Why air’s torque is negative and water’s torque is positive?
 

Hello, everybody.
I‘m working on a stirred vessel with gas and water. I would like to know the torque value of my stirred vessel.
As we know, the CFD-post can give us the torque value by air and water separately. But I find that, in one of my simulations, the torque of air is negative and the water of torque is positive. And the total torque is the sum of air and water’s torque.

But surprising me, in some of my other simulations, the torque values of air and water are both positive.

Could you tell me why did these happen？

PS:
1. My model is a stirred vessel with diameter 700mm, Height 530mm, total volume 0.2 m3, and mesh grid 530,000.
2. The Euler-Euler model, MRF model, k-εmodel .etc were chosen.
3. The drag coefficient is 0.44.

I cannot see how the torque of the individual phases is physically meaningful in any way. If you can imagine how it could be meaningful let me know.

If it is not meaningful then don't worry about it and ignore it. The total torque (which obviously is physically meaningful) is the parameter to look at.

yes.
The total torque is meaningful. Actually, my simulation on torque disagrees with the experiment.

In single phase(water only), the torque by CFD which is 11NM agrees with experiment by torque sensor which is 10.5 NM.

But, in gas liquid phase simulation, the torque by cfd which is only 4 NM is much less than experiment which is 6.5 NM.

And what is confused me is that the torque by cfd was decreased sharply in step way. For example, the inlet velocity of air is 5 m/s or 2.5m/s are the similar torque around 4 NM.

However, the experiment datas are decreased linearly. For example, when air inlet is 5 m/s, the torque is 6 NM. When air inlet is 2.5 m/s, the torque is 8 NM.

I would like to find the reasons. So I fond this strange phenomenon on each phase.

So, how could I improve our accuracy of our gas liquid simulation?

Thanks very much

For general comments on accuracy see the FAQ: http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F

In your case it appears your multiphase model is not modelling something which is important. But without more specific details of what you are trying to do I cannot help you much.