Neural Networks and Statistical based Models for Surface Roughness Prediction

D. Villaseñor, R. Morales-Menéndez, C. Rodríguez (Mexico), and J.R. Alique (Spain)


Surface Roughness, Modelling, Highspeed machining, Neural Networks.


This paper describes an approach to predict Surface Rough ness in a High Speed end-milling process. Artificial neural networks and statistical tools were used to develop different surface roughness predictors. The models were developed based on information gathered from a monitoring system with accelerometers and data acquisition cards. Process pa rameters such as depth of cut, spindle speed, feed per tooth and machine dynamical vibration values have been used to build different models which predict, with high efficiency, the resultant surface quality of 6061 aluminum parts. Both surface roughness models are compared showing promis ing results.

Important Links:

Go Back