利用製程能力指標Cpm建立品質特性分析圖並用之改善產品品質

機構典藏 > College of Business and Management > Department of Management Sciences > Thesis > Item 987654321/105327

Please use this identifier to cite or link to this item: https://tkuir.lib.tku.edu.tw/dspace/handle/987654321/105327

Title:	利用製程能力指標Cpm建立品質特性分析圖並用之改善產品品質
Other Titles:	Using process capability index Cpm to construct a quality characteristic analysis chart for product quality improvement
Authors:	楊俊明;Yang, Chun-Ming
Contributors:	淡江大學管理科學學系博士班歐陽良裕;陳坤盛
Keywords:	製程能力指標;準確度;精確度;品質特性分析圖;熵法;理想解類似度順序偏好法;模糊理論;Process Capability Index;accuracy;precision;QCAC;Entropy;TOPSIS;Fuzzy Theory
Date:	2015
Issue Date:	2016-01-22 14:53:19 (UTC+8)
Abstract:	製程能力指標是一種有效且方便用以衡量產品品質水準的方法，然而單一製程能力指標並無法明確地指出造成不合格品質特性的真正原因。另一方面，雖然多品質特性分析圖已被廣泛地用以衡量產品具多個不同型式品質特性的品質水準。但由於多品質特性分析圖是由多個製程能力指標所構成，可能會增加品質管理者在計算上的複雜度和錯誤率，並且它也無法辨識望大型或望小型不合格品質特性是由於製程偏移或/和變異過大所造成。此外，對於企業而言，它可能會同時考量改善不合格品質特性所需耗用的資源以及改善後可獲得的效益，和決策者主觀判斷上具有不確定性的情況。是故，如何運用一套有效的方法去衡量具單一品質特性或多個不同型式品質特性的產品品質水準，並分析造成不合格品質特性的原因及決定何者為被優先考慮改善的項目是非常重要的課題。有鑒於此，本論文將上述研究問題分為五個章節進行探討。第一章首先說明本論文之研究動機與目的、相關文獻探討和論文架構。第二章將製程能力指標Cpm、最小個別品質能力指標、準確度及精確度做結合，建立一個品質特性分析圖。藉由資料在圖形上的分布情況，便可找出不合格品質特性並判斷不合格品質特性產生缺失的原因是由於製程偏移或/和變異過大所造成。接著，計算各不合格品質特性的判別距離。當品質管理者考量總改善成本有限制的情況下，可根據判別距離值的大小進行排列不合格品質特性改善的先後順序。第三章延續第二章，將六標準差概念加入品質特性分析圖，建構一個新的品質特性分析圖並將之與熵法和理想解類似度順序偏好法做結合。品質管理者便可在考量改善每個不合格品質特性所需耗用的資源以及改善後可獲得的效益不同情況下，決定不合格品質特性改善的先後順序。第四章利用變數變換法將望目型、望小型和望大型的品質特性之樣本觀測資料做適當變化，進而建立一個多品質特性分析圖並將之與模糊理想解類似度順序偏好法做結合。品質管理者可衡量望目型、望小型和望大型的品質特性是否合格，並在考量改善每個不合格品質特性所需耗用的資源以及改善後可獲得的效益不同，且為語意變數的情況下，進行排列不合格品質特性的改善先後順序。最後，第五章則對本論文各章所提出的方法和研究結果做一總結，同時建議後續研究者未來可進行的研究方向。 Process capability indices (PCIs) are an effective and easy–to–use approach for measuring the quality level of products in the manufacturing industry. However, using a single PCI cannot effectively reveal the causes of deficient quality characteristics in the manufacturing process. On the other hand, some multi–quality characteristic analysis charts (MQCACs) combining two or more PCIs have been proposed to overcome the problem products with two or more different types of quality characteristics. However, if quality managers use these existing MQCACs to evaluate whether the quality capability of a product meets the acceptance standard, the numerical analysis and calculation of the PCIs will be quite complex, which will directly affect the outcome of the analysis and entail a greater expenditure of time, cost and manpower. Also, these existing MQCACs cannot be useful in identifying problems with all substandard quality characteristics in larger–the–better and smaller–the–better situations due to process shift and/or variability. In practice, manufacturers should consider how to prioritize improvements that need to be made in all substandard quality characteristics in light of resource requirements, the potential for performance improvements and the uncertainty represented in decision data. Therefore, a key issue faced by the manufacturing industry is determining how to measure single or multiple quality characteristics and prioritize improvements to be made to all substandard quality characteristics of a product with respect to resource requirements and performance improvement potential. This dissertation is organized into five chapters in order to analyze and discuss the abovementioned issues. Chapter one covers the motivation, objectives, framework of this dissertation, and a literature review on PCIs. Chapter 2 illustrates how to combine the process capability index Cpm, minimum individual quality capability index C0, accuracy A and precision P to construct a quality capability analysis chart (QCAC). From the scatter diagram of related data on the QCAC, we can find all of the substandard quality characteristics, and identify the causes of all the substandard quality characteristics in a product owing to process shift and/or variability. Meanwhile, the values of the discrimination distance (DD) of all the substandard quality characteristics are calculated. Quality managers can use the values of the DD to rank all the substandard quality characteristics slated for improvement in order of priority if the total budget for all the substandard quality characteristics improvements is limited. Chapter 3 adds to the QCAC in Chapter 2 with the concept of six sigma to construct a new QCAC. Next, the new QCAC is tailored to combine entropy and technique for order preference by similarity to ideal solution (TOPSIS) methods into a QCAC–Entropy–TOPSIS approach. Quality managers can categorically prioritize improvement options for all substandard quality characteristics with respect to resource requirements, and consider the potential for performance improvements. Chapter 4 first uses the change–of–variable technique to transfer all of the sample data of nominal–the–best, larger–the–better and smaller–the–better quality characteristics into new evaluation data. Meanwhile, an MQCAC can be established as a powerful tool for finding all substandard quality characteristics of a product. Next, a novel hybrid method is presented that integrates a new MQCAC and a fuzzy technique for order preference by similarity to ideal solution (fuzzy TOPSIS). Quality managers can clearly prioritize improvement options for all substandard nominal–the–best, larger–the–better and smaller–the–better quality characteristics with respect to resource requirements and performance improvement potential under a fuzzy environment. Finally, Chapter 5 provides a summary of the main findings and conclusions of this dissertation, and offers suggestions for the direction of future work.
Appears in Collections:	[Department of Management Sciences] Thesis

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	181	View/Open

Loading...