Low Order-Value Multiple Fitting for supercritical fluid extraction models

Abstract

Low Order-Value Optimization (LOVO) is a useful tool for nonlinear estimation problems in the presence of observations with different levels of relevance. In this paper LOVO is associated with a Multiple Fitting strategy for the estimation of parameters in supercritical fluid extraction models. Experimental data of supercritical CO2 extraction of peach almond oil are considered. Multiple fitting makes it possible to impose constraints on the estimation procedure that improve the physical meaning of the parameters. A novel combination of minimization methods is used to solve problems in the LOVO setting. Numerical results are reported.