Resumo:
We show that it is possible to estimate kinetic parameters for complex mechanistic polymerization models from available industrial data. A methodology is developed for efficient handling and reconciliation of industrial data and is then applied to allow estimation of kinetic parameters for industrial ethylene polymerizations performed in reactor trains using soluble Ziegler-Natta catalysts. The parameter estimation procedure is formulated as a nonlinear optimization procedure subject to hard and soft model and process constraints. Parameter estimates obtained for the catalyst system allow a very good description of actual industrial data used during the estimation process and also allow very good prediction of process performance when completely new operating conditions are considered. It is concluded that complex phenomenological models can be successfully fitted to actual industrial processes without the need to carry out extensive experimental tests in the laboratory.
