IC封裝廠短期生產規劃系統之構建

Abstract

半導體產業在台灣蓬勃發展，帶動下游IC封裝業的商機，形成眾家廠商爭食產業大餅 的現象。在競爭激烈的環境下，需要一套完整的短期生產規劃系統來規劃與整合生產問題 。本文將之分為細部主生產排程模式與現場活動控制模式兩部份，其中包含產能規劃、生 產排程、訂單進度控制方面，以達到客戶訂單如期產出，且充分利用機台產能。 本文所建構IC封裝廠細部主生產排程模組，在產線規模與客戶產品導期已知的情況下 ，透過模擬，求算出產線負荷量容許範圍內，各客戶合理的每天計畫分配量。 在面對動態到臨訂單、製程特性與製造現場變異，利用細部主生產排程模式之輸出結 果與現場即時資訊，構建IC封裝廠之生產活動控制模式。對於實際到臨訂單，從訂單交期 之設定、投料、派工、訂單進度監控與現場控制作一系列探討，其中，現場控制包括產線 微調與緊急訂單評估功能，期能控制訂單進度，以降低現場干擾的影響。 在本文中舉出一個符合現場實際狀況的範例以進行實例驗證工作，比較有無進行產線 微調功能時，訂單延遲的情況，結果得知本文所發展的產線微調功能，除了可平穩各產線 的負荷外，還能降低訂單延遲個數的效果；之後，再針對緊急訂單到臨，進行緊急訂單承 接模式，驗證結果得知；可在儘量不影響一般訂單產出的前提下，使得所允諾的緊急訂單 均能如期產出。

With the development of semiconductor industry in Taiwan, IC packaging factories are growing up quickly and many enterprises intend to participate in this field. Therefore, it needs a system to handle the production planning problems. A short-term planning system thus is designed including the detail MPS (DMPS) and PAC modules. The capacity planning, scheduling and order progress control are covered so as to make the production output meeting the corresponding due dates and to improve the utilization of workstations. In the DMPS stage, the daily capacity of each production line is allocated to each customer with committal leadtime through simulation. Then, applying the output of DMPS module and real-time information in the shop, the PAC module is constructed for the IC packaging factory. With the actual arriving orders, the PAC module executes the due-date assignment, order releasing, dispatching, monitoring and shop-floor controlling functions. To make the shop-floor control better, the production line adjustment and rush order justification functions are designed. With these designs, the PAC module can control the orders' progress and decrease the influence of disruptions validly. An example is used to verify the performance of the proposed short-term production planning system. We find that by using the production line adjustment function, the load of each line can be balanced so as to decrease the number of tardiness orders. Then, the rush order justification function is implemented to measure whether there exists enough capacity for an arriving rush order. For each agreed rush order, its due-date can be met on the premise that the normal order's due date will not be influenced preferably.