Abstract
A method of formulation and solution is presented for an optimal design problem of the measuring subsystem in a class of fault diagnosis systems where the signal-processing and information-processing subsystems are specified and the relationship among a measuring, signal-processing and information-processing subsystems are connected in series. The optimal measuring subsystem is so defined that minimizes its cost satisfying the conditions for the diagnostic reliability and the diagnostic accuracy.
The usefulness of this method is demonstrated by applying it to a fault diagnosis system for a test plant, where the signals from the plant are classified into “too high”, “normal” and “too low” and the fault diagnosis is performed using the signed digraph.
