Date of Award
1-1991
Document Type
Dissertation - Restricted
Degree Name
Doctor of Philosophy (PhD)
Department
Electrical and Computer Engineering
First Advisor
Robert Stango
Second Advisor
Vikram Cariapa
Third Advisor
Stephen M. Heinrich
Fourth Advisor
James A. Heinen
Fifth Advisor
Ronald H. Brown
Abstract
Compliant tools, such as filamentary brushes, are being increasingly used in modern manufacturing facilities, especially in automated deburring and finishing operations. Although numerous advantages of brushing tools for deburring applications have been cited, the material removal mechanics and dynamic characteristics of such compliant tools are not well understood. It is recognized that there is an urgent need to formulate a process model for deburring with compliant tools. Furthermore, the control strategies for automatic deburring need to be developed in order to carry out advanced automatic deburring operation. In this research, process models have been obtained from brush/workpart dynamic systems and brush material removal characteristics. A process model for the dynamic system is established by employing a viscoelastic representation and Dynamic Data System (DDS) analysis. Parameters of the systems are identified by experimentally obtained dynamic response. The viscoelastic and first order DDS dynamic process models yield close agreement for poles of the system. Material removal characteristics are described by first order differential equations. Parameters of the system are evaluated by experimentally obtained "Master Curves" of dynamic force response. In addition, a multiple regression method and an Neural Network model are applied in modelling the burr height reduction. Control strategies have been developed for cases of burr with known constant height, known varying height, and unknown height. The closed-form solutions of transverse feed rate along the X axis have been obtained for both known constant height burr and known varying height burr. The feed rate is increased proportionally with increase of initial penetration depth. An equal-distance sampling technique was employed in the development of the closed-loop control system for burr with unknown height. The use of the concept of dividing the brush into many differential brushes has shown the success in the analysis for burr with unknown height. Computer simulations show the validity of control law developed as a result of burr totally removed, and indicate that Ka = 1.25 seems to be desired proportional gain value. Over control strategies developed show promise in the future implementation because of the simplicity of the strategies.