Non-parametric neuro-model of a flexible beam structure

In this paper, the development of dynamic model of flexible cantilever (fixed-free) beam in transverse motion using FD approach is presented. Validation using theoretical value is carried out in order to ensure the reliability of the developed FD algorithm. Next, system identification using non-parametric Neural Network (NN) methods: Multilayer Perceptron (MLP) and ELMAN networks are developed. To suppress the unwanted vibration, simulated case studies of AVC using P and PI control schemes are investigated. Results demonstrate that PI control methods outperformed P controller in cancelling the vibration.