This project is automating the diagnosis of faults in semiconductor manufacturing equipment. Statistical methods indicate when a machine is malfunctioning. The causes of malfunctions are then determined by a combination of numerical simulation of the machine and rule-based inference. The method is being tested on a vapor deposition furnace in the Berkeley Microfabrication Laboratory.
