Prediction of surface roughness and dimensional deviation by measuring cutting forces and vibrations in turning process

Prediction of surface finish and dimensional deviation is an essential prerequisite for developing an unmanned turning center. It is also important for optimization of turning process. In this work, it is found that, using neural network, surface finish can be predicted within a reasonable degree of accuracy by taking the acceleration of radial vibration of tool holder as a feedback. It is also possible to utilize the fitted network for predicting the surface finish in turning with a tool of same material but different geometry provided coolant situation is same. For that purpose, only few experiments are needed with the new tool to modify the neural network predicted results. However, different neural network models have to be fitted for dry and wet turning, as well as for turning by HSS and carbide tools. It was observed that while turning the steel rod with TiN coated carbide tool, surface finish improves with increasing feed up to some feed where from it starts deteriorating with further increase of feed. This type of behavior is not observed in turning with HSS tool. Dimensional deviation is significant only in the case of turning of a slender work-piece. Hence, neural network prediction models are developed separately for that. Radial component of cutting force and acceleration of radial vibration were taken as a feedback to predict dimensional deviation. The performance of the developed neural network models is assessed by carrying out a number of experiments involving dry and wet turning of mild steel rods using HSS and carbide tools.