JP7225896B2 - battery monitoring system - Google Patents

Description

The present invention relates to a battery monitoring system for diagnosing failures of secondary batteries.

2. Description of the Related Art Conventionally, a battery monitoring system for diagnosing a failure of a secondary battery is known (see Patent Document 1 below). In this battery monitoring system, for example, the open circuit voltage (OCV: Open Circuit Voltage) of the secondary battery, the charging rate (SOC: State Of
Charge) are measured periodically and these measurements are stored cumulatively. Then, when the newly measured relationship between OCV and SOC changes significantly from the previously measured relationship between OCV and SOC, it is determined that the secondary battery has failed.

Another example of fault determination is a method of measuring the capacity of a secondary battery. In this case, the secondary battery is first completely discharged and then fully charged. Then, the capacity of the secondary battery is measured by measuring the charge amount required for charging. If the newly measured value of capacity changes significantly from the value measured in the past, it is determined that the secondary battery has failed.

JP 2016-152704 A

However, in the method of measuring OCV and SOC and judging whether or not the relationship between them has changed compared to past measurements, it is necessary to cumulatively store the measured values of OCV and SOC. Therefore, the amount of data to be stored is large, and a large storage device is required. In addition, failure determination is difficult until the secondary battery has completely failed. That is, it is difficult to detect the initial failure of the battery. In addition, the initial failure means the initial stage of the failure when the failure occurs in the battery.

Moreover, when measuring the capacity of a secondary battery and determining a failure, there is a problem that it takes a long time to measure the capacity.

The present invention has been made in view of such problems, and provides a battery monitoring system capable of detecting an initial failure of a secondary battery, reducing the amount of data to be stored, and making failure determination in a short time. is trying to provide.

One aspect of the present invention is a data acquisition unit (3) that acquires a plurality of types of monitoring data (X ₁ to X _n ) for monitoring the state of a secondary battery (2);
A failure determination unit (4) that determines whether the secondary battery has failed,
The failure determination unit
A matrix calculation unit (41) for calculating a partial correlation coefficient matrix (Λ) of the monitoring data by performing sparse regularization using the monitoring data as a variable;
The degree of abnormality for calculating the amount of change in the partial correlation coefficient (λ), which is the component of the partial correlation coefficient matrix, between the two partial correlation coefficient matrices calculated at different times as the degree of abnormality (Δ) a calculator (42),
A battery monitoring system (1) configured to determine that the secondary battery has failed when the calculated degree of abnormality exceeds a predetermined threshold value (Δ _TH ).

The failure determination unit of the battery monitoring system performs sparse regularization using the monitoring data of the secondary battery as variables to calculate a partial correlation coefficient matrix. Then, the amount of change in the partial correlation coefficients between the two partial correlation coefficient matrices acquired at different times is calculated as the degree of abnormality. If the degree of abnormality exceeds the threshold, it is determined that the secondary battery has failed.
By doing so, an initial failure of the secondary battery can be detected. That is, when sparse regularization is performed, two types of monitoring data that are highly related to each other can be selected from a plurality of types of secondary battery monitoring data. That is, the absolute value of the partial correlation coefficient approaches 1 for two types of monitoring data that are highly related. Also, the partial correlation coefficients of two types of monitoring data with low relevance approach zero. Therefore, two partial correlation coefficient matrices are calculated at different times, and if the partial correlation coefficients contained in these matrices change significantly, it is assumed that the correlation between the two types of highly related monitoring data has collapsed. means. Therefore, in this case, it can be determined that some kind of failure has occurred in the secondary battery. In particular, since the partial correlation coefficient of a secondary battery changes significantly even when an initial failure occurs, the change in the partial correlation coefficient can be used to detect the initial failure of the secondary battery.

Further, the above battery monitoring system does not need to store all the monitoring data acquired in the past, and can perform failure detection by storing the calculated partial correlation coefficient matrix. Moreover, it is not necessary to store all past partial correlation coefficient matrices, and it is only necessary to store the most recently obtained partial correlation coefficient matrix. Therefore, the amount of data to be stored can be reduced. Furthermore, the above-mentioned battery monitoring system can determine the failure of the secondary battery in a short period of time compared to the conventional case of measuring the capacity of the secondary battery.

As described above, according to the above aspect, it is possible to provide a battery monitoring system that can detect an initial failure of a secondary battery, can reduce the amount of data to be stored, and can perform failure determination in a short period of time. .
It should be noted that the symbols in parentheses described in the claims and the means for solving the problems indicate the corresponding relationship with the specific means described in the embodiments described later, and limit the technical scope of the present invention. not a thing

1 is a circuit diagram of a battery monitoring system according to Embodiment 1. FIG. 4 is a graph showing changes over time in monitoring data according to the first embodiment; A visualization of the partial correlation coefficient matrix Λ ₁ in Embodiment 1. FIG. A visualization of the partial correlation coefficient matrix Λ ₂ in Embodiment 1. FIG. A visualization of the partial correlation coefficient matrix Λ _n in Embodiment 1. FIG. 4 is a graph of the degree of abnormality determined from two partial correlation coefficient matrices Λ ₁ and Λ ₂ in Embodiment 1. FIG. 4 is a graph of the degree of anomaly determined from two partial correlation coefficient matrices Λ _n-1 and Λ _n in Embodiment 1. FIG. 4 is a conceptual diagram of the secondary battery during discharge in Embodiment 1. FIG. 1 is a conceptual diagram of a secondary battery being charged according to Embodiment 1. FIG. 4 is a flowchart of a control unit according to the first embodiment; 11 is a flowchart following FIG. 10; 6 is a conceptual diagram of a battery monitoring system in which a storage unit is arranged outside the vehicle according to Embodiment 2. FIG. FIG. 7 is a conceptual diagram of a battery monitoring system in which a failure determination unit is arranged outside the vehicle according to Embodiment 2; FIG. 7 is a conceptual diagram of a battery monitoring system in which a storage unit and a failure determination unit are arranged outside the vehicle according to Embodiment 2; 4A and 4B are conceptual diagrams showing temporal changes in the battery performance of a secondary battery in the case where a normal state continues and in the case where a failure occurs.

(Embodiment 1)
An embodiment of the battery monitoring system will be described with reference to FIGS. 1 to 11. FIG. The battery monitoring system 1 of this embodiment includes a data acquisition section 3 and a failure determination section 4 as shown in FIG. The data acquisition unit 3 acquires a plurality of types of monitoring data X ₁ to X _n (see FIG. 2) for monitoring the state of the secondary battery 2 . Failure determination unit 4 determines whether or not secondary battery 2 has failed.

The failure determination unit 4 includes a matrix calculation unit 41 and an abnormality degree calculation unit 42 . The matrix calculator 41 performs sparse regularization using the monitoring data as variables. Thereby, the partial correlation coefficient matrix Λ of the monitoring data is calculated (see FIGS. 10 and 11). In addition, the degree-of-abnormality calculator 42 calculates, as the degree of abnormality Δ, the amount of change in the partial correlation coefficient λ between the two partial correlation coefficient matrices Λ obtained at different times. The failure determination unit 4 determines that the secondary battery 2 has failed when the calculated degree of abnormality Δ exceeds a predetermined threshold Δ _TH .

A battery monitoring system 1 of the present embodiment is an in-vehicle battery monitoring system to be installed in a vehicle such as an electric vehicle or a hybrid vehicle. As shown in FIG. 1, the secondary battery 2 is connected with a load 10 and a charging device 11 . The load 10 of this embodiment is an inverter. Using this inverter, the DC power supplied from the secondary battery 2 is converted into AC power to drive a three-phase AC motor (not shown). This allows the vehicle to run.

A discharge switch 12 is arranged between the load 10 and the secondary battery 2 . A charging switch 13 is arranged between the charging device 11 and the secondary battery 2 . The on/off operations of these switches 12 and 13 are controlled by the controller 6 . The control unit 6 turns on the charging switch 13 when charging the secondary battery 2 and turns on the discharging switch 12 when driving the load 10 .

The data acquisition unit 3 of the present embodiment obtains CCV (closed circuit voltage), charge/discharge current I, SOC (state of charge), environmental temperature T _E around the secondary battery 2, battery temperature as monitoring data of the secondary battery 2. Obtain T _B , integrated temperature stress ΣT, integrated charging time Σt _c , integrated discharging time Σt _d , and integrated current ΣI.

The data acquisition unit 3 includes a current sensor _30A that measures the charging/discharging current I, a voltage sensor _30V , a battery temperature sensor _30TB that measures the battery temperature _TB , and an environmental temperature sensor _30TE that measures the environmental temperature _TE. , an integrator 30 _I and an SOC calculator 30 _S . A voltage sensor 30 _V measures the CCV and OCV of the secondary battery 2 . The SOC calculator 30 _S calculates the SOC of the secondary battery 2 using the measured OCV. Further, the accumulator 30 _I calculates the accumulated temperature stress ΣT, the accumulated charging time Σt _c , the accumulated discharging time Σt _d , and the accumulated current ΣI. The accumulated charging time Σt _c is the accumulated value of the charging time t _c of the secondary battery 2 , and the accumulated discharging time Σt _d is the accumulated value of the discharging time t _d of the secondary battery 2 . The integrated current ΣI is the integrated value of the charge/discharge current I. Also, the integrated temperature stress ΣT is an integrated value of time at each temperature during use. The integrated temperature stress ΣT can be calculated, for example, as {10° C.×time}+{15° C.×time}+ . . . +{45° C.×time}. Also, since the higher the temperature, the more stress is applied to the secondary battery 2, each time can be weighted. Note that the method for calculating the integrated temperature stress ΣT is not limited to this. In addition, for example, a method of counting and accumulating only the time spent at 40° C. or higher can be adopted.

Moreover, the battery monitoring system 1 of the present embodiment includes a storage unit 5 . The storage unit 5 stores monitoring data necessary for calculating the partial correlation coefficient matrix Λ and the calculated partial correlation coefficient matrix Λ.

Next, the structure of the secondary battery 2 will be described. As shown in FIG. 8, the secondary battery 2 includes a positive electrode 21 _P and a negative electrode 21 _N , a separator 24 interposed therebetween, and an electrolytic solution 25 . The positive electrode 21 _P and the negative electrode 21 _N include metal conductive portions 23 (23 _P , 23 _N ) and active materials 22 (22 _P , 22 _N ) attached to the conductive portions 23 . The secondary battery 2 of this embodiment is a lithium ion secondary battery.

As shown in FIG. 8, when the SOC of the secondary battery 2 is approximately 100%, most of the lithium ions are present in the active material _22N of the negative _{electrode 21N} . When discharged, lithium ions move to the active material _22P of the positive _{electrode 21P} . Moreover, as shown in FIG. 9, when the SOC of the secondary battery 2 is approximately 0%, most of the lithium ions are present in the active material _22P of the positive _{electrode 21P} . During charging, lithium ions move to the active material _22N of the negative electrode _21N .

If the secondary battery 2 is subjected to external impact or the like, the secondary battery 2 may fail. For example, the conductive portions 23 (23 _P , 23 _N ) of the individual electrodes 21 (21 _P , 21 _N ) may come into contact with each other, or the active material 22 may separate from the conductive portions 23 . In addition, when the secondary battery 2 is used for a long period of time, metal lithium may be deposited in the electrolyte 25 and the pair of electrodes 21 may be short-circuited. The failure determination unit 4 of this embodiment determines whether or not such a failure has occurred in the secondary battery 2 .

Next, a method of diagnosing a failure of the secondary battery 2 will be described with reference to FIGS. 2 to 7. FIG. Here, as shown in FIG. 2, the case of using six types of monitoring data X ₁ to X ₆ will be described. For example, the first monitoring data X ₁ is the closed circuit voltage, the second monitoring data X ₂ is the charging/discharging current I, the third monitoring data X ₃ is the charging rate, the fourth monitoring data X ₄ is the environmental temperature T _E , and the fifth monitoring data The data X ₅ can be the battery temperature T _B , and the sixth monitoring data X ₆ can be the cumulative charging time Σt _c . Note that another type of monitoring data X may be used, and the arrangement order thereof may be arbitrary. In addition, six types of monitoring data X ₁ to X ₆ are used in FIG. 2 and the like, but the present invention is not limited to this, and two or more types may be used.

As shown in FIG. 2, each of the monitoring data X ₁ -X ₆ changes over time. The failure determination unit 4 first performs sparse regularization using the monitoring data X ₁ to X ₆ acquired in the first period t ₁ as variables to calculate the first partial correlation coefficient matrix Λ ₁ . The first partial correlation coefficient matrix Λ ₁ can be expressed as follows, for example.

In the above formula, λ ₁₂ means the partial correlation coefficient between monitoring data X ₁ and monitoring data X ₂ . Also, since the partial correlation coefficient matrix Λ is a symmetric matrix, part of the partial correlation coefficient λ is omitted in the above equation. Furthermore, since all the components on the main diagonal of the partial correlation coefficient matrix Λ are 1, the description is omitted.

The partial correlation coefficient λ approaches 1 or −1 when the correlation between two types of monitoring data is high. Also, when the correlation is low, it approaches 0. A visualization of the first partial correlation coefficient matrix Λ ₁ is shown in FIG. As shown in the figure, two types of highly correlated monitoring data (for example, _X2 and _X6 , _X5 and _X6, etc.) are connected by relatively thick lines. Monitoring data with lower correlation (for example, X ₃ and X ₆ ) are connected by a slightly thinner line. Monitoring data with even lower correlation (eg, X ₂ and X ₃ , X ₃ and X ₄ , etc.) are connected by thinner lines.

Next, the failure determination unit 4 performs sparse regularization again using the monitoring data X ₁ to X ₆ acquired in the second period t ₂ (see FIG. 1) as variables, and obtains the second partial correlation coefficient matrix Λ ₂ as calculate. The second partial correlation coefficient matrix Λ ₂ can be expressed, for example, by the following formula.

A visualization of the second partial correlation coefficient matrix Λ ₂ is shown in FIG. If the correlations of the monitoring data X ₁ to X ₆ do not change significantly, the two partial correlation coefficient matrices Λ ₁ and Λ ₂ will be substantially the same. Therefore, graphs (see FIGS. 3 and 4) that respectively visualize the two partial correlation coefficient matrices Λ ₁ and Λ ₂ are substantially the same.

After calculating the two partial correlation coefficient matrices Λ ₁ and Λ ₂ in this manner, the failure determination unit 4 determines the change in the partial correlation coefficient λ between these two partial correlation coefficient matrices Λ ₁ and Λ ₂ . The amount is calculated as the degree of anomaly Δ. The degree of anomaly Δ can be expressed, for example, as follows.

Δ ₁₂ means the degree of abnormality Δ of the partial correlation coefficient λ ₁₂ between the first monitoring data X ₁ and the second monitoring data X ₂ .

FIG. 6 shows a graph of individual anomaly degrees Δ. As described above, the two calculated partial correlation coefficient matrices Λ ₁ and Λ ₂ are almost the same during the period t ₁ -t ₂ when the correlation of the monitoring data X ₁ -X ₆ does not change significantly. Therefore, the change in the partial correlation coefficient λ is small, and the degree of abnormality Δ becomes a small value. If all the degrees of abnormality Δ are smaller than the threshold value Δ _TH , the failure judgment unit 4 judges that the correlation of the monitoring data X ₁ to X ₆ has not changed, that is, the secondary battery 2 has not failed. .

Next, FIG. 5 shows a graph visualizing the partial correlation coefficient matrix Λ when the secondary battery 2 fails. In this graph, the correlation between monitoring data _X3 and _X5 and the correlation between monitoring data _X3 and _X6 are lower than in FIGS. 3 and 4, and are almost zero. FIG. 7 shows a graph of the degree of abnormality Δ at this time. As shown in the figure, the abnormality degree _Δ35 of the monitoring data _X3 and _X5 and the abnormality degree _Δ36 of the monitoring data _X3 and _X6 are high. When the calculated degree of abnormality Δ exceeds the threshold value Δ _TH , the failure judgment unit 4 judges that the correlation between the monitoring data X ₁ to X ₆ has greatly changed, that is, the secondary battery 2 has failed. More specifically, the failure determination unit 4 determines that the secondary battery 2 has failed when at least one of the calculated degrees of abnormality Δ ₁₂ to Δ ₅₆ exceeds the threshold Δ _TH .

The failure determination unit 4 of this embodiment periodically (for example, once a day) calculates the partial correlation coefficient matrix Λ. Then, using two partial correlation coefficient matrices Λ _n and Λ _n-1 , the latest partial correlation coefficient matrix Λ _n calculated and the partial correlation coefficient matrix Λ _n-1 calculated immediately before, The degree of abnormality Δ is calculated. If at least one of the calculated degrees of abnormality Δ ₁₂ to Δ ₅₆ exceeds the threshold Δ _TH , it is determined that the secondary battery 2 has failed.

Next, the flow chart of the failure determination section 4 will be described. As shown in FIG. 10, the failure determination unit 4 of this embodiment first performs step S1. Here, the data acquisition unit 3 is used to acquire monitoring data X ₁ to X _n for a certain period of time. After that, the process moves to step S2. Here, the measured monitoring data X ₁ to X _n are used to calculate the first partial correlation coefficient matrix Λ ₁ .

After that, the process moves to step S3 and counts for a predetermined period. Next, in step S4, the monitoring data X ₁ to X _n are measured again for a certain period. After that, the process moves to step S5. Here, newly measured monitoring data X ₁ to X _n are used to calculate the second partial correlation coefficient matrix Λ ₂ .

After that, the process moves to step S6. Here, the degree of anomaly Δ is calculated using two partial correlation coefficient matrices Λ ₁ and Λ ₂ . After that, the process moves to step S7, and it is determined whether or not any one of the plurality of degrees of abnormality Δ is higher than the threshold Δ _TH . If the determination is Yes (see FIG. 7), the process moves to step S8, and it is determined that the secondary battery 2 is out of order.

Further, when it is determined No in step S7, step S9 and subsequent steps shown in FIG. 11 are performed. That is, the monitoring data X ₁ to X _n are acquired periodically (step S9), and the partial correlation coefficient matrix Λ _n calculated using the latest monitoring data and the partial correlation coefficient matrix Λ _n calculated immediately before ₋₁ is used to calculate the degree of abnormality Δ (steps S10 and S11). When the degree of abnormality Δ exceeds the threshold Δ _TH (step S12), it is determined that the secondary battery 2 has failed (step S13). If it is determined No (the degree of abnormality Δ does not exceed the threshold Δ _TH ) in step S12, the process returns to step S9.

Next, the effect of this form is demonstrated. The failure determination unit 4 of this embodiment performs sparse regularization using the monitoring data X ₁ to X _n of the secondary battery 2 as variables to calculate the partial correlation coefficient matrix Λ. Then, the amount of change in the partial correlation coefficient λ between the two partial correlation coefficient matrices Λ _n-1 and Λ _n calculated at different times is calculated as the degree of abnormality Δ. When the degree of abnormality Δ exceeds the threshold Δ _TH , it is determined that the secondary battery 2 is out of order.
By doing so, an initial failure of the secondary battery 2 can be detected.
FIG. 15 is a diagram conceptually showing temporal changes in the battery performance of the secondary battery 2 in the case where the normal state continues and in the case where a failure occurs. When the secondary battery 2 maintains a normal state, the battery performance gradually declines from L0 to L1 in FIG. On the other hand, when a failure occurs in the secondary battery 2, the battery performance is degraded so as to shift from L0 to L2 in FIG.
Therefore, when determining failure by measuring deterioration of battery performance (for example, decrease in capacity, increase in resistance) as in the prior art, it is difficult to determine failure until battery performance has significantly decreased. Therefore, even if a failure occurs in the secondary battery 2, in the initial stage of failure indicated by Fi in FIG. 15, the change in battery performance is small, making failure determination difficult. In other words, it is difficult to determine the failure until the failure approaches complete failure indicated by Fc in the figure.
On the other hand, in the present embodiment, as described above, failure determination can be performed at the initial failure stage Fi by using the degree of abnormality Δ. That is, since the degree of abnormality Δ can be detected even at the initial stage Fi of the failure, it is possible to determine the failure at the initial stage Fi before the secondary battery 2 drops significantly. In other words, the initial stage of failure of the secondary battery 2 can be detected. In this specification, "initial failure" means the initial stage of this failure.
That is, as described above, when sparse regularization is performed, two types of monitoring data highly related to each other can be selected from the plurality of types of monitoring data X ₁ to X _n of the secondary battery 2 . That is, the absolute value of the partial correlation coefficient λ is close to 1 for two types of monitoring data that are highly related. Also, two types of monitoring data with low relevance have partial correlation coefficients λ close to zero. Therefore, two partial correlation coefficient matrices Λ _n-1 and Λ _n are calculated at different times, and if the partial correlation coefficients λ and λ ^' included in these matrices change significantly, two This means that the correlation between the types of monitoring data has broken down. Therefore, in this case, it can be determined that some kind of failure has occurred in the secondary battery 2 . In particular, since the secondary battery 2 undergoes a large change in the partial correlation coefficient λ even when an initial failure occurs, the initial failure of the secondary battery 2 can be detected by using the change in the partial correlation coefficient λ. can.

Moreover, the battery monitoring system 1 of the present embodiment does not need to store all the monitoring data acquired in the past, and can perform failure detection by storing the calculated partial correlation coefficient matrix Λ. Moreover, it is not necessary to store all past partial correlation coefficient matrices Λ, and it is only necessary to store the most recently obtained partial correlation coefficient matrix Λ. Therefore, the amount of data to be stored can be reduced. Furthermore, the battery monitoring system 1 of this embodiment can determine the failure of the secondary battery 2 in a short time compared to the conventional case of measuring the capacity of the secondary battery 2 .

Also, the matrix calculator 41 of the present embodiment periodically calculates the partial correlation coefficient matrix Λ. The degree-of-abnormality calculation unit 42 calculates two partial correlation coefficient matrices Λ _{n - 1} , Λ Calculate the degree of anomaly Δ using _n .
By doing so, it is possible to calculate the degree of anomaly Δ by storing only the previously acquired partial correlation coefficient matrix Λ _n−1 . Therefore, the previously acquired partial correlation coefficient matrix Λ need not be stored in the storage unit 5 and can be deleted. Therefore, the storage unit 5 having a small storage capacity can be used, and the manufacturing cost of the battery monitoring system 1 can be reduced.

In addition, the matrix calculation unit 41 of the present embodiment uses CCV ( _closed circuit voltage ₎ , charging/discharging current I, SOC (charging rate), battery temperature T _B and are used at least to calculate the partial correlation coefficient matrix Λ.
These monitoring data are highly correlated with each other. Therefore, when these monitoring data are used, the partial correlation coefficient λ changes greatly when the secondary battery 2 fails. Therefore, it is easy to detect a failure of the secondary battery 2 .

As described above, according to the present embodiment, it is possible to provide a battery monitoring system capable of detecting an initial failure of a secondary battery, reducing the amount of data to be stored, and determining a failure in a short period of time. .

In the present embodiment, the degree of abnormality Δ is calculated using the latest partial correlation coefficient matrix Λ _n and the previously calculated partial correlation coefficient matrix Λ _n-1 , but the present invention is limited to this. isn't it. That is, the degree of abnormality Δ may be calculated using the latest partial correlation coefficient matrix Λ _n and the partial correlation coefficient matrix Λ calculated two or more times before. For example, when calculating the partial correlation coefficient matrix Λ once a day, the latest partial correlation coefficient matrix Λ _n and the partial correlation coefficient matrix Λ obtained a few days ago are used to calculate the degree of abnormality Δ. can do. In this way, even if the failure of the secondary battery 2 progresses gradually, the degree of anomaly Δ tends to increase because comparison is made with the partial correlation coefficient matrix Λ from several days ago. Therefore, failure of the secondary battery 2 can be reliably detected.

Also, in FIG. 2, etc., a 6×6 partial correlation coefficient matrix Λ is calculated using six types of monitoring data X ₁ to X ₆ , and a plurality of anomaly degrees Δ ₁₂ to Δ ₅₆ (see FIG. 7) are calculated from this. is calculated, and it is determined that the secondary battery 2 has failed when at least one of the plurality of anomaly degrees Δ ₁₁ to Δ ₅₆ exceeds the threshold value Δ _TH , but the present invention is limited to this. isn't it. That is, a 2×2 partial correlation coefficient matrix Λ may be calculated using two types of monitoring data X ₁ and X ₂ , and only one abnormality degree Δ ₁₂ may be calculated therefrom.

In the following embodiments, among the reference numerals used in the drawings, the same reference numerals as those used in the first embodiment represent the same components and the like as those of the first embodiment, unless otherwise indicated.

(Embodiment 2)
This embodiment is an example in which the configuration of the battery monitoring system 1 is changed. In this embodiment, at least one of the storage unit 5 and the failure determination unit 4 is arranged outside the vehicle. For example, as shown in FIG. 12, the secondary battery 2, the data acquisition unit 3, and the failure determination unit 4 can be arranged in the vehicle 7, and the storage unit 5 can be provided in the external device 8 (eg, server). The external device 8 can be installed, for example, at a dealer of the vehicle 7 or the like. Further, the vehicle 7 is provided with a transmitter/receiver 71 . The transmitter/receiver 71 is used to transmit the monitoring data and the partial correlation coefficient matrix Λ to the storage unit 5 and receive them from the storage unit 5 .

Similarly, as shown in FIG. 13, the secondary battery 2, the data acquisition unit 3, and the storage unit 5 can be arranged in the vehicle 7, and the failure judgment unit 4 can be arranged outside the vehicle. Also, as shown in FIG. 14, both the failure judgment section 4 and the storage section 5 can be arranged outside the vehicle.

The effects of this embodiment will be described. With the above configuration, the storage unit 5 and the failure determination unit 4 can be arranged outside the vehicle, so the number of parts mounted on the vehicle 7 can be reduced, and the weight of the vehicle 7 can be easily reduced. Further, the failure judgment unit 4 can be configured by a device with a high calculation speed, such as a server. Therefore, failure determination can be performed in a shorter time.
In addition, it has the same configuration and effects as those of the first embodiment.

The present invention is not limited to the above embodiments, and can be applied to various embodiments without departing from the scope of the invention.

1 battery monitoring system 3 data acquisition unit 4 failure determination unit 41 matrix calculation unit 42 abnormality degree calculation unit X ₁ to X _n monitoring data Λ partial correlation coefficient matrix λ partial correlation coefficient Δ degree of abnormality

Claims (4)

a data acquisition unit (3) for acquiring a plurality of types of monitoring data (X ₁ to X _n ) for monitoring the state of the secondary battery (2);
A failure determination unit (4) that determines whether the secondary battery has failed,
The failure determination unit
A matrix calculation unit (41) for calculating a partial correlation coefficient matrix (Λ) of the monitoring data by performing sparse regularization using the monitoring data as a variable;
The degree of abnormality for calculating the amount of change in the partial correlation coefficient (λ), which is the component of the partial correlation coefficient matrix, between the two partial correlation coefficient matrices calculated at different times as the degree of abnormality (Δ) a calculator (42),
A battery monitoring system (1) configured to determine that the secondary battery has failed when the calculated degree of abnormality exceeds a predetermined threshold value (Δ _TH ). The matrix calculation unit periodically calculates the partial correlation coefficient matrix, and the abnormality degree calculation unit calculates the latest partial correlation coefficient matrix (Λ _n ) and the partial correlation calculated immediately before that 2. The battery monitoring system according to claim 1, wherein the degree of abnormality is calculated using two partial correlation coefficient matrices and a number matrix ([Lambda] _n-1 ). A storage unit (5) for storing the calculated partial correlation coefficient matrix is further provided, and the secondary battery and the data acquisition unit are mounted on a vehicle (7), respectively, and the storage unit and the failure determination unit 3. The battery monitoring system according to claim 1, wherein at least one of and is arranged outside the vehicle. The matrix calculation unit uses at least the CCV of the secondary battery, the charge/discharge current (I), the SOC, and the battery temperature (T _B ) of the secondary battery as the monitoring data, and uses the phase deviation 4. The battery monitoring system according to any one of claims 1 to 3, configured to calculate a relational coefficient matrix.

Images