ASA 125th Meeting Ottawa 1993 May

1pSA1. Monitoring, diagnosing, and actively controlling of complex systems using acoustic signatures and artificial neural networks.

A. Unal

Cogswell Polytech. College, 10420 Bubb Rd., Cupertino, CA 95014

Artificial neural network-based technology enables us to monitor complex systems in real time with reasonable efficiency. Acoustic signatures and/or their transforms can be utilized for diagnosis with less false alarms by exploiting the robust feature extraction abilities of the artificial neural nets. Active controlling requires the instantaneous sensing and actuating abilities that cannot be obtained otherwise. Only the use of artificial neural nets enables us to annihilate the excessive vibrational fields in complex systems. Examples from real life will be given.