ERROR #004100018 has occurred in subroutine FINMES.

Farough · October 16, 2015, 18:57

Hi everyone,

I am currently using CFX to solve a conjugate heat transfer (fluid-porous) problem with heat source. It is steady state problem; however, I am using the physical timescale of 10 and 100 for the fluid and porous media, respectively.

Unfortunately, i am getting this error after 84 iteration:
ERROR #004100018 has occurred in subroutine FINMES. | | Message: | | Fatal overflow in linear solver.

It seems like my results can not be converged.

The geometry is quite simple and i only have structured mesh, so i don't think it will be from the mesh. The fluid is passing on a porous media with heat source and i am trying to find the temperature distribution of the porous region. I also increased the memory allocation factor in the Define Run Advanced Settings from 1 to 2 and it doesn't make any changes.

Can anyone help me out here? Thank you

ghorrocks · October 18, 2015, 20:53

FAQ: http://www.cfd-online.com/Wiki/Ansys...do_about_it.3F

Sasquatch · August 23, 2016, 07:30

I am also facing this problem now.

I am familiar with the link you provided, however, my simulation is running for 3D meshes and converges well.

For quasi 2D, it fails after 8 iterations. Since it happens after 8iternations, i expect mesh to be ok.

The flow physics remained the same.

Double precision is required and is ok, smaller timesteps have been checked too, it just last longer and does not go below RMS = 10^-1.

The quality of the mesh is higher than in 3D case.

What can be the reason then?

ghorrocks · August 23, 2016, 07:51

Is your mesh 2D planar or axisymetric? What is the thickness or wedge angle?

What flow physics are you modelling?

Sasquatch · August 23, 2016, 10:23

Dear Glenn,

It is 2D planar, 2 cells high, 1e-04 m thick (yes, desirable range is 0.2-200 but all the previous stages work well).

It is a small, 10cm long, 3-stage, radial outflow turbine, organic rankine cycle based.

I started debugging, I noticed that 2.5 stages quasi 2D works well, then I simulated only 3rd rotor with the 'inflow' conditions taken from the '2.5 stage case outlet' - works well too.

Then It looks like my mixing plane is the source but I am puzzled why.

The domain is a bit prolonged at the 3rd rotor outlet so should not cause problem.

My idea now is to use expert parameter and check the 0-back up file to see what is actually being initialized there...

P.S. Interesting, I am trying to check the 0 backup file but I cannot: "ERROR Unknown CCL syntax found: *INFO."

ghorrocks · August 23, 2016, 21:13

Also try making the thickness of your domain equal to the smallest element size. This will improve mesh quality.

Sasquatch · August 24, 2016, 05:18

Temporarily problem solved.

The highly unsteady behavior in the beginning can be changed by initialization from similar case and coarser mesh, here: 3D for example.

The solver went through the first initial 'bumps' and gave a convergence between 10^-4 and 10^-5.

That is in the link above, indeed, but, the error:

"ERROR Unknown CCL syntax found: *INFO."

can be handled by accessing CFD-Post from CFX solver Manager instead of directly when opening backup files apparently.

October 16, 2015, 18:57	ERROR #004100018 has occurred in subroutine FINMES.	#1
Farough New Member Farough Join Date: Jul 2015 Posts: 5 Rep Power: 11	Hi everyone, I am currently using CFX to solve a conjugate heat transfer (fluid-porous) problem with heat source. It is steady state problem; however, I am using the physical timescale of 10 and 100 for the fluid and porous media, respectively. Unfortunately, i am getting this error after 84 iteration: ERROR #004100018 has occurred in subroutine FINMES. \| \| Message: \| \| Fatal overflow in linear solver. It seems like my results can not be converged. The geometry is quite simple and i only have structured mesh, so i don't think it will be from the mesh. The fluid is passing on a porous media with heat source and i am trying to find the temperature distribution of the porous region. I also increased the memory allocation factor in the Define Run Advanced Settings from 1 to 2 and it doesn't make any changes. Can anyone help me out here? Thank you

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
ERROR #001100279 has occurred in subroutine ErrAction.	aniseh	CFX	6	October 5, 2015 12:01
ERROR #002100004 has occurred in subroutine Out_Scales_Flu -> Reynolds-Number = 0???	thda7	CFX	2	June 21, 2014 06:53
ERROR #001100279 has occurred in subroutine ErrAction.	P9408	CFX	1	August 19, 2009 08:56
ERROR #001100279 has occurred in subroutine ErrAct	Carl	CFX	2	July 16, 2005 15:39
ERROR #004100018 has occurred in subroutine FINMES	San Chang	CFX	1	May 26, 2004 19:30

October 18, 2015, 20:53		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,854 Rep Power: 144	FAQ: http://www.cfd-online.com/Wiki/Ansys...do_about_it.3F

August 23, 2016, 07:30		#3
Sasquatch Member Join Date: Jul 2016 Posts: 33 Rep Power: 10	I am also facing this problem now. I am familiar with the link you provided, however, my simulation is running for 3D meshes and converges well. For quasi 2D, it fails after 8 iterations. Since it happens after 8iternations, i expect mesh to be ok. The flow physics remained the same. Double precision is required and is ok, smaller timesteps have been checked too, it just last longer and does not go below RMS = 10^-1. The quality of the mesh is higher than in 3D case. What can be the reason then?

August 23, 2016, 07:51		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,854 Rep Power: 144	Is your mesh 2D planar or axisymetric? What is the thickness or wedge angle? What flow physics are you modelling?

August 23, 2016, 10:23		#5
Sasquatch Member Join Date: Jul 2016 Posts: 33 Rep Power: 10	Dear Glenn, It is 2D planar, 2 cells high, 1e-04 m thick (yes, desirable range is 0.2-200 but all the previous stages work well). It is a small, 10cm long, 3-stage, radial outflow turbine, organic rankine cycle based. I started debugging, I noticed that 2.5 stages quasi 2D works well, then I simulated only 3rd rotor with the 'inflow' conditions taken from the '2.5 stage case outlet' - works well too. Then It looks like my mixing plane is the source but I am puzzled why. The domain is a bit prolonged at the 3rd rotor outlet so should not cause problem. My idea now is to use expert parameter and check the 0-back up file to see what is actually being initialized there... P.S. Interesting, I am trying to check the 0 backup file but I cannot: "ERROR Unknown CCL syntax found: *INFO."

August 23, 2016, 21:13		#6
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,854 Rep Power: 144	Also try making the thickness of your domain equal to the smallest element size. This will improve mesh quality.

August 24, 2016, 05:18		#7
Sasquatch Member Join Date: Jul 2016 Posts: 33 Rep Power: 10	Temporarily problem solved. The highly unsteady behavior in the beginning can be changed by initialization from similar case and coarser mesh, here: 3D for example. The solver went through the first initial 'bumps' and gave a convergence between 10^-4 and 10^-5. That is in the link above, indeed, but, the error: "ERROR Unknown CCL syntax found: *INFO." can be handled by accessing CFD-Post from CFX solver Manager instead of directly when opening backup files apparently.