Predictive Model for Ultrasonic Slitting of Glass Using Feed Forward Back Propagation Artificial Neural Network

The prediction of process performance is essential to select the control parameters for obtaining the goals of production. Ultrasonic machining is popular material removal process brittle materials like glass, ceramics etc. Glass is a widely used engineering material in number of engineering applications like microscopy, optics etc. In this paper, experiments are conducted to obtain data regarding the effect of process parameters on ultrasonic slitting in common glass. Amplitude, pressure and thickness of the glass sheet are chosen as control parameters. Three levels of each of these parameters are selected giving 33 = 27 trials. Material removal rate (MRR), overcut (OC), taper produced on the slits are determined as response parameters. Artificial Neural Network (ANN) model is developed to capture relationship between control and response parameters as a predictive tool to predict the performance of the process.