Using strip-block designs as an alternative to reduce costs of experimentation in robust product design and multistage processes.

摘要

Although strip-block designs have been applied in the agricultural field since the late thirties, its use in industrial settings has been very limited. Strip-block designs provide not only an economical, but also an operational advantage in executing experiments. These designs are appropriate in three basic scenarios: robust product experimentation, presence of hard-to-change factors, and multistage processes. In this research, we explore the use of strip-block configurations by presenting a novel arrangement aiming at reduction of experimentation costs. An experiment on a battery production process is discussed to illustrate the application of strip-block configurations in industry.; Our goal is to provide theoretical properties and applied characteristics of a new layout, called post-fractionated strip-block design. We initially present results to facilitate the analysis of standard two-level strip-block designs. Specifically, we develop general expressions for the variance of the contrasts, followed by a rule to identify the contrasts associated with each of the three error strata in these designs. Then, we illustrate how to construct post-fractionated two-level strip-block designs. The correct classification of the contrast strata is crucial to the analysis of ordinary strip-block designs as well as post-fractionated strip-block designs. Separate analyses need to be conducted for each stratum. Therefore, we extend the rule of associating the contrasts with the appropriate stratum to the post-fractionated case. This procedure makes the task of separating contrasts simple and straightforward. Consequently, the practitioner can without difficulty perform the proper analysis of the experiment.; Furthermore, we provide catalogs of post-fractionated strip-block designs as an aiding tool in the selection of appropriate plans. These catalogs cover designs with eight, sixteen, and thirty-two runs, involving up to eleven factors.; This research is a contribution to the expansion of the use and understanding of strip-block designs in industrial contexts. This new approach to strip-block designs represents a way to reduce costs of experimentation. The proposed framework will provide organizations scientific means to acquire critical knowledge of processes and products essential to improve quality, leading to competitive advantage in the market.