The influence of cutting parameters on the surface quality of routed paper birch and surface roughness prediction modeling.

The objective of this study was to characterize the routing process to better understand the machining conditions that affect surface finish. Experiments were designed to determine the impact of cutting depth, feed speed, and grain orientation of the workpiece on the surface quality of paper birch wood. Statistical analysis showed that the cutting depth did not influence surface finish. Roughness depended greatly on feed speed and grain orientation, increasing linearly as the feed speed increased. The roughest surfaces were obtained by routing against the grain between 120 and 135° grain orientation, depending on the feed speed. Two models able to predict the surface finish based on initial cutting parameters were developed and compared. Both the statistical regression and neural network models were subjected to a validation procedure in which their performance was confirmed using data that were not used for the learning process. Results indicated that the neural network system estimates the surface roughness with less error than the statistical regression model.