'Modeling of Gas Distribution in an Indoor Space' simulation project by vgon_alves


I created a new simulation project called 'Modeling of Gas Distribution in an Indoor Space':

The purpose of this project is to validate the accuracy of the solver for transport and distribution of Gas (passive scalar) in an indoor environment.The airflow pattern (velocity) and concentration results from CFD are to be compared with experimental data obtained from full-scale tests.

More of my public projects can be found here.




Mining is known as one of the world’s most dangerous occupations. Many serious accidents have taken place in several parts of the world, that resulted in loss of life. Emissions and accumulations of Methane gas at a working face in underground coal mines can result in a major explosive event if not taken care of effectively [1]. CFD here will be used to predict the spread and distribution of a gas in a proposal ventilation setup and vising to design more effective ventilation.

The purpose of this project is to validate the accuracy of the SimScale solver for transport and distribution of Gas (passive scalar) in an indoor environment.The airflow pattern (velocity) and concentration results from CFD will be compared with the data obtained.


Geometry was created by using Ansys SpaceClaim v17.2. According to the source, the air inlet and the outlet are separated by a curtain, that in this case exposed by SimScale has 25 ft (7.6 m) of distance from methane inlet.

Fig 1. Dimensions of the geometry


The mesh used in the simulations of this case were created in SimScale using first order Tet-Dominant with 13285 points and 68764 elements.

Fig 2. Visualization of the mesh with a cross section of the mesh

Simulation Setup

Using Multiphase model and Passive Scalar Transport to validate the simulation:

  1. Was raised a new simulation run where the properties (density and viscosity) of methane are same as that of air. This way we can make sure if it is an expected physical phenomenon that we see (in the video below) - the flow in case of same material properties should be more or less equal to that of incompressible flow. To compare this, was used the next suggestion;

  2. Was raised a passive scalar simulation with properties of air and defining methane as another passive scalar. We can see the movement of methane only with respect to flow properties, leaving alone the material differences (streamlines below).

The first fluid used in the simulations was incompressible Air with a kinematic viscosity υ = 0.000015295[m²/s] and a density ρ = 1.22[kg/m³]. The same way, the next fluid is methane, with the same properties υ = 0.000015295[m²/s] and a density ρ = 1.22[kg/m³]. The turbulence models used were k-epsilon at transient wich uses MULES phase fraction solver.

The accuracy of the results is highly dependent of the flow profiles of mean air speed and methane quantities that are applied in the inlet plane. These profiles should be fully developed and representative at the domain, according to the source.


Fig 3. Boundary conditions

Results and discussion

At the end of the 90 seconds to which the simulation runs were submitted, the results for methane velocity and quantity (alpha_phase1), as close to the source as satisfactory. To get even closer and eliminate errors, it would take a race with a higher time step, but due to the instabilities with the Courant number, the 90 seconds were considered very close to the comparative level.

Valocity U

Fig 4. Air streamlines

Methane Gas Distribution

The coefficieint of determination R²


0.7 to 0.9 positive or negative indicates a strong correlation [3]. After that, the results could be compared with the source and indicates wichs points this one needs to be improved to abtain more fine result aproximation with the source.

Comparison with experimental data in the source:
The velocity at the source shows as 2.3 [m/s], which make this simulation valid with 2.23 [m/s]. The distribution of methane from the source has similar behavior to what the video shows, with the values ​​of the Pearson’s correlation coefficient next to the latter also.


  • [1] Zhou Lihong, Pritchard Christopher, Zheng Yi “CFD modeling of methane distribution at a continuous miner face with various curtain setback distances”

  • [2] Moriasi DN, Arnold JG, Van Liew MW, Bingner RL, Harmel RD, Veith TL. Am Soc
    Agric Biol Eng 2007;50(3):885–900

  • [3] Taylor CD, Chilton JE, Goodman GVR. Guidelines for the control and monitoring
    mining operations of methane gas on continuous. Depart Health Hum Serv, Inform Circulat 2010;3:322–40

@bdelatti, @sjesu_rajendra and @jousefm for all support and patience.

Solve it with SimScale - Hall of Fame

Fantastic job Vinicius!! Congrats!


Nice job, Vinicius! :slight_smile: