Boundary effects for cyclic boundary conditions
Tamara Tromp
---Sorry, the images were not visible in my previous message so let me try this again.---
Dear dr. Fang,
Recently I was trying out some different resolutions when I noticed something remarkable, namely that there can be boundary effects even when using cyclic boundary conditions (cc_cc_).
For example, running a simulation with a test slab with LengthUnit 0.002 mm yields the following CCD image, where no boundary effects are visible:
The corresponding mcfg is:
{'dim': [425, 200, 384],
'prop': array([[ 0. , 0. , 1. , 1.35],
[ 0.5 , 150. , 0.9 , 1.35]]),
'issrcfrom0': 0,
'unitinmm': 0.002,
'tstart': 0,
'tend': 5e-09,
'tstep': 5e-09,
'srctype': 'planar',
'srcpos': [0, 0, 0],
'srcdir': [0, 0, 1, 0],
'srcparam1': [425, 0, 0, 0],
'srcparam2': [0, 200, 0, 0],
'nphoton': 100000000.0,
'respin': 1,
'isreflect': 1,
'bc': 'cc_cc_001000',
'isnormalize': 1,
'maxdetphoton': 100000000.0,
'autopilot': 1,
'outputtype': 'flux',
'savedetflag': 'pxv',
'debuglevel': 'P',
'issavedet': 1,
'issaveseed': 1,
'verb': 8}
However, when running the same simulation but with LengthUnit 0.008 mm, I obtain the following CCD image, which displays a clear intensity dropoff at the boundaries:
The corresponding mcfg is:
{'dim': [106, 50, 96],
'prop': array([[ 0. , 0. , 1. , 1.35],
[ 0.5 , 150. , 0.9 , 1.35]]),
'issrcfrom0': 0,
'unitinmm': 0.008,
'tstart': 0,
'tend': 5e-09,
'tstep': 5e-09,
'srctype': 'planar',
'srcpos': [0, 0, 0],
'srcdir': [0, 0, 1, 0],
'srcparam1': [106, 0, 0, 0],
'srcparam2': [0, 50, 0, 0],
'nphoton': 100000000.0,
'respin': 1,
'isreflect': 1,
'bc': 'cc_cc_001000',
'isnormalize': 1,
'maxdetphoton': 100000000.0,
'autopilot': 1,
'outputtype': 'flux',
'savedetflag': 'pxv',
'debuglevel': 'P',
'issavedet': 1,
'issaveseed': 1,
'verb': 8}
I did some small test simulations to investigate this issue (all with a single slab) and the boundary effects seem to be more pronounced for a coarser resolution (larger LengthUnit), for a smaller g, for a larger mu_s and for a smaller mu_a, as you can see in the figure below. In that figure, I used [mean(I_middle) - mean(I_edge)]/[max(I_middle) - min(I_middle)] as a boundary effect metric, where I is the array with CCD image intensity values, the middle of the CCD image is taken from 0.25* the height and width to 0.75 * the height and width, and the edge is taken as all the outer pixels. (For comparison, the boundary effect metric of the 0.002 mm image above is 0.02, while the metric of the 0.008 mm image is 0.30.)
One possible explanation I could think of is the source (planar in this case) being non-cyclic. Therefore, in the x and y direction there might occur a similar effect as Li 2015 (doi: 10.1117/1.JBO.20.9.095014., section 2.1.2) describes for the w-direction. However, I am not sure whether this makes sense.
I was wondering whether you have encountered this before and if you maybe know what is happening and why. Any help is greatly appreciated!
Kind regards,
Tamara Tromp
Qianqian Fang
hi Tamara,
you should set cfg.issrcfrom0 to 1 instead of 0.
when you set it to 0, the domain origin is actually located at [1,1,1] instead of [0,0,0]. because of this, your source and the domain are not aligned.
see this line in the cyclic boundary demo script, this line is critical
https://github.com/fangq/mcx/blob/v2025.10/mcxlab/examples/demo_infinite_slab_cyclic_bc.m#L13
Qianqian
--
You received this message because you are subscribed to the Google Groups "mcx-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mcx-users+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/mcx-users/555fab7b-7ad5-4f9d-8f75-fb36630a02b9n%40googlegroups.com.
Qianqian Fang
just to explain what happened -
every user's first post will be moderated, which means I have to
manually approve your first post before your post can appear. This
way we can prevent spams.
As a matter of fact, I approved your post about 1 hour before you repost, likely you did not refresh and saw the approved post.
For future users, please wait with patience, it may take up to a
day or two for me to see the notification and approve your post,
but I will handle it eventually, no need to repost (unless you
want to add/correct something)
--
Tamara Tromp
Thanks a lot, this solves the issue indeed!
Kind regards, Tamara