Abstract
This study is focused on several stages of an identification methodology and consists of selection of the parameters of influence and their identification considering state observations. A partial differential equations system describing the temperature evolution in a plasma-assisted chemical vapor deposition process is investigated. A preliminary study based upon the numerical design of experimental method leads to determination of the parameters which have to be carefully estimated since their uncertainties sharply reduce the adequacy of the model. Then, a sensitivity analysis is performed and the sensitivity problem derived from the direct problem is solved. Optimal observation strategy is briefly discussed in order to obtain state observations for the unknown parameters identification, and a conjugate gradient method is implemented for the resolution of the ill-posed inverse problem.