Doctor of Philosophy (PhD)
A seam-welded steel pipe manufacturing process has mainly four distinct major design and/or operational problems dealing with buffer inventory, cutting tools, pipe sizing and inspection-rework facility. The general objective of this research is to optimally solve these four important problems to improve the throughput and yield of the system at a minimum cost.
The first problem of this research finds the optimal buffer capacity of steel strip coils to minimize the maintenance and downtime related costs. The total cost function for this coil feeding system is formulated as a constrained non-linear programming (NLP) problem which is solved with a search algorithm. The second problem aims at finding the optimal tool magazine reload timing, magazine size and the order quantity for the cutting tools. This tool magazine system is formulated as a mixed-integer NLP problem which is solved for minimizing the total cost. The third problem deals with different type of manufacturing defects. The profit function of this problem forms a binary integer NLP problem which involves multiple integrals with several exponential and discrete functions. An exhaustive search method is employed to find the optimum strategy for dealing with the defects and pipe sizing. The fourth problem pertains to the number of servers and floor space allocations for the off-line inspection-rework facility. The total cost function forms an integer NLP structure, which is minimized with a customized search algorithm.
In order to judge the impact of the above-mentioned problems, an overall equipment effectiveness (OEE) measure, coined as monetary loss based regression (MLBR) method, is also developed as the fifth problem to assess the performance of the entire manufacturing system. Finally, a numerical simulation of the entire process is conducted to illustrate the applications of the optimum parameters setting and to evaluate the overall effectiveness of the simulated system. The successful improvement of the simulated system supports this research to be implemented in a real manufacturing setup. Different pathways shown here for improving the throughput and yield of industrial systems reflect not only to the improvement of methodologies and techniques but also to the advancement of new technology and national economy.
Hossain, Md Shahriar J., "Throughput and Yield Improvement for a Continuous Discrete-Product Manufacturing System" (2018). LSU Doctoral Dissertations. 4661.
Available for download on Wednesday, June 30, 2021