Statistical design for material and process optimization

CIM Bulletin, Vol. 85, No. 961, 1992

G. Pouskouleli and T.A. Wheat, Mineral Sciences Laboratories, CANMET-EMR, Ottawa, Ontario

The role and benefits of using statistical methodology in the development and optimization of materials and processes is examined. It is shown that considerable benefits accrue from the use of an appropriate design in which a number of factors (process parameters) are varied simultaneously while the system response such as percent recovery, degree of separation, purity, etc. is monitored and ultimately optimized. Compared with the commonly encountered one-factor-at-a-time approach, this schema exploits reduced experimentation (thereby requiring reduced resources [staff, time, dollars, etc. ]) yet yields quantitative information on the significance of each factor and their combinations on the measured response.

Mots Clés: Statistical design, Research and development, Process optimization, Statistical process control.