Assessment method of dynamic, stochastic process models with an application to papermaking

The purpose of this paper is to present a method for assessing whether optimized process design and operations, based on simplified simulation models, can be carried towards real life. The proposed method relies on a more realistic, or if available, a set of more and more realistic simulation models being compared with the original model utilized in the optimization problem. The models which are utilized are dynamic and stochastic in nature, necessitating statistical analyses of the results. The method is applied recursively to a generic papermaking process model with the addition of increasingly more realistic modifications. It is shown how the proposed comparison method detects the effects of such modifications and thus provides information on how well the optimization results are applicable to real life situations. The purpose of the presented method is to add one element to an overall framework of simulation-aided process design.