Bio
Konstantinos obtained his MEng in Chemical Engineering at the National Technical University of Athens (NTUA), where during his diploma thesis he worked on self-consistent field theory to predict the surface properties of polymer melts. Fascinated by the idea of bridging the gap between chemical scientists and chemical engineers to address the emerging industrial challenges, he became a member of the first cohort of the IMSE MRes Programme in Molecular Science and Engineering at Imperial College London (ICL).
During his research project, he developed data generation procedures for the accurate parameterisation of reactive SAFT EoS, pursuing a 3-month industrial placement at Procter & Gamble Company in Cincinnati and PSE Ltd in London. Being well-experienced in the molecular modelling, he chose to "scale up" and be involved with the challenging world of fluid dynamics, carrying out his PhD in Matar Fluids Group at ICL. The objective of his research is to develop the proper numerics (CFD and sensitivity analysis/machine learning) that are required to understand the physics governing the spray formation featuring complex (i.e. viscoelastic) fluids.
Research interests
The jetting processes of viscoelastic fluids, known by their complex rheology due to the presence of dissolved macromolecules, are determined by various parameters (i.e. relaxation time, polymeric viscosity, injection flowrate) representing the effect of all the acting forces (elastic, inertial, viscous and capillary forces) on the flow behavior. Therefore, it is essential to explore and quantify how much each of the underlying input parameters influences specific outcomes that we are interested in about viscoelastic spray systems (i.e. rate-of-thinning, droplet size distribution, length of breakup).
Hence, the objective of the current research project at Turing is the implementation of sensitivity analysis and uncertainty quantification tools coupled with proper surrogate models and classification techniques (polynomial chaos expansions, Gaussian processes, deep neural networks) for the fast exploration of the entire parameter space of interest which finally rules the emergence of all the different flow regimes.