Dealing with Uncertainty in Polymer Manufacturing by Using Linear Regression Metrics and Sensitivity Analysis

Abstract

In the polymers industry, models are heavily parameterizised, and the effect of each parameter on model outputs has not been extensively studied. A wide range of values for most of model's parameters can be found in literature [1], a thorough analysis regarding the model's sensitivity to the parameters' values is needed to find the set of parameters that have the most impact in the output results and consequently deserve an extra effort and care during their estimation. In this work, a global sensibility analysis of a styrene emulsion polymerization reactor model is carried out in order to determine the set of critical parameters.