Estimation system, estimation device and estimation method for estimating charging time

Abstract

The estimation system includes an estimation device and a charging management system. The charging management system transmits the current battery state data and current charging power detected from the current charging device to the estimation device. In the data preprocessing operation, the estimation device is used to: segment historical charging curves corresponding to multiple charging powers of multiple charging devices according to multiple energy ranges to generate multiple battery state segment data corresponding to multiple energy ranges; determine whether each battery state segment data belongs to normal or abnormal data; and merge the battery state segment data that belongs to normal data to generate multiple charging curves corresponding to multiple charging powers. In the charging time estimation operation, the estimation device calculates the estimated charging time based on the current charging power, current battery state data, and multiple charging curves.

Claims

1. A prediction system, comprising: a prediction device for performing a data preprocessing operation and a charging time prediction operation; and a charging management system connected between a current charging device and the prediction device among a plurality of charging devices, for transmitting current battery state data and a current charging power detected from the current charging device to the prediction device, wherein the plurality of charging devices have a plurality of charging powers, wherein in the data preprocessing operation, the prediction device is configured to: segment a plurality of historical charging curves corresponding to the plurality of charging powers of the plurality of charging devices according to a plurality of power ranges to generate a plurality of battery state segment data corresponding to the plurality of power ranges respectively; determine whether each of the plurality of battery state segment data belongs to normal data or abnormal data; and merge the plurality of battery state segment data belonging to the normal data to generate a plurality of charging curves corresponding to the plurality of charging powers of the plurality of charging devices respectively, and wherein in the charging time prediction operation, the prediction device is configured to: Based on the current charging power, the current battery status data, and the multiple charging curves, the estimated charging time of the current charging device is calculated and output to the charging management system.

2. The estimation system as described in claim 1, wherein in the data preprocessing operation, the estimation device is further configured to: divide the plurality of battery state segment data into a plurality of battery state segment data groups based on a segment charging time of each of the plurality of battery state segment data, wherein the segment charging times of the plurality of battery state segment data in the same group of the plurality of battery state segment data are similar to each other; and determine the plurality of battery state segment data of one of the plurality of battery state segment data groups as the normal data, and determine the plurality of battery state segment data of the others in the plurality of battery state segment data groups as the abnormal data, wherein the segment charging time of the plurality of battery state segment data of one of the plurality of battery state segment data groups is less than the segment charging time of the plurality of battery state segment data of the others in the plurality of battery state segment data groups.

3. The prediction system as described in claim 2, wherein in the data preprocessing operation, the prediction device is further configured to: calculate a time median and a time standard deviation based on the charging time of the segment of the plurality of battery state segment data from one of the plurality of battery state segment data groups; classify the plurality of battery state segment data whose segment charging time is within two of the time standard deviations from the time median as normal data; and classify the plurality of battery state segment data whose segment charging time is greater than or equal to two of the time standard deviations from the time median as abnormal data.

4. The prediction system as described in claim 1, wherein in the data preprocessing operation, the prediction device is further configured to use a Lasso regression method to generate the plurality of charging curves based on the plurality of battery state segment data.

5. The estimation system as described in claim 1, wherein in the charging time estimation operation, when a current charging time of the current charging device is less than the initial charging time, the estimation device is further configured to perform a static estimation operation, including: selecting one of the plurality of charging curves based on a device charging power of the current charging device; and calculating the estimated charging time based on the current battery state data and one of the plurality of charging curves; and when the current charging time of the current charging device exceeds the initial charging time, the estimation device is further configured to perform a dynamic estimation operation, including: selecting two of the plurality of charging curves based on the current charging power of the current charging device; and calculating the estimated charging time using interpolation or extrapolation based on the current battery state data and two of the plurality of charging curves.

6. The estimation system as described in claim 5, wherein the charging management system is used to detect the current charging power and the current battery status data of the current charging device once every detection interval, and the estimation device is further used to perform the charging time estimation operation once every detection interval.

7. The estimation system as described in claim 1, wherein the charging management system is configured to simultaneously generate a notification message related to the estimated charging time to a user device upon receiving the estimated charging time.

8. A prediction device for performing a data preprocessing operation and performing a charging time prediction operation on a current charging device among a plurality of charging devices to calculate a predicted charging time for the current charging device, wherein the plurality of charging devices have a plurality of charging powers, wherein in the data preprocessing operation, the prediction device is configured to: read a plurality of historical charging curves corresponding to the plurality of charging powers of the plurality of charging devices; segment the plurality of historical charging curves according to a plurality of power ranges to generate a plurality of battery state segment data corresponding to the plurality of power ranges respectively; determine whether each of the plurality of battery state segment data belongs to normal data or abnormal data; and merge the plurality of battery state segment data belonging to the normal data to generate a plurality of charging curves corresponding to the plurality of charging powers of the plurality of charging devices respectively, and wherein in the charging time prediction operation, the prediction device is configured to: calculate the predicted charging time based on a current charging power of the current charging device, a current battery state data and the plurality of charging curves.

9. The estimation apparatus as described in claim 8, wherein in the data preprocessing operation, the estimation apparatus is further configured to: divide the plurality of battery state segment data into a plurality of battery state segment data groups based on a segment charging time of each of the plurality of battery state segment data, wherein the segment charging times of the plurality of battery state segment data in the same group of the plurality of battery state segment data are similar to each other; and determine the plurality of battery state segment data of one of the plurality of battery state segment data groups as the normal data, and determine the plurality of battery state segment data of the others in the plurality of battery state segment data groups as the abnormal data, wherein the segment charging time of the plurality of battery state segment data of one of the plurality of battery state segment data groups is less than the segment charging time of the plurality of battery state segment data of the others in the plurality of battery state segment data groups.

10. The estimation apparatus as described in claim 9, wherein in the data preprocessing operation, the estimation apparatus is further configured to: calculate a time median and a time standard deviation based on the charging time of the segment of the plurality of battery state segment data from one of the plurality of battery state segment data groups; classify the plurality of battery state segment data whose segment charging time is within two of the time standard deviations from the time median as normal data; and classify the plurality of battery state segment data whose segment charging time is greater than or equal to two of the time standard deviations from the time median as abnormal data.

11. The estimation apparatus as described in claim 8, wherein in the data preprocessing operation, the estimation apparatus is further configured to use a Lasso regression method to generate the plurality of charging curves based on the plurality of battery state segment data.

12. The estimation device as described in claim 8, wherein in the charging time estimation operation, when a current charging time of the current charging device is less than the initial charging time, the estimation device is further configured to perform a static estimation operation, including: selecting one of the plurality of charging curves based on a device charging power of the current charging device; and calculating the estimated charging time based on the current battery state data and one of the plurality of charging curves; when the current charging time of the current charging device exceeds the initial charging time, the estimation device is further configured to perform a dynamic estimation operation, including: selecting two of the plurality of charging curves based on the current charging power of the current charging device; and calculating the estimated charging time using interpolation or extrapolation based on the current battery state data and two of the plurality of charging curves.

13. The estimation device as claimed in claim 12, wherein the estimation device is configured to perform the charging time estimation operation based on the current charging power of the current charging device, the current battery state data and the plurality of charging curves at each detection interval.

14. A prediction method applicable to a prediction system including a prediction device and a charging management system, comprising the following steps: performing a data preprocessing operation by the prediction device to generate a plurality of charging curves; detecting a current charging power and a current battery state data of a current charging device among a plurality of charging devices by the charging management system, wherein the plurality of charging devices have a plurality of charging powers; and performing a charging time prediction operation by the prediction device to output a predicted charging time, wherein the prediction device performing the data preprocessing operation comprises the following steps: reading a plurality of historical charging curves corresponding to the plurality of charging powers of the plurality of charging devices by the prediction device; and dividing the plurality of historical charging curves according to a plurality of power ranges by the prediction device to generate a plurality of battery state segment data corresponding to the plurality of power ranges respectively; The estimation device determines whether each of the multiple battery state data segments belongs to normal data or abnormal data; and the estimation device merges the multiple battery state data segments belonging to the normal data to generate multiple charging curves corresponding to the multiple charging powers of the multiple charging devices. The charging time estimation operation performed by the estimation device includes the following steps: receiving the current charging power and the current battery state data from the charging management system; and calculating and outputting the estimated charging time based on the current charging power, the current battery state data, and the multiple charging curves.

15. The estimation method as described in claim 14, wherein the step of segmenting the plurality of historical charging curves according to the plurality of power ranges by the estimation device includes the following steps: The estimation device divides the plurality of battery state segment data into a plurality of battery state segment data groups based on a segment charging time of each of the plurality of battery state segment data groups, wherein the segment charging times of the plurality of battery state segment data in the same group are similar to each other, and the step of determining whether the plurality of battery state segment data belongs to the normal data or the abnormal data by the estimation device includes the following steps: The estimation device determines the plurality of battery state segment data of one of the plurality of battery state segment data groups as the normal data; and the estimation device determines the plurality of battery state segment data of the others in the plurality of battery state segment data groups as the abnormal data. The charging time of one of the multiple battery status segment data groups is less than the charging time of the multiple battery status segment data of the other multiple battery status segment data groups.

16. The estimation method as described in claim 15, wherein the step of determining whether the plurality of battery state segment data belongs to the normal data or the abnormal data by means of the estimation device further comprises the following steps: by means of the estimation device, calculating a time median and a time standard deviation based on the segment charging time of the plurality of battery state segment data in one of the plurality of battery state segment data groups; by means of the estimation device, determining the plurality of battery state segment data corresponding to the segment charging time being within two time standard deviations from the time median as the normal data; and by means of the estimation device, determining the plurality of battery state segment data corresponding to the segment charging time being greater than or equal to two time standard deviations from the time median as the abnormal data.

17. The estimation method as described in claim 14, wherein the step of merging the plurality of battery state segment data belonging to the normal data by the estimation device to generate the plurality of charging curves corresponding to the plurality of charging powers of the plurality of charging devices includes the following steps: generating the plurality of charging curves based on the plurality of battery state segment data using a Lasso regression method by the estimation device.

18. The estimation method as described in claim 14, wherein the step of performing the charging time estimation operation by the estimation device further comprises the following steps: In response to a current charging time of the current charging device not exceeding the initial charging time, performing a static estimation operation by the estimation device, comprising the following steps: selecting one of the plurality of charging curves by the estimation device based on a device charging power of the current charging device; and calculating the estimated charging time by the estimation device based on the current battery state data and one of the plurality of charging curves; and in response to a current charging time of the current charging device exceeding the initial charging time, performing a dynamic estimation operation by the estimation device, comprising the following steps: selecting two of the plurality of charging curves by the estimation device based on the current charging power of the current charging device; and calculating the estimated charging time by the estimation device based on the current battery state data and two of the plurality of charging curves using interpolation or extrapolation.

19. The estimation method as described in claim 18, wherein the step of detecting the current charging power and the current battery status data of the current charging device by means of the charging management system includes: detecting the current charging power and the current battery status data of the current charging device once every detection interval by means of the charging management system; and the step of performing the charging time estimation operation by means of the estimation device further includes the following step: performing the dynamic estimation operation once every detection interval by means of the estimation device.

20. The estimation method as described in claim 14 further includes the following steps: In response to the charging management system receiving the estimated charging time, the charging management system synchronously generates a prompt message related to the estimated charging time to a user device.

Images