Date of Award
Doctor of Philosophy (PhD)
Hoda A. Elmaraghy
Launching new manufacturing systems for production is a difficult task which involves several people and resources, and is time-consuming and capital intensive. Flexible manufacturing systems (FMS), which are a new generation of manufacturing systems, further complicate this launching process because they consist of several interacting components such as workstations, pallets, automated material handling systems, and buffers. The lack of knowledge about the interaction between these components has often resulted in the poor performance of many FMS designs. Major corporations today are adopting a proactive philosophy rather than reactive one, and for a complex and capital intensive manufacturing system such as an FMS, the design process is extremely critical for its successful performance.
The design process, which precedes the launching of an FMS, is iterative and consists of planning, model development, and output analysis. While several computer-based tools are available for modeling and evaluating FMS designs, here is a dearth of such tools for synthesizing new design and analyzing output from simulation models to improve these designs. System designers have traditionally performed these two activities using their knowledge and experience. These are cumbersome and time-consuming activities and consequently increase the design cycle time.
The objective of this thesis is to reduce the FMS design cycle time by automating the design process. A knowledge-based system called FMX, Flexible Manufacturing Expert, was developed in this thesis to generate and evaluate designs of manufacturing systems with a high degree of automation. The object-oriented FMX is an intelligent system which combines expert systems and simulation modeling to design and evaluate flexible manufacturing systems. It consists of several modules, the major ones including an expert design synthesizer to generate initial system designs, a simulation models, and an expert analyzer to analyze simulation output, identity design deficiencies, and recommend changes. The current implementation of FMX was developed using the Knowledge Engineering Environment (KEE) expert system shell and the SimKit simulation package, and it runs on a Sun workstation under a UNIX operating system.
FMX has been applied for designing flexible manufacturing systems in the machining domain. Two case studies have been included in this thesis to demonstrate its capabilities-one of which includes industrial system used by a major automotive manufacturer in Ontario. The first case study took two iterations while the second one took three iterations to fine tune a synthesized design that satisfied various performance measures such as production volumes, equipment utilizations, equipment blocking, and queue length of automatic pallet changers. For both case studies, the investment cost per part index decreased for each iteration thus proving that the design changes suggested by the expert analyzer improved the performance of the initial system design while reducing the cost. The final output from FMX is a list of components that comprise the system, its graphical layout, and its performance measures.
FMX is a comprehensive decision support system which integrates all phases of a flexible manufacturing system design into a single software framework. Its modules such as the expert design synthesizer, simulation model developer, and expert analyzer each address a specific phase of the system design process. FMX generates consistent designs with the minimum intervention from the user and hence reduces the design cycle time significantly.
Ravi, Thiruvengadam, "A Knowledge-Based Approach to The Design, Simulation, and Evaluation of Flexible Manufacturing Systems" (1994). Open Access Dissertations and Theses. Paper 3010.