Gear noise emitted from transmission units for automobiles should be evaluated by an expert of gear noise. Although on some production lines, quietness performance estimates from measured noise levels of the units, the estimation could be severe. The reason is that it is almost impossible to determine a definite relationship between the measured noise levels and the evaluations by the experts. Therefore, for standing on the safe side, the estimation should be severe. As a result, such an automatic gear noise diagnosis system must yield transmission units with over-quality, and then must induced high cost. The present study deals with a new gear noise diagnosis system to which an artificial intelligence is applied; i.e., the neural net is applied). In order to use the new gear noise diagnosis system, the system must be trained with the evaluations by gear noise experts. In the training process, the system iteratively changes connection intensity between neurons in its own neural net for the system evaluations reaching to agreement with those by the experts. The convergence of this process depends on input data for the system, the structure of the neural net, and so on. In the present paper, conditions were determined for the convergence. Then, the judgement performance of the developed system was discussed. As a result, the system can realize over 70% performance of noise evaluation experts.
Gear noise diagnosis system for automobile transmission using artifical intelligence (Convergence of training process)
Getriebegeräuschdiagnosesystem für Fahrzeugantriebe unter Einsatz künstlicher Intelligenz (Konvergenz des Lernprozesses)
2001
8 Seiten, 11 Bilder, 4 Tabellen, 3 Quellen
Aufsatz (Konferenz)
Englisch
SAE Technical Papers | 2001
|British Library Conference Proceedings | 2001
|Artifical intelligence in the factory today
Kraftfahrwesen | 1987
|Artifical intelligence applications to traffic engineering
TIBKAT | 1994
|