Negative fission rates and Error: "Too many rejections on Maxwell fission spectrum"

User Support
1

Hi all,

In a calculation where I asked for fission rates, in some places that I expected zero fission rates I obtained sth like that: -7.9752E-10 +/- 7.9752E-10. Why?

And secondly, what is the meaning of the following error message: “Too many rejections on Maxwell fission spectrum” ?

Thanks in advance.

Antonis

2

For the first question (negative fission rates), I’m going to guess that you are experiencing these only during a mesh tally, is that right?
I suspect that the negative reaction rates come from negative distances to mesh boundaries which occur when the code asks “which mesh cell is this particle in?” and “how far until the surface of this mesh?”. Sometimes these questions don’t always give consistent answers. For example the distance to a surface may actually indicate being outside of the mesh cell that the particle is supposedly in.
This is a function of floating point operations on computers, there are some things programmers can do to work around this, but they probably wont ever be perfect until computers start working in base-10 :-).

I know my answer is only the “why” and not a “here is how you or we can fix it”, maybe someone else has a suggestion for the fix? (Though that might be better for a GitHub issue discussion, if it makes it that far).
For now its probably safe to disregard those very small and negative numbers.

On to the second question:
You received this error because OpenMC rejects samples from energy distributions which result in non-physical outgoing energies above some threshold defined by the data; if it has to do this more than 100,000 times when sampling a reaction, it decides something is wrong and quits. This is the message you get when it has to quit because of this.
I am surprised you received that error. Which ACE data set are you using, and do you have any idea which nuclide is causing that? It sounds like something in the data (or our reading of the data) is incorrect.

Adam

3

Dear Adam

Thank you for your answer.

Yes, I have experienced that during a mesh tally. I also supposed that it is probably due to floating points operations because apart from negative fission rates, I have also noticed some very small positive reaction rates in regions where I would expect zero ones.

As regards the second question, I used JEFF3.1.2 ACE data. I will investigate this problem a little bit, and I will inform you for what I discovered.

Finally, I have also experienced the following error message:“Too many rejections on evaporation spectrum”. It is also a matter of nuclear data, isn 't it?

Thanks again.

Best regards,

Antonis

Τη Παρασκευή, 25 Ιουλίου 2014 3:06:45 μ.μ. UTC+3, ο χρήστης Adam Nelson έγραψε:

4

Antonis,

Do you know what version of NJOY was used to process the JEFF 3.1.2 data you’re using?

Yes, the message about the evaporation spectrum is also related to the nuclear data.

Thanks,
Paul

5

Paul,

They have used NJOY 99.364 with some slight modifications.

Thanks,
Antonios

Τη Παρασκευή, 25 Ιουλίου 2014 10:11:49 μ.μ. UTC+3, ο χρήστης Paul Romano έγραψε: