Comparison of low- and high-fidelity approach in model based design in the case of a portable motion platform
Editor: Udo Lindemann, Srinivasan V, Yong Se Kim, Sang Won Lee, John Clarkson, Gaetano Cascini
Author: Ellman, Asko Uolevi; Krus, Petter; Jouppila, Ville
Institution: 1: Tampere University of Technology, Finland; 2: Linköping University, Sweden; 3: Tampere University of Technology, Finland
In this paper a low- and high-fidelity modeling approach was applied and compared in the case of design of a portable motion platform. In the low-fidelity design approach, a Design Analysis Tool was used for generating information about design sensitivity analysis and correlation of design parameters. In the high-fidelity design approach, the motion platform was designed using Matlab/Simulink/SimMechanics including the mechanical and pneumatic subsystem of the platform. These approaches were compared in terms of required modeling effort to a relative error of predicted and measured system properties. The results of a low-fidelity model were achieved within one day whereas high-fidelity approach required three weeks of work. The relative error in low-fidelity approach was about 21-25 % and 4-6 % in high-fidelity modeling. Based on experiences achieved in this design case both of these modeling approaches were justified. Low-fidelity model can be seen especially important in early design phase when design alternatives were speculated with design constrains. High-fidelity model was useful on detailed design when dynamic properties of the system were considered more detailed.