not getting a uniform temperature in hot water tank

[kamal.issa]

Hello, i have water vapor heating a water tank. it's heating it but the temperature in the tank is not uniform. the solution says that the fluid which is being heated is 120 celsius at a certain point and 30 at another point far from the first point. that's not possible in real life. if you are heating a water tank, the temperature inside the tank is gonna be uniform everywhere. increases over time but uniform as increasing. my objective is to find out how much time it will take to heat the water from 10 degree celsius to 100 celsius. i have attached the CFD files in case someone would like to try it. the following picture explains the problem very well. this is after 2 hours of heating. time step 360 seconds. total time 21600 seconds (2 hours)
thanks

[Jon.Wilde]

I would suggest that your timestep is way too large to capture the proper flow field here. Something like 5-10s might be more sensible.

To get good heat transfer, try the first SST turbulence model and then modify the boundary layer mesh to achieve a Y+ down near 0-2 (you can plot Y+ once you enable 'Stream Function' under Solve-> Result Quantities).

Here is a starting point for your boundary layer:

I agree that over time, you will get everything to 120C. Check the scale as you go though - you might still see what looks like a range of temperatures even with less than a 1C delta throughout the domain 🙂

Also - to get a good thermal gradient through the metal, you need at least 2 elements through its thickness. It might be easier to split the part in CAD so that CFD is forced to do with with little effort on your part.

[Jon.Wilde]

I would also suggest your mesh is too coarse to properly capture the fluid flow within the water, the results do not look smooth, they are controlled by the mesh:

[kamal.issa]

I tried refining the mesh, setting the time step to 10 seconds and i checked all the result quantities before solving but got the same results. I'm not sure what you mean by plotting Y+. i found a solution for it. not a very good solution but enough i think. after 1 hour, i delete all the boundary conditions. and then solve from where it finished last time. not from t=0. in the end the temperature in the water tank will become uniform. it became uniform with a temperature of 74 celsius. that means after 1 hour the water temperature will become 74 celsius. what do you think about this ?

[Jon.Wilde]

Is it possible that you have the correct result already and that is just takes this long to heat up?

Also, I am not sure what you mean by 'deleted all boundary conditions', this does not sound like a good idea to me?

[kamal.issa]

by deleting the boundary conditions, i mean that after solving for 2 hours for example, u have the option to solve again from time=0 and time= when the solution ended the last time u solved it. i can choose to solve again from time=when the solution ended the last time but i delete the boundary conditions before doing that. this way i let the temperature in the water settle down until reaching equilibrium without adding or removing heat

[kamal.issa]

now i really found the solution. if you set the direction of the gravity, it will give you a uniform temperature of the water as it increases. you wont get a high value at a certain point and a small value at another point. use the file that i shared. solve it the first time by keeping the gravity direction as 0,0,0. then solve it again by setting the gravity direction to 0,0,-1. you will get a completely different answer. i don't know why the gravity has a big effect on the heat transfer 

[Jon.Wilde]

With gravity you have natural convection and movement of the water, without you pretty much have only conduction 🙂