User login

You are here

Ph.D. position available in France/Belgium: Optimization methods supported by multi-fidelity full-field meta-models and enrichment strategies dedicated to turbine engine design

breit's picture

Safran Aircraft Engines is a world-class engine manufacturer in the aerospace propulsion market - civil and military - and space. At CFM International, Safran Aircraft Engines develops and produces the CFM56 "bestseller", the most reliable engine of its generation in the single-aisle aircraft category. LEAP, its successor, was designed to equip new generations of single-aisle aircraft. Incorporating the most advanced technologies, the LEAP engine will offer operators a two-digit reduction in fuel consumption and CO2 emissions, compared to the best CFM engines currently in service. It will enter service in 2016 on the Airbus A320neo. Safran Aircraft Engines is also a major player in the field of military propulsion: the company has developed and produced the M53 and M88 engines which respectively equip the Mirage 2000 and Rafale fighters. Finally, Safran Aircraft Engines offers its civilian and military customers a complete range of supports and services, in order to optimize the availability of aircraft.

 * CFM56 engines are produced by CFM International, a 50/50 joint venture between Safran Aircraft Engines and GE.

 Context

 Current environmental issues are making the design of aircraft engines more and more constrained. Thus, multidisciplinary optimization tools based on numerical simulations of increasing fidelity constitute a key technology for a motorist as Safran Aircraft Engines is part of a process of innovation and search for breakthrough solutions. A possible solution to the exorbitant cost of optimizations based on 3D numerical models consists in the exploitation of meta-models built on the basis of information at different levels of fidelity and different calculation costs [Ken2000]. At the same time, the reduced-order methods have shown their effectiveness in taking into account complex physical phenomena that have a strong impact on machine performance [Coe2008]. Effective use of meta-models combining paradigms "multi-fidelity" and "reduced order model" [Ben2017] [Mif2016] requires rethinking of optimization loop aided by meta-models as well as adaptive enrichment criteria.

Objective

 The objective of this thesis is the development of multi-fidelity enrichment criteria allowing the selection of points of interest in the design space. The enrichment strategies developed will, among other things, aim to improve the quality of the meta-models by selecting the level (s) of fidelity that maximizes the gain/cost ratio [Rei2010]. These developments will have to be integrated into an online optimization scheme, with strongly constrained dynamics and aiming at large design spaces. For this purpose, a work on the sampling methods will have to be realized as well for the global construction ab initio as for the progressive (global/local) sampling during the optimization.

The developments will be realized within the Minamo optimization platform (developed by Cenaero) to take advantage of meta-models and evolutionary optimization algorithms commonly used in the context of constrained multi-disciplinary designs within the Safran group. After demonstrating the methods developed in the framework of this thesis on benchmark cases (of the literature or to be defined), these approaches will be validated on an industrial application still to be defined with the various stakeholders of this project.

Description of the mission

 The thesis will be jointly supervised by Safran Aircraft Engines, Cenaero (Gosselies, Belgium) and Université de Technologie de Compiègne, in particular, the Roberval laboratory (Compiègne, France). At Safran Aircraft Engines, the Ph.D. student will be part of the optimization team at the Methods Office. The thesis will be supervised by the Minamo development team at the Cenaero research center and will take place largely on their premises. Periods of the presence of varying durations within the Roberval laboratory of UTC (Compiègne) and on the Villaroche Safran site will be planned.

Candidate Profile 

You have graduated in mechanics and have solid skills in applied mathematics.

You have completed an internship in one of these areas.

You have a pronounced taste for programming and methodological development.

You have a good level of English (French is a plus).

Pragmatic and rigorous, your scientific curiosity and your technical background will allow you to quickly grasp the industrial stakes of this thesis. Beyond your technical skills, your analytical, synthesis and communication skills are real assets to carry out this thesis work.

Contract length: 3 years

Contact

If you are interested, please send a cover letter as well as your CV to: piotr.breitkopf@utc.fr

http://roberval.utc.fr/Breitkopf-Piotr

http://www.cenaero.be

https://www.safran-aircraft-engines.com 

Subscribe to Comments for "Ph.D. position available in France/Belgium: Optimization methods supported by multi-fidelity full-field meta-models and enrichment strategies dedicated to turbine engine design"

Recent comments

More comments

Syndicate

Subscribe to Syndicate