[fekalnik]

Hello,

i have this pretty easy simulation where i study the heat transfer between solid region with a heat source and fluid cooling system (picture)

a simple equation: Q= m * cp * delta T where Q is my heat source (75W), m mass flow (0,35) c specific heat (4181) and delta T temperature diference between inlet and outlet of the water channel.

delta t = Q/m*cp = 0,0512 K

but simulation always show higher temperature difference and it seems the better simulation, the higher difference.

But all regions are set as adiabatic, there is no other heat source than 75W and it all goes to water, how can the temperature be higher than 0,0512?

i know that the difference is just 0,003 something.. But everything i do to make the simulation better (better mash, adding prism layers,...) only makes this difference higher.

and i am like, what the hell?

i also checked the mass flow - its 0,35 also in simulation.

can anyone please tell me how is it possible?

thanks

[ping]

i imagine you don't have convergence and the fact that the error gets worse with more cells tend to confirm this since more cells will take longer to converge with the seg solver
so check your residuals are very level and zoom in on the temperature plots too
also set the solid energy under relaxation factor to 0.99999 to speed up solid energy convergence
add a monitor or two of solid temperature using point probes and ensure these are very flat before assuming convergence

[fekalnik]

i actually had relaxation factor set to 0,99 and residuals look fine.

the temperature channel outlet is 303.0057 K with 75 W coming in.

thats 12% mistake in comparison with theoretical solution.

isnt that little high with such a easy simulation..?

[fekalnik]

this is a report of probe temperature in solid.

[ping]

your energy residual is a long way off converged as I expected.
.99 for that urf will cause very slow solid energy convergence and why you should be using .99999

[fekalnik]

now its done after 70 iterations and it really stops at the right temperature.

thanks a lot pinq, the world of star ccm never stop to amaze me

[ping]

good to hear
you could have also added a heat transfer report and monitor on the solid-fluid interface and i am sure it would have shown that 75W was not exiting the solid until convergence was achieved