
Postdoctoral fellow
Department of Aerospace Engineering
Universidad Carlos III de Madrid
Av. de la Universidad, 30, Leganés, 28911 Madrid, Spain
Mail : gcornejo[at]ing.uc3m.es
GitHub: https://github.com/gycm134
Google Scholar: https://scholar.google.com/citations?user=1TFIbWcAAAAJ&hl=fr
My research interest is flow control, and I am especially interested in exploring/discovering noteworthy dynamics thanks to machine learning techniques.
My collaboration with Profs. Bernd R. Noack and François Lusseyran led us to the development a machine learning software — xMLC, Machine Learning Control — to tame nonlinear complex systems. The purpose of xMLC is to solve non-convex minimization problems, whether in simulation and experiment, that are typically at the heart of engineering control problems.
Our latest advancement in MLC is the acceleration of the automated learning for multiple-input multiple-output feedback laws by one order of magnitude with the introduction of gradient-based techniques. Our “gradient-enriched machine learning control” (gMLC), published in the Journal of Fluid Mechanics, Volume 917, A42 (freely available on https://doi.org/10.1017/jfm.2021.301), has been successfully employed to control numerical plants and low to high-Reynolds number experiments.
Our software and tools have been developed to analyze and control the Fluidic Pinball, a fluid control benchmark that is geometrically simple yet presents remarkable dynamics.
Our closest collaborators are Prof. Marek Morzynski of Poznan University of Technology for the numerical simulations and Prof. Luc Pastur of ENSTAParis and Nan Deng, postdoctoral fellow at HIT Shenzhen, for the analysis and modeling of the Fluidic Pinball.