在部分因子實驗中,把實驗徑以適當的順序排列之後,我們能夠使可估計效應不會被由於實驗徑被執行的時間、所在位置不同所造成的趨勢效應所影響。在應用上,所有的實驗徑可能無法在相同的時間或是位置被執行,因而造成系統性誤差,為了消除這種誤差而使得實驗能夠在更一致的環境下進行,我們必須把這些實驗徑放入不同的集區進行實驗,這樣的情形就叫做集區設計。然而,在各個集區裡面仍可能存在著趨勢效應,如同在部分因子實驗中一樣,如何使可估計效應具有抗趨性是非常重要的,但大部分的研究並未考量當集區中含有趨勢效應時,如何設計出能夠消除趨勢效應的實驗。 本篇文章將因子水準數設定為三,在假設所有集區中的趨勢效應均相同的前提下,找出了具有抗趨性因子的集區設計方法。Wang (2006)在其尚未發表的論文中,提出了以直交表建構出水準數為二的抗趨性因子集區(trend-free block)設計的方法,本文仿照Wang的方式,在集區個數不同的情況下,利用直交表的特性,找出了具抗趨性因子的集區設計,如此一來即使設計實驗者沒有艱深的設計理論背景,只要具備直交表的基本知識,就可以透過本研究提出的結果來安排實驗。 An appropriate run order of experimental runs makes estimable effects trend free in fractional factorial experiments. In practice, there is possible that not all the experimental runs in the fractional factorial experiments can be executed in the same time or place. Thus, we might need to put them in different blocks. Unfortunately, there might have some trend effects within each block. The research on this issue is as important as that on trend problems in fractional factorial experiments. Most of researches dealing with this topic have not considered block factorial experiments with trend effects in blocks. Wang (2006) proposed the solution for designing fractional factorial experiments with the assumption of the same trend effects in blocks. In this article, we focus on the fractional factorial experiments at three levels, and explore the solution of constructing the trend-free block factorial designs.