can ALTair products be used to predict heavy gas (Hydrogen Sulfide) concentrations distribution in outdoor environment?

Gamal2023

Hydrogen Sulfide emits from tailrace, spillway, and reservoir of a hydroelectric power plant (HEP), disperses in ambient air, and tends to accumulate in low-lying areas. Its emission rate will be measured and its accumulative concentrations in surrounding area of HEP are to be estimated. The study area is a complex terrain which contains  one-level buildings, trees, hills, a river, and dam constructions, but it will be assumed as a fixed region. The meteorological data will be collected from a weather station nearby the HEP. The main aim of the project is to develop a dispersion model to predict H2S concentration in HEP surrounding and investigate the influence of wind (speed and direction) and humidity on H2S dispersion/distribution in the surrounding area.

In Fact, I am a new to ALTair products  and I do not have the knowledge to decide whether ALTair products (CFD) can be utilized or not for such project. I would like to know if I can use these products to achieve the objectives of my research. I start learning some stuff of hyperworks to acclimate myself with ALTair products.

your assistance highly appreciated 

Gamal2023

Hi Acupro

I have read that article and started learning Acusolve. But, when I want replicate some previous work which done in Ansys Fluent, I face some difficulties in defining the properties of H2S and related it as species. Also defining the mixture properties (H2S and Air). I learnt how to setup simple simulation (as in tutorials) although I still find myself can not mimic a simple species transport in flat terrain from a ground or elevated source. 

I would kindly request your guidance and assistance.

Thank you very much

acupro

Perhaps this publication can be of use.  It may not cover everything on your list, but it can give you some ideas.

https://www.researchgate.net/publication/220405622_Modeling_a_Hypothetical_Chlorine_Release_on_a_College_Campus

Gamal2023

acupro_21778 said:

Perhaps this publication can be of use.  It may not cover everything on your list, but it can give you some ideas.

https://www.researchgate.net/publication/220405622_Modeling_a_Hypothetical_Chlorine_Release_on_a_College_Campus

Thank you very much for replying

That  is really helpful 

Gamal2023

Hi Acupro

I have read that article and started learning Acusolve. But, when I want replicate some previous work which done in Ansys Fluent, I face some difficulties in defining the properties of H2S and related it as species. Also defining the mixture properties (H2S and Air). I learnt how to setup simple simulation (as in tutorials) although I still find myself can not mimic a simple species transport in flat terrain from a ground or elevated source. 

I would kindly request your guidance and assistance.

Thank you very much

acupro

Gamal2023 said:

Hi Acupro

I have read that article and started learning Acusolve. But, when I want replicate some previous work which done in Ansys Fluent, I face some difficulties in defining the properties of H2S and related it as species. Also defining the mixture properties (H2S and Air). I learnt how to setup simple simulation (as in tutorials) although I still find myself can not mimic a simple species transport in flat terrain from a ground or elevated source. 

I would kindly request your guidance and assistance.

Thank you very much

Perhaps you've already seen the basic species tutorial - but here it is again:
https://help.altair.com/hwcfdsolvers/acusolve/topics/tutorials/acu/acu_4300_intro_cfd_r.htm#acu_4300_intro_cfd_r

You can follow that at least for describing the various material properties as piecewise-linear, as a function of the species concentration.

It sounds like you'll need one inlet for the air and another for the inlet of the H2S - with the appropriate species value - one being 1.0 and the other 0.0 - depending on how you define the fit based on species.

Gamal2023

Hi Mr. Acupro

I have been trying to simulate a simple set up problem for species transport as described in this paper " characteristics of unorganized hydrogen sulfide dispersion for industrial building layout optimization" in which the computational domain contains no obstacle (flat terrain) with surface source of H2S on ground. I set one inlet for wind (4m/s) with zero H2S (species value =0) and another inlet for H2S (species value =1) (mass flow rate 2.23 kg/s). I have tried several simulation trials but I couldn't replicate the same work as described in the abovementioned paper. most of the time I encountered this kind of warning "WARNING: inflow at outflow boundary: Outlet (mass/time: in=-5.153964e+03, out=5.249148e+03)" whenever I activate the species to simulate H2S dispersion. Meanwhile, when simulating the wind only (no species) the simulation converges. I am currently uncertain about potential omissions or inaccuracies in my simulation setup. For your review, I have included two pictures of simulation trials one for velocity profile simulation (steady state)  and the other one is for H2S transport simulation "transient".  Your insights and guidance on how best to proceed would be immensely valuable and greatly appreciated.

acupro

Gamal2023 said:

Hi Mr. Acupro

I have been trying to simulate a simple set up problem for species transport as described in this paper " characteristics of unorganized hydrogen sulfide dispersion for industrial building layout optimization" in which the computational domain contains no obstacle (flat terrain) with surface source of H2S on ground. I set one inlet for wind (4m/s) with zero H2S (species value =0) and another inlet for H2S (species value =1) (mass flow rate 2.23 kg/s). I have tried several simulation trials but I couldn't replicate the same work as described in the abovementioned paper. most of the time I encountered this kind of warning "WARNING: inflow at outflow boundary: Outlet (mass/time: in=-5.153964e+03, out=5.249148e+03)" whenever I activate the species to simulate H2S dispersion. Meanwhile, when simulating the wind only (no species) the simulation converges. I am currently uncertain about potential omissions or inaccuracies in my simulation setup. For your review, I have included two pictures of simulation trials one for velocity profile simulation (steady state)  and the other one is for H2S transport simulation "transient".  Your insights and guidance on how best to proceed would be immensely valuable and greatly appreciated.

If you can post the input files (.inp) along with some images of your simulation model/domain - including where your inlet(s) and outlet(s) are, perhaps we can help.

Gamal2023

Thank you very much for your response

The domain size is 40m x 30m x 3m and the H2S source surface is 20m x 10m.
I have attached the files and please let me know if requires other files.

I highly appreciate your assistance 

acupro

Gamal2023 said:

Thank you very much for your response

The domain size is 40m x 30m x 3m and the H2S source surface is 20m x 10m.
I have attached the files and please let me know if requires other files.

I highly appreciate your assistance 

The convergence really isn't that bad.  You do have reverse flow - flow entering the exit - which is not uncommon.  (How do the results look?)  I do have some suggestions.

On the outflow boundary:
1.  Activate hydrostatic pressure (you can leave origin at 0,0,0).  This makes the outlet pressure consistent with the fluid 'weight' since you have gravity active.
2. Activate back-flow conditions.  Set the turbulence quantities and species back-flow-type to exiting-average or exiting-area-average - depending on what's available.  This allows you to set the boundary conditions on whatever flow enters at the outlet.  (Another option would be to extend the outlet farther away, and maybe add a contraction so the outlet is smaller area than the inlet.  This helps to prevent reverse flow.)

You don't need convergence-tolerance that low (1.e-5).  The default 1.e-3 is typically plenty.  (This is roughly equivalent to 1.e-5 or 1.e-6 in other solvers)

Johan Dahlberg_20306

Hi Gamal2023
I took a look at your model (did remesh the BL somewhat, used dt=0.02, Spalart-Allmaras and initial condition on velocity)

section cut at y=0 @ 10sek ( maybe good to run it somewhat longer time or Steady State)

 

Still some backflow, but less compared to your simulation
 *** WARNING: inflow at outflow boundary: Outlet (mass/time: in=-4.639508e+01, out=4.887392e+02)

 

So, play around with mesh (refine a bit) and try some different turbulence models!
I did attached my file!
But as mentioned above, you can start with a steady-state solution and for transient 4m/s inlet it might be good to run a couple of more seconds.

Hope this helps, Interesting case!

/johan

Gamal2023

Johan Dahlberg_20306 said:

Hi Gamal2023
I took a look at your model (did remesh the BL somewhat, used dt=0.02, Spalart-Allmaras and initial condition on velocity)

section cut at y=0 @ 10sek ( maybe good to run it somewhat longer time or Steady State)

 

Still some backflow, but less compared to your simulation
 *** WARNING: inflow at outflow boundary: Outlet (mass/time: in=-4.639508e+01, out=4.887392e+02)

 

So, play around with mesh (refine a bit) and try some different turbulence models!
I did attached my file!
But as mentioned above, you can start with a steady-state solution and for transient 4m/s inlet it might be good to run a couple of more seconds.

Hope this helps, Interesting case!

/johan

Hi Mr. Johan;

Thank you very much for your help

In the case you simulated, you applied boundary layers on slip walls and mass flow rate intake  (H2S inlet), I wonder if it is possible to do so.  I am currently using a personal laptop to run the simulation in which it can not handle refined mesh.

In Fact, I have been trying different simulation set up following Mr. Acupro's instructions to prevent backflow. Although the convergence criteria is met the backflow occurs at the beginning of the simulation and  the residuals starts with high non-zero magnitude.

 

Gamal2023

acupro_21778 said:

The convergence really isn't that bad.  You do have reverse flow - flow entering the exit - which is not uncommon.  (How do the results look?)  I do have some suggestions.

On the outflow boundary:
1.  Activate hydrostatic pressure (you can leave origin at 0,0,0).  This makes the outlet pressure consistent with the fluid 'weight' since you have gravity active.
2. Activate back-flow conditions.  Set the turbulence quantities and species back-flow-type to exiting-average or exiting-area-average - depending on what's available.  This allows you to set the boundary conditions on whatever flow enters at the outlet.  (Another option would be to extend the outlet farther away, and maybe add a contraction so the outlet is smaller area than the inlet.  This helps to prevent reverse flow.)

You don't need convergence-tolerance that low (1.e-5).  The default 1.e-3 is typically plenty.  (This is roughly equivalent to 1.e-5 or 1.e-6 in other solvers)

Hi Mr. Acupro

I ran some simulation trials with Spalart-Allmaras turbulent model for extended domain (11m, 15m, and 20m) beyond the outlet of 40m x 30m x 3m domain in small time step (0.001, 0.0005, and 0.0001) for less than one second and maximum stagger 10. All simulation trials solution ratio start with high value specially the pressure and backflow encountered except for 0.0001 time step for different domain extension. when increasing the maximum stagger to 20 the backflow no longer exists within the domain extension of 20 m  and time step of 0.001, however the starting values of solution are large values (I couldn't provide result of this simulation, the software keep crashing when trying to open the simulation results). I have attached pictures to illustrate the addressed simulation issues. I do not know how to overcome such challenges as I am still in the  learning stages. 

Regarding the other properties of the mixture (H2S-Air) such the specific heat and thermal conductivity, I would like to know how to add such properties to materials model. 

Thank you very much for your kind cooperation

Gamal2023

Gamal2023 said:

Hi Mr. Acupro

I ran some simulation trials with Spalart-Allmaras turbulent model for extended domain (11m, 15m, and 20m) beyond the outlet of 40m x 30m x 3m domain in small time step (0.001, 0.0005, and 0.0001) for less than one second and maximum stagger 10. All simulation trials solution ratio start with high value specially the pressure and backflow encountered except for 0.0001 time step for different domain extension. when increasing the maximum stagger to 20 the backflow no longer exists within the domain extension of 20 m  and time step of 0.001, however the starting values of solution are large values (I couldn't provide result of this simulation, the software keep crashing when trying to open the simulation results). I have attached pictures to illustrate the addressed simulation issues. I do not know how to overcome such challenges as I am still in the  learning stages. 

Regarding the other properties of the mixture (H2S-Air) such the specific heat and thermal conductivity, I would like to know how to add such properties to materials model. 

Thank you very much for your kind cooperation

Here are the set ups ; I forgot to include them in previous reply;