時至今日’科技的發展變化非常快速，使得科技 產品的製程手法也日新月異，在於高標準的品質 要求下，如何降低成本及提高產品品質是相當重 要的課題，尤其在激烈的商業競爭環境下，如何 有效使用統計製程管制工具進行監控整個製程 狀態，以及在穩定的製程當中提供更精確的品質 評級方法’都是產業界必須面對的課題。
本硏究的目的在於探討使用模糊數建模方法所 建立的新評級規則，應用於太陽能多晶矽晶錠長 成品質的判讀程度與效益，硏究結果顯示，透過 模糊規則庫的建立，可大幅降低傳統管制圖當中 造成誤差且低敏感度的缺點，避免在品質管制過 程造成誤判，造成無謂的成本損失，此結果除了 可以提供品管人員作爲擬定改善策略的參考 外，亦有助於業界提升生產線管控品質的能力。

Up to date, the change of science development and technology changes are very fast that causes the manufacturing process of high-tech product changes with each passing day more and more. According to the high standard requirements of product quality, how to increase product quality and reduce cost becomes an important task. Particularly, in the intense commercial competition surrounding, how to effectively use statistical process control tools to monitor the entire process, as well as in the stability manufacturing process provides more precise quality rating methodology, are topics that industry must face.
The purpose of this study is to construct the new evaluation rules via the fuzzy number modeling and apply to analysis the interpretation of the extent and effectiveness of polysilicon ingot quality. Results of this study show that the new evaluation rules are able to greatly reduce the error judgment and the low sensitivity phenomena and to avoid the unnecessary cost loss in the quality control process that belongs to the shortcomings of the traditional control charts. These results not only can provide quality control personnel prepared to improve the strategy reference but also can assist the industry to upgrade quality control ability to control the quality under requirement.