This is the website of the research team on Computational Methods in Fluid Engineering.
Our main interest is the development of computational simulations, tools, algorithms, analysis, and designs in the field of fluid engineering, a branch of mechanical engineering, by using mathematical methods.
This involves a wide range of applications, from thermal engineering, aerospace or aerodynamics to microfluidics or oil&gas industry.
Our research team has leading-edge publications and solid expertise on:
Our main interest is the development of computational simulations, tools, algorithms, analysis, and designs in the field of fluid engineering, a branch of mechanical engineering, by using mathematical methods.
This involves a wide range of applications, from thermal engineering, aerospace or aerodynamics to microfluidics or oil&gas industry.
Our research team has leading-edge publications and solid expertise on:
- Computational Fluid Dynamics (CFD). This is one of our most prolific fields of research. We simulate fluid flows for aerodynamics, heat transfer, ventilation, mixing or fluid-structure interaction using different methods (RANS, URANS, Lattice Boltzmann, LES, etc.) with different softwares/codes (ANSYS Fluent, Openfoam, Nek5000, in-house codes, etc).
- Multi-objective optimization. We have conducted many optimization studies and even developed new methods (e.g. Machine Learning-Aided Design Optimization [MLADO] or Additive Aerodynamic Shape Optimization [AASO]). We create optimization algorithms or methods that fit special needs and adapt to specific problems.
- Uncertainty Quantification & Sensitivity analysis. This is one of our favourites. We have applied probabilistic (Stochastic Collocation Method, Polynomial Chaos, Gaussian Processes pdf sampling) and physics-based non-probabilistic (Eigenspace Perturbation Method) Uncertainty Quantification methods to CFD simulations of aircraft exhaust jet flows, rotating pipe flows, swirling jet flows for heat transfer, active flow control on airfoils and jet stability analysis.
-
Data-driven methods and control engineering. Another favourite one! We work with Reduced-Order Modelling techniques (Proper Orthogonal Decomposition [POD], Dynamic Mode Decomposition [DMD], DMD with control [DMDc], etc.) in complex scenarios such as Fluid Structure Interaction (FSI) and sensor data from Oil&Gas drilling.
We also work with Machine Learning algorithms in design optimization, PINNs, porous media modelling, etc.
- Reliability and risk design. Pretty much linked to the previous UQ&SA skill. We quantify reliability and risk by using probabilistic methods and classification algorithms.
- Numerical PDEs. Solution to partial differential equations by using similarity transformations, uniform/non-uniform finite difference methods and PINNs.
- Mathematical modelling. We have developed, for instance, the first closure porous medium characterization model of wire woven screens (e.g. insect-proof screens for ventilation openings or fluid filters) that includes 3D modelling of porosity, constriction factor and aerodynamic tortuosity from experimental data and CFD simulations. This work was entirely semi-analytical.
- Algorithm&code development. We have developed some numerical codes for shape optimization with (additive) LBM [not yet publicly available], vortex shedding-based fluid mixing [private software], 3D porosity and geometry characterization of woven screens for ventilation (Poro3D v1.0) and full 3D porous medium characterization of woven screens (AeroScreen v2.0). We have also developed optimization methods such as the Machine Learning-Aided Design Optimization (MLADO), protected as Utility Model in the development of micromixers and selected as Editor's Pick by the Physics of Fluids journal.
[Picture]
Uncertainty Quantification and global Sensitivity Analysis in under-expanded jet flow instabilities