Design of experiments: a tool for continuous process improvement

Abstract

The author outlines an optimal approach needed to implement the process of continuous process improvement through the use of proper statistical methods. Industrial systems are usually affected by high noise and very weak signals. Active data collection systems lead to statistical design of experiments that are invaluable in all stages of process improvement, both for reduction of variability and for improvement in output variables, such as yield and reliability. The most critical issue is to decide where to perform experiments to obtain the maximum knowledge with the fewest tests. Continuous improvement processes are emphasized in order to achieve a reduction in noise and an improvement in the signal. A stepwise process which has been used successfully in industries to achieve this goal is described.<>