chapter  9
75 Pages

Design of Experiments and Application to Product Design

WithNicholas P. Cheremisinoff

The purpose will heavily influence both the experimental procedure and the specification of independent and dependent variables. Bias error is experimental error whose numerical value tends to remain constant or to follow a consistent pattern over a number of experimental runs. A source of experimental error may follow a consistent pattern, or cycle, depending on the hour within a day, season of the year, catalyst age, age of equipment, and so on. In designing an experiment it is important to come as close to the ideal case as possible by careful choice of experimental apparatus, good experimental controls, and so on. The requirement of “uniform error” means that the experimental error has approximately the same magnitude at all experimental points. The prediction variance should also be relatively constant at varying distances from the center of the design out to the boundary of the experimental region.