ERROR in opening *.h5 file

Hello!

I am modelling detailed reactor in openmc. But I am having some problems with simulations.

First I am having error when I want to get the tallys from .h5 file. The error is:

Traceback (most recent call last):

File ˝test.py˝, line 3, in

sp = openmc.StatePoint(‘statepoint.2.h5’)

File ˝/usr/lib/python2.7/dist-packages/openmc/statepoint.py˝, line 143, in init

su = openmc.Summary(path_summary)

File ˝/usr/lib/python2.7/dist-packages/openmc/summary.py˝, line 47, in init

self._read_geometry()

File ˝/usr/lib/python2.7/dist-packages/openmc/summary.py˝, line 96, in _read_geometry

self._read_cells()

File ˝/usr/lib/python2.7/dist-packages/openmc/summary.py˝, line 300, in _read_cells

region, {s.id: s for s in self.surfaces.values()})

File ˝/usr/lib/python2.7/dist-packages/openmc/region.py˝, line 121, in from_expression

tokens.append(+surfaces[abs(j)])

KeyError: 1

And the second problem is that when the simulation is running there appear warning: Particle (number of particle) underwent maximum number of events. I have a limited model with vacuum boundary conditions and the same problem was with reflective boundary conditions. With reflective boundary condition there was another problem that I was losing some particles which couldn’t be found after reflecting.

Has anyone already had a similar problem and can help me?

Best regards,

Ingrid Vavtar

Hi Ingrid,

What version of OpenMC are you using, and how did you install it?

Best,
Paul

Hi Paul,

I am using OpenMC version 0.8.0 and I am runing it via virtual machine on »Ubuntu« (Release 16.04.2 LTS (Xenial Xerus) 64-bit, Kernel Linux 4.10.0-32-generic x86_64, MATE 1.12.1). OpenMC was installed by default commands:

sudo apt-add-repository ppa:paulromano/staging

sudo apt-get update

sudo apt-get install openmc

Best regards,

Ingrid Vavtar

Hi Ingrid,

A newer version of OpenMC should be available from the PPA. One thing you can do is to try updating your version and see if that fixes the problem. Note that some things have changed in version 0.9.0.

Another thing you can try in your script is to not automatically link in geometry information when reading the statepoint. When you run:

sp = openmc.StatePoint(‘statepoint.2.h5’)

It will look also for a file named summary.h5 that contains information about your model and try to read that in too. If instead you run:

sp = openmc.StatePoint(‘statepoint.2.h5’, autolink=False)

it will not look for that summary file (which seems to be what’s causing problems based on the traceback.

Best,
Paul

Hello Paul,

Thank you for your recommendations. After reviewing my code I also found an overlapping where I used a quadratic function. After implementing improvements and removing overlapping my code worked.

Best regards,

Ingrid Vavtar

Dne četrtek, 24. avgust 2017 13.31.01 UTC+2 je oseba Paul Romano napisala: