Laser cutting quality control of melamine using artificial neural networks

Experimental analysis has been carried out to seek the optimum combination (cutting speed, laser power, assist pressure of air and standoff distance) of input controllable variables in the process of laser cutting in order to improve the laser cutting quality on non-metallic such as Urea formaldehyde (Melamine). Furthermore, the values of edge quality, kerf widths, percent overcut and material removal rate were measured for calculating quality. Taguchi method was used in experimental design using orthogonal array. The effect of input parameters on output quality variation was assessed by analysis of variance to determine the optimum combination of input. Artificial neural network can measure and improve the quality of cutting by training on aggregation data, using feed-forward back-propagation to predict overall cutting quality. Simulation of aggregated function can be used for better optimization than ANOVA technique because it provides the overall quality prediction, as against single quality prediction.