Flex 2016 Error Codes

[Anonymous]

Hi there,

It seems a very common issue amongst myself and others to run into what I refer to as The error message of death:

I call it this because it is a generic, unhelpful, fatal and frequent error message. Which is just about the worst kind you can get. Furthermore, after reading the solver logs, I can usually always find an exit code but I have not been able to find a reference table for these codes so I can't even attempt to resolve the issue myself. Is this issue currently being worked on? Has this been fixed in Flex 2017? (We are currently running 2016). And if not, is there a public resource somewhere containing the details of these exit codes so I can at least attempt to resolve them myself rather than heading over to the forums every time it happens?

In a related matter, a simulation of mine just ran into this error message again. Exit code 102. I cannot figure out the issue 😞 cfz file is attached.

Regards,

Matthew Gent

[Anonymous]

Hi Matthew,

from my users perspective: there are mostly two reasons for this message:

1. some issues in the "physical" setup. Like a viscosity of 0 and so on. In these cases you will get the message at the very beginning of the solution and you should be able to fix it by correcting the physics.
2. "act of nature beyond control"... then it is what it states: something killed the solver process during the solution... Sometimes randomly... there's no quick answer for this, I'm afraid...

Marco

[Anonymous]

Hi Marco,

An "act of nature beyond our control" are really not the words of a engineer there Marco!

But yes I have encountered some of the common issues, for example the "MaxRadMatrixSize" flag, and I agree with your first point that it is usually down to an issue with the physical setup. However, I believe this problem to be reportable. I'm a programmer myself, so I know that any error that a program encounters can be debugged and reported, no matter what the situation. It's the way programs are supposed to work. A generic error message is a sign of an incomplete program. Given how frequently the "error message of death" occurs, I think this needs a high priority. I'm hoping someone from Autodesk could weigh in on the matter.

[Jon.Wilde · Accepted solution]

I would guess your issue is the fans and BC's close by.

Can flow enter where you have P=0? What is the temp here?

Are the centrifugal fans pumping air in or pulling it out? Either way they have a lot of inlet/outlet surfaces selected. From what I see there is one inlet surface (where the fan touches the cylinder of air) and two outlet surfaces (where the flow finally leaves the pump), any more will confuse CFD for sure 🙂

Check you do not have flow recirculating over and outlet - if flow re-enters and CFD has no reference temp you will likely see an error.

What is the actual planned air path here?

A note on the reporting also - it is something we are looking to improve so that the messages are far more specific (and therefore helpful).

Hope that helps Matt!

[marwan_azzam]

Another cause for this error is the inability to create a mesh.

To verify if that is the case:

Right-click on the scenario in the browser (model tree) and select "Generate mesh"

Note that you are not using more, or wasting any, time by doing this step since the mesh has to be generated anyway.

[Anonymous]

Hi Jon,

Flow can enter where the P0 is correct. The temperature at the start of the simulation is 27 degrees celsius. The centrifugal fans are pulling air out of the P0 and pushing it into the room. I remodelled the fans, selecting each fan individually and selecting the inlet and outlets that specific fan only. This seemed to fix the problem. I guess you can't model multiple centrifugal fans under one material like you can with internal fans.

And also I'm super glad to hear that! That error message really is a pain in flex. As long as there is a specific error message to the problem then I am happy because it allows me fix the issue myself 🙂

Cheers

[Anonymous]

Thanks for the tip marwan azzam!

[Jon.Wilde]

Correct  🙂

Be sure to have a temp anywhere flow can enter though.

If you have recirculation at the outlet (I guess you might), really you need to extend it so that flow only exits - might need a little rethink.

Thanks!

Jon

[Anonymous]

Hi Jon,

Yes I have a temperature of every inlet. I do this wherever I can on my simulations. That outlet thing will be tricky. I'll have to think of a 3D design to make that work. But cheers Jon.