以實驗設計方法尋找影響線放電加工之表面粗糙度的次要因子

Abstract

各產業對產品的品質要求越來越高，田口法和實驗設計方法因而被廣泛的運用在各種製程中。依據其規劃的實驗步驟能找出主要加工因子與製程間的關係，建立製程的數學模型，用以改善製程，提高製程品質。然而田口法與實驗設計方法只能找出影響製程的重要加工因子，若想再進一步提高製程品質，勢必要找出製程中的次要因子。 本研究選用線放電加工來進行實驗，根據實驗設計法找出線放電加工製程中影響工件表面粗糙度的主要因子為放電時間，並建立放電時間與表面粗糙度的數學模型。利用區集劃分的概念將主要因子放電時間的效應排出製程中，找出了影響表面粗糙度的兩個次要因子，分別為電弧放電時間和進給速度，亦建立了次要因子與表面粗糙度的數學模型。在模型比較中，具有次要因子的數學模型的工件表面粗糙度值較只有考慮主要因子的數學模型的工件表面粗糙度值減少5％~至8％，說明次要因子對品質的提昇是有效用的。

Quality is the minimum requirement for today’s products. In order to improve the quality of a product, Taguchi method and the experiment design were often applied to different industry processes. The main effect factors and their effects of a specific process can thus be found and estimated. The mathematic model describing the process can then be established. Based on the model, people can well control the manufacturing process and improve their product quality. However, both Taguchi method and the experiment design can find the most important factors that affect a process, and have difficulties to obtain the secondary effect factors of the process to further improve the quality. This study intends to find the secondary effect factors of a process, and investigate how much improvement can be reached by adding the secondary effect factors into the model. The wire electrical discharge machining is used to perform the experiment, and the work piece surface roughness is the quality index. It is known that the pulse-on time is the main factor affecting the work piece surface roughness. Finding main factor effects and establishing the main factor mathematic model are performed, firstly, to ensure the experiment setup. Then, the block design by choosing the main factor level and fixing it to exclude its effect from the experiment results. After performing five block designs, it is found that the arc pulse-on time and feed speed are the secondary effect factors, and the mathematic model including the secondary factors are established. This study concludes that the work piece surface roughness can be improved about 5％ to 8％ by adding the secondary effect factors in the model.