적응형뉴로퍼지시스템을 이용한 인코넬718 밀링가공시 표면상태감시

Inconel 718 is a typical difficult-to-cut material, has low machinability. The condition of the machined surface is easily deteriorated by rapid tool wear caused by high temperature. Therefore, it is important to monitor the state of the machined surface by considering abnormal phenomena in machining. In this study, surface roughness is predicted using an ANFIS (Adaptive Neuro Fuzzy Inference System). In wet machining, there is a limit to using sensors for signal acquisition. Thus, the cutting force and acceleration signals are obtained using only a dynamometer and acceleration sensor in the cutting process. The cutting condition and acquisition of signals were selected as input variables and the learning data for ANFIS was set according to these parameters. The ANFIS algorithm was optimized by back-propagation learning and the surface roughness value was predicted through this. The results of this study can be used to monitor surface roughness in real-time machining. KCI Citation Count: 0