[selvam2487]

Dear Friends,

I am performing LES of T-junction mixing using CFX. My inlet conditions are

Hot inlet: T - 421 K, v - 0.16 m/s
Cold inlet: T - 298 K, v - 0.084 m/s

bulk velocity in mixing region is 0.18 m/s. I am analyzing velocity distribution downstream of T-junction in the mixing region. Attached along with this thread is the velocity profile in the vertical direction (z axis) at 6 diameters (6D, with D - 71.8 mm) downstream of T-junction. Velocity is higher near the upper region where no mixing occurs (marked by red contours) than in the lower region (where mixing occurs). The reason for this behavior (in my opinion) is the low density of hot fluid which does not mix due to low velocity of cold fluid (0.084 m/s) coming from branch pipe. Also, Richardson number in mixing region is calculated to be Ri = 1.7, indicating buoyancy effects.

Could you guys help me in understanding the reason for higher velocity near upper region of pipe ? Thank you for your help in advance.

Regards,
Karthick

[ghorrocks]

An image of what you are modelling would be useful.

[selvam2487]

Dear Glenn,

I have attached the pictures of my T-junction model. Hope this gives you the required information that you were looking for. Thanks for your help.

Regards,
Karthick

[ghorrocks]

I bet the side inlet flow is corkscrewing around the main flow. This will reduce bulk mixing and allow these flows to stay separate for longer.

Put streamlines on the side inlet and I bet they corkscrew around.

[selvam2487]

Dear Glenn,

This LES is based on experiments performed at my institute and I am performing numerical validation for a particular test case.

1. Yes, the flow is not completely mixed through out the computational domain due to low velocity of branch fluid (Reynolds number - 3670) whereas the hot fluid in main pipe has relatively higher velocity (Reynolds number - 38400) along with buoyancy effects (Richardson number - 2.93 in the mixing region).

2. Also there is turbulent penetration of hot fluid into branch pipe and vice versa and is confirmed by experimental measurements.

3. So, there is incomplete mixing of fluids and velocity in the upper region (where no fluid mixing occurs) is higher than in mixing region. So am I correct in thinking that the low density of hot fluid near upper region along with mixing of fluids (which introduces fluid from branch pipe into main pipe) results in higher velocity near the upper region of pipe than in mixing region?

I have attached velocity and temperature streamlines for your reference.
Thank you.
Regards,
Karthick

[ghorrocks]

Your images show the flow is not corkscrewing around, it is the hot and cold fluids separating and oscillating slightly from an equilibrium position.

Just as the incomplete mixing can lead to a thermal gradient it can lead to a velocity gradient as well.

Note that I do not know what fluid model you are using for this, if the fluid is compressible then there will be density effects as well.

[selvam2487]

Hello Glenn,

It is incompressible fluid model and variation of physical properties of water with temperature is taken into account using IAPWS libraray

Regards,
Karthick