제어로봇시스템학회:학술대회논문집
- 2001.10a
- /
- Pages.157.1-157
- /
- 2001
Diagnosis of Processing Equipment Using Neural Network Recognition of Radio Frequency Impedance Matching
Abstract
A new methodology is presented to diagnose faults in equipment plasma. This is accomplished by using neural networks as a pattern recognizer of radio frequency(rf) impedance match data. Using a realtime match monitor system, the match data were collected. The monitor system consisted mainly of a multifunction board and a signal flow diagram coded by Visual Designer. Plasma anomaly was effectively represented by electrical match positions. Twenty sets of fault-symptom patterns were experimentally simulated with experimental variations in process factors, which include rf source power, pressure, Ar and O
Keywords