Abstracts

ABSTRACTS

 

Tabulation of Dynamic Adaptive Chemistry for simulation complex chemical problems

Nicolas Bourgeois (UCL)

Due to their highly non linear and very stiff behaviors, the systems of ODEs resulting from the chemical kinetics equations have to be solved using specific solvers that usually scales as O(N^2) with N the number of species of the kinetic mechanism.  Coupling CFD simulations to detailed kinetic mechanisms (i.e. with more than 1000 chemical species) is unaffordable in terms of CPU cost as such systems have to be solved in each cell. Reduction techniques have to be employed to tackle this problem. Such strategies act on two main aspects: first, the number of species can be dynamically decreased by selecting the deemed active species; the second option is to tabulate the results of previously computed solutions and to interpolate them in cells where conditions are assessed to be sufficiently similar. The TDAC (Tabulation of Dynamic Adaptive Chemistry) algorithm combines those two reduction techniques. In the present study, the low-compressible reactive RANS equations have been coupled to the TDAC algorithm and are solved using OpenFOAM. TDAC will also most probably be part of the next official release of OpenFOAM. Our current studies concern simulations of Rapid Compression Machines (RCM). A RCM is an experimental device dedicated to the investigation of auto-ignition phenomena. A reactive mixture is placed inside a reactive chamber that is rapidly compressed by a single-stroke piston that stops at top-dead-center, essentially reproducing the first part of the cycle of an internal combustion engine (compression followed by fuel ignition). A specific solver dedicated to RCM simulations has been developed, basically starting from the reactingFoam and engineFoam solvers available in the OpenFOAM library. Together with the TDAC algorithm, this solver allows us to obtain results that are the first to couple high fidelity CFD simulations to detailed chemistry in the context of RCM studies.

 

Using OpenFOAM® for engineering consultancy

Tom Fahner (Actiflow)

Although many people have heard about OpenFOAM as an open source alternative for CFD simulations, there are still some doubts about accuracy, ease of use and there is a general fear for learning a new code. Actiflow has been using OpenFOAM since 2007 for various projects involving flow related engineering problems. During the presentation a wide range of projects will be shown were OpenFOAM was used to solve the fluid problem. Several of the projects involved some development work, for others standard solvers were sufficient. The markets where Actiflow is present are the built environment, the automotive sector, the process industry and the HVAC sector. We show that OpenFOAM can provide solutions to problems in all these markets. We hope that this presentation helps people to think about using OpenFOAM for their CFD work.

 

Wind Resource Assessment by using OpenFOAM® / WRF simulations

O. Temel, J. van Beeck (VKI)

Computational fluid dynamics (CFD) has been used for the modeling of the atmospheric boundary layer for the wind energy assessment studies in the last decades. In spite of the high spatial resolution that can be reach by the CFD models, the inflow conditions of CFD simulations depend on large-scale atmospheric motions that the CFD codes are not designed to work with. Therefore, CFD codes must be coupled with the numerical weather prediction (NWP) codes that can sustain realistic inflow conditions in order to perform more reliable wind energy resource studies. This present study is devoted to the development and application of a numerical methodology for the modelling of atmospheric boundary layer by using coupled OpenFOAM/WRF simulations.

 

 

Dynamic Reduction of Detailed Kinetic Mechanisms

F. D’Ambrosio, A. Cuoci, A. Parente (ULB)

The description of particular combustion phenomena, e.g. flameless combustion, pollutants formation etc., requires the use of large kinetic mechanisms. CFD simulations of a practical system that involve also complex geometry, heat exchange, radiation, and turbulent flows are prohibitive in industrial applications due to the high computational cost. Therefore, reduction of detailed kinetic mechanisms is necessary to allow the resolution of practical problem with the required accuracy and within an affordable computation time. Usually, kinetic mechanism reduction is performed via pre-processing methods based on the analysis of the kinetic mechanism in a range of operating conditions. However, during
the simulation the kinetic scheme is not adapted to the local conditions. The present work aims to couple an on-the-fly reduction method for chemistry with two combustion solvers edcSMOKE, an open solver for turbulent combustion, and laminarSMOKE, an open solver for laminar combustion. Validation is performed through a series of 2D simulations, using large kinetic mechanisms. The sensitivity of the reduction method to the modelling parameters is also presented.

 

 

OpenFOAM® simulations of the influence of breaking logs on the flow pattern in a lock chamber

W.A. Breugem, B. Decrop, P. Rauwoens, K. Verelst and W. Van Hoydonck (IMDC)

Locks are key structures for the accessibility of ports and navigable waterways. The filling of the lock chamber has to be done with special precautions, taking into account the forces experienced by the moored vessels during the levelling process. In case of filling a lock with openings in the lock gate, one of the potential methods to minimize the hydrodynamic forces on the ships, is to insert breaking logs (i.e. energy dissipation bars) at the downstream side of the lock gate, aiming at an enhanced spreading and energy dissipation of the filling jets.
The present investigation studies the influence of breaking logs on the flow pattern in a lock chamber using OpenFOAM. Thereto, two different kinds of simulations were performed:
• Simulations  were performed of a configuration with a single opening in the lock gate with and without breaking logs using simpleFoam. The results were compared with data from physical model experiments of the same configuration.
• Simulations were performed  of the filling process of a complete lock chamber with six openings using interFoam. This was done for two different configurations of openings in the lock gate. Both configurations were simulated both with and without breaking logs.
In the presentation, the results of these simulations will be discussed.

 

Periodic reactive simulations

Jens Dedeyne, Laurien A. Vandewalle, David J. Van Cauwenberge, Kevin M. Van Geem, Guy B. Marin (UGhent)

Even though LES is being used extensively to evaluate flow features in many different applications, RANS simulations remain the industrial standard as it is less computationally demanding. This is especially the case for simulation of steam cracking, where detailed kinetics increase the computational demand even more. To reduce the computational cost of these simulations, the domain size can be decreased drastically by implementing streamwise periodic boundary conditions (SPBC). However, the underlying assumptions of constant property flows limit this to non-reactive systems.
A simulation approach was developed to extend the existing methodology to non-equilibrium kinetics. This novel methodology was validated by comparison with full-scale CFD simulations for an industrially applied propane pyrolysis reactor.

 

Numerical Simulation of MILD Combustion with OpenFOAM®

Zhiyi Li, Alberto Cuoci, Alessandro Parente (ULB)

The present work focuses on the validation with OpenFOAM for the numerical simulation of turbulent combustion under Moderate or Intense Low-oxygen Dilution (MILD) combustion. Results for a Jet-in-Hot-Coflow (JHC) burner using a mixture of CH4 and H2 50/50 on molar basis are presented and discussed, with the objective of showing the sensitivity of the results to modelling parameters and numerical approach. Simulation were performed using Eddy Dissipation Concept (EDC) combustion model combined with edcPimpleSMOKE, a RANS transient solver coupled to the OpenSMOKE tool for stiff chemistry management. A 2D mesh with 30,000 computational cells is used. The sensitivity of the simulation results to different modelling choices is presented. This includes the sensitivity to the choice of the k-epsilon model parameters, the choice of canonical reactor simulating the fine structures and the EDC model coefficient. Moreover, the impact of molecular and turbulent diffusion on the results will be assessed, by investigating the effect of turbulent Schmidt number and considering the unbalanced influence of thermal to mass diffusion with non-unity Lewis Number[5]. Finally, the combination of EDC combustion model with in-situ adaptive tabulation (ISAT) is presented, to assess the potential of computational time reduction.

 

 

Combined Error Estimation and Shape Optimization using adjoint Methods

Joao Duarte Carrilho Miranda, VUB

Ajoint formulations have been extensively used for shape optimization problems as they allow to compute the gradient of the objective at the cost of roughly an extra single CFD simulation, and this independently of the number of design variables. As industrial design problems usually have many design variables, the adjoint approach outperforms the more  traditional approaches to find the sensitivities such as a finite differencing approach. Another application of the adjoint method is in adjoint-based a posteriori error estimation where the adjoint solution is used to estimate the error on the objective and to carry out grid adaptation. We propose to combine both applications of the adjoint into a one-step combined design optimization and a posteriori  error estimation methodology using a single adjoint solution. The developed approach is applied to relevant engineering problems, such as geometrical optimization of pipe flows. The design variables are the shape coordinates and no parameterization is used. The results indicate a significant reduction of the computational costs compared to the standard optimization techniques.

 

Participation fees

50 €

The price includes the lunch and coffee breaks.