Transient statistics missing from results file

turbomax · August 29, 2018, 11:17

Hi all,

I have been running a transient simulation using CFX v17.2 over many time-steps and am now confident the solution has converged. I wish to now compute the time-averaged results.

In CFX-Pre, I went to Output Control > Trn Stats; created a new transient statistics "Transient Statistics 1"; selected the "Arithmetic Average" option; selected the variables I want to average (e.g. Pressure); and finally set the Start/End Iteration List (time-steps 1-240).

I then re-started the simulation, while using the command "-initial-file XYZ.res" where XYZ.res is my previously converged solution.

My problem is that when the simulation completes and I open the new .res file in CFD-Post, none of the time-averaged results are there! For example, in the variables list, there is no sign of "Pressure.Trnavg" or any other transient statistics. Confusingly, when I save backup files (.bak), the transient statistics do appear. But why aren't they saved to the final .res file?

Many thanks for any help!

Max

Opaque · August 30, 2018, 17:15

By any chance, what option did you select for your results file?

Standard, Minimal, Selected Variables, Essential?

turbomax · August 30, 2018, 18:41

Yes, I had been outputting “Selected Variables”. Do you think this is the problem?

Gert-Jan · August 30, 2018, 19:45

It depends on which variables you selected. If you didn't select any in Pre, then the solver manager won't write any variable to the res file.........

turbomax · August 31, 2018, 05:57

Gert-Jan, I actually selected many variables.

Gert-Jan · August 31, 2018, 06:11

I already thought sa, but you never know....... ;-)

Check 1:

In the older versions, the averaged results appeared in the list of variables.
In the latest versions the transient averaged results are grouped in a subcategory. Like there is Geometry, Derived, a subcategory Transient Statistics is created.
Not sure if that is the problem. Meaning that the results are there, but you overlooked them.....?

Check 2: You mentioned that the transient results appear in the back up file, but not in the res file. That is weird. What are the file sizes? Is the backup significantly larger than the res-file?

Check 3: Not sure what happens if you average from 1-240, but perform 250 iterations. I would expect CFX to still save the statistics. But never tried.

Opaque · August 31, 2018, 10:49

I would try Standard, and see.

However, I will also address Gert-Jan concern on the same run. That is, start the run from scratch, accumulate for 1-5 timesteps, and write the results file.

Based on the observations you can decide what the problem was/is, and what to do next.

August 29, 2018, 11:17	Transient statistics missing from results file	#1
turbomax New Member Join Date: Jul 2015 Posts: 13 Rep Power: 11	Hi all, I have been running a transient simulation using CFX v17.2 over many time-steps and am now confident the solution has converged. I wish to now compute the time-averaged results. In CFX-Pre, I went to Output Control > Trn Stats; created a new transient statistics "Transient Statistics 1"; selected the "Arithmetic Average" option; selected the variables I want to average (e.g. Pressure); and finally set the Start/End Iteration List (time-steps 1-240). I then re-started the simulation, while using the command "-initial-file XYZ.res" where XYZ.res is my previously converged solution. My problem is that when the simulation completes and I open the new .res file in CFD-Post, none of the time-averaged results are there! For example, in the variables list, there is no sign of "Pressure.Trnavg" or any other transient statistics. Confusingly, when I save backup files (.bak), the transient statistics do appear. But why aren't they saved to the final .res file? Many thanks for any help! Max

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Custom Thermophysical Properties	wsmith02	OpenFOAM	4	June 1, 2023 15:30
OpenFoam "Permission denied" and "command not found" problems.	iyidaniel@yahoo.co.uk	OpenFOAM Running, Solving & CFD	11	January 2, 2018 07:47
[Other] Adding solvers from DensityBasedTurbo to foam-extend 3.0	Seroga	OpenFOAM Community Contributions	9	June 12, 2015 18:18
[swak4Foam] Error bulding swak4Foam	sfigato	OpenFOAM Community Contributions	18	August 22, 2013 13:41
[swak4Foam] funkySetFields compilation error	tayo	OpenFOAM Community Contributions	39	December 3, 2012 06:18

August 30, 2018, 17:15		#2
Opaque Senior Member Join Date: Jun 2009 Posts: 1,880 Rep Power: 33	By any chance, what option did you select for your results file? Standard, Minimal, Selected Variables, Essential?

August 30, 2018, 18:41		#3
turbomax New Member Join Date: Jul 2015 Posts: 13 Rep Power: 11	Yes, I had been outputting “Selected Variables”. Do you think this is the problem?

August 30, 2018, 19:45		#4
Gert-Jan Senior Member Gert-Jan Join Date: Oct 2012 Location: Europe Posts: 1,928 Rep Power: 28	It depends on which variables you selected. If you didn't select any in Pre, then the solver manager won't write any variable to the res file.........

August 31, 2018, 05:57		#5
turbomax New Member Join Date: Jul 2015 Posts: 13 Rep Power: 11	Gert-Jan, I actually selected many variables.

August 31, 2018, 06:11		#6
Gert-Jan Senior Member Gert-Jan Join Date: Oct 2012 Location: Europe Posts: 1,928 Rep Power: 28	I already thought sa, but you never know....... ;-) Check 1: In the older versions, the averaged results appeared in the list of variables. In the latest versions the transient averaged results are grouped in a subcategory. Like there is Geometry, Derived, a subcategory Transient Statistics is created. Not sure if that is the problem. Meaning that the results are there, but you overlooked them.....? Check 2: You mentioned that the transient results appear in the back up file, but not in the res file. That is weird. What are the file sizes? Is the backup significantly larger than the res-file? Check 3: Not sure what happens if you average from 1-240, but perform 250 iterations. I would expect CFX to still save the statistics. But never tried.

August 31, 2018, 10:49		#7
Opaque Senior Member Join Date: Jun 2009 Posts: 1,880 Rep Power: 33	I would try Standard, and see. However, I will also address Gert-Jan concern on the same run. That is, start the run from scratch, accumulate for 1-5 timesteps, and write the results file. Based on the observations you can decide what the problem was/is, and what to do next.