Robust System Design
Make products insensitive to variation: quality loss, orthogonal array experiments, tolerance design and term projects, with quizzes, solutions and labs.
Products fail in the field because of variation, and robust design aims to make them insensitive to it. Created for working engineers on a split summer schedule, this class teaches quality loss functions, orthogonal array experiments and tolerance design, with paper helicopter and catapult labs. Teams were encouraged to apply the methods to a system at their own company.
What You Will Achieve
- Formulate quality characteristics and identify noise factors
- Run matrix experiments with orthogonal arrays
- Estimate a design’s robustness and select rational tolerances
- Lead product development work using robust design techniques
Program Overview
Principles
The opening lectures make the economic case.
- Probability, statistics and quality loss
- Context of robust design in product development
- Selection of quality characteristics
Experiments
The core is designing and analyzing tests.
- Matrix experiments using orthogonal arrays
- Noise strategy
- Analysis of variance
- Constructing orthogonal arrays
Simulations, Dynamic Systems and Tolerances
Later lectures extend the methods.
- Design of software
- Robust design using simulations
- Design of dynamic systems
- Tolerance design and Mahalanobis Taguchi System
Course Completion
Job Outlook
These methods apply in product development, quality engineering, reliability and manufacturing work across industries.
Good for: Product engineers and quality professionals who can run experiments on a real system they control.
Keep in mind: From summer 1998, and it shows in the format, though the statistics remain useful. Phadke’s book is separate, and the term project works best with access to a system and company data.
Bottom line: Old but solid, with quizzes, solutions and a project brief that translate directly to working engineering teams.
Frequently Asked Questions
What were the hands-on labs?
A paper helicopter lab for matrix experiments and a catapult lab for noise strategy, both of which are easy to imagine repeating at low cost.
What does the term project ask for?
A team picks a system, quantifies the economic cost of its variation, improves it, and confirms the improvement experimentally.
Which textbook do the readings use?
Most lectures assign chapters from Phadke, and the first session reads the article Robust Quality by Taguchi and Clausing.



There are no reviews yet.