A Study of the Gray-ANFIS Tool Life Prediction Model for Machine Differential Compensation

• Main Authors: Yu-Ting Ye, 葉禹廷
• Other Authors: Po-Tsang Huang
• Format: Others
• Language: zh-TW
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/u89546

碩士 === 中原大學 === 工業與系統工程研究所 === 106 === 　The prediction of CNC machining quality is regarded as a very important part in process management.Through forecasting, it is possible to avoid the occurrence of defective products in advance and save the steps of stopping processing and measuring.However, today''s factories have different types of processing machines in the same process, which has the potential to have different process capabilities, resulting in different machine conditions. This study is in response to the current small number of processing types and hopes to achieve immediate results.Therefore, a ANFIS model is constructed with a small number of surface roughness and vibration sensing signals, and the processing quality of other machines is predicted through learning correction. Furthermore, the tool life is predicted by the gray theory.In this study, two sets of experimental parameters of different processing parameters were used to input the Gray-ANFIS model.After learning and compensating , ANFIS predict surface roughness MAPE is 8.415 and 8.613,In addition, the accuracy of the number of records that can be processed is 90.0% and 93.75%.It is confirmed that the prediction model proposed in this study is accurate and feasible.