The Use of a 12‐run Plackett–Burman Design in the Injection Moulding of a Technical Plastic Component

Tyssedal, John; Grinde, Hallgeir; Røstad, Carl Christian

doi:10.1002/qre.805

Cited by 16 publications

(17 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A technique proposed in [8] to pick out 'the vital few' is successfully used with the PlackettBurman design but could prove valuable with every design.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Comparing different fractions of a factorial design: a metal cutting case study

Mønness¹,

Linsley

Garzon

2006

Appl Stoch Models Bus & Ind

View full text Add to dashboard Cite

SUMMARYFull factorial designs of a significant size are very rarely performed in industry due to the number of trials involved and unavailable time and resources. The data in this paper were obtained from a six-factor full factorial ( 2 6 ) designed experiment that was conducted to determine the optimum operating conditions for a steel milling operation. Fractional-factorial designs 2 623 III (one-eighth) and 2 622 IV (one-fourth, using a fold-over from the one-eighth) are compared with the full 2 6 design. Four of the 2 623 III are de-aliased by adding four more runs. In addition, two 12-run Plackett-Burman experiments and their combination into a fold-over 24-run experiment are considered.Many of the one-eighth fractional-factorial designs reveal some significant effects, but the size of the estimates varies much due to aliasing. Adding four more runs improves the estimation considerably. The one-quarter fraction designs yield satisfactory results, compared to the full factorial, if the 'correct' parameterization is assumed. The Plackett-Burman experiments, estimating all main effects, always perform worse than the equivalent regular designs (which have fewer runs). When considering a reduced model many of the different designs are more or less identical. The paper provides empirical evidence for managers and engineers that the choice of an experimental design is very important and highlights how designs of a minimal size may not always result in productive findings.

show abstract

“…A technique proposed in [8] to pick out 'the vital few' is successfully used with the PlackettBurman design but could prove valuable with every design.…”

Section: Resultsmentioning

confidence: 99%

“…(All possible interactions are included in the estimated model.) The smaller the RMSS, the better the model [8]. The best candidates are shown in Table VI.…”

Section: Plackett-burman Designsmentioning

confidence: 97%

Comparing different fractions of a factorial design: a metal cutting case study

Mønness¹,

Linsley

Garzon

2006

Appl Stoch Models Bus & Ind

View full text Add to dashboard Cite

show abstract

“…The functional relationship between the response and the factors may then be further explored afterwards. Standard factor‐based procedures normally start with investigating how well all subsets of m ‐factors, m = 1, 2, … , k , where k seldom exceeds three or four, explain the variation in the data according to some measure(s) (see the works of Box and Meyer and Tyssedal et al). An evaluation of the measure(s) is performed to decide how many factors and which subset of factors are likely active.…”

Section: Motivational Examplesmentioning

confidence: 99%

Assessing some aspects of factor screening with nonnormal responses

Chaudhry

Tyssedal

2019

Appl Stoch Models Bus & Ind

Self Cite

View full text Add to dashboard Cite

Nonnormally distributed response values, such as count data for instance, create challenges for factor screening. One problem is that variances may vary from run to run. Another is the choice of screening design for such responses. In this paper, we assess some screening performances for three popular screening designs: a definite screening design, a minimum resolution IV design, and a Plackett‐Burman design. Four distributions, two binomials, one gamma, and one Poisson are chosen for the response values. For each distribution, we test out if it is best to use the raw data, a variance‐stabilizing transformation of the data, or perform a generalized linear modeling assuming three factors are active. From our investigations, two‐level nonregular designs gave the highest success rate in identifying the subset of active factors and a variance‐stabilizing transformation turned out to perform equally good or better than generalized linear modeling in most cases.

show abstract

“…Expanded applications lead to new problems, requiring new research to develop the techniques for solving these problems. As examples of some of this research, Mønness and Coleman 7 present a new approach to analyzing experiments with multiple responses, Godolphin 8 investigates the impact of missing values in factorial designs, and Tyssedal et al 9 provide a novel application of Plackett-Burman designs in injection molding. These authors demonstrate that designed experiments continue to be one of the most important analytical tools for system improvement.…”

Section: Editorialmentioning

confidence: 99%

Designed experiments in process improvement

Montgomery¹

2006

Quality & Reliability Eng

View full text Add to dashboard Cite

Statistically designed experiments are an important tool for the quality and/or reliability engineer. Their role in improvement and optimization of manufacturing processes has been well documented. Designed experiments have been used in industrial process development for well over 50 years, so one might expect that new and important developments are relatively infrequent. However, experimental design is a vibrant field of research. This is being spurred on by applications beyond the traditional manufacturing process arena, and by the discovery that variations of old techniques have new and modern applications.An example of the latter scenario is split-plot designs (SPDs). A SPD involves two kinds of factors, those that are easy to change and those that are hard to change. An example might involve temperature as the hardto-change factor and feed rate as the easy-to-change factor. The SPD employs two levels of randomization to minimize the number of changes of the hard-to-change factor. The hard-to-change factor is assigned to the whole-plot (WP) experimental units, and then the easy-to-change factors are randomly assigned to experimental units within each WP, which are referred to as sub-plots or split-plots (SPs). The SPD has an agricultural heritage, with the WPs often being large plots of ground and the sub-plots being smaller plots of ground within the WPs.As industrial experiments often involve both easy-and hard-to-change factors, there has been widespread renewed interest in the SPD. However, the WP and SP structure of the design leads to two error structures, with two different variance components, one for the WP and one for the SP. Consequently, generalized least squares must be used to fit the model associated with the design. Parker et al. 1 describe classes of SPDs that have an equivalent estimation property; that is, the ordinary least-squares estimates of the model parameters are identical to the generalized least-squares estimates. This property provides the best linear unbiased estimators and simplifies the estimation of the model parameters because the experimenter can use standard computer software that may not have the generalized least-squares capability to fit the underlying models. Their methodology can be applied to many standard response surface designs, including the central composite design, the small composite design, equiradial designs, and the Box-Behnken design. Liang et al. 2 show how fraction of design space (FDS) plots can be extended to evaluate the properties of SPDs. The FDS plot is a graphical representation of the prediction variance from the model that has been fit to the data from the experimental design. It plots the prediction variance (or the scaled prediction variance) versus the fraction of the design space that has prediction variance values at or below a given value. These plots are very valuable in studying the potential performance of a design, and they have been completely randomized designs have been studied extensively with FDS plots and other graphical methods....

show abstract

The Use of a 12‐run Plackett–Burman Design in the Injection Moulding of a Technical Plastic Component

Cited by 16 publications

References 11 publications

Comparing different fractions of a factorial design: a metal cutting case study

Comparing different fractions of a factorial design: a metal cutting case study

Assessing some aspects of factor screening with nonnormal responses

Designed experiments in process improvement

Contact Info

Product

Resources

About