Learning Forum

[Vihaan Mukherjee]

Hello All, recently I have tried to use 2D axisymmetrical Analysis in Fluent to analyze the Coefficient of Drag as well as the Coefficient of Pressure about different nosecones. Every time I run the simulation, I recieve the error "Error at Node 0: Divergence detected in AMG solver".  How would I go about fixing this issue? When I run mesh check, I do not recieve any errors, and I believe that my boundary zones are set up correctly as well.

Thank you All in advance.

[Rob]

Does the solver diverge immediately or after some iterations? 

[Vihaan Mukherjee]

The solver diverges immediately after starting the calculation. 

[NickFL]

Do you have your axis of rotation aligned to be the x-axis?  And the computational region must be in the +Y quadrant. 

[Vihaan Mukherjee]

yes, I have my axis of rotation aligned to be the x-axis, and the computational region is within the y+ quadrant. I ran solution steering with my calculation, and now it says "Error at Node 0: floating point exception". Where should I go from here?

[NickFL]

When are you recieving this error? If you initialize the solution with hybrid initialization does it appear or is it only after you start to iterate?

[Rob]

The reason for the questions is to figure out why Fluent is struggling. 

Mesh check actually only checks cell quality. It's not looking at gradient resolution, nor does it check domain scale (ie you built the model in mm and Fluent took that as m). So, please post some images (Forum permitting - we're working on that). 

Boundary conditions: are they sensible? Are you using pressure in & out or velocity/massflow in? For the former you need a good initial condition if the flow is complicated and dP is high. 

Model choice: hard to get wrong for turbulence but if you add in radiation & multiphase the number of opportunities to mess up significantly increase.