where Xi is the coded value of each factor; xcp is the value of the xi at
the center point obtained from the preliminary experiments; and Δxi
presents the step change.
Table 3 shows the CCD in the form of a 23 full factorial design with
four additional experimental trials (run numbers 15 to 18) as duplicates of the central point. Eighteen runs were performed sequentially. Each
run was per formed in tri plicate to guarant ee the reliab ility of result s
( er ro r li m it s le ss th an 1 0% ) . I n a dd it io n , a f ul l qu a dr a ti c m od el wi th
regression coeffi cients was selected to fi t t he e xpe ri me nt al da ta . O th er
than intercept, linear, and quadratic terms, this model also considers
two-way interactions: