2000 Volume 66 Issue 650 Pages 3258-3263
Gear drive is one of important parts on plant machinery. However, up to now, the method of failure diagnosis of gear drive has not been clarified theoretically, and the condition monitoring of gear drive has been carried out mainly by statistical way and experience. So the main problems are : (1)the failure signal is hardly extracted from measured signal for early diagnoisis, (2)the failure type can not been easily recognized. In onder to overcome the difficulties, in this study we develop the diagnosis theory for gear drive by clarifying the change of dynamic characteristics between normal and abnormal states, and using theories of signal processing, fuzzy and neural network etc. In the 1st report, we point out the problems of the traditional diagnosis method using only the dynamic model of normal state, and clarify the dynamic characteristics on failure gears by the locus analysis of the meshing contact point, and show the results of analysis and experiment on eccentric gear to verify the efficiency of the method proposed in this paper.