SAS: an advisory framework to support surrogate model based MDAO

Master Thesis (2022)
Author(s)

C.L.A. de Priester (TU Delft - Aerospace Engineering)

Contributor(s)

G. La Rocca – Mentor (TU Delft - Flight Performance and Propulsion)

Faculty
Aerospace Engineering
Copyright
© 2022 Costijn de Priester
More Info
expand_more
Publication Year
2022
Language
English
Copyright
© 2022 Costijn de Priester
Graduation Date
07-07-2022
Awarding Institution
Delft University of Technology
Programme
['Aerospace Engineering']
Faculty
Aerospace Engineering
Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Abstract

This work introduces the Surrogate Advisory System (SAS), a Python package that aims to reduce the existing complexity when working with surrogate-based MDAO systems. SAS automates most of the steps for a CPACS-based MDAO workflow and drastically reduces required efforts to integrate surrogate models within MDAO systems. Key developments enabling this automation, such as a newly developed Python interface with the Process Integration and Design Optimization tool RCE, a comprehensive surrogate modelling engine and a design database structure are discussed. Furthermore, an advisory framework providing insights into the inherent trade-off between computational efficiency and accuracy when working with surrogate models is presented. It is shown that a-priori estimation of a surrogate model's accuracy for a given number of samples is difficult, but that an estimation of the potential surrogate model candidates in a provided workflow can be made.

Files

License info not available