Accuracy improvement for CNC system using wavelet-neural networks

Abstract

Wavelet neural networks are investigated for learning a multidimensional input-output complex nonlinear function. The CNC turning process is modeled using wavelet neural networks. The error on the component is different from the desired dimensions because of the dynamics of the machining system. The error, if predicted, a priori can be used for compensating the same thus improving the accuracy in the part. In this work, wavelet neural networks are employed to predict the error given the process conditions as the input. Simulation studies are carried out to arrive for the selection of a suitable wavelet function for the CNC turning system in particular.