CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   CFX (https://www.cfd-online.com/Forums/cfx/)
-   -   CFX transient solution not giving trn files (https://www.cfd-online.com/Forums/cfx/219218-cfx-transient-solution-not-giving-trn-files.html)

Dale MC July 19, 2019 16:02

CFX transient solution not giving trn files
 
3 Attachment(s)
I am trying to get multiple trn results files for a transient solution but only get 1 res file. Also the solution does not iterate all the timesteps. My geometry is a micro channel with an injection point and am putting a cyclic pulse flow in for the injection with this expression 2*sin(2*3.1416*4/.05) [m s^-1]. I followed tutorial 6 but without setting up the smoke concentration plot. My total time is 4s and my timestep is .05s. I am using a selected variable of Velocity and I have tried almost every trn results option for frequency. the solution solves the SS then does the transient but only gives one .res file only doing 1 timestep. I have attached a picture of when it finishes as well as a picture of the settings for trn results and the ccl file which is zipped. If anyone can give advice it will be greatly appreciated.
Thanks

ghorrocks July 19, 2019 21:47

* You say this is a micro channel but you are using a turbulence model. This is probably incorrect. You should check the Reynolds number of your flow, as it is probably laminar.
* in your transient run you are using backup files. You probably want transient results files, not backup files. Transient results files will give you TRN files.

Dale MC July 19, 2019 22:23

I thought about it being laminar flow but the injection creates turbulence in the channel which is why I am examining it. That is why I was using turbulent. Do you think laminar would work.
Also are you saying do not set backup in order to get trn files? I set the trn results but nothing happens. I don't get trn files just .res file

ghorrocks July 20, 2019 08:21

Make sure you understand the difference between turbulence and an unsteady flow. An unsteady flow can be laminar. So just because you are squirting some fluid in from the side does not make the flow turbulent. You need to work out the Reynolds Number and determine if it is laminar or turbulent from that. Most Micro-fluidics are purely laminar, so I would be surprised if your flow is turbulent.

Yes, do not use backup files. Use transient results files instead. You will get a trn file whenever you define a transient results file to be written.

ghorrocks July 20, 2019 08:21

Make sure you understand the difference between turbulence and an unsteady flow. An unsteady flow can be laminar. So just because you are squirting some fluid in from the side does not make the flow turbulent. You need to work out the Reynolds Number and determine if it is laminar or turbulent from that. Most Micro-fluidics are purely laminar, so I would be surprised if your flow is turbulent.

Yes, do not use backup files. Use transient results files instead. You will get a trn file whenever you define a transient results file to be written.

Dale MC July 21, 2019 23:21

3 Attachment(s)
The flow is technically laminar the Re is less than 1000. My thought was that because of the rotation in the flow I should use the turbulent model. I have attached 3 pictures. One of the contours, one of the vectors and one with the timestep selector. As you can see there is rotation in the flow but I am not sure if that means I should use the turbulent model. What do you think? I know the flow is laminar on down the channel for sure. I will also take off the backup and try that too. I would like your thoughts on laminar vs laminar model though. I can try the laminar model if you think it will help me get the timestep trn files I would like.
Thanks

AtoHM July 22, 2019 03:13

Quote:

I can try the laminar model if you think it will help me get the timestep trn files I would like.

I don't think this is what Glenn is saying. He is giving you an additional hint about wrong physical modeling which will probably influence your results. Getting trn files has nothing to do with that.

karachun July 22, 2019 03:23

Here is von Karman vortices. This is unsteady problem, but at low speed, flow is laminar, in spite of this there are lot of vorticities. Laminar don’t mean steady.
You may confuse laminar flow with potential flow, potential flow don’t have vorticities. Ok?
https://upload.wikimedia.org/wikiped...-animation.gif

Dale MC July 22, 2019 09:16

Thank you for the insight about the flow model. I will switch to the laminar model in order to get better results. I see now that rotation in the flow does not mean that I should use the turbulent model. I will also try removing the backup in just a few and see if that will help me get the trn files I need for the transient part of the analysis and let you if that works.
Thanks

Dale MC July 22, 2019 15:37

Quick question, is it necessary to do a SS solution first when doing a transient analysis? I have tried setting it up different not specifically following tutorial 6 where they do the SS first and then use that as the start of the transient analysis?

ghorrocks July 22, 2019 20:03

You can start a transient simulation with any initial state you like. That initial state can come from a steady state simulation, or it could come from some other state you define.

Dale MC July 22, 2019 23:39

great Thanks! I skipped the SS solution and finally got it to produce trn files. The problem is that I had no pulse using the expression I defined. I changed it now to 2*sin(30*theta) but am getting an error.

ghorrocks July 23, 2019 00:01

If you want us to help you please post your output file.

Dale MC July 23, 2019 10:54

2 Attachment(s)
Thanks, I have attached the .out file and a picture of the error I am getting since I changed the expression for the small velocity inlet to the one in my last post. I am not sure if "theta" is a proper function. What I need to do is get the small inlet to pulse the flow, like push in, suck out, cyclically 5 times a second.
In regards to writing the trn files I think that taking the backup off might have worked. I also got rid of the doing the SS solution first as they did in tutorial 6.

ghorrocks July 23, 2019 19:43

Have you looked in the output file? It tells you exactly what the error is.

Some other comments:
* Why are you using theta in your velocity function? Don't you mean time (variable name "t")? Look in the CFX reference manual for what the variable "theta" is defined as, I don't think you want that.
* You have a max of 5 coeff loops per iteration. Why have you done this? 10 is more normal.
* You have fourth order Rhie Chow activated? Unless you have a good reason just leave complex options like this at the defaults.
* You have large problem option activated. It does not look like a large problem to me, deactivate this option.

Dale MC July 29, 2019 13:16

Sorry I had to work on some other things for a little while. Ok I have followed your advice and I have gotten it to run. I will mention the velocity function last. I changed the coeff loops per iteration to 10, got rid of the Rhie Chow (I thought coupling the pressure and velocity might help) and deactivated large problem. It gives a transient result but not a "in and out" flow at the small inlet. When I try to put "t" in the sine expression it gives me an error in CFX as soon as I apply it. Theta will run with no errors but not give the pulsing in and out flow cycles. Now that I am back on this, I am going to look in the manual. Any tips will be much appreciated.
Thanks

Светлана July 29, 2019 20:43

I believe .trn files are only available with unsteady (=transient) simulations. In steady state simulations only the backup files are available, but transient results (.trn) are not. This is consistent with what you are saying, I just wanted to make it more clear.

Dale MC September 20, 2019 16:01

Got it!
 
Just wanted to say thanks for everyone's advice and help. I was able to get it and I am greatly appreciative. I have attached a little video.https://1drv.ms/v/s!ApLaxjQceotVjtIzrgEWjZYLsw_edA


All times are GMT -4. The time now is 11:28.