Divergence detected in AMG solver and floating point exception error

Masroor · June 7, 2018, 01:53

Dear Experts,
I am simulating the unsteady multiphase flow dynamics of air and water in a bubble column. I am using class method of population balance to study the bubble size distribution. In this study I need to use some drag correlations through the compiled c source files. As I iterate the solution, after several iterations I receive the divergence detected in AMG solver and the floating point exception. Basically, as in that drag I define the sauter mean diameter through this macro "diamb=C_PB_DISCI(cell, thread_b,3)/C_PB_DISCI(cell, thread_b,2)". But when I define it through the "C_PHASE_DIAMETER" (without using above mentioned macro/without PBM), it works. I don't know where the problem exists and how it can be rectified. I need your expert opinion to get rid of this unwanted solution.
Thanking you in advance and looking forward to receive a quick and positive response from your side.

obscureed · June 8, 2018, 10:51

Hi Masroor,

Your formula would correctly calculate a Sauter mean diameter if applied to (length-based, number-weighted) moments of the size distribution, but that is not appropriate for C_PB_DISCI. C_PB_DISCI(c,tc,i) is the mass fraction in bin "i" of the size distribution in cell c, (sub-)thread tc. (I always have to remind myself that i=0 is the largest bin, not the smallest.)

If you can live with using C_PHASE_DIAMETER (which can be defined as the phase's Sauter mean diameter in the GUI), then this is a lot simpler.

Best regards,
Ed

June 7, 2018, 01:53	Divergence detected in AMG solver and floating point exception error	#1
Masroor New Member Masroor Join Date: Nov 2015 Posts: 10 Rep Power: 11	Dear Experts, I am simulating the unsteady multiphase flow dynamics of air and water in a bubble column. I am using class method of population balance to study the bubble size distribution. In this study I need to use some drag correlations through the compiled c source files. As I iterate the solution, after several iterations I receive the divergence detected in AMG solver and the floating point exception. Basically, as in that drag I define the sauter mean diameter through this macro "diamb=C_PB_DISCI(cell, thread_b,3)/C_PB_DISCI(cell, thread_b,2)". But when I define it through the "C_PHASE_DIAMETER" (without using above mentioned macro/without PBM), it works. I don't know where the problem exists and how it can be rectified. I need your expert opinion to get rid of this unwanted solution. Thanking you in advance and looking forward to receive a quick and positive response from your side.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
PEMFC model with FLUENT	brahimchoice	FLUENT	22	April 19, 2020 16:44
Floating Point Exception Error	nyox	FLUENT	11	November 30, 2018 13:31
[ANSYS Meshing] Help with element size	sandri_92	ANSYS Meshing & Geometry	14	November 14, 2018 08:54
fluent divergence for no reason	sufjanst	FLUENT	2	March 23, 2016 17:08
Divergence problem	Smaras	FLUENT	13	February 21, 2013 06:03

June 8, 2018, 10:51		#2
obscureed Senior Member Join Date: Sep 2017 Posts: 246 Rep Power: 12	Hi Masroor, Your formula would correctly calculate a Sauter mean diameter if applied to (length-based, number-weighted) moments of the size distribution, but that is not appropriate for C_PB_DISCI. C_PB_DISCI(c,tc,i) is the mass fraction in bin "i" of the size distribution in cell c, (sub-)thread tc. (I always have to remind myself that i=0 is the largest bin, not the smallest.) If you can live with using C_PHASE_DIAMETER (which can be defined as the phase's Sauter mean diameter in the GUI), then this is a lot simpler. Best regards, Ed