OPTIMIZATION OF ACTIVE VIBRATION CONTROL USING ARTIFICIAL NEURAL NETWORK

Abstract

The cantilever beam is a bolted construction that is supported at one end. This structure finds applications in many fields like aeronautical, bridges, etc. The major limitation with the cantilever beam is the vibration produced. The primary purpose of this effort is analysing the Active Vibration Control (AVC) in cantilever platesusing smart detectors and using Artificial Neural Network (ANN) to minimize the vibration.In this work, the cantilever plate is embedded with small piezoelectric materials thatbehave as sensors and actuators controlling the movement of plates. The features defined for cantilever plates are finite length, breadth, thickness, location, and tip displacement. The cantilever beam is analysed withthefinite element method. The controller used for the reduction of vibration is the LQR controller with a Kalman filter combined with a fuzzy controllerandANN controller. The ANN is trained, tested, and validated for the different combinations of features of the cantilever plate. Thus for a different combination of inputs, the settling time of vibration and amplitude are determined. Thus from the analysis, it can be observed that the combination with ANN controller performance is better compared to the other controllers; hence, reducing the vibration efficiently.