Development of a fuzzy-nets-based surface roughness prediction system in turning operations

This paper discusses the development of a surface roughness prediction system for a turning operation, using a fuzzy-nets modeling technique. The goal is to develop and train a fuzzy-nets-based surface roughness prediction (FN-SRP) system that will predict the surface roughness of a turned workpiece using accelerometer measurements of turning parameters and vibration data. The FN-SRP system has been developed using a computer numerical control (CNC) slant-bed lathe with a carbide cutting tool. The system was trained using feed rate, spindle speed, and tangential vibration data collected during experimental runs. A series of validation runs indicate that this system has a mean accuracy of 95%.