The Parallel CFD & Optimization Unit (PCOpt) of the School of Mechanical Engineering of NTUA comprises 10 research engineers. Its research activities are related to the development of (CFD-based) analysis and optimization (via stochastic or gradient-based algorithms) methods/software for application in internal (including turbomachinery) and external (aircraft, automotive) aerodynamics.
Pioneering research is conducted in the field of continuous adjoint methods, especially for the computation of higher-order derivatives, including the differentiation of turbulence models, for creating optimization methods with reduced turnaround time or for solving robust design problems in aerodynamics; the development of adjoint methods and tools is based on both the in-house CFD code and OpenFOAM. PCOpt has developed a very efficient Metamodel-Assisted Evolutionary Algorithm (MAEA) based optimization platform EASY (Evolutionary Algorithm System, http://velos0.ltt.mech.ntua.gr/EASY). EASY is used by several academic groups and companies in a variety of applications.