Method for predicting life of rotary machine and determining repair timing of rotary machine

1. A method for predicting life of rotary machine comprising:

reading a load recipe of loading conditions of a rotary machine;

determining whether changes exist for said loading conditions by comparing said load recipe with an already existing load recipe for the process;

employing an already existing determination reference if no changes exist for said loading conditions, and reading in and employing a determination reference accommodating said process conditions if changes exist for said loading conditions instead of said already existing determination reference;

processing time series data by reading in detected characterizing feature data for said rotary machine, which correspond to said determination reference; and

calculating life expectancy of said rotary machine in conformity with said time series data and said detennination reference.

2. The method of claim 1 , wherein said loading conditions are defined by process conditions of a chamber where gas is introduced and a semiconductor manufacturing process is performed, and said rotary machine is a vacuum pump.

3. The method of claim 2 , wherein another semiconductor manufacturing process having different object from said semiconductor manufacturing process is further performed in said chamber.

4. A method of claim 2 , wherein said process conditions include kind of said gas and flow rate of said gas.

5. The method of claim 2 , wherein said characterizing fcature data include one of voltage, current, power, temperature, vibration, and sound of said vacuum pump.

6. The method of claim 2 , wherein said calculated life expectancy is fed to a local area network.

7. The method of claim 2 , wherein based on said calculated life expectancy, an emergency shutoff signal for said process is provided.

8. The method of claim 2 , wherein calculation of said life expectancy uses time series data of said characterizing feature data.

9. The method of claim 8 , wherein calculation of said life expectancy uses any one of average value, standard deviation, auto correlation coefficient, lag width of an auto correlation coefficient, and statistical analysis value of a Mahalanobis-Taguchi distance of a limited time segment in said time series data for said characterizing feature data.

10. The method of claim 8 , wherein said calculation of life expectancy is derived from speed of change of said statistical analysis value, a simple regression or multiple regression analysis.

11. A method for determining timing of rotary machine repair comprising:

reading a load recipe of loading conditions of a rotary machine;

determining whether changes exist for said loading conditions by comparing said load recipe with an already existing load recipe for the process;

employing an already existing determination reference if no changes exist for said loading conditions, and reading in and employing a determination reference accommodating said process conditions if changes exist for said loading conditions instead of said already existing determination reference;

processing time series data by reading in detected characterizing feature data for said rotary machine, which correspond to said determination reference;

calculating life expectancy of said rotary machine in conformity with said time series data and said determination reference;

finding stand-by times of the process in a time period until said calculated life expectancy is reached, by a semiconductor production simulator; and

determining a stand-by time, of said found stand-by times, which least affects said process or a time including said stand-by time, to be the replacement or repair time of said rotary machine.

12. The method of claim 11 , wherein said loading conditions are defined by process conditions of a chamber where gas is introduced and a semiconductor manufacturing process is performed, and said rotary machine is a vacuum pump.

13. The method of claim 12 , wherein another semiconductor manufacturing process having different object from said semiconductor manufacturing process is further performed in said chamber.

14. The method of claim 12 , wherein said process conditions include kind of said gas and flow rate of said gas.

15. The method of claim 12 , wherein said characterizing feature data include one of voltage, current, power, temperature, vibration, and sound of said vacuum pump.

16. The method of claim 12 , wherein calculation of said life expectancy uses time series data of said characterizing feature data.

17. The method of claim 16 , wherein calculation of said life expectancy uses any one of average value, standard deviation, auto correlation coefficient, lag width of an auto correlation coefficient, and statistical analysis value of a Mahalanobis-Taguchi distance of a limited time segment in said time series data for said characterizing feature data.

18. The method of claim 16 , wherein said calculation of life expectancy is derived from speed of change of said statistical analysis value, a simple regression or multiple regression analysis.