Design and manufacturing tolerance optimization with quality loss function

Abstract

In the competitive global market, the needs and expectations of the customers change over time. To satisfy the needs of the customers as well as to remain competitive in the current world market, manufacturers are forced to manufacture high quality products faster with low cost. It is a challenging task for the design engineer to design the product according to the needs of the customers. After identifying the needs of the customers, the design engineer translates the information into specifications. Tolerance is defined as the variation of the nominal dimension between upper and lower limit. Tighter tolerance results in high quality products with increase in manufacturing cost, whereas looser tolerance results in poor quality products with reduction in manufacturing cost. It implies that the selection of tolerance has a profound impact on the product cost and quality. Over a few decades, optimal tolerance has been the focus of extensive research. Earlier, tolerance optimization problem has been solved using traditional and non-traditional optimization techniques without considering design and manufacturing tolerance simultaneously and quality loss In this work, tolerance optimization problem is formulated as a non-linear integer model by considering both the design and manufacturing tolerance simultaneously to minimize the manufacturing cost and quality loss. Here, the tolerance optimization of the components of the assembly is carried out when the components are manufactured by alternate sequence of manufacturing processes, alternate machines and fixed sequence of manufacturing process. It is very difficult to find the optimal solution of nonlinear tolerance optimization problem by using traditional approaches. Hence, in this thesis, the tolerance optimization problem is solved by using nontraditional optimization techniques like Genetic algorithm, patricide Swarm Optimization and Differential evaluation. Finally the optimal solutions obtained by these techniques are analyzed. newline newline newline