液晶面板組裝廠產能配置模組之構建

Abstract

液晶顯示器在近年來廣為流行，形成今日液晶顯示器市場的繁榮景象，其中又以薄膜電晶體液晶顯示器（Thin Film Transistors Liquid Crystal Display；簡稱TFT-LCD）應用最為廣泛。因此，液晶面板組裝廠無不以「產出最大化」作為生產規劃之目標。 而產能配置的意義是指如何在滿足各產品別之基本需求數量和機台可用產能受限之前提下，有效率地去將機台產能做適當之安排，以減少不必要之換線次數，達到最大化的產出。 為使液晶面板組裝廠達成最大化產出之規劃目標，本文分別構建瓶頸工作站、非瓶頸工作站機台產能配置模式。瓶頸工作站機台產能配置模式考量以下的限制條件，包括滿足所給定各產品別之基本需求數量、批量加工、設置時間、機台產能等，建構線性規劃模式來求解瓶頸工作站各機台每期產品加工數量與順序表，以達到最大化的產出；在非瓶頸工作站機台產能配置模式中，則依據瓶頸資源的產出量，以最小換線頻率為首要目標，亦構建一線性規劃模式來求解最適之機台配置。為使生產系統達成模式所求解出來的產出量，本文亦設計一個即時派工法則，藉由兩者互補的效應，以達到良好之生產規劃。 實例驗證結果顯示，本文所構建之機台產能配置模組，不僅可充分利用瓶頸資源，而且可達到產出最大化之規劃。

In recent years, the rapid growth in Liquid Crystal Display (LCD) has made market demand flourishing in today’s Taiwan, especially the TFT-LCD (Thin Film Transistor-Liquid Crystal Display). By the reason of its importance, the TFT-LCD cell assembly factory chooses“maximum throughput”as the target of the production planning. The meaning that capacity allocation is how to properly arrange the machine capacity to save setup frequency and to maximize system throughput simultaneously under the constraints of satisfying the product basic demand and of the machine usable capacity. This thesis proposes two capacity allocation models: the bottleneck workstation and the non-bottleneck workstation respectively. For bottleneck workstation, we construct a linear programming (LP) model which considers the demand of each product type, the batch process characteristics, the setup time and the machine capacity restrictions so as to maximize system throughput. The planning results contain the processing quantity and the production sequence at every time period. For non-bottleneck workstations, we also construct a linear programming model to minimize setup frequency for achieving the bottleneck’s throughput. Finally, we design a dispatch rule to help the operation of the non-bottleneck workstation match with the programming result. Experimental studies reveal that not only the capacity constrainted resource is fully utilized, but also the maximum throughput plan can be achieved through the proposed capacity allocation module.