A Novel Multiple RBF-NN Model for Rolling Force Prediction Based on Anti-Aliasing Wavelet Analysis

Abstract

Aiming at the problem of rolling force prediction, an anti-aliasing wavelet method utilizing FFT and IFFT is first proposed to decompose the rolling force signal and reconstruct it as a serial of sub-components. Then several RBF networks, each with different input and output parameters, are established for the modeling of each sub-component, their output values are added up as the rolling force value. Simulation results show that this proposed model can reduce the system dimension, and improve the learning ability of the network. The error rate of rolling force prediction is reduced from 10% by BP-NN model to 4% by the proposed model.