CFD simulation/ dust distribution at pedestrian level

Hello everyone,

  • Overview about the project:
    my project is to investigate the dust flow and concentration at different points in outdoor space of small community before and after adding different types of trees as buffer zone around the community.
    the dust is presented as PM10 which is any particles less than or equal to 10 micrometers in diameter , these particles are one of the major pollutants in the air and have many bad effects on human health.
    first i want to clarify that PM10 are solids but once they are lifted by the air they are converted from a static solid-like state to a dynamic fluid-like state under the phenomena of fluidization.
    the study area surrounded by desert from three sides, which i will consider as the source of the dust in my simulation.

    with trees

i will simulate the windest day in three months before and after adding three types of plants where i will define the wind speed and direction in that day as will as the dust flux value as measured in the field in microgram/m2/s.

my questions are:
1- is incompressible analysis with k-omega SST Turbulence model/ Steady state / 1 Passive species, the right simulation?
2- what is the right way to define dust as passive scalar and the 3 surrounding area as the source of dust
3- how can i define the trees as porous media
4- is the program capable to present the dust concentration values and distribution in the results ?

i did a test simulation but all the results were showing only wind
flow

the link for the project

thank you for your help,

Hi @Mo_Elsebaei - very nice project idea! :slight_smile:

  1. Sounds good but it depends if you want to see transient structural behavior on the buildings or is steady-steate fine for you? Did you already decide? For the turbulence model maybe @1318980 or @ananthu_ajit can comment as I have not tested k-omega SST myself, all I know is that k-epsilon underpredicts in the strong wind and wake regions, maybe they can comment on that.

  2. Maybe this approach here helps you: Passive Scalar Sources - for the calculation of the concentration of dust you could use the approach from @tcakir here: Pollution exhaust in a garage - #4 by tcakir

  3. You can define the porosity ( P ) of a (coniferous) tree by

P = 1-\frac{Dc}{4*(LAI/h)}

where

  • LAI = Leaf Area Index (LAI)
  • Dc = Needle diameter
  • h = Height of the canopy

  1. Not if you want to use the approach mentioned in 2) I would say as you need the calculator filter which is not yet implemented (but on the roadmap)

Hope I could help you a bit!

Jousef


Relevant articles: Pekin, B.; Macfarlane, C. Measurement of Crown Cover and Leaf Area Index Using Digital Cover Photography and Its Application to Remote Sensing. Remote Sens. 2009 , 1 , 1298-1320.doi:10.3390/rs1041298

Source for the porosity calculation: Porosity of Trees -- CFD Online Discussion Forums

4 Likes

I’ve never done anything like this so not much input other than pay attention to your mesh, especially those trees. Use mesh refinements otherwise you will end up with a massive number of elements. Those tiny features (the trees) are not ideal.

hi @jousefm, thank you for your fast answer!
i’ve been reading and looking at all the possible related topics and projects but i couldn’t make even any general idea of how to start my project… it may seems easy to you but for me unfortunately i don’t find it user friendly program… so i will highly highly appreciate more detailed or step by step answer,
1- my concern is only the outdoor pedestrian comfort level, so i only need to see how the air circulates dust between the building and on the streets. that being said, is transient behavior will give me more accurate result? i sow a project that used Laminar/ Transient, but i really don’t no what the differences.
2- so i have to define areas surrounding the study area which are the sources of the dust as geometry primitives and then select them as passive scalar source and define the flux, when i did so in a test simulation the highest value of T was 1 where the flux was set to 50/s, so how this is possible? i think i missed something
3- can you please explain how can i define porosity? and which model support this? (Darcy forchheimer medium, Fixed coeff… ?)

thank you

thank you roy, unfortunately many trail of refinements ended by failing because out of memory error
Capture
do you think i should select a smaller area for the study?

You could do, and you can simplify your model so it has less small features especially small rounded features

hello @jousefm, can you please tell me after i calculate the porosity , where to put the value in the simulation? (Darcy forchheimer medium, Fixed coeff… ?)

Hi @Mo_Elsebaei!

Could you find additional information about the coefficients first before entering them? Here is a doc entry for porous media: Porosity & Porous Media | Advanced Concepts | SimScale.

@varsey, could you help our user here to clarify if anything is not clear? The use of the porous media module is straightforward but we would have to know the coefficients as previously mentioned. I am not sure what you could find in the literature.

Cheers,

Jousef

hi @Mo_Elsebaei

if you want to apply porous media for the tree, first decided porosity for the tree (in term of percentage).

when velocity flow through the tree it will change its nature and become nonlinear and if we consider the steady-state condition than you can only use Darcy-Forchheimer equation for this case.

you can apply porosity and particles diameter only in Darcy-Forchheimer equation (pressure-flux equation) and calculated the value of Darcy coefficients and Forchheimer coefficients.

follow above formula and used your value for calculation.
if you still face any problem contact me at any time.:wink:

Cheers,
Rohit.

2 Likes

thank you so much :slight_smile:

@ROHIT_SR , Good morning! i’ve been looking in the literature to find an easy way to calculate the porosity of a tree because the information needed to solve this (P=1−Dc4∗(LAI/h)) is not available for the plants that i will use in the simulation (not for most of the plants in fact!)
so i found an article that have used a program “SideLook” to calculate the porosity, the program depends on high resolution digital photo of the tree and converts it to black and white image by defining the blue spectrum then it defines the characteristic based on the ratio of white and black pixels. (Velarde, Jacinto Garrido et al. “Estimation of Optical Porosity or Canopy Structure of Two Species of Tree with Hemispherical and Vertical Images.” (2018))
and the below describes in details the process of defining the vegetation characteristic using the program (Multiparameter analysis of vertical vegetation structure based on digital image processing - ScienceDirect).

i wanted to share with you this information in case you are interested in this topic and i also want to ask you what we mean by particles diameter (for Darcy-Forchheimer equation) in the case of a tree? and after i calculate both of Darcy coefficients and Forchheimer coefficients, are they the same for x, y, z or it depends on the porosity of the material in each direction? and what unit vector is used for?

Thank you, i highly appreciate your help

1 Like

hi @Mo_Elsebaei,

Always interested …!
happy to see your progress because you done lots of research and trying to understand from basic.:clap:

Particles dia. means its spacing between two solid pieces. it was different from material to material. In your case ( Tree) I think the gap between two branches but it was in a circle every where. so I would suggest you, don’t apply porous media for the tree( its too complicated for this simulation.)

yes. it was same for all direction.
NOTE:- it was also dependent on your B.C. beacuse Darcy coefficients & Forchheimer coefficients was different.
read my post for more details

Unit vector applied in principle direction ( fluid flow direction). used your geometry x, y and z value for e1 and e2.
if you have any other doubt, contact me at any time. :email:

Best,
Rohit.

1 Like

thank you so much for your support!! :smiley:

so it is actually refers to the void… well, SideLook also can estimate the average diameter of the holes within the tree with reference to a distance the user defines at the image. i will use this value as particle diameter. i know that the result will not be accurate 100% but at least within an acceptable range :slight_smile:
thank you so much again

1 Like

interesting to see your post-processing image. :+1:

yes, you are right.

best,
Rohit.

1 Like

Hello!
i have been looking for the formula that you have used to calculate the inertial resistance forchheimer coefficient in the literature ; there is several formulas but none of them have “3.5” , can you please give me the reference of this formula?

Best regards,

hi @mkayed

yes, you are right. no one mention 3.5.
but,
if you know the equation of Ergun than you can understand all term involve in inertial resistance (forchheimer coefficient). in Ergun equation 1.7 was mention & if you solve Ergun equation than you get 3.5 at the end.

follow my post for more information:

reference,

if you face any problem contact me any time. :wink:

cheers!
Rohit.

3 Likes

thanks!

1 Like