نوع مقاله : پژوهشی
گروه مهندسی صنایع، دانشکده فنی و مهندسی، دانشگاه قم
عنوان مقاله [English]
Statistical Process Monitoring (SPM) and maintenance are two major concepts to
decrease the number of non-conforming products. In this regard, an integrated
model of SPM and maintenance for imperfect high-quality processes is presented.
Furthermore, in studies that consider these two concepts simultaneously, it is assumed that there is only one assignable cause in the production process. This simplifying assumption is unlikely to occur in real production processes due to
the usual complexity of manufacturing systems. It may lead to a poor performance in both economic and statistical criteria if the assignable cause originating the shift is different from the one anticipated at the design of the chart. To make the model more adapted to real manufacturing situations, the process under consideration can turn into an out-of-control state due to several types of assignable causes. The particle swarm optimization algorithm is used to maximize the expected profit per time unit, subject to statistical quality constraints. Then, the effects of parameters lambda_in (the event occurrence rate when the process is in the in-control state),
A_in(the false alarm cost), A_out(the search and repair cost), V_in (the profit
per time unit when the process is in-control state), V_out (the profit per time unit when the process is out-of-control state), and C_e (the cost incurred by the occurrence of the event) on the profit per time unit are investigated. Finally, to show the effectiveness of the suggested approach, two comparative studies are presented. In the first comparative study, the integrated model is compared to a similar model without maintenance activities. The results confirm that implementation of maintenance activities leads to a significant increase
in the manufacturer's profit. In the second comparative study, the presented model is compared to a similar model in which statistical constraints are eliminated. The results indicate that the proposed model extremely improves average time to signal in both in-control and out-of-control states while the profit per time unit decreases slightly.