Abstract
This paper proposes a compact expert system with fuzzy inference on insulation diagnoses of rotating machines for elevators by means of a personal computer. The hypotheses used in this system for backward reasoning are derived from the results of the fuzzy inference, which specifies the deterioration grades from the automatically measured data for some diagnosis factors. The hypotheses are verified by human visual inspection, and the system synthetically diagnoses the objective system by the results of the verified hypotheses. The results of the diagnosis are shown as the messages how to treat the machines on the display. This system has the advantage as follows: breakdown voltages can be estimated with the regression line which means the correlation between deterioration grades obtained through the fuzzy inference and breakdown voltages. It is finally shown that the resultant diagnosis of the system agrees well with that of the experts. Then, It is concluded that the system is promising for assising insulation diagnosis of the rotating machines.
