Normalised Imbalance Summary for a transient simulation

Thomasbr · July 27, 2016, 03:33

Hello,

at the end of a transient simulation, a .out file is generated. The normalised imbalance summary for mass, momentum, H-Energy is indicated in the file.

But in a transient simulation, the imbalance can be calculated for each time step, so I wanted to know about the values for the normalised imbalance summaries in the .out file:

is it the values for the last time step
or is it an average of the values of each time step?

Thomas

Opaque · July 27, 2016, 09:16

You can monitor the equation imbalances in the ANSYS CFX Solver Manager.

Create a New Plot, Select IMBALANCE, Select equation of interest, Press Apply..

If you really want to see the summary at every timestep, check the expert parameters panel in ANSYS CFX-Pre and look for "monitor balances" or similar.

Hope the above helps,

Thomasbr · July 27, 2016, 13:46

Hello Opaque,

thanks for your answer. I already monitor the H-Energy Imbalance (the only physical quantity with which I have significant imbalances so far).

For most of the simulation time, it remains between -1% and 1%, and usually gets totally crazy (between -30% and 30%) at the end of the simulation. I know I can control this with adaptive time stepping but the simulation time is then multiplied by approximately 15.

So, with what you said, I deduce that the normalised imbalances that are given in the .out file are actually the imbalances of the last time step. So I will have no choice but to apply a conservation target of 0.01 and wait 10/11 days till the simulation finishes (because with this condition, the time step won't go over 1,1e-7s, and the duration of my simulation is a couple ms).

July 27, 2016, 03:33	Normalised Imbalance Summary for a transient simulation	#1
Thomasbr Member Thomas B Join Date: Apr 2016 Location: Germany Posts: 30 Rep Power: 10	Hello, at the end of a transient simulation, a .out file is generated. The normalised imbalance summary for mass, momentum, H-Energy is indicated in the file. But in a transient simulation, the imbalance can be calculated for each time step, so I wanted to know about the values for the normalised imbalance summaries in the .out file: is it the values for the last time step or is it an average of the values of each time step? Thomas

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Transient RANS Simulation	Bazinga	Main CFD Forum	5	June 21, 2017 11:21
Question on transient simulation in OpenFOAM and FLUENT	nicklj	OpenFOAM Running, Solving & CFD	4	May 8, 2014 22:30
Exporting data of transient simulation DURING a simulation, at user locations ?	Milan2013	CFX	0	April 18, 2014 03:47
Imbalance in transient simulation	Roland R	CFX	3	December 9, 2009 05:39
how to identify transient simulation converged	littlelz	CFX	5	January 27, 2009 18:13

July 27, 2016, 09:16		#2
Opaque Senior Member Join Date: Jun 2009 Posts: 1,804 Rep Power: 32	You can monitor the equation imbalances in the ANSYS CFX Solver Manager. Create a New Plot, Select IMBALANCE, Select equation of interest, Press Apply.. If you really want to see the summary at every timestep, check the expert parameters panel in ANSYS CFX-Pre and look for "monitor balances" or similar. Hope the above helps,

July 27, 2016, 13:46		#3
Thomasbr Member Thomas B Join Date: Apr 2016 Location: Germany Posts: 30 Rep Power: 10	Hello Opaque, thanks for your answer. I already monitor the H-Energy Imbalance (the only physical quantity with which I have significant imbalances so far). For most of the simulation time, it remains between -1% and 1%, and usually gets totally crazy (between -30% and 30%) at the end of the simulation. I know I can control this with adaptive time stepping but the simulation time is then multiplied by approximately 15. So, with what you said, I deduce that the normalised imbalances that are given in the .out file are actually the imbalances of the last time step. So I will have no choice but to apply a conservation target of 0.01 and wait 10/11 days till the simulation finishes (because with this condition, the time step won't go over 1,1e-7s, and the duration of my simulation is a couple ms).