Overview

This project aimed to provide economical methods of investigating, through experimentation, the design of a product and its manufacturing process in order to identify the key factors for achieving product improvement. Many conventional methods are available for such experiments and have achieved success in a wide range of applications. This project was directed primarily towards experiments on mechanical products where current methods are uneconomical or infeasible, although the techniques developed have wider application.

The project had two main strands of research, namely group screening and semi-controlled experimentation. Group screening addresses the problems that arise when a large number of factors has to be considered. Here the few key factors that affect product performance, and any important interactions between them, need to be identified within achievable experimental resource. Semi-controlled experiments concern products assembled from several components where the factors in an experiment cannot be readily set to pre-assigned values and components cannot be reused.

For each of the two strands of research, the project developed the necessary underlying theory, produced research and user-friendly software to implement the ideas and demonstrated the methods in industrial environments. Three industrial collaborators provided experiments to validate the ideas and considerable input into making the methodologies into practical tools: Hosiden Besson Ltd, Jaguar Cars Ltd and Goodrich Engine Control Systems Ltd.

---

Last Updated: 14th October 2008