Battery model construction method and battery degradation prediction device

1 . A battery model construction method for constructing a battery model associating values of a plurality of first input parameters and a plurality of second input parameters correlated with a degradation indicator for a battery with a predicted value of the degradation indicator, the battery model construction method comprising:

a data acquiring step of acquiring data related to measured values of the first and second input parameters and the degradation indicator by using a plurality of sample batteries;

a constructing step of constructing the battery model by using at least a portion of the data acquired in the data acquiring step as training data;

an error trend determining step of evaluating, for each sample battery, a prediction error between the measured value of the degradation indicator and the predicted value of the degradation indicator according to the battery model, and determining whether there is significant bias in the prediction error inherent for each sample battery;

a first correlation determining step of using the training data to construct a first error prediction model associating an explanatory variable defined on a basis of the first input parameters with an objective variable corresponding to a predicted value of the prediction error acquired in the error trend determining step, and determining whether a first correlation exists between the measured value of the prediction error and the predicted value of the prediction error according to the first error prediction model; and

a first reconstruction step of reconstructing the battery model in a case where it is determined in the error trend determining step that there is the significant bias and it is determined in the first correlation determining step that the first correlation exists.

2 . The battery model construction method according to claim 1 , further comprising:

a second correlation determining step of using the training data to construct a second error prediction model associating an explanatory variable defined on a basis of the second input parameters with an objective variable corresponding to a predicted value of the prediction error, and determining whether a second correlation exists between the measured value of the prediction error acquired in the error trend determining step and the predicted value of the prediction error according to the second error prediction model; and

a second reconstruction step of reconstructing the battery model in the case where it is determined in the error trend determining step that there is the significant bias and it is determined in the second correlation determining step that the second correlation exists.

3 . The battery model construction method according to claim 2 , wherein

the first input parameters are usage history parameters defined on a basis of time series data about a current, a voltage, and a temperature of the battery, and

the second input parameters are manufacturing history parameters defined on a basis of data from a time of manufacture of the battery.

4 . The battery model construction method according to claim 3 , wherein in the first reconstruction step, the battery model is reconstructed by using the first error prediction model.

5 . The battery model construction method according to claim 4 , wherein in the second reconstruction step, the battery model is reconstructed by using the second error prediction model.

6 . The battery model construction method according to claim 3 , wherein in the second reconstruction step, the battery model is reconstructed by using the second error prediction model.

7 . The battery model construction method according to claim 5 , wherein

the time series data related to measured values of the first and second input parameters and the degradation indicator acquired in the data acquiring step is divided into the training data that belongs to a prescribed training period and verification data that belongs to a verification period subsequent to the training period, and

in the error trend determining step, the presence or absence of the significant bias is determined on a basis of an evaluation result, for each sample battery, regarding a distribution of the prediction error in the training period and an evaluation result regarding a correlation between the prediction error in the training period and the prediction error in the verification period.

8 . The battery model construction method according to claim 5 , further comprising:

a third reconstruction step of reconstructing the battery model by adding an offset term that outputs a constant value according to a cumulative use time of the battery in the case where it is determined in the error trend determining step that there is the significant bias and it is determined in the first and second correlation determining steps that neither the first nor the second correlation exists.

9 . The battery model construction method according to claim 4 , wherein

the time series data related to measured values of the first and second input parameters and the degradation indicator acquired in the data acquiring step is divided into the training data that belongs to a prescribed training period and verification data that belongs to a verification period subsequent to the training period, and

in the error trend determining step, the presence or absence of the significant bias is determined on a basis of an evaluation result, for each sample battery, regarding a distribution of the prediction error in the training period and an evaluation result regarding a correlation between the prediction error in the training period and the prediction error in the verification period.

10 . The battery model construction method according to claim 4 , further comprising:

a third reconstruction step of reconstructing the battery model by adding an offset term that outputs a constant value according to a cumulative use time of the battery in the case where it is determined in the error trend determining step that there is the significant bias and it is determined in the first and second correlation determining steps that neither the first nor the second correlation exists.

11 . The battery model construction method according to claim 3 , wherein

the time series data related to measured values of the first and second input parameters and the degradation indicator acquired in the data acquiring step is divided into the training data that belongs to a prescribed training period and verification data that belongs to a verification period subsequent to the training period, and

in the error trend determining step, the presence or absence of the significant bias is determined on a basis of an evaluation result, for each sample battery, regarding a distribution of the prediction error in the training period and an evaluation result regarding a correlation between the prediction error in the training period and the prediction error in the verification period.

12 . The battery model construction method according to claim 3 , further comprising:

a third reconstruction step of reconstructing the battery model by adding an offset term that outputs a constant value according to a cumulative use time of the battery in the case where it is determined in the error trend determining step that there is the significant bias and it is determined in the first and second correlation determining steps that neither the first nor the second correlation exists.

13 . A battery degradation prediction device comprising:

a processor configured to:

acquire each value of a plurality of first input parameters and a plurality of second input parameters correlated with a degradation indicator for a battery; and

calculate a predicted value of the degradation indicator by inputting the values of the first and second input parameters into a battery model constructed according to the battery model construction method according to claim 3 .

14 . A battery degradation prediction device comprising:

a processor configured to:

acquire each value of a plurality of first input parameters and a plurality of second input parameters correlated with a degradation indicator for a battery; and

calculate a predicted value of the degradation indicator by inputting the values of the first and second input parameters into a battery model constructed according to the battery model construction method according to claim 1 .

15 . A battery degradation prediction device comprising:

a processor configured to:

acquire each value of a plurality of first input parameters and a plurality of second input parameters correlated with a degradation indicator for a battery; and

calculate a predicted value of the degradation indicator by inputting the values of the first and second input parameters into a battery model constructed according to the battery model construction method according to claim 2 .