OpenFOAM Workshop 6 September 2024

OpenFOAM

06 September 2024 13:00 till 16:45 - Location: Faculty of EEMC, Lecture room G, HB00.230, Mekelweg 4, Delft | Add to my calendar

On September 6, 2024 there will be an OpenFOAM Workshop at the TU Delft organized by the Dutch OpenFOAM user group. Four speakers will give a talk on how OpenFOAM can be used for various applications.

Presenters

Daniel van Odyck

Simulation of an industrial scale reducing electric furnace

The production of steel accounts for around 7% of global greenhouse gas emissions and 11% of global carbon dioxide emissions [1]. To gradually reduce CO2 emissions to zero, a challenging decarbonization route has been mapped out for the steel production at Tata Steel Netherlands. One of the proposed routes is to produce hot metal (liquid iron) by producing reduced iron ore pellets in a Direct Reduced Iron (DRI) plant in combination with a Reducing Electric Furnace (REF) as a smelter. The produced hot metal then goes into consecutive processing steps to produce liquid steel and finally, coiled steel sheet. In this research, the focus is on the REF. The REF consists of a hot metal layer with on top a slag layer, both in the liquid phase. The REF is equipped with electrodes and heat is added via Joule heating. In between the electrodes the feed piles of DRI and other fluxing and trimming agents are positioned. Modelling this process on industrial scale presents considerable challenges. Therefore, several models have been developed, each considering specific aspects of the process. The current study focuses on the effects of buoyancy and magneto-hydrodynamics on fluid flow and heat distribution inside the REF.

Tyler Buchanan

Data-Driven Turbulence Modeling for Complex Flows

Computational Fluid Dynamics (CFD) is crucial in engineering applications, particularly in turbulence modeling. This presentation focuses on addressing the challenges of accuracy and efficiency in CFD simulations through data-driven techniques. Reynolds-averaged Navier-Stokes (RANS) simulations are widely used for their speed and cost-effectiveness, but they often struggle to accurately predict flows with massive separation. This research explores the development and calibration of closure models for RANS equations using machine learning (ML), deep learning, and artificial intelligence (AI). Despite RANS models being essential for many engineering applications, they face significant inadequacies for complex, non-equilibrium flows. The increasing availability of high-fidelity datasets from Large Eddy Simulation (LES) and Direct Numerical Simulation (DNS) enables the use of ML techniques to derive improved closure models with minimal computational overhead. By bridging the gap between RANS and reality, this research aims to enhance the predictive capability of RANS simulations, ultimately contributing to more accurate and efficient turbulence modeling. Overall, the objective is to develop accurate and interpretable turbulence models for complex flow problems using data-driven approaches.

Praveen Doddugollu

A Multiphase Mixture Model for Mud in OpenFOAM

A numerical three-phase model is crucial for accurately simulating fluid flow in engineering applications involving mud as a seabed in free-surface flows. In these scenarios, water and sediment (mud) mix together, while the air and the water-sediment mixture remain immiscible. Current numerical models often simplify by treating water-sediment as a single phase, employing Reynolds-Averaged Navier-Stokes (RANS) equations supplemented by a sediment transport equation. However, these approaches often fail to adequately capture the complex two-way and four-way interactions between phases, thereby neglecting relative velocity and momentum transfers. Mixt3MudSedFoam model has been developed specifically for cohesive sediments (mud) within the OpenFOAM framework, adapting from an initial three-phase model developed for non-cohesive sediments (sand). This model directly integrates relative velocity and momentum transfer between the phases into the continuity and momentum equations. The mud behavior is incorporated through closures derived from rheological experiments and consolidation experiments. The rheology experiments are used to include mud's non-Newtonian properties and time-dependent thixotropic effects through a viscosity closure. Additionally, closures from consolidation experiments are employed to address permeability and effective stress considerations. This presentation will introduce the mathematical formulation of the three-phase mixture model and discuss the implementation of these closures within OpenFOAM. Validation results of the model against benchmark experiments involving a cylinder and a hydrofoil moving in a muddy environment will be presented. Finally, the presentation will conclude by highlighting the solver's limitations, advantages and future improvements.

TBD

Program

13:00 - 13:30
Arrival with coffee and tea

13:30 - 14:15
Daniel van Odyck (Tata Steel) - Simulation of an industrial scale reducing electric furnace

14:15 - 14:45
Tyler Buchanan (TU Delft) - Data-Driven Turbulence Modeling for Complex Flows

14:45 - 15:15
Coffee/Tea

15:15 - 15:45
Praveen Doddugollu - A Multiphase Mixture Model for Mud in OpenFOAM

15:45 - 16:15
TBD

16:15 - 16:45
Wrap up

Directions

Faculty of EEMC, Lecture room G, HB00.230
Mekelweg 4
Delft 2628CD
Netherlands