Experimental design for improved ceramic processing, emphasizing the Taguchi Method

Ceramic processing often requires substantial experimentation to produce acceptable product quality and performance. This is a consequence of ceramic processes depending upon a multitude of factors, some of which can be controlled and others that are beyond the control of the manufacturer. Statistical design of experiments is a procedure that allows quick, economical, and accurate evaluation of processes and products that depend upon several variables. Designed experiments are sets of tests in which the variables are adjusted methodically. A well-designed experiment yields unambiguous results at minimal cost. A poorly designed experiment may reveal little information of value even with complex analysis, wasting valuable time and resources. This article will review the most common experimental designs. This will include both nonstatistical designs and the much more powerful statistical experimental designs. The Taguchi Method developed by Grenichi Taguchi will be discussed in some detail. The Taguchi method, based upon fractional factorial experiments, is a powerful tool for optimizing product and process performance.