Automated Execution Process Formulation using Sequencing and Decomposition Algorithms for Collaborative MDAO
A.M.R.M. Bruggeman (TU Delft - Aerospace Engineering)
Imco van van Gent – Mentor (TU Delft - Flight Performance and Propulsion)
G. La la Rocca – Graduation committee member (TU Delft - Flight Performance and Propulsion)
Leo Veldhuis – Graduation committee member (TU Delft - Flight Performance and Propulsion)
M B Zaayer – Graduation committee member (TU Delft - Wind Energy)
More Info
expand_more
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
With Multidisciplinary Design Analysis and Optimization (MDAO) a fully automated aircraft design analysis is set up and optimization algorithms are used to obtain better designs by balancing the synergy between components. In the EU project AGILE, a new methodology and framework were developed to make the MDAO approach more accessible to industry. A key component of this framework is the KADMOS package. KADMOS is used to formulate large, heterogeneous MDAO problems and their execution process before they are implemented as executable workflows. This thesis focuses on the automation of a key step in the problem formulation for MDAO systems: the execution process definition, i.e. the order and grouping of the disciplines. Several sequencing and decomposition algorithms are developed to optimize the execution order of the disciplines and their division over multiple processors for parallel execution. The algorithms are verified and validated on thousands of MDAO systems using a scalable mathematical test case. Furthermore, the conceptual design of a conventional aircraft is performed using a novel implementation of the Initiator toolbox in KADMOS to test the algorithms in a realistic aircraft design problem. This showed that the algorithms resulted in a setup time reduction due to the automation of the execution process formulation and a reduced convergence time thanks to the improved usage of computational resources.