Design For Experiments

General Information

NameDesign For Experiments
AbbreviationDFE
Learning Cost160
Playing Cost500
Suggested Phases2,3,4
Engineers
Mechanical Engineer Industrial Design System Engineer Electrical Engineer Production Engineer Software Engineer
Technique and Issue Views
BusinessNeeds Stakeholder Stakeholder Needs System Requirements System Structure Architecture
System Functional Architecture Detail Hardware Design Detail Service Design Detail Software Design Manufacturing Operations
Technique Traits
Identify Stakeholders Elicit Needs Remove Ambiguity Layman's Terms Technical Terms Teamworkings
0 1 2 0 3 1
Traceability Prioritizing Exploring Breadth Inside the Box Outside the box V&V
3 1 4 1 1 4
Verification and Validation
Analysis Calculus Inspection Demonstration Test

Description

Design of experiments, referred to as DOE, is a systematic approach to understanding how process and product parameters affect response variables such as processability, physical properties, or product performance. Unlike quality, mechanical, or process tools, DOE is a purely mathematical tool used to define the importance of specific processing and/or product variables, and how to control them to optimize the system performance while maximizing properties [1][2].

Invented by Ronald A. Fisher in 1935 in his book “The design of experiments” [3], DOE uses statistical methodology to analyze data and predict product property performance under all possible conditions within the limits selected for the experimental design. Next to understanding how a particular variable affects product performance, interactions between different process and product variables are identified. DOE is a technique or procedure to generate the required information with the minimum amount of experimentation [1]. A DoE approach can greatly improve the efficiency in screening for suitable experimental conditions, for example, for cell culture, nutrient complement, factor level, optimization of a process, or robustness testing [4].

Design of experiments is best used to answer questions such as: “What factor contributes most to the problem?” and “what is the most excellent arrangement of aspect values to reduce dissimilarity in response?” [5].