System and method for predicting remaining useful life of component of equipment
A system and a method for predicting a remaining useful life (RUL) of a component of an equipment are provided. The system for predicting the RUL of the component of the equipment includes a data acquisition unit, a feature capturing unit, a mapping function generating unit, a similarity analyzing unit and a RUL calculating unit. The feature capturing unit obtains an estimation feature according to a real time sensing record, and obtains a plurality of training features according to a set of history sensing records. The similarity analyzing unit obtains k similar features which are similar to the estimation feature according to the training features. The RUL calculating unit obtains at least one of k predicting information via a mapping function according to the k similar features and calculates an estimation RUL according to at least one predicting value.
1. A system for predicting a remaining useful life (RUL) of a component of an equipment, comprising:
a data acquisition (DAQ) unit for obtaining a real time sensing record;
a feature capturing unit for obtaining an estimation feature according to the real time sensing record, and obtaining a plurality of training features according to a plurality of history sensing records;
a mapping function generating unit for obtaining a mapping function according to the training features, wherein the mapping function includes a relationship between the training features and at least one predicting value;
a similarity analyzing unit for obtaining k similar features which are similar to the estimation feature according to the training features; and
a RUL calculating unit for obtaining the at least one predicting value via the mapping function according to the k similar features and calculating an estimation RUL according to the at least one predicting value;
wherein the equipment is a semiconductor equipment.
2. The system according to claim 1 , wherein the feature capturing unit further comprises:
a dictionary training machine for generating at least one basis set according to the history sensing records; and
a sparse coder for generating the training features according to the history sensing records by linearly combining the at least one basis set, and generating the estimation feature according to the real time sensing record by linearly combining the at least one basis set.
3. The system according to claim 1 , wherein the at least one predicting value includes a remaining number of runs.
4. The system according to claim 1 , wherein the at least one predicting value includes a remaining life percentage.
5. The system according to claim 1 , wherein the at least one predicting value includes a cycle identification which represents one operation cycle.
6. The system according to claim 1 , wherein the RUL calculating unit calculates the estimation RUL by a weighted average calculation on the at least one predicting value.
7. The system according to claim 6 , wherein a weighting in the weighted average calculation is inversely proportional to the at least one predicting value.
8. The system according to claim 6 , wherein a weighting in the weighted average calculation is directly proportional to the at least one predicting value.
9. The system according to claim 1 , wherein the RUL calculating unit further calculates an estimation confidence which is related to a variability of the at least one predicting value.
10. The system according to claim 9 , wherein the RUL calculating unit adjusts the estimation RUL according to the estimation confidence.
11. A method for predicting a remaining useful life (RUL) of a component of an equipment, comprising:
obtaining a real time sensing record;
obtaining an estimation feature according to the real time sensing record, and obtaining a plurality of training features according to a plurality of history sensing records;
obtaining a mapping function according to the training features, wherein the mapping function includes a relationship between the training features and at least one predicting value;
obtaining k similar features which are similar to the estimation feature according to the training features; and
obtaining the at least one predicting value via the mapping function according to the k similar features and calculating an estimation RUL according to the at least one predicting value.
12. The method according to claim 11 , wherein the equipment is a semiconductor equipment.
13. The method according to claim 11 , further comprising:
generating at least one basis set according to the history sensing records; and
generating the training features according to the history sensing records by linearly combining the at least one basis set, and generating the estimation feature according to the real time sensing record by linearly combining the at least one basis set.
14. The method according to claim 11 , wherein the at least one predicting value includes a remaining number of runs.
15. The method according to claim 14 , wherein a weighting in the weighted average calculation is inversely proportional to the at least one predicting value.
16. The method according to claim 14 , wherein a weight in the weighted average calculation is directly proportional to the at least one predicting value.
17. The method according to claim 16 , further comprising:
adjusting the estimation RUL according to the estimation confidence.
18. The method according to claim 11 , wherein the at least one predicting value includes a remaining life percentage.
19. The method according to claim 11 , wherein the at least one predicting value includes a cycle identification which represents one operation cycle.
20. The method according to claim 11 , wherein in the step of calculating the estimation RUL, the estimation RUL is calculated by performing a weighted average calculation on the at least one predicting value.
21. The method according to claim 11 , further comprising:
calculating an estimation confidence which is related to a variability of the at least one predicting value.