Optimization of Modelica Models Utilizing Design of Experiments and Robust Design Methods

Abstract

This thesis describes the implementation of Robust Design and Design of Experiments methodologies with the FMI Microsoft Excel Add-in (FMIE). FMIE allows engineers to use Microsoft Excel in collaboration with simulation models created in other software and exported as FMUs. A choice of pre- and post processing methods are implemented as Excel macros for use with the FMIE. Their functionality are demonstrated with a non trivial parameter optimization of a truck model created with Dymola. The development of the truck model and integral parameters are explained within this thesis. The Excel macros are used to find a design point that is sufficiently good in aspect to two response factors. The stability of this design point with aspect to a noise factor and small variance of the design factors is determined.