Integration of part quality and tooling information for effective process control and maintenance planning

This dissertation develops new methodologies by integration of product quality and tooling information for effective process control and maintenance planning, which are essential to improve product quality, increase process productivity, and reduce production cost. Specifically, three research efforts have been conducted: (1) multivariate mean shift diagnosis and adjustment by using simultaneous modified and acceptance control charts considering design tolerance: The relationship between part quality and process faults is described by a linear model obtained from product/process design. The least square estimation and linear mixed model estimation are used to estimate and diagnose the mean shifts. A simultaneous test is developed to test the mean shifts with the ensured specified total false alarm rate. A new set of simultaneous modified and acceptance control charts is first time proposed to reduce the number of unnecessary tooling adjustments with consideration of the specifications of product design and tooling tolerances. (2) The optimal process adjustment methodology for multistage assembly processes: A fixture adjustment methodology is developed to determine the optimal adjustment strategy that improves the final product quality and reduces the overall production costs in multistage assembly processes. From the proposed observation model, Bayesian estimation is used to estimate and predict the tooling deviations. The optimal adjustment strategies are determined by dynamic programming which consider quality cost, tooling adjustment cost, and product life cycle. (3) The optimal selection of maintenance strategies considering variation of process availability: A systematic procedure is developed for the optimal selection of the maintenance strategies among reactive, preventive, and predictive maintenance. The proposed optimal decision considers not only the expected process availability but also the variation of the process availability, which has not been found in the existing literature. The expectation and variation of the process availability is analyzed under different maintenance strategies. Simulation studies are conducted to illustrate the effectiveness of the developed methodologies in manufacturing processes. Those studies demonstrated its significant impact on production cost, productivity and product quality.