This reference has been designed to introduce you to Design of Experiments (DOE) concepts, and it accompanies ReliaSoft's Design of Experiments software package, DOE++. The treatment of the subject and most of the examples included in this reference assume that you have installed and can refer to the DOE++ software.
The purpose of this reference is to provide you with a general overview of the DOE subject, as well as outline specific items, settings, rules and assumptions in DOE++. This book includes the underlying theory and principles, relevant calculations and derivations, as well as step-by-step calculations to provide an insight and understanding of the analysis and the calculations performed by the software. It also provides examples and sample problems that can be analyzed using ReliaSoft's DOE++.
Examples in this reference use ReliaSoft's DOE++ software. Use of the software is described in-depth in the DOE++ User's Guide.
This reference assumes a familiarity with the software and uses DOE++ to reinforce and expand on the fundamentals of the analysis.
This reference is one in a series of reliability engineering reference works created by ReliaSoft. Some concepts and examples related to life data analysis and accelerated life testing that have been extensively covered in other references, such as ReliaSoft's Life Data Analysis Reference, and ReliaSoft's Accelerated Life Testing Reference, have been omitted from this reference.
Interested readers can obtain these references from ReliaSoft or view this material on-line at www.Weibull.com.
The following list provides a summary of the chapters in this reference book.
Chapter 2: Overview presents an introduction to DOE concepts.
Chapter 3: Statistical Background provides a review of the principles and terminology used in this reference. The chapter gives brief descriptions of the distributions used in calculations related to experiment analysis, and it explains the concept of hypothesis testing. Hypothesis testing finds direct application in the analysis of experiments to conduct significance tests and is widely used in later chapters.
Chapter 4: Simple Linear Regression Analysis introduces regression analysis using linear regression models with a single factor. Gaining a clear understanding of regression analysis is important in interpreting the results obtained from the analysis of designed experiments. The procedure to conduct significance tests using the analysis of variance (ANOVA) is explained in this chapter.
Chapter 5: Multiple Linear Regression Analysis expands on the analysis of simple linear regression models and discusses the analysis of linear regression models with more than one factor, namely, multiple linear regression models. Significance tests on individual factors are discussed in this chapter. The chapter sets the stage to gain an understanding of ANOVA models, which are a category of multiple linear regression models used to analyze data obtained from experiments.
Chapter 6: Analysis of Experiments uses the concepts covered in Chapters 4 and 5 to illustrate the analysis of single factor and factorial experiments (multiple factors) using ANOVA models. The chapter explains how the analysis objective affects the choice of the underlying models, namely regression or ANOVA. The concepts of randomization and blocking are also covered in this chapter.
Chapter 7: Two Level Factorial Experiments concentrates on a special case of the factorial experiments discussed in Chapter 6, where each factor under investigation is run at two levels. The concept of blocking, introduced in Chapter 6, is expanded in this chapter to include experiment designs that use incomplete blocks. Unreplicated two level experiments are also covered in this chapter. The later portions of the chapter discuss fractional factorial experiments and related concepts such as aliasing, folding over and design resolution.
Chapter 8: Other Factorial Designs covers the Plackett-Burman designs and Taguchi's orthogonal arrays.
Chapter 9: Response Surface Methods discusses methodologies and experiment designs that help achieve an optimum response value. Optimization of single and multiple responses is covered in this chapter.
Chapter 10: Taguchi's Robust Parameter Design Method presents a brief discussion of Taguchi's philosophy to achieve a robust design.
Chapter 11: Reliability DOE introduces a new concept in DOE, where the disciplines of reliability and design of experiments are integrated, and illustrates how design of experiments can be used to "build" reliability into the products. Reliability DOE analysis is an essential portion of the methodology of Design for Reliability, which is a process to design robust products that operate with minimal failures. The chapter illustrates the analysis of experiments with censored observations and introduces the use of the Weibull, lognormal and exponential distributions as the underlying models of the response.
See Also:
