positive electrode capacity _Qc , and the positive and negative electrode SOC deviation capacity _QLi of the secondary battery 21. The capacity prediction unit 110 acquires the CCP _a and CCP _c calculated by the CCP calculation unit 109 in step S7, the temperature T of the secondary battery 21 acquired by the temperature acquisition unit 101 in step S1, and the ΔDOD calculated by the change amount calculation unit 105 in step S3. The capacity prediction unit 110 calculates the negative electrode capacity _Qa , the positive electrode capacity _Qc , and the positive and negative electrode SOC deviation capacity _QLi of the secondary battery 21 based on at least one of the CCP _a and CCP _c , the temperature T of the secondary battery 21, the current value I of the secondary battery 21, and the ΔDOD. The negative electrode capacity _Qa corresponds to the number of sites of the negative electrode into which lithium ions can be inserted. The positive electrode capacity _Qc corresponds to the number of sites of the positive electrode into which lithium ions can be inserted. The positive and negative electrode SOC deviation capacity _QLi is the deviation of the usable capacity ranges of the positive electrode and the negative electrode in the secondary battery 21. The positive and negative electrode SOC deviation capacity _QLi corresponds to the number of lithium ions that can move between the positive electrode and the negative electrode, and the ease of movement of all the lithium ions.

The capacity prediction unit 110 expresses the negative electrode capacity _Qa by the same theory as that used when the electrode resistance acquisition unit 107 calculates the negative electrode resistance R _a . That is, the negative electrode capacity _Qa can be expressed as the following formula (9) using a function i _A (T, CCP _a ) that takes into account the formation of a coating on the surface of the active material, a function i _B (T, CCP _a , ΔDOD, I) that takes into account the cracking of the coating formed on the surface of the active material, and a function i _C (T, CCP _a , ΔDOD, I) that takes into account the cracking of the active material itself. That is, the negative electrode capacity Qa is expressed by a function of the temperature T, CCP _a , the change amount ΔDOD (i.e., the pulverization function f(t, ΔDOD)) of the secondary battery ₂₁ , and the charge/discharge current value I.

In addition, the capacity prediction unit 110 expresses the positive electrode capacity _Qc by the same theory as that in the case of calculating the positive electrode resistance _Rc in the electrode resistance acquisition unit 107. That is, the positive electrode capacity _Qc can be expressed as the following formula (10) using a function _jA (T, _CCPc ) that takes into consideration the deterioration of the surface of the active material, a function _jB (T, _CCPc , ΔDOD, I) that takes into consideration the cracking of the deteriorated surface of the active material, and a function _jC (T, _CCPc , ΔDOD, I) that takes into consideration the cracking of the active material itself. That is, the positive electrode capacity Qc is expressed by a function of the temperature T of the secondary battery 21, _CCPc , the amount of change ΔDOD (i.e., the pulverization function f(t, ΔDOD)), and the charge/discharge current value _I.

The positive and negative electrode SOC deviation capacity Q _Li correlates with the consumption of lithium ions due to the formation of a coating (SEI: Solid Electrolyte Interface) at the negative and positive electrodes. The consumption of lithium ions is a chemical reaction, and the positive and negative electrode SOC deviation capacity Q _Li follows the Arrhenius law. Therefore, the positive and negative electrode SOC deviation capacity Q _Li can be expressed as a function of temperature T.

The consumption of lithium ions due to the formation of films on the negative and positive electrodes is an oxidation-reduction reaction, and therefore follows Tafel's law. Therefore, the positive and negative electrode SOC shift capacity Q _Li can be expressed as a function of CCP _a and CCP _c .

In addition, the repetition of charge and discharge cycles of the secondary battery 21 causes repeated expansion and contraction of the active material of each electrode (i.e., negative electrode, positive electrode), and cracks in the surface coating of the active material of each electrode progress. As a result, the surface of each electrode is eventually exposed through the cracks in the coating. The consumption of lithium ions increases as a new coating is formed on this exposed surface. The larger ΔDOD is, the greater the degree of expansion and contraction of the active material is. Therefore, the positive and negative electrode SOC deviation capacity Q _Li can be expressed as a function of ΔDOD.

In addition, in each electrode, the active material itself is cracked and the diameter becomes smaller due to repeated expansion and contraction of the active material as described above. The cracking of the active material itself has both an element that increases the positive and negative electrode SOC deviation capacity Q _LI and an element that decreases the positive and negative electrode SOC deviation capacity Q _LI . First, due to the cracking of the active material itself, a new surface (i.e., a surface on which a film is not formed) is formed on the active material, so that lithium ions are easily moved to the active material of each electrode, which is a factor of increasing the positive and negative electrode SOC deviation capacity Q _Li . On the other hand, when a new surface is formed on the active material, the formation of a film is promoted on the new surface, and lithium ions are consumed, which is a factor of decreasing the positive and negative electrode SOC deviation capacity Q _Li . In consideration of the above, the positive and negative electrode SOC deviation capacity Q _Li can be expressed by an equation including a pulverization function f (t, ΔDOD), that is, a function of ΔDOD, from the same theory as the negative electrode resistance R _a and the positive electrode resistance R _c .

In addition, the cracking of the active material itself of each electrode described above also depends on the charge/discharge current I. That is, the larger the charge/discharge current value I, the more the current tends to flow concentratedly through the low resistance parts of the active material, which can cause differences in the degree of expansion and contraction depending on the part of the active material. This makes it easier for distortion to occur in the active material, causing the active material itself to crack. Therefore, the cracking of the active material itself of each electrode can be expressed as a function of the charge/discharge current value I, or a function of the C rate, which is correlated with the charge/discharge current value I.

To summarize the above, the positive and negative electrode SOC deviation capacity Q _Li can be expressed as the following equation (11) using function k _A (T, CCP _a ) that takes into account the formation of a coating on the surface of the negative electrode active material, function k _B (T, CCP _a , ΔDOD, I) that takes into account the cracking of the coating formed on the surface of the negative electrode active material, and function k _C (T, CCP _a , ΔDOD, I) that takes into account the cracking of the negative electrode active material itself, as well as function l _A (T, CCP _c ) that takes into account the formation of a coating on the surface of the positive electrode active material, function l _B (T, CCP _c , ΔDOD, I) that takes into account the cracking of the coating formed on the surface of the positive electrode active material, and function l _C (T, CCP _c , ΔDOD, I) that takes into account the cracking of the positive electrode active material itself.

As described above, the positive and negative electrode SOC deviation capacity Q _Li can be expressed as a function of the temperature T, CCP _a , CCP _c , the amount of change ΔDOD, and the charge/discharge current value I of the secondary battery 21 .

In step S9, the capacity prediction unit 110 obtains the battery capacity _QB by _QB = min ( _Qa , _Qc , _QLi ). That is, the capacity prediction unit 110 judges the smallest one of the negative electrode capacity _Qa , the positive electrode capacity _Qc , and the positive and negative electrode SOC deviation capacity _QLi of the secondary battery 21 as the battery capacity (i.e., full charge capacity) of the secondary battery 21. As described above, the negative electrode capacity _Qa corresponds to the number of sites of the negative electrode into which lithium ions can be inserted, and the positive electrode capacity _Qc corresponds to the number of sites of the positive electrode into which lithium ions can be inserted. The positive and negative electrode SOC deviation capacity _QLi corresponds to the number of lithium ions that can move between the positive electrode and the negative electrode. Therefore, the smallest one of the negative electrode capacity _Qa , the positive electrode capacity _Qc , and the positive and negative electrode SOC deviation capacity _QLi corresponds to the battery capacity _QB of the secondary battery 21.

Here, a secondary battery 21 through which a large current flows, such as that used for driving an automobile, is often used only in a region where the positive and negative electrode SOC deviation capacity Q _Li is the smallest among the negative electrode capacity Q _a , the positive electrode capacity Q _c , and the positive and negative electrode SOC deviation capacity Q _Li of the secondary battery 21. That is, in a secondary battery 21 through which a large current flows, the battery capacity Q _B is often the positive and negative electrode SOC deviation capacity Q _Li . As described above, the positive and negative electrode SOC deviation capacity Q _Li can be expressed by the above formula (11). As described above, the positive and negative electrode SOC deviation capacity Q _Li is calculated with high accuracy because it takes into account the formation of a coating on the surface of each active material of the negative electrode and the positive electrode, the cracking of the coating formed on the surface of each active material of the negative electrode and the positive electrode, and the cracking of each active material of the negative electrode and the positive electrode itself. Accordingly, when the battery capacity Q _B is expressed by the positive and negative electrode SOC deviation capacity Q _Li , the battery capacity Q _B is also calculated with high accuracy. This will be explained.

Here, the following simulation was performed assuming two secondary batteries of the same type (hereinafter, for convenience, they will be distinguished as a first battery and a second battery, but they are batteries of the same type). The results of the simulation are shown in FIG. 5. The horizontal axis of the graph in FIG. 5 indicates the square root of the number of days, and the vertical axis indicates the capacity maintenance rate of the secondary battery. The capacity maintenance rate of the secondary battery is the ratio of the capacity of the secondary battery at the current time to the capacity of the secondary battery in the initial state. In addition, in FIG. 5, the results for the first battery are shown by line L1, and the experimental results for the second battery are shown by line L2. In each of the results of the first battery and the second battery shown in FIG. 5, the positive and negative electrode SOC deviation capacity Q _Li is the smallest among the negative electrode capacity Q _a , the positive electrode capacity Q _c , and the positive and negative electrode SOC deviation capacity Q _Li , and the battery capacity Q _B = the positive and negative electrode SOC capacity Q _Li . And, the capacity maintenance rate in FIG. 5 is the ratio of the positive and negative electrode SOC capacity Q _Li after deterioration to the capacity of the initial secondary battery.

As for the first battery, the first battery was degraded in storage at 45°C from a capacity retention rate of 100% to a capacity retention rate of 92%. When the first battery was degraded from a capacity retention rate of 100% to a capacity retention rate of 92%, the decrease in capacity was 7.2% due to the formation of a coating on each electrode, 0.4% due to cracking of the coating formed on the surface of the active material of each electrode, and 0.4% due to cracking of the active material of each electrode itself.

As for the second battery, the second battery was cycle-degraded in an environment of 45°C from a state of 100% capacity retention rate, and the capacity retention rate was reduced to 92%. When cycle-degraded from a capacity retention rate of 100% to a capacity retention rate of 92%, the capacity reduction of the second battery was 4.0% due to the formation of a coating on each electrode, 1.6% due to cracking of the coating formed on the surface of the active material of each electrode, and 2.4% due to cracking of the active material of each electrode itself.

In other words, even if the first battery and the second battery have the same capacity maintenance rate and positive and negative electrode SOC capacity _QLi , the values of functions _kA , _kB , _kC , _lA , _lB , and _lC in equation (11) that configure the positive and negative electrode SOC capacity _QLi will differ between the first battery and the second battery depending on the usage conditions up to that point.

Then, the first battery and the second battery, which had a capacity maintenance rate of 92%, were deteriorated under the same conditions by combining cycle deterioration and storage deterioration. As a result, as shown in FIG. 5, it was found that the second battery, which was cycle-degraded first, deteriorated faster than the first battery, which was storage-degraded first (i.e., the slope of the graph in the region below 92% capacity maintenance rate in FIG. 5 becomes larger). From this, it can be seen that even if the secondary batteries have the same capacity maintenance rate, the degree of deterioration of the secondary battery thereafter differs depending on the usage condition of the secondary battery up to that point. From this, it can be seen that the battery capacity Q _B can be calculated with high accuracy by calculating the positive and negative electrode SOC deviation capacity Q _Li based on the functions k _A and l _A taking into account the formation of the film on each electrode, the functions k B and l _B taking into account the cracking of the film formed on the surface of the active material of each electrode, and the functions k _C and l _C taking into account the cracking of the active material _{itself of each electrode. The same applies when the battery capacity Q B is} the negative electrode capacity Q _a or the positive electrode capacity Q _c .

Next, with reference to Figures 2 and 6, a step will be described in which the battery deterioration prediction system 1 controls the current flowing to the secondary battery 21 by using the negative electrode capacity _Qa , positive electrode capacity _Qc , and positive/negative electrode SOC deviation capacity _QLi of the secondary battery 21 calculated in step S9 of Figure 3.

First, in step S10, the OCV characteristic calculation unit 112 calculates an updated OCP characteristic. In step S10, the initial OCP characteristic stored in the storage unit 115 is updated based on the negative electrode capacity _Qa , the positive electrode capacity _Qc , and the positive/negative electrode SOC deviation capacity _QLi calculated in the above-mentioned step S8. The initial OCP characteristic indicates the relationship between the SOC and _OCPa and the relationship between the SOC and _OCPc of the secondary battery 21 in the initial state. The method of updating the initial OCP characteristic is not particularly limited, and for example, a known method can be adopted. An example of the updated OCP characteristic is shown in FIG. 7.

2 and 6, in step S11, the OCV characteristic calculation unit 112 calculates an updated OCV characteristic. The updated OCV characteristic is calculated by taking the difference between OCP _c and OCP _a from the updated OCP characteristic calculated in step S10, thereby calculating an updated OCV characteristic, which is the relationship between the SOC and OCV of the secondary battery 21 after deterioration. An example of the updated OCV characteristic is shown in FIG. 7.

As shown in Fig. 2 and Fig. 6, in step S12, the differential characteristic calculation unit 113 calculates a differential characteristic based on the updated OCV characteristic calculated by the OCV characteristic calculation unit 112 in step S11. The differential characteristic is a relationship between the SOC of the secondary battery 21 and a differential value obtained by first-order or second-order differentiation of the OCV of the secondary battery 21 with respect to the SOC. As shown in Fig. 6, in step S12 of this embodiment, the differential characteristic calculation unit 113 calculates a ^d2V / ^dQ2 curve. The ^d2V / ^dQ2 curve is a graph showing a relationship between the SOC of the secondary battery 21 and a differential value obtained by second-order differentiation of the OCV, and is a type of differential characteristic. Fig. 8 shows a curve obtained by second-order differentiation of the OCV with respect to the SOC.

2 and 6, in step S13, current control unit 114 calculates a current control region. That is, in step S13, the SOC region (e.g., the SOC region surrounded by a dashed line in FIG. 8) in which the peak in the ^d2V / ^dQ2 curve shown in FIG. 8 is equal to or greater than a threshold value is calculated.

2 and 6, in step S14, the current control unit 114 controls the current value flowing through the secondary battery 21 to be reduced when the SOC of the secondary battery 21 is in a region of the SOC where the peak calculated in step S13 exceeds a threshold value. That is, when the secondary battery 21 is at the SOC calculated in step S13, the current control unit 114 controls the current value flowing through the secondary battery 21 to be smaller than the current value flowing through the secondary battery 21 in a region other than the SOC region. This makes it possible to suppress deterioration of the secondary battery 21. This will be explained next.

When the peak in the curve showing the differential characteristic is equal to or greater than a predetermined threshold, it is considered to be a region where the OCV fluctuates relatively greatly and the expansion and contraction are large. That is, in the SOC region where the peak is equal to or greater than the threshold, the expansion and contraction of the active material is large, so that the higher the current applied, the more likely it is that strain will occur in the active material, and the more likely it is that cracking of the active material will be accelerated. Therefore, the current control unit 114 controls to limit the current flowing to the secondary battery 21 in the SOC region where the peak is equal to or greater than the threshold. This makes it possible to suppress the acceleration of cracking of the active material. In addition, by detecting peaks equal to or greater than the threshold, the current control unit 114 prevents the current value from being limited in the noise region (i.e., the SOC region where a relatively small peak appears).

As described above, the current flowing through the secondary battery 21 can be controlled by utilizing the negative electrode capacity Q _a , the positive electrode capacity Q _c , and the positive/negative electrode SOC deviation capacity Q _Li of the secondary battery 21 .

As shown in FIG. 2, the battery deterioration prediction system 1 of this embodiment includes a learning unit 111 as described above. The learning unit 111 can be configured to update a prediction formula for predicting at least one of the negative electrode resistance R _a , positive electrode resistance R _c , negative electrode capacity Q _a , positive electrode capacity Q _c , positive and negative electrode SOC deviation capacity Q _Li , and battery capacity Q _B of the secondary battery by learning. The prediction formula updated by the learning unit 111 by learning is not limited, but in this embodiment, the learning unit 111 updates the constants A, B, and C of the pulverization function f (t, ΔDOD) used in the calculation of the negative electrode resistance R _a and the positive electrode resistance R _c in the electrode resistance acquisition unit 107, and the calculation of the negative electrode capacity Q _a , the positive electrode capacity Q _c , and the positive and negative electrode SOC deviation capacity Q _Li in the capacity prediction unit 110 by learning. The learning unit 111 can be corrected by, for example, fitting. For example, the learning unit 111 calculates an actual measurement value of the battery capacity _QB of the secondary battery 21 based on the integration of the secondary battery 21 acquired by the current value acquisition unit 102, and calculates constants A, B, and C that minimize the error between this and the battery capacity _QB calculated by the capacity prediction unit 110, thereby learning A, B, and C. It is also possible to learn the constants A, B, and C so as to minimize the error between other actual measurements and theoretical values.

The learning unit 111 is configured by a BMU mounted on the vehicle, but may also be configured by an information processing system arranged outside the vehicle. In this case, for example, the vehicle may be equipped with a DCM (Data Communication Module) that is a mobility communication device, and the BCU and ECU inside the vehicle may communicate with the information processing system outside the vehicle via the DCM. In addition, instead of the DCM, for example, a mobile terminal (smartphone, etc.) may be used. In addition, various short-distance wireless transmission methods such as WiFi (registered trademark) and Bluetooth (registered trademark) may be used as communication means between the BCU and ECU inside the vehicle and the information processing system outside the vehicle. Furthermore, the information processing system outside the vehicle may be in a fixed form or in a mobility form. In addition, it may be a retrofit control device connected to a connector provided on the ECU or BMU of the vehicle.

Then, the electrode resistance acquisition unit 107 and the capacity prediction unit 110 use the updated constants A, B, and C to calculate the negative electrode resistance R _a , the positive electrode resistance R _c , the negative electrode capacity Q _a , the positive electrode capacity Q _c , and the positive and negative electrode SOC deviation capacity Q _Li .

In this embodiment, at least one of the functions _gA , gB, gC, hA, hB, _hC , _iA , _iB , iC, jA, _jB , _jC , _kA , _kB , _kC , lA, lB, and _lC included in the above formulas (7) to (11) for calculating the negative electrode _{resistance R a ,} the positive electrode resistance _{R c} , the negative electrode capacity Q _a , the positive _electrode capacity _{Q c ,} and the positive and negative electrode SOC deviation capacity Q _Li may include _{, as} a further variable, any of the _elapsed time t from when the secondary battery was started to be used, the number of charge and discharge cycles N, and the throughput Ah indicating the current integrated value in charge and discharge. The further variables included in the above functions can be appropriately selected and may be any combination.

Next, the effects of this embodiment will be described.
The battery deterioration prediction system 1 of the above embodiment includes a CCP calculation unit 109 that calculates a CCP _a based on the OCP _a and the polarization ΔV _a , and calculates a CCP _c based on the OCP _c and the polarization ΔV _c . The capacity prediction unit 110 predicts the negative electrode capacity Q _a , the positive electrode capacity Q _c , and the positive and negative electrode SOC deviation capacity Q _Li of the secondary battery 21 based on the CCP _a and CCP _c , and predicts the battery capacity Q B of the secondary battery 21 based on the predicted negative electrode capacity Q _a , the positive electrode capacity Q _c , and the positive and negative electrode SOC deviation capacity Q _Li . By predicting the battery capacity Q _B of the secondary battery 21 while taking into account the polarization ΔV _a and ΔV _c , the battery capacity Q _B can be predicted taking into account the polarization that becomes apparent as the secondary battery ₂₁ deteriorates, and the prediction of the battery capacity Q _B becomes highly accurate.

In addition, the electrode resistance acquisition unit 107 and the capacity prediction unit 110 calculate the negative electrode resistance R a and the negative electrode capacity Q a using a plurality of functions g _A , g _B , g _C , i _A , i _B , and i _C that are correlated with a plurality of deterioration factors in the negative electrode of the secondary battery 21. In addition, the electrode resistance acquisition unit 107 and the capacity prediction unit 110 calculate the positive electrode resistance R _c and the positive electrode capacity Q _c using a plurality of functions h _A , h _B , h _C , j _A , j _B , and j _C that are correlated with a plurality of deterioration factors in the positive electrode of the secondary battery 21. In addition, the capacity prediction unit 110 calculates the positive and negative _electrode SOC deviation capacity Q _Li using a plurality of functions k _A , k _B , k _C , l _A , l _B , and l _C that are correlated with a plurality of deterioration factors in the electrolyte of the secondary battery ₂₁ . This makes it possible to accurately predict the negative electrode capacity Q _a , positive electrode capacity Q c, positive and negative electrode SOC deviation capacity Q _Li , negative electrode resistance R _a and positive electrode resistance R _c taking into account various deterioration factors of each part of the secondary battery 21 (i.e., the aforementioned coating _formation , coating cracking, cracking of the active material itself, etc.).

Further, the electrode resistance acquisition unit 107 and the capacity prediction unit 110 predict at least one of the negative electrode capacity Q _a , the positive electrode capacity Q _c , the positive and negative electrode SOC deviation capacity Q _Li , the negative electrode resistance R _a and the positive electrode resistance R _c based on the pulverization function f(t, ΔDOD)=A×exp{−B×exp(C×ΔDOD)×t}. As a result, at least one of the negative electrode capacity Q _a , the positive electrode capacity Q _c , the positive and negative electrode SOC deviation capacity Q _Li , the negative electrode resistance R _a and the positive electrode resistance R _c can be predicted while reflecting the pulverization of the electrodes of the secondary battery 21 due to repeated charging and discharging of the secondary battery 21, and therefore these can be predicted with high accuracy.

The battery deterioration prediction system 1 also includes a learning unit 111 that updates the constants A, B, and C of the pulverization function by learning. Therefore, the pulverization function can be updated according to the deterioration state of the secondary battery 21, and a timely pulverization function can be obtained. This can improve the calculation accuracy of the negative electrode capacity Q _a , the positive electrode capacity Q _c , the positive/negative electrode SOC deviation capacity Q _Li , the negative electrode resistance R _a , and the positive electrode resistance R _c predicted using the pulverization function.

In addition, at least a part of the battery deterioration prediction system 1 other than the learning unit 111 is mounted on the vehicle, and the learning unit 111 is provided outside the vehicle. Therefore, there is no need to provide hardware inside the vehicle for performing processing by the learning unit 111, which makes it easier to reduce the load on the system inside the vehicle.

The battery degradation prediction system 1 also includes a current control unit 114 that limits the current flowing through the secondary battery 21 in a current amount region where the peak of the differential value is equal to or greater than a threshold value, based on the differential characteristic calculated by the differential characteristic calculation unit 113. That is, as described above, the current control unit 114 limits the amount of current flowing through the secondary battery 21 in a current amount region where cracking of the active material of the electrodes of the secondary battery 21 is likely to be accelerated. This makes it possible to suppress the progression of cracking of the active material of the electrodes of the secondary battery 21, thereby suppressing the progression of deterioration of the secondary battery 21.

As described above, this embodiment provides a battery degradation prediction system that can predict the degree of degradation of a secondary battery with high accuracy.

In the above embodiment 1, the above formulas (7) to (11) are used to calculate the negative electrode resistance Ra, the positive electrode resistance Rc, the negative electrode capacity Qa, the positive electrode capacity Qc, and the positive and negative electrode SOC deviation capacity QLi. However, instead of these, the operators between the functions in each formula may be changed to +. That is, instead of the above formulas (7), (8), (9), (10), and (11), the following formulas (7A), (8A), (9A), (10A), and (11A) may be used, respectively.

In addition, without being limited to these, any operator can be used between functions in the above formulas (7) to (11), and the operators between functions in the same formula may be the same or different. Regardless of which operator is used in each of the above formulas, the same effect as in embodiment 1 can be achieved.

(Embodiment 2)
In the above-described first embodiment, the learning unit 111 updates the constants A, B, and C of the pulverization function by learning, but the updating by learning in the learning unit 111 is not limited to this, and the prediction formula for the battery state of the secondary battery 21 can be updated by learning based on the usage history. The battery state of the secondary battery 21 can be, for example, the negative electrode resistance R _a , positive electrode resistance R _c , negative electrode capacity Q _a , positive electrode capacity Q _c , positive and negative electrode SOC deviation capacity Q _Li , and battery capacity Q _B of the secondary battery.

The prediction formula can be a prediction formula for predicting at least one of the battery information. For example, the formulas (7) to (11) or (7A) to (11A) in the first embodiment can be used. The pulverization function in the first embodiment can be understood to be a part of the prediction formula.

The learning unit 111 can update the prediction formula, for example, as follows. First, the learning unit 111 uses the state estimation amount obtained from the usage history as an explanatory variable and calculates a relational formula using machine learning in which the battery state of the secondary battery 21 is the objective variable. The form of the relational formula is not limited, and can be expressed in the form of, for example, a linear or multi-order model function, a map, etc.

Then, the learning unit 111 calculates an estimated value for the battery state of the secondary battery 21 as a true value from the above relational equation. Thereafter, the learning unit 111 compares the estimated value with a predicted value of the battery state obtained from a prediction equation for the secondary battery 21 prepared in advance, and can update the prediction equation based on the comparison result. For example, when the above comparison result indicates that the difference between the estimated value and the predicted value is larger than a preset reference value, the learning unit 111 can update the prediction equation so that the difference becomes smaller.

The usage history is not limited to anything related to the use of the secondary battery 21, but if the secondary battery 21 is mounted on a vehicle, the vehicle's driving data can be used as the usage history. Examples of state estimation quantities obtained from the usage history include the voltage behavior of the secondary battery 21 at the end of charging or discharging, the voltage relaxation behavior after a charging or discharging pause, and the AC impedance at a specific frequency. The usage history can be stored in the memory unit 115. The usage history can also be saved in an external storage device provided outside the battery deterioration prediction system 1, and the learning unit 111 can read the usage history saved in the external storage device as necessary.

As a machine learning method for calculating the above relational equation based on the usage history, a known method can be used, such as a regression method, a support vector machine, or a neural network. Examples of regression methods include linear regression, multiple regression, logistic regression, and Gaussian process regression.

In addition, the object of update through learning in the learning unit 111 is not limited, but can be a constant in the deterioration prediction formula, for example, at least one constant in the deterioration prediction formula shown in formulas (7) to (11) or (7A) to (11A) in embodiment 1.

In this embodiment 2, vehicle driving data is used as the usage history, and the voltage behavior at the end of charging the secondary battery 21 is used as the state estimation quantity obtained from the usage history. Note that configurations not specifically mentioned in this embodiment 2 are assumed to have the same configurations as those in embodiment 1 shown in FIG. 2, and the same reference numerals as in embodiment 1 are used and the description thereof is omitted.

In the second embodiment, the DC resistance of a specific time constant in the charging state is acquired as the voltage behavior at the end of charging. The end of charging refers to a period in which the SOC is charged to 90% or more. The difference R _4sec -R _2sec between the DC resistance value R _4sec when the secondary battery 21 is discharged at a constant current for 4 seconds and the DC resistance value R _2sec when the secondary battery 21 is discharged at a constant current for 2 seconds is acquired as the DC resistance. In the second embodiment, although not shown, a voltage sensor that detects the voltage value of the secondary battery 21 and a voltage acquisition unit that acquires the voltage value detected by the voltage sensor are provided. The DC resistance value R _4sec can be calculated as ΔV _4sec /I from the voltage change amount ΔV _4sec during constant current discharge for 4 seconds acquired by the voltage acquisition unit and the current value acquisition unit 102 and the current value I flowing through the secondary battery 21. Similarly, the DC resistance value R _2sec can be calculated as ΔV _{2sec /I from the voltage change amount ΔV 2sec} during constant current discharge for 2 seconds acquired by the voltage acquisition unit and current value acquisition unit ₁₀₂ and the current value I flowing through the secondary battery 21. The calculation of the DC resistance value R _4sec may be performed by the learning unit 111 or by another calculation unit.

The update flow for updating the prediction formula in the second embodiment will be described below.
First, in step S21 shown in FIG. 9, the learning unit 111 acquires, for example, the difference R 4sec -R _2sec and the battery capacity _QB in the deterioration process in which the secondary battery 21 is subjected to repeated charge and discharge cycles between SOC ₁₀ % and 90% at a battery temperature of 25°C as the first usage history, and calculates a first relational expression by machine learning from the relationship between the two. The first relational expression is a regression line obtained based on the least squares method from the actual measurement values of the relationship between the difference R _4sec -R _2sec and the capacity maintenance rate. FIG. 10(a) shows a scatter diagram of the actual measurement values that are the relationship between the difference R _4sec -R _2sec and the capacity maintenance rate in the first usage history, and the regression line that is the first relational expression is indicated by a dashed line L1. Note that the capacity maintenance rate is the ratio of the battery capacity _QB at each time point to the battery capacity of the secondary battery 21 in the initial state.

9, the learning unit 111 acquires, for example, the difference R4sec-R2sec and the capacity maintenance rate in a deterioration process in which the secondary battery 21 is subjected to repeated charge and discharge cycles between SOC 0% and 100% at a battery temperature of ₄₅ ° _C as the second usage history, and calculates a second relational equation by machine learning from the relationship between the two. As in the case of the first usage history, FIG. 10(b) shows a scatter plot of actual measured values, which is the relationship between the difference _R4sec - _R2sec and the capacity maintenance rate in the second usage history, and the regression line, which is the second relational equation, is indicated by the dashed line L2.

9, the learning unit 111 calculates an estimated value of the battery capacity _QB of the secondary battery 21 as a true value from a first relational expression calculated by machine learning based on the first usage history, for example. Then, in step S23 shown in FIG. 9, the capacity prediction unit 110 calculates a predicted value of the battery capacity _QB of the secondary battery 21 based on the prediction expressions shown in the above-mentioned formulas (7) to (11) or (7A) to (11A).

Then, in step S24 shown in Fig. 9, the learning unit 111 compares the estimated value of the battery capacity _QB with the predicted value of the battery capacity _QB , and then in step S25, determines whether or not the prediction formula needs to be updated based on the comparison result. In the second embodiment, the need for updating is determined by determining whether or not the comparison result indicates that the difference between the estimated value and the predicted value is greater than a reference value. In step S25, if it is determined that the comparison result indicates that the difference between the estimated value and the predicted value is greater than a reference value, it is determined that the prediction formula needs to be updated, and the process proceeds to Yes in step S25.

Then, in step S26, the learning unit 111 updates at least one of the constants included in the prediction formula shown in formulas (7) to (11) or formulas (7A) to (11A), and ends the update flow. Note that the comparison of the estimated value and the predicted value in step S24 and the determination of whether or not an update is required in step S25 may use both the estimated value based on the first relational expression based on the usage history and the estimated value based on the second relational expression based on the usage history, or may use only one of them.

On the other hand, in step S25 shown in FIG. 9, if it is determined that the comparison result does not indicate that the difference between the estimated value and the predicted value is greater than the reference value, it is determined that updating of the prediction formula is unnecessary, and the process proceeds to No in step S25 to end the update flow. Note that the update flow in this embodiment 2 can be performed at any desired timing.

According to the second embodiment, the learning unit 111 can update the prediction formula through learning in accordance with the deterioration state of the secondary battery 21, which changes from moment to moment with use, thereby further improving the prediction accuracy. This makes it possible to provide a battery deterioration prediction system 1 that can predict the degree of deterioration of the secondary battery 21 with high accuracy. Note that the second embodiment also achieves the same effects as the first embodiment.

In the above-described second embodiment, the voltage behavior of the secondary battery 21 at the end of charging is used as the state estimated quantity obtained from the usage history, but instead, in the first modified embodiment, the voltage relaxation behavior of the secondary battery 21 after charging is suspended is used. In the first modified embodiment, for example, a voltage change ΔV _600sec for 600 seconds after charging of the secondary battery 21 is suspended is acquired by the voltage acquisition unit as the voltage relaxation behavior after charging is suspended. Note that the voltage relaxation period is the period from after charging is suspended until the battery voltage reaches a predetermined steady state.

Then, the learning unit 111 acquires, for example, the voltage change ΔV _{600 sec} and the battery capacity QB in the deterioration process in which the secondary battery ₂₁ is subjected to repeated charge and discharge cycles between SOC 10% and 90% at a battery temperature of 25° C. as the first usage history, and calculates a first relational expression by machine learning from the relationship between the two. The first relational expression is a regression line obtained based on the least squares method from the actual measured values of the relationship between the voltage change ΔV _{600 sec} and the battery capacity _QB . FIG. 11(a) shows a scatter diagram of the actual measured values, which are the relationship between the voltage change ΔV _{600 sec} and the capacity maintenance rate in the first usage history, and the regression line, which is the first relational expression, is indicated by a dashed line L3.

Similarly, the learning unit 111 calculates the second relational expression by machine learning from the relationship between the voltage change ΔV _600sec and the battery capacity _QB in the deterioration process in which the secondary battery 21 is repeatedly charged and discharged at a battery temperature of 45° C. and the SOC is between 0% and 100% as the second usage history. As in the case of the first usage history, FIG. 11B shows a scatter diagram of actual measurements that are the relationship between the voltage change ΔV _600sec and the capacity maintenance rate in the second usage history, and the regression line that is the second relational expression is shown by the dashed line L4. The first modified embodiment also has the same effects as the second embodiment.

In a second modified embodiment, the secondary battery 21 employs an AC impedance of a specific frequency as a state estimation quantity obtained from the usage history. In the second modified embodiment, the secondary battery 21 employs LiNiCoMnO ₂ (NCM) as a positive electrode material and graphite (Gr) as a negative electrode material. The AC impedance of the specific frequency can be calculated by an AC impedance calculation unit (not shown) mounted on the vehicle. In this modified embodiment, the AC impedance calculation unit calculates the AC impedance of the specific frequency based on the change in the current value and the change in the voltage value in the secondary battery 21 during charging and discharging. The specific frequency can be set as appropriate, and the value of the specific frequency is not limited.

In the second modified embodiment, the learning unit 111 calculates a first relational expression by machine learning from the relationship between the AC impedance and the battery capacity _QB at a frequency of 630 Hz under the condition of a battery temperature of -10° in the deterioration process after storing the secondary battery 21 at a battery temperature of 60° C. and an SOC of 90% as the first usage history. Also, the second relational expression is calculated by machine learning from the relationship between the AC impedance and the battery capacity _QB at a frequency of 630 Hz under the condition of a battery temperature of 25°. The first relational expression and the second relational expression are regression lines obtained based on the least squares method from the actual measured values of the relationship between the AC impedance and the battery capacity _QB . FIG. 12(a) shows a scatter diagram of the actual measured values that are the relationship between the AC impedance and the capacity maintenance rate under the condition of a battery temperature of -10° in the first usage history, and the regression line that is the first relational expression is shown by a dashed line L5. FIG. 12(b) shows a scatter diagram of actual measured values representing the relationship between AC impedance and capacity retention rate at a battery temperature of 25° C. in the first usage history, and also shows the regression line, which is the second relational expression, by a dashed line L6.

Similarly, the learning unit 111 calculates a third relational expression by machine learning from the relationship between the AC impedance and the battery capacity _QB at a frequency of 630 Hz under the condition of a battery temperature of -10° in the deterioration process after storing the secondary battery 21 at a battery temperature of 55° C. and an SOC of 100% as the second usage history. Also, a fourth relational expression is calculated by machine learning from the relationship between the AC impedance and the battery capacity _QB at a frequency of 630 Hz under the condition of a battery temperature of 25°. The third relational expression and the fourth relational expression are regression lines obtained based on the least squares method from the actual measured values of the relationship between the AC impedance and the battery capacity _QB . FIG. 12(c) shows a scatter diagram of the actual measured values that are the relationship between the AC impedance and the capacity maintenance rate under the condition of a battery temperature of -10° in the second usage history, and the regression line that is the third relational expression is shown by a dashed line L7. 12(d) shows a scatter diagram of actual measured values representing the relationship between the AC impedance and the capacity retention rate at a battery temperature of 25° in the second usage history, and the regression line, which is the fourth relational expression, is shown by a dashed line L8. The modified embodiment 2 also has the same effects as the above-mentioned embodiment 2.

In addition, in the third modified embodiment as a further modified embodiment, as in the second modified embodiment, the AC impedance of a specific frequency in the secondary battery 21 is adopted as a state estimation quantity obtained from the usage history. In the third modified embodiment, the learning unit 111 calculates a first relational expression by machine learning from the relationship between the AC impedance of a frequency of 630 Hz and the battery capacity _QB under the condition of a battery temperature of -10° in a deterioration process in which the secondary battery 21 is subjected to repeated charge and discharge cycles between SOC 0% and 100% at a battery temperature of 45° C. as the first usage history. In addition, the learning unit 111 calculates a second relational expression by machine learning from the relationship between the AC impedance of a frequency of 630 Hz and the battery capacity _QB under the condition of a battery temperature of 25°. The first relational expression and the second relational expression are regression lines obtained based on the least squares method from the actual measured values of the relationship between the AC impedance and the battery capacity _QB . FIG. 13(a) shows a scatter diagram of the actual measured values that are the relationship between the AC impedance and the capacity maintenance rate under the condition of a battery temperature of -10° in the first usage history, and the regression line that is the first relational expression is shown by a dashed line L9. FIG. 13(b) shows a scatter plot of actual measured values representing the relationship between AC impedance and capacity retention rate at a battery temperature of 25° C. in the first usage history, and also shows the regression line, which is the second relational expression, by the dashed line L10.

Similarly, the learning unit 111 calculates a third relational expression by machine learning from the relationship between the AC impedance and the battery capacity _QB at a frequency of 630 Hz under the condition of a battery temperature of -10° in the deterioration process after storing the secondary battery 21 at a battery temperature of 55° C. and SOC of 100% as the second usage history. Also, a fourth relational expression is calculated by machine learning from the relationship between the AC impedance and the battery capacity _QB at a frequency of 630 Hz under the condition of a battery temperature of 25°. The third relational expression and the fourth relational expression are regression lines obtained based on the least squares method from the actual measured values of the relationship between the AC impedance and the battery capacity _QB . FIG. 13(c) shows a scatter diagram of the actual measured values that are the relationship between the AC impedance and the capacity maintenance rate under the condition of a battery temperature of -10° in the second usage history, and the regression line that is the third relational expression is shown by a dashed line L11. 13(d) shows a scatter diagram of actual measured values representing the relationship between the AC impedance and the capacity retention rate at a battery temperature of 25° in the second usage history, and the regression line, which is the fourth relational expression, is shown by a dashed line L12. The third modified embodiment also has the same effects as the second embodiment.

In addition, in the fourth modification as a further modification, as in the second and third modifications, the AC impedance of a specific frequency in the secondary battery 21 is adopted as a state estimation quantity obtained from the usage history. In the fourth modification, the learning unit 111 calculates a first relational expression by machine learning from the relationship between the AC impedance of a frequency of 630 Hz and the battery capacity _QB under the condition of a battery temperature of -10° in a deterioration process in which the secondary battery 21 is subjected to repeated charge and discharge cycles between SOC 10% and 90% at a battery temperature of 10° C. as the first usage history. In addition, the learning unit 111 calculates a second relational expression by machine learning from the relationship between the AC impedance of a frequency of 0.005 Hz and the battery capacity _QB under the condition of a battery temperature of 25°. In addition, the learning unit 111 calculates a third relational expression by machine learning from the relationship between the AC impedance of a frequency of 50 Hz and the battery capacity _QB under the condition of a battery temperature of -10°.

The first relational expression, the second relational expression, and the third relational expression are regression lines obtained based on the least squares method from the measured values of the relationship between the AC impedance and the battery capacity _QB . FIG. 14(a) shows a scatter diagram of the measured values of the relationship between the AC impedance at a frequency of 630 Hz and the capacity maintenance rate under the condition of a battery temperature of -10° in the first usage history, and the regression line which is the first relational expression is shown by the dashed line L13. FIG. 14(b) shows a scatter diagram of the measured values of the relationship between the AC impedance at a frequency of 0.005 Hz and the capacity maintenance rate under the condition of a battery temperature of 25° in the first usage history, and the regression line which is the second relational expression is shown by the dashed line L14. FIG. 14(c) shows a scatter diagram of the measured values of the relationship between the AC impedance at a frequency of 50 Hz and the capacity maintenance rate under the condition of a battery temperature of -10° in the first usage history, and the regression line which is the third relational expression is shown by the dashed line L15. The fourth modified embodiment also has the same effects as the second embodiment.

(Embodiment 3)
15, the battery deterioration prediction system 1 of the third embodiment has a mechanical change calculation unit 116 in addition to the configuration of the first embodiment described above. The mechanical change calculation unit 116 calculates a relation value related to the mechanical change of the secondary battery 21 based on the positive and negative electrode SOC deviation capacity _QLi predicted by the capacity prediction unit 110. The mechanical change calculation unit 116 is configured by, for example, a BMU mounted on the vehicle, and includes a processor capable of executing a program for calculating a relation value related to the mechanical change of the secondary battery 21.

In the third embodiment, examples of mechanical changes in the secondary battery 21 include changes in cell thickness, volume, and spring constant in the secondary battery 21. The mechanical changes in the secondary battery 21 are caused by the expansion of the negative and positive electrodes of the secondary battery 21. The expansion of both electrodes occurs due to a film formed on both electrodes. The film is formed by consuming the electrolyte containing lithium ions in the battery reaction system. On the other hand, the positive and negative electrode SOC deviation capacity Q _Li depends on the number of lithium ions that can move between the positive and negative electrodes. Therefore, based on the predicted value of the positive and negative electrode SOC deviation capacity Q _Li , a relational value related to the mechanical changes in the secondary battery 21 caused by the formation of the film can be calculated. The relational value related to the mechanical changes in the secondary battery 21 can be the amount of change in cell thickness, the amount of expansion, and the amount of change in spring constant in the secondary battery 21.

In this embodiment 3, the mechanical change calculation unit 116 calculates relational values related to the mechanical changes of the secondary battery 21, thereby making it possible to predict rapid deterioration of the secondary battery 21 due to electrolyte depletion or breakage of the current collector, and to predict lithium precipitation due to high-rate deterioration caused by charging at a high C rate. This makes it possible to diagnose the reliability of the secondary battery 21 from a structural point of view. Note that this embodiment 3 also provides the same effects as the above-mentioned embodiment 1.

The mechanical change calculation unit 116 may predict gas generation in the secondary battery 21 as a mechanical change in the secondary battery 21. Gas generation in the secondary battery 21 is caused by oxidation-reduction decomposition of the electrolyte in the secondary battery 21, and the oxidation-reduction decomposition of the electrolyte consumes lithium ions in the battery reaction system. Therefore, similar to the case of the formation of the coating described above, the amount of gas generation can be calculated as a relational value regarding the mechanical change in the secondary battery 21 based on the predicted value of the positive and negative electrode SOC deviation capacity Q _Li .

(Embodiment 4)
The battery deterioration prediction system 1 of the present embodiment 4 has a lithium precipitation prediction unit 117 as shown in Fig. 16 instead of the mechanical change calculation unit 116 in the above-mentioned embodiment 3. The lithium precipitation prediction unit 117 predicts the timing of lithium precipitation occurring in the negative electrode of the secondary battery 21 based on the negative electrode resistance R _a calculated by the electrode resistance acquisition unit 107. The lithium precipitation prediction unit 117 compares the negative electrode resistance R _a with a preset reference voltage that is a reference for lithium precipitation, and can predict that lithium precipitation will start to occur in the negative electrode of the secondary battery 21 at the timing when the comparison result indicates that the negative electrode resistance R _a is lower than the reference voltage.

In this embodiment 4, the lithium precipitation prediction unit 117 predicts the timing of lithium precipitation, thereby further improving the accuracy of the deterioration prediction of the secondary battery 21. Note that this embodiment 4 also provides the same effects as the above-mentioned embodiment 1.

(Embodiment 5)
In the above-described first embodiment, the secondary battery 21, the current sensor 26, the temperature sensor 27, and the entire configuration of the battery deterioration prediction system 1 are mounted on the vehicle, and as another embodiment, only the learning unit 111 is disposed outside the vehicle, and the configuration other than the learning unit 111 is mounted on the vehicle. This is not limiting, and the secondary battery 21 may be mounted on the vehicle, and at least a part of the battery deterioration prediction system 1 may be disposed outside the vehicle. For example, in the fifth embodiment shown in FIG. 17, the secondary battery 21, the current sensor 26, and the temperature sensor 27 are mounted on the vehicle 50, and the entire configuration of the battery deterioration prediction system 1 is disposed outside the vehicle 50.

In this embodiment 5, the battery deterioration prediction system 1 is arranged on a server 60 arranged outside the vehicle 50. The server 60 can be configured by a computer. The server 60 may be configured by a single computer, or may be configured by a network in which multiple computers are connected by wire or wirelessly. The server 60 may also be configured as a cloud server via the Internet.

As shown in FIG. 17, in this embodiment 4, the vehicle 50 has a wireless communication unit 51, and the server 60 also has a wireless communication unit 61. The current value and battery temperature of the secondary battery 21 detected by the current sensor 26 and the temperature sensor 27 in the vehicle 50 can be transmitted to the battery deterioration prediction system 1 on the server 60 via both wireless communication units 51, 61.

In addition, in this embodiment 5, as shown in FIG. 17, the vehicle 50 may be provided with a display unit 52 that displays the prediction results by the battery deterioration prediction system 1. The prediction results by the battery deterioration prediction system 1 can be transmitted from the battery deterioration prediction system 1 on the server 60 to the vehicle 50 via the wireless communication units 51, 61 and displayed on the display unit 52 of the vehicle 50. This allows the prediction results by the battery deterioration prediction system 1 to be confirmed on the vehicle.

In this embodiment 5, since there is no need to install the configuration of the battery deterioration prediction system 1 on the vehicle 50, the configuration of the BMU of the vehicle 50 is prevented from becoming complicated. In addition, since the battery deterioration prediction system 1 is arranged on a server 60 arranged outside the battery deterioration prediction system 1, management of the battery deterioration prediction system 1 is easy. Furthermore, if the battery deterioration prediction systems 1 of multiple vehicles are arranged on the server 60, workability is improved by managing them collectively. Note that this embodiment 5 also achieves the same effects as the above-mentioned embodiment 1.

(Embodiment 6)
In this embodiment 6, as shown in FIG. 18, the battery deterioration prediction system 1 has an estimation unit 118 that estimates the data that could not be acquired _during a _data unacquired period ΔtN in _which data _{necessary to} acquire the values of variables included in the above- _mentioned multiple functions _gA , _gB , _gC , hA, _hB , _hC , _iA , iB, iC, jA _{, jB} , _jC , kA, _kB , kC, lA, _lB , _lC could not be acquired. Then, the electrode resistance acquisition unit 107 and the capacity prediction unit 110 calculate the negative electrode resistance R _a , the positive electrode resistance R c, the negative electrode capacity Q a, the positive electrode capacity Q _c , and the positive and negative electrode SOC deviation capacity _{Q Li during} the data unacquired _period using the values of the variables acquired based on the above data estimated by the estimation unit 118. In this embodiment 6, the same symbols are used for the configurations equivalent to those of embodiment 1, and their description is omitted.

In this embodiment 6, the data required to obtain the values of the above variables may be, for example, battery information such as the temperature T, current value I, voltage value V, and SOC of the secondary battery 21, and the data may be the variables themselves in the above function, or the values of the above variables may be calculated based on the data.

The estimation unit 118 shown in FIG. 18 is composed of a calculation device that executes a program for making the above estimation, and in this embodiment 6, is composed of a BMU mounted on the vehicle. In this embodiment 6, the data non-acquired period can be, for example, a period in which the above data cannot be acquired because no power is supplied to the battery deterioration prediction system 1, or a period in which the above data cannot be acquired due to a malfunction in the battery deterioration prediction system 1. In this embodiment 6, the secondary battery 21 constitutes a battery mounted on the vehicle, the battery deterioration prediction system 1 is composed of a BMU, and the period during which the vehicle is parked with the ignition off corresponds to the data non-acquired period.

In the sixth embodiment, the estimation unit 118 estimates the temperature T, current value I, voltage value V, and SOC of the secondary battery 21 during the data non-acquired period ΔtN. These estimation processes by the estimation unit 118 are performed before step S1 of the process of predicting the battery capacity _QB by the battery deterioration prediction system 1 shown in Fig. 3. After the estimation process is completed, the process of predicting the battery capacity _QB is performed using the values estimated by the estimation process.

Next, the estimation process in the sixth embodiment will be described with reference to the flow chart shown in FIG.
First, in step S101 shown in Fig. 19, the estimation unit 118 determines whether or not the vehicle has an unprocessed parking period in the history. In this embodiment, it determines whether or not the vehicle has a period during which the ignition Ign was off as shown in Fig. 20(a). If it is determined that there is a parking period ΔtN during which the Ign was off as shown in Fig. 20(a), the process proceeds to Yes in step S101, and if not, the process proceeds to No in step S101, and the estimation process ends.

Next, if the process proceeds to Yes in step S101 shown in FIG. 19, the estimation unit 118 reads data immediately before parking begins and data immediately after parking ends. The data immediately before parking begins is the temperature T1, current value I1, voltage value V1, and SOC1 immediately before the start timing t1 of the parking period ΔtN shown in FIG. 20(a) to FIG. 20(e). The data immediately after parking ends is the temperature T2, current value I2, voltage value V2, and SOC2 immediately after the end timing t2 of the parking period ΔtN shown in FIG. 20(a) to FIG. 20(e). All of this data is stored in the memory unit 115.

After that, in step S103 shown in FIG. 19, the estimation unit 118 acquires the ambient temperature Ts of the secondary battery 21 during the parking period ΔtN. The ambient temperature Ts of the secondary battery 21 is not the temperature of the secondary battery 21 itself, but the temperature of the surrounding area where the secondary battery 21 is placed. When another device that operates during the parking period ΔtN is placed around the secondary battery 21, the temperature detected by a temperature sensor mounted on the other device can be used as the ambient temperature Ts. In addition, if it is difficult to acquire the ambient temperature using a temperature sensor mounted on the other device, the outside air temperature at the location and time when the vehicle is parked during the parking period ΔtN can be used as the ambient temperature of the secondary battery 21 by referring to weather data provided by the Japan Meteorological Agency or the like. In this embodiment 6, the ambient temperature Ts of the secondary battery 21 is the temperature acquired by a temperature sensor mounted on a device provided around the secondary battery 21, and can be expressed as the ambient temperature Ts shown in FIG. 20(b).

Then, the process proceeds to step S104 shown in FIG. 19, where the estimation unit 118 performs a provisional estimation of the battery temperature during parking. This provisional estimation is performed by acquiring in advance the outside air heat dissipation resistance to the ambient temperature in the secondary battery 21, and estimating the temperature change in the outside air environment of the ambient temperature Ts from the temperature T1 at the start timing t1 of the parking period ΔtN to the end timing t2 of the parking period ΔtN. For example, the provisional estimation value by the estimation unit 118 can be expressed as the provisional estimated temperature Ta shown in FIG. 20(b).

Then, the process proceeds to step S105 shown in FIG. 19, where the estimation unit 118 corrects the provisional estimated temperature, and calculates the corrected value as the estimated temperature. The correction is performed so that the temperature Tas of the provisional estimated temperature Ta at the start timing t1 of the parking period ΔtN coincides with the temperature T1 immediately before parking, and the temperature Tae of the provisional estimated temperature Ta at the end timing t2 of the parking period ΔtN coincides with the temperature T2 immediately after parking. In this embodiment 6, a time-proportional correction is performed as the correction. That is, as shown in FIG. 20(b), the provisional estimated temperature Tae at the end timing t2 of the parking period ΔtN is compared with the temperature T2 of the secondary battery 21 at the end timing t2 of the parking period ΔtN. If the two values are different, the provisional estimated temperature Ta is reduced in its entirety in the vertical direction of the graph so that the provisional estimated temperature Tae coincides with the temperature T2 while the provisional estimated temperature Tae at the start timing t1 coincides with the temperature T1, and the corrected estimated temperature Tb is calculated. The calculated corrected estimated temperature Tb is then stored in the memory unit 115. Note that the correction for making the tentative estimated temperature Tae coincide with the temperature T2 is not limited to this, and can be performed by other methods.

Next, the process proceeds to step S106 shown in FIG. 19, where the estimation unit 118 estimates the current value I, the voltage value V, and the SOC during the parking period ΔtN. First, since no current actually flows through the secondary battery 21 during the parking period ΔtN, the estimated current value Ib, which is an estimated value of the current value I, is estimated to be 0 A as shown in FIG. 20(c). Since the change in the voltage value V is small during the parking period ΔtN, the estimated voltage value Vb is estimated by linearly interpolating from the voltage value V1 immediately before the start of parking to the voltage value V2 immediately after the end of parking as shown in FIG. 20(d). Since the change in the SOC is also small during the parking period ΔtN, the estimated SOC value SOCb is estimated by linearly interpolating from the SOC1 immediately before the start of parking to the SOC2 immediately after the end of parking as shown in FIG. 20(e). Then, the estimated current value Ib, the estimated voltage value Vb, and the estimated SOC value SOCb are also stored in the storage unit 115. The method of estimating the current estimate value Ib, the voltage estimate value Vb, and the SOC estimate value SOCb is not limited to this, and other methods may be used. Then, the flow of the estimation process is ended, and the process proceeds to step S1 of the process of predicting the battery capacity _QB shown in FIG.

According to the sixth embodiment, in the process of predicting the battery capacity _QB , even during a period in which the values of variables included in a plurality of functions for calculating the negative electrode resistance R _a , the positive electrode resistance R _c , the negative electrode capacity Q _a , the positive electrode capacity Q _c , and the positive and negative electrode SOC deviation capacity Q _Li cannot be obtained, the negative electrode resistance R _a , the positive electrode resistance R _c , the negative electrode capacity Q _a , the positive electrode capacity Q c , and the positive and negative electrode SOC deviation capacity Q _Li can be calculated by using the above-mentioned temperature estimate value Tb, current estimate value Ib, voltage estimate value _Vb , and SOC estimate value SOCb. This makes it possible to further improve the prediction accuracy of the battery capacity _QB .

Furthermore, in the sixth embodiment, the temperature estimate Tb is calculated by correcting the provisional estimated temperature Ta, which is based on the outside air heat radiation resistance of the secondary battery 21, by using the ambient temperature Ts. Therefore, the temperature estimate Tb more accurately reflects the parking situation of the vehicle, and the prediction accuracy of the battery capacity _QB can be further improved.

The present invention is not limited to the above-described embodiments and modifications, and can be applied to various embodiments without departing from the spirit of the present invention. For example, the present invention can be configured to combine the mechanical change calculation unit 116 in embodiment 3 and the lithium precipitation prediction unit 117 in embodiment 4.

REFERENCE SIGNS LIST 1 Battery degradation prediction system 102 Current value acquisition unit 106 OCP acquisition unit 107 Electrode resistance acquisition unit 108 Polarization calculation unit 109 CCP calculation unit 110 Capacity prediction unit 111 Learning unit 116 Mechanical change calculation unit 117 Lithium precipitation prediction unit 21 Secondary battery

Claims (11)

an electrode resistance acquisition unit (107) that acquires a negative electrode resistance (R _a ) and a positive electrode resistance (R _c ) of the secondary battery (21);
A current value acquisition unit (102) that acquires a current value (I) of the secondary battery;
an OCP acquisition unit (106) that acquires an open circuit potential (OCP _a ) of a negative electrode of the secondary battery and an open circuit potential (OCP _c ) of a positive electrode of the secondary battery;
a polarization calculation unit (108) that calculates a negative electrode polarization (ΔV _a ) and a positive electrode polarization (ΔV _c ) from the negative electrode resistance and the positive electrode resistance acquired by the electrode resistance acquisition unit and the current value of the secondary battery acquired by the current value acquisition unit;
a CCP calculation unit (109) that calculates a closed circuit potential (CCP _a ) of the negative electrode based on the open circuit potential of the negative electrode acquired by the OCP acquisition unit and the negative electrode polarization calculated by the polarization calculation unit, and calculates a closed circuit potential (CCP _c ) of the positive electrode based on the open circuit potential of the positive electrode acquired by the OCP acquisition unit and the positive electrode polarization calculated by the polarization calculation unit;
and a capacity prediction unit (110) that predicts the negative electrode capacity (Q _a ), positive electrode capacity (Q _c ), and positive/negative electrode SOC deviation capacity (Q _Li ) of the secondary battery based on at least one of the negative electrode closed circuit potential and the positive electrode closed circuit potential calculated in the CCP calculation unit, and predicts the battery capacity (Q _B ) of the secondary battery based on the predicted negative electrode capacity, positive electrode capacity, and positive/negative electrode SOC deviation capacity. the electrode resistance acquisition unit and the capacity prediction unit calculate the negative electrode resistance and the negative electrode capacity, respectively, using a plurality of functions ( _gA , _gB , _gC , _iA , _iB , _iC ) correlated with a plurality of deterioration factors in a negative electrode of the secondary battery, and calculate the positive electrode resistance and the positive electrode capacity, respectively, using a plurality of functions ( _hA , _hB , _hC , _jA , _jB , _jC ) correlated with a plurality of deterioration factors in a positive electrode of the secondary battery;
2. The battery degradation prediction system according to claim 1, wherein the capacity prediction unit calculates the positive and negative electrode SOC deviation capacities using a plurality of functions ( _kA , _kB , _kC , _lA , _lB , _lC ) correlated with a plurality of degradation factors in an electrolyte of the secondary battery. an estimation unit that estimates data that could not be acquired during a data unacquired period in which data necessary for acquiring values of variables included in the plurality of functions could not be acquired;
3. The battery degradation prediction system according to claim 2, wherein the electrode resistance acquisition unit and the capacity prediction unit calculate the negative electrode resistance, the negative electrode capacity, the positive electrode resistance, the positive electrode capacity, and the positive and negative electrode SOC deviation capacity during the data no-acquisition period, using values of the variables acquired based on the data estimated by the estimation unit. a change amount calculation unit (105) that calculates a change amount (ΔDOD) of a state of charge (SOC) of the secondary battery based on the current value of the secondary battery acquired by the current value acquisition unit,
The electrode resistance acquisition unit and the capacity prediction unit predict at least one of the negative electrode capacity, the positive electrode capacity, the positive and negative electrode SOC deviation capacity, the negative electrode resistance, and the positive electrode resistance based on a pulverization function f (t, ΔDOD) = A × exp {-B × exp (C × ΔDOD) × t}, in which the elapsed time t from when the secondary battery was started to be used and the change amount ΔDOD of the charging rate of the secondary battery acquired by the change amount calculation unit are variables, and A, B, and C are constants. The battery deterioration prediction system according to any one of claims 1 to 3. The battery deterioration prediction system according to claim 4, further comprising a learning unit (111) that updates the constants A, B, and C of the pulverization function through learning. The battery degradation prediction system according to any one of claims 1 to 4, further comprising a _learning unit ( ₁₁₁ ) that updates, through learning, a prediction formula for calculating or predicting at least one of the negative electrode resistance (R _a ), positive electrode resistance (R _c ), negative electrode capacity (Q a ), positive electrode capacity (Q _c ), positive/negative electrode SOC deviation capacity (Q Li ), and battery capacity (Q _B ) of the secondary battery. The battery deterioration prediction system according to claim 5 or 6, wherein the learning unit performs the learning using a usage history of the secondary battery. The battery deterioration prediction system according to any one of claims 5 to 7, wherein at least a part of the battery deterioration prediction system other than the learning unit is mounted on a vehicle and used, and the learning unit is provided outside the vehicle. A memory unit (115) that stores initial OCP characteristics of the secondary battery in an initial state, the initial OCP characteristics being a relationship between a charging rate (SOC) and the open circuit potential of the negative electrode, and a relationship between the charging rate and the open circuit potential of the positive electrode;
an OCV characteristic calculation unit (112) that calculates an updated OCV characteristic, which is a relationship between the charging rate and an open circuit voltage (OCV) of the secondary battery after deterioration, based on the negative electrode capacity, the positive electrode capacity, and the positive and negative electrode SOC deviation capacity predicted by the capacity prediction unit and the initial OCV characteristic stored in the storage unit;
a differential characteristic calculation unit (113) that calculates a differential characteristic that is a relationship between the charging rate of the secondary battery and a differential value obtained by first-order or second-order differentiation of the open circuit voltage of the secondary battery with the charging rate of the secondary battery, based on the updated OCV characteristic calculated by the OCV characteristic calculation unit;
The battery degradation prediction system according to any one of claims 1 to 8, further comprising a current control unit (114) that limits a current flowing through the secondary battery in a region of the charging rate where the peak of the differential value is equal to or greater than a threshold value, based on the differential characteristic calculated by the differential characteristic calculation unit. The battery degradation prediction system according to any one of claims 1 to 9, further comprising a mechanical change calculation unit ( ₁₁₆ ) that calculates a relationship value related to a mechanical change of the secondary battery based on the positive and negative electrode SOC deviation capacity (Q Li ) predicted by the capacity prediction unit. The secondary battery is mounted on a vehicle,
The battery deterioration prediction system according to any one of claims 1 to 10, wherein at least a part of the battery deterioration prediction system is provided outside the vehicle.

Images