Session
| ||

Turbulent flow
| ||

Presentations
| ||

11:30am - 11:50am
A lattice-Boltzmann solver coupled to an immersed boundary method with applications to turbulent flows Aix Marseille Univ, France An immersed boundary method is coupled to a turbulent wall model and Large Eddy Simulation, with a Lattice-Boltzmann solver based on the Hybrid Recursive Regularized (HRR) collision model. The method is able to handle arbitrarily moving objects immersed at high Reynolds number flows and to accurately capture the shear layer and near wall effects. When dealing with high Reynolds number turbulent wall-bounded flows, modeling near wall turbulence, and more precisely the inner layer of the turbulent boundary layer, becomes a primary concern in many industrial fields, since this part of the flow governs most of the friction and heat/mass transfers at the wall. In this work, the power wall law, which is an explicit wall model without any iterative procedure for the determination of the friction velocity, is combined with the immersed boundary method (IBM). The IBM consists of defining a volume force term on the boundary between the fluid and the solid, calculated to satisfy the no-slip boundary condition at the wall. For the LBM, the recently developed Hybrid Recursive Regularized (HRR) model is used to enhance numerical stability and the algorithms are implemented within the ProLB software. We perform a thorough numerical study to validate and characterize the efficiency, stability and accuracy of the numerical framework using a set of test-cases of increasing complexity. The robustness and accuracy of the method are assessed first in a static laminar configuration, then in a mobile laminar case, and finally in a static and oscillating turbulent simulation. In all cases, the proposed method shows good results compared to the available data in the literature. 11:50am - 12:10pm
Large-Eddy Simulation of an axisymmetric and isothermal jet developing in a room based on the Lattice Boltzmann Method
Convective flows induced by HVAC (Heating, Ventilation and Air Conditioning) systems play an important role in the thermal comfort of buildings. However, they are particularly difficult to predict because they are not fully turbulent and interact with their environment. In order to better predict and understand these complex flows, this contribution presents the validation of Large Eddy Simulations (LES) [1] of an axisymmetric and isothermal jet developing in a room near a wall. Although rarely used in building applications, mainly due to its high computational cost, LES provides more detailed and accurate predictions of interior flows compared to statistical models (RANS - Reynolds Averaged Navier-Stokes) [2]. Moreover, this study is based on the Lattice-Boltzmann Method (LBM) [3] and the ProLB software [4]. This framework allows for massively parallel computations and easy preprocessing of complex geometries. Therefore, the approach chosen allows to efficiently perform detailed and high-fidelity simulations of complex environments. The simulations are validated and calibrated using an extensive experimental data set of a full-scale mechanically ventilated test room called MINIBAT [5]. To this end, numerical results are compared with experimental data in terms of vertical and lateral jet expansion rates, mean velocity profiles within the jet and turbulent quantities. The results show a good qualitative and quantitative agreement between numerical and experimental data: the anisotropy of the jet expansion is well recovered and the shapes of the profiles as well as the maximum values of the mean velocity are consistent with experiment. Also, a qualitative analysis of the jet turbulence distribution is performed through the observation of Lumley triangles and the visualization of turbulent structures with vorticity contour plots. The analysis reveals that the main turbulent mechanisms in the jet development zone are well captured. Hence, the modelling approach adopted is suitable for the intended building application. This work should be extended to anisothermal jet simulations to study thermal effects coupled with turbulence. It is also intended to couple the simulation of the hot jet with a high resolution building energy simulation [6]. The coupling will enable more realistic dynamic boundary conditions at the room walls to evaluate the dynamics of heat transfers and thermal comfort. [1] P. Sagaut, Large Eddy Simulation for Incompressible Flows: An Introduction, Springer, 2006. [2] B. Blocken, “LES over RANS in building simulation for outdoor and indoor applications: A foregone conclusion?,” Building Simulation, 2018. [3] T. Krüger, H. Kusumaatmaja, A. Kuzmin, O. Shardt, G. Silva and E. M. Viggen, The Lattice Boltzmann Method - Principles and Practice, Springer, 2016. [4] “http://www.prolb-cfd.com/,” [Online]. [5] F. Kuznik, “Etude expérimentale des jets axisymétriques anisothermes horizontaux se développant près d’une paroi: application à la modélisation numérique des cavités ventilées,” PhD Thesis, INSA, Lyon, 2005. [6] T. Gresse, L. Merlier, J.-J. Roux and F. Kuznik, “Development of a 3D and high resolution dynamic thermal model of a room with sun patch evolution for thermal comfort applications,” Building Simulation Conference Proceedings, 2021. 12:10pm - 12:30pm
Why is it still difficult to simulate turbulent two-phase flow at high density ratios with the lattice Boltzmann method? TU Braunschweig, Germany The lattice Boltzmann method is often claimed to be particularly well suited for the simulation of two-phase flows but simulating flows at the technically most relevant density ratio of the water/air system has remained difficult even after many promising proposals. Even though significant density ratios can be obtained in some models, their range of admissible Reynolds numbers is usually drastically reduced when compared to single phase models. Recent proposals to eliminate the strong gradients in momentum by switching to a velocity-based formulation lead to mixed results. In the velocity-based formulation the pressure is replaced with the normalized pressure (i.e. the pressure divided by density) which introduces a new discontinuity into the equation. Proposed solutions include the complete elimination of the pressure from the lattice Boltzmann equation which then needs to be supplemented by an external pressure solver. Another solution is to limit the simulations to very small pressure differences to keep the gradients in normalized pressure small enough to compensate for them by simple finite difference corrections. However, these corrections deteriorate quickly if the pressure is not extremely small. The reason for this is that the pressure term in the velocity-based Navier-Stokes equation is not conservative and that the implementation of non-conservative operators is generally difficult in the LBM. In this contribution we explore a new way of implementing non-conservative operators in the lattice Boltzmann framework and apply it to the pressure term in the velocity-based phase field LBM. |