Automatic generation method of optimum symptom parameters for condition diagnosis of plant machinery by genetic algorithms

When using computers for the automatic condition diagnosis of plant machinery, symptom parameters (SP) extracted from some signals are indispensable. Currently, however there is no acceptable method for attracting the optimum SP. In order to overcome this difficulty and ensure highly accurate condition diagnosis, a new method called the “automatic generation of symptom parameters” is proposed by using genetic algorithms (GA). The authors have applied the method to many diagnoses of plant machinery and, in each case, the optimum SP has been quickly discovered. In this paper, they show the example of gear equipment diagnosis to verify the efficiency of this method