Experimental_Design

Experimental_Design - Experimental Design Author: John M....

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
Experimental Design Author: John M. Cimbala, Penn State University Latest revision: 19 September 2007 Introduction When setting up an experiment, it is important to first take some time to carefully design the experiment. In this module we discuss how to design an experiment . A particularly helpful reference book for this material is Elmer E. Lewis, Introduction to Reliability Engineering , 2 nd ed., Wiley, TA169.L47 1996, Ch. 4. The basics of experimentation Below are listed the basic steps for conducting an experiment. Each of these steps is critical to a successful experiment. Define the problem o This may seem obvious, but it is really the most important and most critical step. Namely, the required output of the experiment must be clearly defined. o Examples include: ± Determine the operating temperature range of a voltmeter such that the overall error is less than some value. ± Determine the optimum low-pass filter for an audio system that minimizes cost while still meeting performance specifications. Design the experiment o This step involves many components, including (but not limited to): ± Conduct a literature survey . ± Perform an analytical analysis (as far as possible). ± Select the variables to be measured. Note : Dimensional analysis is useful here, since it can potentially decrease the number of required independent variables to be measured . ± Select the instruments to be used in the experiment. ± Estimate the experimental uncertainties . ± Select or design a test matrix . (Test matrices are discussed in detail in this module.) ± Design the test rig and the experimental procedure . Construct the experiment Gather data Analyze the data Do confirmation experiments and/or follow-up experiments [ if necessary ]. Interpret and report results/conclusions Other issues There are other issues that enter into consideration when conducting an experiment, such as o cost o schedule o personnel These issues will not be addressed here. In this learning module, emphasis is placed on one very critical aspect of experimental design, namely choosing a test matrix . Choosing a test matrix By way of introduction, suppose some experiments must be conducted in order to determine an optimum value of parameter X , which is a function of parameters a , b , c , . .., i.e., X = X ( a , b , c , . ..) . The optimum value can either be a maximum (for example best efficiency, longest life, or highest strength) or a minimum (for example shortest time, minimum cost, lowest pressure drop, or smallest surface nonuniformity). In all the examples below, it is assumed that the optimum value of X is the maximum value of X . The analysis is similar if X is to be minimized instead. In this learning module, you will learn how to choose a test matrix so that the number of necessary experimental runs (and therefore cost and time) is kept to a minimum .
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 2
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 7

Experimental_Design - Experimental Design Author: John M....

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online