運用統計方法於品質事件風險評估與決策制定：以TFT-LCD驅動IC之分屏不良問題為例

當重大品質事件發生時，若遇到資訊不足以判斷風險並作成決策時，此時有必要建立一適當的科學化風險評估方法，否則，將造成企業相當龐大的損失。本研究以薄膜電晶體液晶顯示器（thin-film transistor liquid-crystal display, TFT-LCD）驅動積體電路（integrated circuit, IC）之H公司的玻璃覆晶（chip on glass, COG）產品分屏不良品質問題為例，提出一套運用統計方法的推論來進行風險評估，以降低不良的風險與損失，獲得客戶同意進入試投驗證與放量管控。研究結果顯示，本研究提出之風險評估方法得以將H公司原先由於品質不良事件可能發生的報廢損失美金55.7萬元，有效降低至最終賠付客戶不良品面板維修費用美金8千元。這個方法可提供給產業界當面臨品質不良問題發生，可能遭客戶整批退貨報廢的鉅額損失時，一套相當具有實務參考價值的風險評估，以有效降低損失。

關鍵字

統計方法；品質事件分析；風險評估； TFT-LCD驅動IC

並列摘要

When a major quality event occurred and the information is insufficient to judge the risk and make correct decision, it is essential to establish a suitable risk assessment method. Otherwise it will result in giant loss for the company. This research proposes a risk assessment method that applies statistical inference to reduce the risk of poor quality and further acquire customer's agreement to release production verification and volume control. The case of abnormal display problem in the chip on glass (COG) product of thin-film transistor liquid-crystal display (TFT-LCD) driver integrated circuit (IC) at H-company is used as the case study example. Results reveal that the proposed risk assessment method could effectively reduce the loss from discarding the entire returning lot of USD$ 557,000 to compensating the repair charge of USD$ 8,000 when the poor quality event of abnormal display problem happens in the H-company. This method could provide to the industry a valuably practical reference of risk assessment when facing a quality event that may result in huge loss of returning and discarding the entire lot.