Heat Exchanger, steady state converges but transient does not

ringel_f1 · December 2, 2021, 05:46

Hello everybody,
I am using Ansys Academic 2021 R2. I am am calculating a transient heat exhanger with two separate flow curcuits within a solid domain. The hotter flow curcuit has a transient temperature profile at the inlet. The colder flow curcuit has a constant temperature at the inlet. I used massflow inlet and outflow for both curcuits as boundary conditions and k-epsilon model.

When I do a steady state simulation with a constant temperature (steady hot inlet temperature is the same as the transient hot inlet temperature for the first time step; so that boundary conditions for the first time step and the steady simulation are the same) at both inlets the simulation runs smoothly and converges after 80 iterations. But when I switch to transient simulation the simulation runs very unsmooth. It is not diverging but the residuals don't get lower and have a jagged course. Does anybody know what I have to change to get a smoother simulation for the transient calculation?
The transient calculation runs to the end, but results sometimes look not plausible because residuals during calculation are too high.

I haven't changed anything in Methods or Controls because it worked fine in steady state simulation...

Thanks for your help.

LuckyTran · December 2, 2021, 12:52

The residuals should look jagged during a transient simulation. If it doesn't then you have a problem.

Just pay attention to which iteration corresponds to which timestep. The residuals should increase on the very first iteration of any timestep, the rising edge. Then they decrease again as the solution converges for that timestep, the falling edge. If you don't see this pattern then your solution is not converging and then the results would indeed not be plausible.

Congratulations on doing your very first transient simulation by the way! That's an important milestone in your personal development.

ringel_f1 · December 2, 2021, 13:32

No no it looks randomly jagged within one timestep, so the single timesteps do not converge. They stay on a constantly high level for every timestep and always run to the maximum number of iterations per timestep. And I want to know how to improve convergence for the transient case within one timestep, because the steady case converges smoothly after just 80 iterations. And the only difference between steady and transient case, is the transient inlet temperature for the hot curcuit. So I don't see a problem, why the transient case has so many problems.
That's not my first transient simulation by the way

That's because I was wondering about that convergence behavior.

ringel_f1 · December 14, 2021, 05:23

Thanks for nothing but I found the problem by my own. Seems like the solver always changed the Controls when switching from steady (where everything worked fine) to transient. Now I took over the same Control Setting as I had for the steady case and now everthing works fine for the transient case as well.

December 2, 2021, 05:46	Heat Exchanger, steady state converges but transient does not	#1
ringel_f1 New Member Maximilian Ringel Join Date: Jul 2020 Posts: 6 Rep Power: 6	Hello everybody, I am using Ansys Academic 2021 R2. I am am calculating a transient heat exhanger with two separate flow curcuits within a solid domain. The hotter flow curcuit has a transient temperature profile at the inlet. The colder flow curcuit has a constant temperature at the inlet. I used massflow inlet and outflow for both curcuits as boundary conditions and k-epsilon model. When I do a steady state simulation with a constant temperature (steady hot inlet temperature is the same as the transient hot inlet temperature for the first time step; so that boundary conditions for the first time step and the steady simulation are the same) at both inlets the simulation runs smoothly and converges after 80 iterations. But when I switch to transient simulation the simulation runs very unsmooth. It is not diverging but the residuals don't get lower and have a jagged course. Does anybody know what I have to change to get a smoother simulation for the transient calculation? The transient calculation runs to the end, but results sometimes look not plausible because residuals during calculation are too high. I haven't changed anything in Methods or Controls because it worked fine in steady state simulation... Thanks for your help.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Implicit and Explicit finite difference discretization of steady state heat conductio	granzer	Main CFD Forum	14	May 3, 2019 16:24
Steady state solution as an initial condition for a transient problem	adnanakhtar	FLUENT	7	November 25, 2016 06:16
Error - Solar absorber - Solar Thermal Radiation	MichaelK	CFX	12	September 1, 2016 06:15
Simple piston movement in cylinder- fluid models	arun1994	CFX	4	July 8, 2016 03:54
Mass Diffusion: Transient and Steady State BC	rval	CFX	3	November 19, 2008 01:52

December 2, 2021, 12:52		#2
LuckyTran Senior Member Lucky Join Date: Apr 2011 Location: Orlando, FL USA Posts: 5,761 Rep Power: 66	The residuals should look jagged during a transient simulation. If it doesn't then you have a problem. Just pay attention to which iteration corresponds to which timestep. The residuals should increase on the very first iteration of any timestep, the rising edge. Then they decrease again as the solution converges for that timestep, the falling edge. If you don't see this pattern then your solution is not converging and then the results would indeed not be plausible. Congratulations on doing your very first transient simulation by the way! That's an important milestone in your personal development.

December 2, 2021, 13:32		#3
ringel_f1 New Member Maximilian Ringel Join Date: Jul 2020 Posts: 6 Rep Power: 6	No no it looks randomly jagged within one timestep, so the single timesteps do not converge. They stay on a constantly high level for every timestep and always run to the maximum number of iterations per timestep. And I want to know how to improve convergence for the transient case within one timestep, because the steady case converges smoothly after just 80 iterations. And the only difference between steady and transient case, is the transient inlet temperature for the hot curcuit. So I don't see a problem, why the transient case has so many problems. That's not my first transient simulation by the way That's because I was wondering about that convergence behavior.

December 14, 2021, 05:23		#4
ringel_f1 New Member Maximilian Ringel Join Date: Jul 2020 Posts: 6 Rep Power: 6	Thanks for nothing but I found the problem by my own. Seems like the solver always changed the Controls when switching from steady (where everything worked fine) to transient. Now I took over the same Control Setting as I had for the steady case and now everthing works fine for the transient case as well.