[lyro]

I've modeled a 4 riser solar water heater with a 5l water tank above it. Using coupled with 2nd order methods. No initial velocity or anything on the system, I'm trying to simulate thermosyphon action with 1000 w/m2 heat flux on my collector panel. I've been running the iterations for few hours now and the residuals look interesting. Yet nowhere on the console is any error message and solution converges for each time step (1s) for now. Are the scaled residuals normal, or can you guys spot anything funny about the system.

https://imgur.com/gv3sYc1

[LuckyTran]

The residuals look like there isn't any flow

[lyro]

I also think that's the case, I want to create a thermosyphon action and seems I didn't properly set the boussinesq paramter so I hardly got any flow at all. How exactly did you deduce that btw, no change in energy? sorry for simple questions, I have close to zero experience :P

[vinerm]

For using Boussinesq approximation, you need to set density under Material panel. Same value of the density should be set as Operating Density under Operating Conditions panel. Also set a positive value for volumetric expansion coefficient under Material panel. Set a Boussinesq temperature under Operating Conditions. This temperature should correspond to the density value used. Gravity must be in correct direction. 1 s could be rather large time-step for natural convection but depends on the turbulence time-scales. Most likely, you would require a smaller time-step, of the order of 5-10 ms.

[lyro]

I managed to get some solutions, the water in my riser pipes are getting higher and they get into my reservoir, but I can not observe any water flowing out the outlet side of the reservoir yet even some water get's into the reservoir through the outlet pipe. shouldn't the water just rotate clockwise ?
I've added few pics of the contours I get with an isometric of the geometry

https://imgur.com/a/CPy1M3F

[vinerm]

Fluid will recirculate due to two reasons, due to mass conservation and due to cooling, if cooling happens. For mass conservation, do monitor the total mass of liquid in the domain. In case you see that it is changing by more than 1%, then there is some problem with the setup.

[lyro]

how do I set a total mass monitor for a domain ? and since the domain is a closed loop, how can I lose any mass there

[vinerm]

Set up a volume monitor and select mass. Physically, the system cannot lose any mass but, alas, this is numerical and there is always a numerical error. So, the system will lose or gain mass; the only important point is how much.