Implementation of ACE laws, sampling nu, and SIGMA1 doppler broadening. All codes are inspired by OpenMC. #814
Conversation
…ation regions(NR!=0) is encounted in sample(); 2.discrete lines (ND!=0) in _get_energy_distribution()\nAccount isotropic mu case in law 61; Integrate sample function
…sample function for complete read ace data; write tabular_sample function for neat
Wenkai-Fu
changed the title
| @@ -25,6 +25,7 @@ from collections import OrderedDict | |||
|
|
|||
| cimport numpy as np | |||
| import numpy as np | |||
| rand = np.random.rand | |||
There was a problem hiding this comment.
It is more normal to do
from numpy.random import rand
| INTERPOLATE_TAB1 interpolates a function between two points based on | ||
| particular interpolation scheme. The data needs to be organized as a ENDF TAB1 | ||
| type function containing the interpolation regions, break points, and | ||
| tabulated x's and y's. |
There was a problem hiding this comment.
Docstring content should align with quotes
|
Looks like CI is failing for the same reason as in my PR #826 - I will have a play to see if I can figure out what it is. I think that maybe minconda has its own gfortran version which is older than the gfortran we load at the begining of .travis.yml |
Wenkai-Fu
changed the title
Wenkai-Fu
changed the title
|
Hi @scopatz , the implementation of the ACE Laws governing neutron interactions has been written, together with the Doppler broadening of the cross section by SIGMA1 method (inspired by OpenMC). Thanks. |
| # At sample stage, use nd to determine if there is | ||
| # discrete lines error | ||
| edist.intt.append(INTTp % 10) | ||
| edist.nd.append(INTTp / 10) |
There was a problem hiding this comment.
This should probably be INTTp // 10
|
Would be good to have a new test for the |
|
Hi @paulromano , do you have any idea about verifying the broadened cross section, e.g., compare the broadened value with your OpenMC results, because I am not familiar with Fortran? Thanks. |
|
@Wenkai-Fu The doppler module that is in OpenMC has not been thoroughly tested and I know that others have discovered issues with it, so I wouldn't consider comparison with OpenMC to be of much value. It would be better to compare to NJOY or, better yet, Fudge (which is open source). |
|
Thanks for updating this @Wenkai-Fu, I'll let @paulromano keep reviewing this. I am a little overwhelmed at the moment. |
1, In NeutronTable class, a function called sample_nu(self, e_in) is added to sample outgoing nu.
Input: e_in, user-specified incident neutron energy
Outcome: [nu_total, nu_prompt, nu_delay], if no data is contained in this neutron table, return None, e.g., if only total nu data is presented, return [nu_total, None, None]
Input: e_in, user-specified incident neutron energy
Output: [mu, E_out], secondary neutron's angle and energy. In sampling E_out, the applied ace laws are: 3(Inelastic level scattering), 4(Continuous Tabular Distribution), 44(Kalbach-87 Formalism), 61(correlated_sample). Other ace laws can be implemented in the future.