Large error file generation

Dear FLUKA expert,

In an input, sources are read from an external file. If the sources are considered as volume source (sampling uniformly through out the volume), then it is running properly. If I change it to a surface source (sampling over the surface only), then a large error file is getting generated. Here is a screenshot of it.

Here is the surface sampling source routine.
source.f (13.2 KB)

Can anything be concluded from the error file regarding the reason behind this?

Regards,
Riya

Could you please share also your flair file?

Dear experts,

Do I need to upload any other file?

Regards,
Riya

Dear Riya,

there are some geometry error in your simulation where the grains overlap. You can use the coordinates printed in the error file to eliminate them.

Since you not uploaded the points_source.txt file you read in your source routine, I can only assume, it contains the location and radius of these grains. In this case sampling only on the surface increases the probability for encountering the geometry errors.

Cheers,
David

Dear @horvathd ,

When I checked the geometry after creating the input, it always shows me green [1], but nothing is shown in the drop down menu. Can you tell me how to visualize the overlap and why it is not visible in the error tab?

Regards,
Riya

Dear Riya,

the easiest way to visualize the error is discussed on slides 25 and 26 of this presentation:

(Use the 2D views with the Lattice layer.)

Cheers,
David

Thank you for the link @horvathd ,

I tried 2D views and copied the positions, directions and length from the .err file and added an arrow. But still I am unable to see any geometry error (it should show something in red color near the overlap). Here are the screen shots. Also, I am attaching the .err file and point_source for your perusal.


points_source.txt (161.1 KB)
test001.err (421.4 KB)

Regards,
Riya Dey

Dear Riya,

you may see the errors by moving one of the viewport over the geometry, see:

But it seems the problem is when you want to start a primary on a boundary. In this case sometimes FLUKA can’t decide which side of the boundary is the starting point, but it can recover the error.
To eliminate the issue, you may want to increase or decrease the radius used in sampling the starting position of the primary.

Cheers,
David

1 Like

Thank you @horvathd for the suggestion. I subtracted a very small value from the radius while sampling, and now no such errors are generating for 1000 histories (I hope this will work too for higher histories as well).

But I am unable to get the nice plots (where we can see cross symbols in overlapped region). Can you please explain what you meant by

? Do I have to change x,y,z origins manually as well to see if it is happening in some other plane? Or is there any easier way to do that? Also, in your lattice view, only primary lattices are shown, rest of the box are empty white (which is cleaner to visualize). Can you please tell how to do that? Have you set vaccum in those regions?

This is what I am getting in all the view ports:

Also one more thing:

While adding arrow, I put values of x,y,z,dx,dy,dz from the .err file. Then automatically 0.999 was set as length. But as soon as I am putting length value from the .err output, it is changing dx,dy,dz. Is it normal to happen?

Regards,
Riya

Dear Riya,

yes, I mean moving the viewports manually, because Flair only sees the errors when they are showing in one of them. I don’t think there is an easier way to do it, as FLUKA may not even give any errors during the run, as it can’t detect overlapping volumes.

In my plot the lattices are disabled, as I wanted to see the geometry errors in the real volume.

In FLUKA the dx, dy and dz vectors are normalized to the unit length (1.0), while in Flair they are multiplied with the absolute length of the vector.

Cheers,
David

1 Like

Thank you @horvathd for the explanation.

Regards,
Riya