Physical resource management in large computer-controlled manufacturing systems
Abstract
Although the advantages of flexible automation in the medium and small batch production sectors of the metal cutting industry have come to be well recognized over the past two decades, traditional machine shops, comprised of hundreds of conventional stand-alone machines, today still account for a large share of manufacturing output. Automation of these facilities would transform them into large computer-controlled manufacturing systems (LCMSs) which consist of 50 and more intelligent CNC machines integrated by a multi-faceted material handling system. One of the primary reasons for the lack of LCMSs today is the complexity of the control system required to operate large systems in a dynamic environment. This research, therefore, aims at providing a better understanding of the operation of LCMSs. Earlier work on LCMSs has shown that auction-based control schemes not only reduce the complexity of coordinating shop-floor activities, but also facilitate reconfiguration and make the system robust to failure. However, a notable deficiency in extant research on auction-based control is the lack of consideration of cutting tool related issues. This research, therefore, investigates the impact of cutting tool management strategies on the operation of LCMSs. In particular, existing negotiation schemes have been modified to include cutting tool entities and tool carrier entities, and to allow multi-level negotiation among system entities for resource sharing (for instance, inter-machine negotiation for sharing of cutting tools). Under these modified schemes, cutting-tool availability and tool procurement delays are important factors that influence participation in auctions and bid construction. Consequently, the tool management strategy directly influences the performance of the manufacturing system. This research, therefore, examines the effectiveness of various tool allocation schemes, tool sharing policies, tool handling system configurations, and bid construction methods. This research also clarifies the variety of negotiation schemes that are possible and provides a concise grammatical representation of these alternative schemes. A graph-theoretic methodology for tool cluster formation is developed. The requirements and an architecture of the information system necessary to support auction-based control are also identified.
Degree
Ph.D.
Advisors
Barash, Purdue University.
Subject Area
Industrial engineering|Mechanical engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.