Designing experiments for precise estimation of all or some of the constants in a mechanistic model

TLDR: In this paper, statistical design procedures for estimating all the constants in a mechanistically based mathematical model are reviewed and a new procedure is illustrated for precisely estimating only some constants which may be of more interest than the others.

Abstract: An important experimental problem in chemical engineering is to collect data in order to estimate constants such as rate constants and activation energies which are needed, for example, in equipment and plant design. For a given amount of experimental effort, the engineer wants these constants to be estimated as precisely as possible. In this paper, we discuss statistical design procedures that might be used to accomplish this goal. Available procedures are reviewed for estimating precisely all the constants in a mechanistically based mathematical model. In addition, a new procedure is illustrated for precisely estimating only some of the constants which may be of more interest than the others. Un important probleme experimental, dans le domaine du genie chimique, est de recueillir des donnees pour estimer les constantes (comme celles qui ont trait a la vitesse et aux energies d'activation) qui sont necessaires, par exemple, lors de la conception de l'outillage et de l'usine. Pour un certain degre de travail experimental, l'ingenieur desire determiner ces constantes avec autant de precision que possible; on traite, dans le present travail, des procedes de calcul statistique qu'on peut utiliser pour atteindre ce but. On examine les procedes disponibles pour determiner avec precision toutes les constantes dans un modele mathematique du genre mecanistique On illustre egalement un nouveau procede pour estimer avec precision quelques constantes seulement qui peuvent presenter un interet parti-culier.