Learning Forum

[Lutz Rohlfing]

Hello,

i use ANSYS Fluent 2022 R2 to simulate two rotor blades passing each other closely. I use dynamic meshing with a profile too move my rotor blades. The profile specifies the rotational speed at different time steps. When i run my case it sometimes randomly crashes. Fluent closes mostly without giving any error. I once got this message:

Since my computer language is set to german, the message is in german. It says:

 

"The instruction at ... referenced memory at ... The read operation could not be executed in memory.

Click OK to terminate the program. Click Cancel to debug the program."

 

Other than that i can find a cortex error file and a general error file in my fluent folder:

Error [cortex] [time 7/11/24 12:40:36] Cortex received a fatal signal (SEGMENTATION VIOLATION).

Error [node 999999] [time 7/11/24 12:40:32] Abnormal Exit!

 

I ran my case on two different computers, both randomly crash on some simulations. If i drasticlly change my element size in my model, the model runs through completly fine, so i do not think my file is fundamentally broken. The crash can be repeated on the exact same time step, either by completly re running the simulation or running from a previous autosave. But again, some models with some meshing variations work.

 

Help here would be really apreaciated. I can provide further information if needed.

 

With kind regards

Lutz Rohlfing

[Rob]

If there's a rotor close pass how do your meshing settings compare with the tip gap? Have a look in the transcript to see if there's a more diagnostic message (mesh quality being the first one I'd look for). 

[Lutz Rohlfing]

Hello Rob,

 

thanks for your answer. I do not think that this is the issue. The model crashes sometimes way before the two blades meet. I had on the other hand instances where the model ran completly fine even when the blades past really closely. I actually got another error message.

If i load the last auto save (i save every five time steps) and let it run. it crashes again. i can bypass the crash if i change some settings. For example if i change the skewness threshold for the dynamic meshing, the model runs further then the previous crash time step.

[Rob]

How fast are the rotors spinning relative to the tip cell size? Ie how many time steps before the tip moves more than one cell size?  Are the blades interlocking or spaced? Ie are you modelling something like a kitchen blender? 

[Lutz Rohlfing]

The rotors pass the size of an average cell in 4 steps. They are not interlocking, but spaced. I am modelling to wind turbine blades, that move really close to each other, but do not interlock.

[Rob]

If they're spaced you'll find sliding mesh to be far more efficient. Deforming mesh is only necessary when you can't sweep the blades in a unique zone, such as for a kitchen blender.