The Effective Shortening Handling Time Dispatching Method of Automatic Material Handling System in 300mm Semiconductor Factory

碩士 === 國立高雄應用科技大學 === 工業工程與管理系 === 98 === At present, 300mm wafer has become the mainstream. The processing of wafer fab is complicated and with convection phenomenon, which requires constant moving and results in overload of manpower. And we must depend on AMHS at this time. Precedence problem for...

Full description

Bibliographic Details
Main Authors: Hsiu-Lan Chuang, 莊琇蘭
Other Authors: Dr. Chia-Nan Wang
Format: Others
Language:zh-TW
Published: 2010
Online Access:http://ndltd.ncl.edu.tw/handle/97117211507605962508
Description
Summary:碩士 === 國立高雄應用科技大學 === 工業工程與管理系 === 98 === At present, 300mm wafer has become the mainstream. The processing of wafer fab is complicated and with convection phenomenon, which requires constant moving and results in overload of manpower. And we must depend on AMHS at this time. Precedence problem for different products exists due to the intense market competition and product variety. Under the consideration of economic effect, Hot Lot should be handled in priority in order to satisfy customers’ needs. For optimal single loop, OHT will easily cause obstructions while transporting to and fro. The obstruction frequency of OHT and delivery efficiency of products is the main cause that directly impact overall performance of wafer fab. Thus, this study explored Loading Differentiated Preemptive Dispatching (LDPD) as a comparative work of Liao and Wang’s DPD [2006]. By modularizing and testing actual data of 300mm wafer fab with simulation software, forecast and analysis on moving time in accordance with different priority of products are also provided in this research. The result of this study shows that with the use of LDPD rule, the average variable time of Normal Lot is decreased 13.36% while that of Hot Lot is decreased 12.92%. The result also shows that LDPD rule can effectively shorten the non-profit waiting time and reduce cycle time through moving lot quickly as well as preventing OHT from blocking up. Then, not only the bottle neck problem will be effectively solved, but the quantity of output will be improved. Besides, this research found out the best Multiple Regression Model in connection with Normal Lot and Hot Lot of LDPD. After adjusting, R2 is individually 0.984 and 0.917, which can precisely forecast the moving time for precedence of different products. This can provide production scheduling and dispatching for further analysis and forecast.