Vibration cancellation at multiple locations using neurocontrollers with real-time learning
- Resource Type
- Conference
- Authors
- Bozich, D.J.; MacKay, H.B.
- Source
- IJCNN-91-Seattle International Joint Conference on Neural Networks Neural Networks, 1991., IJCNN-91-Seattle International Joint Conference on. ii:775-780 vol.2 1991
- Subject
- Computing and Processing
Components, Circuits, Devices and Systems
Signal Processing and Analysis
Neurocontrollers
Vibration control
Artificial neural networks
Adaptive control
Programmable control
MIMO
Machinery
Instruments
Damping
Time domain analysis
- Language
The simultaneous control of vibration at multiple locations on a vibrating machine or structure has been accomplished through the application of artificial neural networks. Examples are discussed which demonstrate the simultaneous control of up to four vibration generators on a structure to reduce vibrations at up to four arbitrary locations. The real-time neurocontroller described utilizes multiple-input-multiple-output (MIMO) neural networks. Vibration sensors mounted at the desired cancellation locations are not necessarily colocated with the vibration generators. The neurocontroller adaptively learns to control the generators such that equal and opposite waveforms are applied at the sensor locations, reducing the existing vibration waveforms at the sensors. This results in true time-domain waveform cancellation. Reductions of up to 20 dB+or-10 dB of both discrete-frequency and swept-frequency vibrations have been observed.ETX