晶圓針測廠生產規劃模式之構建

Abstract

本文探討對象為半導體廠之專屬晶圓針測廠。因為針測廠之產品輸入來源 為晶圓廠， 而晶圓廠傾向以產出最大化為生產目標，產品在流出晶圓廠 進入針測廠後，因為交期 的接近而以交期的達成為最大目標。所以針測 廠是居於從以最大產出為生產目標，轉 換成以交期滿足為目標的首衝位 置。在以交期滿足為前提而試圖增加針測廠之產出下 ，本文所構建之晶 圓針測廠生產規劃模式，是在配合晶圓製造廠的生產排程規劃架構 下， 考量晶圓製造廠的產出特性與針測廠測試機台的排程特性來設計。 在晶 圓製造廠細部排程四週的規劃幅度中，本生產規劃模式就晶圓廠細部排程 第一週 每天的產出計劃，構建一生產作業控制模組，以規劃測試機台未 來一週的排程﹔在批 量等級劃分與同等級批量投料策略的搭配下，本生 產作業控制模組的確可以在追求產 品交期績效的同時，有效降低機台的 設置時間。 針對細部排程第二、三和四週的產出規劃，本文構建一產能 負荷評估模組，以檢核在 該三週測試機台的產能負荷狀況，確保針測廠 不致在某些週有大量的閒置產能，在某 些週卻因為產能不敷使用而延誤 交期。而對於生產作業控制與產能負荷評估二模組所 得的結果，本文構 建一排程資訊回饋模組，將該二模組的規劃結果回饋給晶圓廠，藉 由晶 圓廠細部排程來調整產品的產出數量與時點，以提高產品的交期績效。 實證中比較各可行方案，其結果顯示：透過針測廠的生產排程規劃架構， 和對晶圓廠 生產資訊回饋之設計，可增加整體積體電路製造的生產效益 ，提升半導體廠商的整體 競爭力。 Because the probing factory is due-date oriented while the fabrication factory is throughput oriented in the production planning, this thesis thus considers such a transition to do the production plans. This proposed production planning model takes the throughput schedule of the fabrication factory as well as the characteristics of test machines in the probing factories into consideration. It is designed to increase the throughput of the probing factory under the constraint of the due date satisfaction. Use the four-week planning results of the detail production planning of the fabrication factory, the proposed model is divided into three modules: production control module, capacity planning module, and the feedback module. The production control module takes the first week's result to schedule the test machines in the probing factory. To satisfy the due-date requirement, each lot is graded to one of the four levels according to its arrival time, its due date, and the capacity availability of the required machine type. Each level has different priority for releasing the lots to shop-floor. However, all the lots of the same level are considered simultaneously so as to reduce the set up time. The capacity planning module takes the planning results of the other three weeks to check whether the required probing capacity can be fulfilled or not. The purpose of this module is to avoid test machines having unbalanced capacity allocation such that dates are missed. The information obtained from the above two modules will be feedback to the fabrication factory by the feedback module. The fabrication factory thus can adjust its quantity and its throughput time by detail schedule planning so as to improve the due date performance. The results of the case studies show that the production planning model indeed achieves the due date performance and controls the setup time of test machines simultaneously. The competitive advantages of the semiconductor factory thus is enhanced.