Statistical methods in manufacturing

Abstract

Statistical techniques are extensively used in process monitoring and control in research development and manufacturing of industrial products. Statistical process control (SPC) methods have become routine in manufacturing. The applications of acceptance sampling techniques and their modification are standard practices in the development of military standards (MILSTD), as well as international quality standards such as the ISO 9000 series. The application of design of experiments in manufacturing has been proposed and developed by Genichi Taguchi for quality improvement. These methods have not been extensively used in U.S. manufacturing, especially in electronic industries, due to the non-availability of easy to use software for engineers. An automated design of experiment package (ADOE) has been developed by the CPQ Group at IBM San Jose for this purpose. The paper describes the scope and application of ADOE in manufacturing development and research, particularly for high technology industries. This package differs from other available packages in that it deals with both classical and Taguchi methods for quality improvements. Experimental data can be analyzed in several alternative ways, including graphical techniques. Extension of Taguchi's signal-to-noise ratio has been made for experiments with several responses having different performance objectives. The package allows the engineer to bypass analysis, if so desired for obtaining decisions for process optimization.