Optimisation of Manufacturing Processes : A Response Surface Approach

Many engineering companies around the world are currently undergoing a quality control and improvement revolution that originally started in Japan many decades ago and this book provides a brief overview of this revolution. Robust design is a central component of the modern approach to quality improvement and is a phrase used to describe any engineering activity whose objective is to develop high quality products (and processes) at low cost. A key characteristic of robust design is the use of statistically planned (designed) experiments to identify those process variables that determine product quality. Robust design was developed in Japan by G. Taguchi in the early 1950s and its widespread use throughout Japanese industry is one of the main reasons why the country has emerged as a major producer of relatively cheap high quality products, especially in the automobile, home electronics and microprocessing sectors. Despite its early success in Japan, robust design remained virtually untried in the United States and Europe until the early 1980s. However, the realisation that quality is a vital ingredient required for success in today's highly global and competitive markets has since prompted Western companies to embrace the robust design concept.

This book explores the planning, implementation and analysis of experiments designed both to improve existing manufacturing process and to create newer and better processes and products.

Mark Evans - My research interests have led to the publication of over 75 peer reviewed journal papers and two postgraduate textbooks. My research has involved the applications of statistics to three main fields of study: i. the lifeing of materials at high temperatures, ii. process optimisation in manufacturing and ii. techno-economic forecasting.

1. My current research in this field is advancing the Wilshire methodology to life assessment by expressing the Wilshire equations within a statistical hierarchical model so that batch to batch variation is modelled correctly to enable meaningful confidence intervals to be made for the predicted failure times on any material.

2. I have worked with TATA Steel to improve the surface finish of their manufactured strip by applying various data mining techniques. For TIMET my work has involved the computer modelling of their processes to identify optimal production schedules and for Rolls Royce my work has involved combining numerical and statistical techniques to optimize a disc test for determining the remaining life of materials used for generating electricity

3. My research has involved the development and estimation of novel models for providing reliable forecast of short to medium term movements in both the price and the demand of various basic metals.