Fuzzy-nets based approach to using an accelerometer for an in-process surface roughness prediction system in milling operations

A fuzzy-nets based in-process surface roughness prediction (FN-ISRP) system based on the fuzzy-nets training scheme has been developed for predicting the surface roughness generated in milling operations while the machining process is taking place. In addition to the consideration of cutting parameters, such as spindle speed, feed rate, and depth of cut as fuzzy-nets input variables, this paper also describes the use of vibration in the FN-ISRP system. This cutting vibration was measured using an accelerometer and a proximity sensor. Five steps of the fuzzy-nets training scheme were implemented throughout the experiments, followed by the fuzzy rule bank, which was created based on physical experimentation. After the fuzzy rule bank was established, tests were conducted in a real-time fashion to evaluate the performance. In the fuzzy-nets model, Ra was predicted with a 96% accuracy rate, and the system could respond to the prediction value within 0.5 seconds during the end-milling process.