Date of Award
Mechanical, Automotive, and Materials Engineering
CC BY-NC-ND 4.0
An expert system was developed to monitor the condition of drill bits to permit proper scheduling of tool replacement. Vibration was used as the sensing medium. The basic structure of the system consisted of the following five modules: design, monitoring, pattern and diagnosis, knowledge base, and update and learning. The design module provides the recommended operating conditions and the predicted tool-life. The monitoring module captures, digitizes and transforms the vibration signals from the drilling process. The pattern and diagnosis module is the most essential component of the system. It includes a routine that calculates four primary descriptors (features) and two secondary descriptors. It also implements two fuzzy algorithms to assess the tool condition. It was found that only one of the four primary descriptors is sufficient to develop the system and the tool replacement is made when the grade of 'unnormal' drilling is predominant. The two fuzzy algorithms are both capable of classifying the vibration signals with a percentage of success of greater than 83 and 90, respectively. The knowledge base contains all the essential knowledge needed in the system. The update and learning module updates the knowledge base when a tool change occurs. With the recent drilling data and through a reinforcement learning scheme, the system relearns the parameters of the fuzzy algorithms and the coefficients of the tool-life equations, so that the performance of the system can be gradually improved. Experimental measurements with a large number of drills of three sizes were performed on several machine tools. Each drill was run until failure. Vibration measurements were continuously obtained in order to relate observed changes to the progressive wear and ultimate failure of these tools. After the expert system was developed, its functioning effectiveness was tested by means of additional experimental test runs. Eight major computer programs were developed for the operation of the system.Dept. of Mechanical, Automotive, and Materials Engineering. Paper copy at Leddy Library: Theses & Major Papers - Basement, West Bldg. / Call Number: Thesis1992 .L385. Source: Dissertation Abstracts International, Volume: 54-05, Section: B, page: 2711. Thesis (Ph.D.)--University of Windsor (Canada), 1992.
Lau, Hui Kong., "An expert system to monitor drill condition." (1992). Electronic Theses and Dissertations. 1055.