[Bisht]

Hi,

I am simulating a low-pressure system (liquid and gas) with VOF method and porous media formulation. Please find the attached image to look into computational domain.

The operating pressure is 1 Bar, pressure inlet A is at -3400 Pa (gauge pressure), Pressure inlet B is at 0 Pa (gauge pressure) and for velocity outlet, I implemented a transient velocity profile (linear increment with time).

The domain over the porous screen is initialized with gas at 1 bar Atm pressure and rest of the domain with liquid at -3400 Pa.

But during simulation, vector contours show that the liquid is flowing towards the pressure inlet A instead of moving to the velocity outlet. I checked the mass flow rate at both inlet A and outlet which reports the negative values.

The static pressure contour in the channel is giving weird values. For the first few time-steps, the pressure is constant but then the pressure on the downstream side becomes equal to the atmospheric pressure.

The same case is working fine if I change the pressure value at the inlet A in the range of 0 Pa to -5 Pa.

ANy idea how to resolve the case for low-pressure inlet value?

[Kushal Puri]

Originally Posted by Bisht

Hi,

I am simulating a low-pressure system (liquid and gas) with VOF method and porous media formulation. Please find the attached image to look into computational domain.

The operating pressure is 1 Bar, pressure inlet A is at -3400 Pa (gauge pressure), Pressure inlet B is at 0 Pa (gauge pressure) and for velocity outlet, I implemented a transient velocity profile (linear increment with time).

The domain over the porous screen is initialized with gas at 1 bar Atm pressure and rest of the domain with liquid at -3400 Pa.

But during simulation, vector contours show that the liquid is flowing towards the pressure inlet A instead of moving to the velocity outlet. I checked the mass flow rate at both inlet A and outlet which reports the negative values.

The static pressure contour in the channel is giving weird values. For the first few time-steps, the pressure is constant but then the pressure on the downstream side becomes equal to the atmospheric pressure.

The same case is working fine if I change the pressure value at the inlet A in the range of 0 Pa to -5 Pa.

ANy idea how to resolve the case for low-pressure inlet value?

Why you are expecting flow from low pressure to high pressure? Can you explain with some more explanation.

[Bisht]

Actually, I want to test the capability of the porous screen to block the gas from entering into the liquid channel. The screen has a bubble point pressure of 3600 Pa, which means when pressure difference across the screen becomes equal or more than 3600 Pa, the gas will enter through it. Below this bubble point pressure, the screen is non-permeable to any kind of gas flow.

My computational domain is very small (its a microcapillary channel) which has a length of 220 mm, the height of 150 mm and width of 5 mm. Due to the geometrical constraint, it cannot sustain such a high flow rate which creates so much pressure difference across the screen.

Therefore I reduced the pressure at the channel inlet to -3400 Pascal w.r.t. the atmosphere in order to test the bubble point pressure of screen. As I said in my previous post, the operating pressure is 1 bar, pressure inlet A is a liquid inlet at -3400 Pa, velocity outlet has a transient velocity profile and pressure inlet B is a gas inlet at 0 pa which represents the atmosphere or an open system problem.

I initialized my problem with -3400 Pa gauge pressure and liquid volume fraction 1. Then I patched the volume above the screen with a gauge pressure of 0 Pa (w.r.t. operating pressure) and liquid volume fraction 0. But during simulation, the liquid flow is in reverse direction towards pressure inlet A and the pressure value look weird as you can see in the attached image of the previous post.

I have the same setup for my experiment and due to some constraints I can only reduce the pressure at the liquid inlet rather than increasing the pressure in the gas domain over the screen to create the pressure difference.

[Kushal Puri]

Originally Posted by Bisht

Actually, I want to test the capability of the porous screen to block the gas from entering into the liquid channel. The screen has a bubble point pressure of 3600 Pa, which means when pressure difference across the screen becomes equal or more than 3600 Pa, the gas will enter through it. Below this bubble point pressure, the screen is non-permeable to any kind of gas flow.

My computational domain is very small (its a microcapillary channel) which has a length of 220 mm, the height of 150 mm and width of 5 mm. Due to the geometrical constraint, it cannot sustain such a high flow rate which creates so much pressure difference across the screen.

Therefore I reduced the pressure at the channel inlet to -3400 Pascal w.r.t. the atmosphere in order to test the bubble point pressure of screen. As I said in my previous post, the operating pressure is 1 bar, pressure inlet A is a liquid inlet at -3400 Pa, velocity outlet has a transient velocity profile and pressure inlet B is a gas inlet at 0 pa which represents the atmosphere or an open system problem.

I initialized my problem with -3400 Pa gauge pressure and liquid volume fraction 1. Then I patched the volume above the screen with a gauge pressure of 0 Pa (w.r.t. operating pressure) and liquid volume fraction 0. But during simulation, the liquid flow is in reverse direction towards pressure inlet A and the pressure value look weird as you can see in the attached image of the previous post.

I have the same setup for my experiment and due to some constraints I can only reduce the pressure at the liquid inlet rather than increasing the pressure in the gas domain over the screen to create the pressure difference.

Can you just check with changing your operating pressure to 0 and defining the absolute values of the pressure at inlets. How your simulation is behaving..

[sridhar.d009]

Hi Bisht

I think it's a boundary conditions problem. Instead of velocity out let use pressure outlet ( static pressure as total pressure) define backflow. Finally define flow direction at inlets.

[Bisht]

Originally Posted by sridhar.d009

Hi Bisht

I think it's a boundary conditions problem. Instead of velocity out let use pressure outlet ( static pressure as total pressure) define backflow. Finally define flow direction at inlets.

Hi Sridhar,

The reason of velocity outlet is due to the pump controlled flow which is located on the downstream side of the channel.

The problem with pressure outlet is that it doesn't allow to specify the targeted mass flow rate in case of multiphase flow.

Because of this reason, I am left with only two BC at the outlet i.e. Velocity outlet or mass flow outlet.

[Bisht]

Originally Posted by Kushal Puri

Can you just check with changing your operating pressure to 0 and defining the absolute values of the pressure at inlets. How your simulation is behaving..

I tried this without any change in the results.

I also simulated with operating pressure equal to inlet pressure and specify a higher value for the gas inlet pressure (3400 Pa). But still, it's not working.

[Kushal Puri]

Originally Posted by Bisht

I tried this without any change in the results.

I also simulated with operating pressure equal to inlet pressure and specify a higher value for the gas inlet pressure (3400 Pa). But still, it's not working.

What are the original boundary conditions, means experimental.

[Bisht]

The experiments boundary condition is -3400 Pa inlet pressure, transient velocity at the outlet which is pump controlled. Before starting the pump, the pressure in the channel is -3400 Pa. The box over the porous screen is actually to mimic the atmosphere during CFD. In the real system, it doesn't exist. The side of the box I defined as symmetry boundary and top as pressure inlet to allow the gas exchange.

I did the simulation by changing the boundary condition of pressure inlet B to pressure outlet, it shows the same result. But when I changed it to symmetry or wall, the pressure in the channel is as I expected it to be. However, the pressure over the screen also changes from 0 Pa to -3400 Pa, which is again weird and wrong.

[Kushal Puri]

Originally Posted by Bisht

Hi,

I am simulating a low-pressure system (liquid and gas) with VOF method and porous media formulation. Please find the attached image to look into computational domain.

The operating pressure is 1 Bar, pressure inlet A is at -3400 Pa (gauge pressure), Pressure inlet B is at 0 Pa (gauge pressure) and for velocity outlet, I implemented a transient velocity profile (linear increment with time).

The domain over the porous screen is initialized with gas at 1 bar Atm pressure and rest of the domain with liquid at -3400 Pa.

But during simulation, vector contours show that the liquid is flowing towards the pressure inlet A instead of moving to the velocity outlet. I checked the mass flow rate at both inlet A and outlet which reports the negative values.

The static pressure contour in the channel is giving weird values. For the first few time-steps, the pressure is constant but then the pressure on the downstream side becomes equal to the atmospheric pressure.

The same case is working fine if I change the pressure value at the inlet A in the range of 0 Pa to -5 Pa.

ANy idea how to resolve the case for low-pressure inlet value?

Can you show the screen shot of the pressure contour with vectors on it.
Also screen shot of the velocity with vector on it

[Bisht]

please find the picture of pressure contour with symmetry BC at top