groupe-dfo-bbo:acteurs:students:doc:solene-kojtych:main-page [Groupe de recherche en DFO-BBO]

Cette page est en lecture seule. Vous pouvez afficher le texte source, mais ne pourrez pas le modifier. Contactez votre administrateur si vous pensez qu'il s'agit d'une erreur.
==== Solène Kojtych ====

<grid>
<col md="2">
{{.:portrait.jpg?200}}
</col>
<col md="10">
  * Background: I arrived at Polytechnique Montreal in September 2016 for a double degree with [[https://www.utc.fr|Université de Technologie de Compiègne]] in France. I began in January 2017 a research Master's in mechanical engineering in the [[http://lava.polymtl.ca/|LAVA]] turbomachinery lab. I was really interested by the black box optimization course I attended and I wanted to combine both mathematical and mechanical aspects in a Ph.D. project. I had the opportunity to follow a Master’s-to-Ph.D. fast track program and I began my Ph.D. in January 2018.

  *  Project: My works focuses on the **development of a methodology for the design of aircraft engine blades, which involves time-consuming simulations and discontinuous functions**. As a consequence, I have developed a **new MADS algorithm revealing and escaping discontinuous areas of the space of variables**.

  * Supervisors:
    * [[https://www.polymtl.ca/expertises/audet-charles|Charles Audet]] (Polytechnique Montréal and GERAD, mathematics)
    * [[https://www.polymtl.ca/expertises/batailly-alain| Alain Batailly]] (Polytechnique Montréal, mechanical engineering) :

  * Research interests:
    * black box optimization, derivative-free optimization
    * design of strongly nonlinear mechanical systems
    * design of aircrat engine blades

  * Contact: <solene.kojtych@gerad.ca>
</col>
</grid>


==== Project overview ====

=== Context ===
<callout color="#fdb120">

<grid>
<col md="6">
  * Design of non-linear mechanical systems
    * **time-consumming simulations** are necessary to account for the nonlinear behaviour of the system (time integration...) => need for blackbox optimization algorithms
    * nonlinear features such as bifurcations cause jump **discontinuities** and **abrupt variations** of the quantities of interest => need for a special treatment of discontinuities

  * Why is it a new topic of interest?
    * maturity of mechanical simulations tools reached within the last decade => reverse design of system made possible 
    * previous successuful use of MADS for the simplified design of turbomachinery blades
</col>

<col md="1">

</col>

<col md="5">

  * An example of discontinuous quantity involved in the design of aircraft engine blades
{{ .frise_image_concours.png?nolink500 }}
</col>
</grid>


</callout>

=== Research axes ===

<callout title="Disco-MADS : revealing and escaping discontinous regions of the space of variables" color="#f66555">
{{ .disco_mads_schema.png?nolink800 }}
  * Supported by a **convergence analysis**
  * Applications :
    * Ensure safe structural design => design of a two-bar truss with buckling behaviour, simplified design of turbomachinery blades
    * Avoid regions where hidden constraints are violated => application to STYRENE
</callout>


<callout title="Optimal design of turbomachinery blades accounting for performance discontinuities" color="#f66555">
  * Developmnent of a blackbox related to the design of turbomachinery blades
    * accurate simulation of blade/casing interaction
    * simplified aerodynamic analysis
 
  * Optimization with DISCO-Mads
 
</callout>

<callout title="3rd project : to be defined" color="#f66555">
 
</callout>




==== Documents ====
<callout title="Presentations" color="#00274C">
    *{{.:autres:presentation_gerad_nov_2019.pdf|presentation at GERAD}} (28/11/2019)

</callout>