#	Pagina
attuale pagina	/open-h2020/projects/205438/results.html

Report

Teaser, summary, work performed and final results

Periodic Reporting for period 1 - UTOPIAE (Uncertainty Treatment and OPtimisation In Aerospace Engineering)

Teaser

Summary

UTOPIAE (Uncertainty Treatment and OPtimisation In Aerospace Engineering) seeks for the ideal compromise between performance and resilience, optimality and reliability in complex engineering systems and processes by embracing the inherent stochasticity of Nature and using optimisation to harness reliability and robustness into a better future.
In an expanding world with limited resources, UTOPIAE aims at developing solutions for a more sustainable and resilient future. By using aerospace as a paradigm of complex engineering system, UTOPIAE directly addresses some of the global challenges for sustainable development and the objectives of Europe in the areas of Transport, Advanced Manufacturing & Processing, and Space.

Work performed

The UTOPIAE programme started on the 1st of Jan 2017 with a kick-off at the University of Ghent on the 20th of Jan 2017.
The first training event, the Opening Training School (OTS), was held on 20th - 24th November 2017 at the University of Strathclyde. More details on the programme of the school can be found at http://utopiae.eu/2-2/utopiae-training/opening-training-school/
The OTS lectures are currently available on a password protected YouTube channel and in pdf format. The same lectures are now being transcribed into a series of chapters that will form a book of lecture published by Springer in 2018.
At the OTS the first batch of Project Working Groups were formed.
The first Local Training workshop (LTW-I) was at the University of Strathclyde in February 2018. LTW-I featured 1 full day of outreach activities, carried out by the ESRs, in collaboration with the Glasgow Science Centre, two days of hands on exercises on optimisation and UQ, one full day of complementary skills, and one day of team work to start the first round of Project Working Groups. More details: http://utopiae.eu/2-2/utopiae-training/local-training-workshop-1/.
In July 2018 the second training school, TS-II, was organised at University of Durham. The school focused on Uncertainty Quantification with a number of lectures and seminars specifically on Imprecise Probabilities. At the TS-II the first batch of Project Working Groups presented their first results. More information: http://utopiae.eu/2-2/utopiae-training/second-training-school/
ESRs have already presented the first results at conferences between May and October 2018. In particular, the following conferences were targeted: modeFRONTIER User conference in May, ECCOMASS conference in June, WCCI2018 in July, SECESA2018 in September and IAC2018 in October.
A UTOPIAE project exists on Research Gate and twitter and the webpage has been used to communicate with the general public and other scholars. At the OTS there was an afternoon session introducing the concept of public engagement. The ESRs in groups developed activities to deliver to school students. Additionally, a public lecture was organised. The lecturer, an expert and book writer from Germany, JÃ¶rg Bewersdorff, author, of Luck, Logic and White Lies, was then interviewed by a local STEM magazine (more information: http://utopiae.eu/2-2/utopiae-training/opening-training-school/ots-public-lecture/).
During LTW-I, the ESRs hosted a day of activities for families and school groups. The event, hosted by Glasgow Science Centre, was very successful and the ESRs managed to explain some fundamental concepts relating to optimisation and uncertainty quantification via a number of games and simple examples to hundreds of people (more information: http://utopiae.eu/2-2/utopiae-training/outreach/outreach/).
During the TS-II, a planning session further developed engagement activities for future events. This session included a presentation by Guy Ross from the Durham University outreach department. A public lecture on Randomness and Imprecision was delivered by Prof Gert de Cooman from Ghent University (more information: http://utopiae.eu/2-2/utopiae-training/second-training-school/tsii-public-lecture/).
ESRs also participated in European Researcherâ€™s Night events (more information: http://www.explorathon.co.uk/events/secret-world-of-space-junk/ & https://twitter.com/JSI_CompSystems/status/1045652478287581185).

Final results

UTOPIAE is addressing how to engineer systems that maximize performance (minimize the use of resources) and are robust and reliable against possible disruptive events.
New models were developed for space systems engineering, in particular a new concept, named Evidence Network Models (ENM), was introduced as a new way to model complex systems in the form of complex networks. Multifidelity computational models have been developed for the aerodynamics of aerofoils, water impingement in aerodynamic profiles and-atmospheric flight based on experimental data.
New results have been obtained on methods for sensitivity analysis based on importance sampling, direct and transparent comparison of Bayesian and frequentist regularization methods, links between regularization, imprecision and credal classification.
Initial results have been obtained to improve the efficiency of reliability analysis, Evidence-Bared Robust Optimisation and sequential filtering with imprecise probabilities. Within the context of standard probability measures and UQ in CFD, the concept of the accelerated Schwarz method has been demonstrated, with a significant acceleration of convergence.
New algorithms have been developed for the solution of constrained min-max problems, which represent a cornerstone of optimization under uncertainty as they allow one to address worst case scenarios and low probability catastrophic events. Some progress has been made in the reduction of computational cost for many-objective optimisation using indicator-based optimisers (SMS-EMOA). Computational cost reduction has been obtained also for mixed-integer problems under uncertainty by using LASSO.
Significant theoretical and algorithmic results have been obtained by the application of clever decomposition techniques to Evidence Network Models of complex space systems. The decomposition approach developed in the context of ENM has been demonstrated on medium scale problems. In addition reliability measures have been included so to tackle a new conceptual problem: resilience optimisation and resilience-based design.
UTOPIAE has already produced significant advances in a number of research areas. With some of our researchers working on their first journal publications and two books of lectures nearly ready to be submitted to the publisher we expect to have a more measurable impact in one-year time when we will be able to quantitatively evaluate how many people in the community have adopted the results of our research.
A first result was to raise awareness on the practical benefits of non-traditional probabilistic approaches to handle uncertainty that emerge naturally in a number of processes, from simple decision making to the tracking of space objects.
The work developed so far within UTOPIAE is still in its infancy given ESRs commenced their activities only 12 months ago. Nonetheless, the introduction of new efficient methods for engineering resilience in aerospace systems and improve performance and reliability will have a substantial impact on the sustainability of the air and space sectors.