turbulence-modeling

CFD-solver-agnostic deep reinforcement learning framework for computational fluid dynamics on HPC platforms

Turbulence modelling in CFD is limited by the tradeoff between accuracy and cost. We propose OT PINNs, Physics Informed Neural Networks with an Optimal Transport based loss, to improve training stability and accuracy under noisy data. With SINDy for interpretability, our method rivals DNS on benchmark flows while cutting computational costs.

This project implements a PINN using TensorFlow to solve a 2D steady-state convection–diffusion PDE on a unit square by minimizing the PDE residual and enforcing Dirichlet boundary conditions. It demonstrates domain sampling, differentiation, constrained training, and inference on unseen test points without requiring labeled solution data.

Automated calibration of RANS turbulence models for hypersonic flows using SciML. Achieved 61.3% RMSE reduction at Mach 14.

2D CFD simulation of NACA0012 at 5° AoA using SST k-ω model. Mesh convergence included.

Industrial CFD and thermal engineering portfolio featuring a vortex-controlled diffuser study for a jet-engine burner, CFD verification, turbulence modelling, parametric analysis, and Python-based engineering automation.

Learning the unresolved-scale closure of two-scale Lorenz '96 - the RANS/URANS averaging operation. Offline skill is a weak proxy for online skill.

OpenFOAM-10 CFD validation lab with four-grid lid-driven-cavity verification and paired k-epsilon/k-omega SST Pitz-Daily diagnostics, including Python post-processing, QoI stability audits, and reproducible public results.

Probabilistic closure model for coarse-grid Navier-Stokes: CNN + deep-kernel SVGP with calibrated uncertainty, deployed online in a MATLAB FV solver (MIT MSEAS)

PhD CFD portfolio: fluid-structure interaction, non-Newtonian rheology, turbulence-model assessment and Lagrangian particle transport in aneurysm hemodynamics.

Computational Analysis of Anisotropic Reynolds Stress Tensor

Real meteorological data (temperature, pressure, precipitation) were obtained from İSKİ. Eddy diffusivity, Monin-Obukhov length, and turbulence intensity were calculated from existing data and added to the dataset. Using 289,000 data points and 27 features, RF, SVM, LSTM, and CNN models were developed. LSTM achieved 98%, CNN 91% accuracy.

NASA Turbulence Modeling Resource NACA0012 vaka setiyle OpenFOAM doğrulama, çoklu AoA taraması

Offline wall-resolution (y⁺) calculator & CFD solver-setup advisor - turbulence models, schemes, inflation layers. Turkish UI.

Mach-4 supersonic-intake shock-wave / boundary-layer transition case study computed with the UniSTAR-CFD solver, validated against three real experimental benchmarks.

Validated NACA 0012 CFD and one-way wind-turbine FSI using ANSYS Fluent and Mechanical.

Rushton turbine stirred-tank CFD in STAR-CCM+: steady MRF and transient sliding mesh validated against the published Power Number (Po 4.85 / 4.89 vs 5.0), after a dimensional audit traced a 2.1× discrepancy to a blade placement defect.