Tolerance design

Background
Variability in the performance of a product or process can often be reduced only by tighter control of process variables or by specifying higher quality components. Similarly, cost reductions are often sought through value engineering where the variability is deliberately relaxed. Statistical tolerance design is an established technique which quantifies the intuitive, engineering feel that the worst case combinations are unlikely to occur. It uses simple rules and formulae to combine tolerances in a realistic way so that performance, yields, failure rates etc. can be predicted.

Course information

Audience
This one day course is aimed at engineers, managers and technicians in design, development, quality, manufacturing and production engineering functions. Delegates do not need any knowledge of statistics, though they should be able to calculate standard deviations with a suitable calculator. ( The formulae are not necessary, just the sequence of button pushes. ) Awareness of Taguchi methods would be beneficial but is not a requirement.

Course content
The emphasis throughout is on practical application and problem solving. This is achieved by providing a set of simple rules which are explained with examples and then practised in group exercises by small teams.
• Worst case analysis vs. statistical tolerancing
• Use of Normal distribution formulae and tables
• Estimating and measuring standard deviations
• How to combine standard deviations in simple systems
• Taguchi methods to analyse tolerances in complex systems
• How to obtain cost effective improvements in
performance by selective tightening of tolerances.

Expected results
After the course the delegates will have good, working knowledge of how to perform tolerance design. They will be able to design experiments to identify critical factors and recommend suitable levels of control for them. They will also be able to predict the variability and pass/fail rates at the chosen settings.