The use of simulation optimization to implement lean system design of TFT-LCD automated material handling systems

碩士 === 國立成功大學 === 製造工程研究所碩博士班 === 96 === More store space is due to the increase of panal’s size in TFT-LCD manufacturing, that means increasing necessity of stocker space and cost from accumulated WIP. This paper proposes a lean pull strategy, combining supermarket with CONWIP, which is applying a...

Full description

Bibliographic Details
Main Authors: Bing-Yuan Cheng, 程炳源
Other Authors: Taho Yang
Format: Others
Language:zh-TW
Online Access:http://ndltd.ncl.edu.tw/handle/04452053361733309949
Description
Summary:碩士 === 國立成功大學 === 製造工程研究所碩博士班 === 96 === More store space is due to the increase of panal’s size in TFT-LCD manufacturing, that means increasing necessity of stocker space and cost from accumulated WIP. This paper proposes a lean pull strategy, combining supermarket with CONWIP, which is applying a viewpoint of shared resources to a re-entrant process in TFT-LCD manufacturing. By using simulation optimization to solve stockers design problem to find more efficient ways of utilizing stocker. Furthermore, using cycle time.to be an appropriate performance measure for answering the improvement of WIP and stocker utilization. This paper also proposes to use Womack’s lean thinking and Rother’s 7 guiding principles procedure as applying the lean pull strategy. The proposed procedure is illustrated to begin defining value, to address the critical improvement point through the current value stream mapping, to solve the optimal problem by using simulation optimization, and to draw the future value stream mapping for implementing the improved alternative. Besides, we implemented the proposed procedure to a real-world TFT-LCD Array process. After we compared current-state map and future-state map, the results indicated that the optimal solution could reduce the average cycle time from 5.08 days to 3.32 days as well as the maximum stocker utilization could be reduced from 62.13% to 18.49% without any extra investment or facilities shifting. In terms of sensitivity analysis, we had found the maximum daily throughput that would be 4500 glasses. It is important to extend capacities for reference.