Doctor of Philosophy
Department of Mechanical Engineering
Welgama, Palitha Sumeda, Computer aided design of manufacturing facilities, Doctor of Philosophy thesis, Department of Mechanical Engineering, University of Wollongong, 1993. https://ro.uow.edu.au/theses/1575
Manufacturing facilities design includes the determination of layout and materials handling system. An optimum facilities design improves the efficiency of manufacturing processes through reduction of materials handling cost. A comprehensive investigation into the use of computer aided techniques in manufacturing facilities design has been carried out during this research.
During the early stages of the study, a real-life industrial facilities design problem heavy manufacturing environment was analysed. This provided an insight into factors considered important in practice, yet ignored by computer aided models and algorithms in literature. The role of Monte-carlo simulation methodology in industrial facilities design was thoroughly investigated, as it is widely used for practical facilities design problems. The simulation methodology was applied to the case-study problem, to study the performance of two alternative layouts, under operating dynamics using SIMAN/CINEMA. The analysis confirmed that simulation methodology is a useful technique which can be used to complement optimisation techniques for industrial facilities design. A new way of modelling batch processing was developed as part of the simulation study.
The main focus of this research was to develop knowledge-based / optimisation algorithms that consider more factors that are important in practice. A new algorithm was developed for the determination of machine layouts based on a bi-criterion optimisation model. The algorithm considers machine dimensions, their configurations and orientations of pick-up and drop-off points. These are very important considerations in a heavy industrial environment for determining an optimum layout. Minimising dead-space in the layout was considered as an objective, in addition to minimising flow-cost, leading to a useful way to obtain compact layouts. This methodology is more appropriate for heavy manufacturing environments. The method is useful for determining layouts when cost of transport is proportional to distance moved.
The graph-theoretic approach for determining layouts was investigated. This enabled a better understanding of the strengths and weaknesses of the approach. A new knowledgebased system was developed to computerise the conversion of a dual graph into a block layout, for which a sound methodology was not available. This system ensures a regular block layout, while attempting to satisfy specified adjacencies as far as possible.
The problem of materials handling equipment selection is an important part of industrial facilities design. This was investigated and resulted in development of a new knowledge-based / optimisation system. The knowledge base developed consists of facts and rules required to determine feasible materials handling equipment for a particular move. The optimisation algorithm attempts to minimise total cost and total aisle space requirements. The system is implemented using LPA PROLOG and integrates optimisation approaches and knowledge-based approaches into a single system.
The highly complex, yet very important, problem of joint determination of layout and materials handling system was attempted. This resulted in a new knowledge-based / optimisation system. The system is an integration of the above two methodologies developed for the determination of layout and materials handling system. This new system provides detailed information on machine layout, machine configurations and orientations of pick-up and drop-off points, materials handling equipment to be used, design load carrying capacities and move assignment. A comparative analysis was made between the joint determination, and the sequential determination of layout and the materials handling systems. The results confirm that the joint determination provides superior solutions in terms of total costs, but at the expense of computer time.
Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.