Diagnosis of tool wear based on regression analysis and fuzzy logic

Tool wear monitoring is important for a number of reasons. Automatic diagnosis of tool wear enables the unmanned use of flexible manufacturing systems and machine tools. Besides, a worn tool if unnoticed could cause a lot of damage, i.e. the machined products could be damaged and unfit for their planned use. As such the machining process is very challenging to monitor due to various reasons. Tool type and cutting parameters may vary resulting in variation of the monitored parameters. Also, there can be a lot of noise in the measured signals. The paper deals with the use of regression analysis techniques together with fuzzy logic in order to overcome the challenges in tool wear monitoring. Regression analysis, based on a higher order polynomial function that emphasizes the most recent measured data and has a limited number of terms, can very well follow and give prognosis of the development of the monitored parameters from such signals as vibration, sound and acoustic emission. The use of fuzzy logic makes it possible to automatically define limits for the monitored parameters and to combine the information from a number of signals. The proposed approach is tested with data from drilling tests.