[Gate-users] Problems with scoring edep and dose in non rectangular geometries and CAD geometries

xela.errac at gmail.com xela.errac at gmail.com
Wed Oct 10 18:17:58 CEST 2018 

Hello,
I had a similar issue with stl file. My geometry was a cylinder with steep corners and when you look at the 3D energy distribution, there are hot spots where energy seems to accumulate.  So I went back to a geometry that didn't use a stl file...

Best regards 


> Le 10 oct. 2018 à 16:21, Maikol Salas Ramirez <mmsalas at gmail.com> a écrit :
> 
> Hi Leonie,
> 
> it is strange that for the sphere works fine and not for the cylinder. Do you have a way to check if your mesh has holes? 
> Freecad for example when you charge a mesh tells you that the mesh has defects.
> 
> Sorry I go again with the possibility that the position is not exactly the same, if you are working with a narrow proton beam, which not cover totally the mesh, the error can be big. If it is the case I would suggest to validate the mesh Vs normal-volume with a broad beam that cover and more the volume, of with small mesh structures.
> 
> Best regards
> Maikol
> 
> 
> 
>> El mié., 10 oct. 2018 a las 14:32, Leonie Brodbek (<leonie.brodbek at uni-oldenburg.de>) escribió:
>> 
>> Hi all,
>> 
>> 
>> 
>> we tried again to play around with the mesh structure but that did not change the results.
>> 
>> We changed the resolution of the mesh structure from 12 to 432 facets for a  box geometry of 1 cm x 1 cm x 0.5 cm. 
>> 
>> In our opinion, for a box geometry the mesh resolution will not have an impact so that is also what we expected.
>> 
>> 
>> 
>> Furthermore we scored a sphere with R = 0.25 mm in a waterphantom in a proton beam and get good results comparable to yours. When we score in a cylinder, we get for the Gate geometry twice the value of the CAD geometry. 
>> 
>> We will keep on searching for a solution of our problem, if anybody has another idea we would be very thankful!
>> 
>> Kind regards 
>> Leonie
>> 
>> 
>> 
>> Von: Maikol Salas Ramirez <mmsalas at gmail.com>
>> Gesendet: Montag, 8. Oktober 2018 15:21:29
>> An: Leonie Brodbek
>> Cc: Maxime Chauvin; gate-users
>> Betreff: Re: [Gate-users] Problems with scoring edep and dose in non rectangular geometries and CAD geometries
>>  
>> Hi all,
>> 
>> I did test (for my, this topic is very important), I put a cylinder (H=1mm, R=0.25mm) and a sphere (R=0.25mm) in a Lu-177 box 1m3 (images attached). The result are consistent:
>> 
>> S-Value (mGy/Mbq-s)
>> Geometry	Cylinder (bone)	Background (source)	Sphere (fat)
>> Solid	9,07	0,01	10,18
>> Mesh	8,87	0,01	9,96
>> %Error	-2,15	0,70	-2,14
>> 
>> For sure there is a small difference, but it comes from the surface triangulation in the mesh.
>> 
>> I calculated the deposited energy by the spectrum.
>> 
>> It will be interesting to see under what conditions there is not consistency between the two geometrical approaches.
>> 
>> Best regards
>> Maikol
>> 
>>> El lun., 8 oct. 2018 a las 13:20, Maikol Salas Ramirez (<mmsalas at gmail.com>) escribió:
>>> Hi Leoni,
>>> 
>>> Just going back to this sentence "When scoring edep we obtain a value of about 3000 MeV in the geometry built with Gate, when scoring in the CAD geometry we obtain a value of about 370 MeV".
>>> 
>>> This difference can come from the "finish" of your mesh, it is important to remember that the mesh representation is not the same that the 3D volume with perfect curves. I would say that it is a good idea to play with the level of detail of the finish of the mesh surfaces. Also depending of the size of your geometries and  if you are simulation electrons, you could have big differences, because electron are very sensible to this surface changes.
>>> 
>>> Best regards
>>> Maikol
>>> 
>>>> El lun., 8 oct. 2018 a las 12:36, Leonie Brodbek (<leonie.brodbek at uni-oldenburg.de>) escribió:
>>>> Dear Maxime,
>>>> 
>>>> 
>>>> 
>>>> Thank you! 
>>>> 
>>>> Unfortunately we already checked this and it seems that both geometries fit well and that this does not cause the problem.
>>>> 
>>>> 
>>>> 
>>>> Kind regards
>>>> 
>>>> Leonie
>>>> 
>>>> Von: Maxime Chauvin <maxime.chauvin at inserm.fr>
>>>> Gesendet: Montag, 8. Oktober 2018 12:01:11
>>>> An: Leonie Brodbek
>>>> Cc: Elia Alessio; gate-users at lists.opengatecollaboration.org
>>>> Betreff: Re: [Gate-users] Problems with scoring edep and dose in non rectangular geometries and CAD geometries
>>>>  
>>>> Dear Léonie,
>>>> 
>>>> something you could check is the position of the geometry. Importing a .stl file can result in a unwanted shift due to a different reference frame from CAD compared to how GATE build a box or a cylinder. And this can cause a difference in Edep and absorbed dose.
>>>> 
>>>> As a rough check you can use the visualisation (Gate --qt) to see if your geometries are identical between your 2 simulations.
>>>> 
>>>> Kind regards,
>>>>   Maxime Chauvin
>>>> 
>>>>> On 8 Oct 2018, at 11:26, Leonie Brodbek <leonie.brodbek at uni-oldenburg.de> wrote:
>>>>> 
>>>>> Hi Alessio,
>>>>> 
>>>>> thanks a  lot!
>>>>> 
>>>>> We tried to create geometries with a CAD program to import more complex geometries and tested it with scoring in a box that is built inside Gate and with the same geometry built in the CAD program. 
>>>>> When scoring edep we obtain a value of about 3000 MeV in the geometry built with Gate, when scoring in the CAD geometry we obtain a value of about 370 MeV. We imported the CAD geometries via a .stl file and the insert specification "tessellated". We obtained the same behavior when scoring Edep in cylinders built in gate and with the help of a CAD programm. 
>>>>> 
>>>>> Kind regards
>>>>> Leonie
>>>>> 
>>>>> Von: Elia Alessio <alessio.elia at medaustron.at>
>>>>> Gesendet: Montag, 8. Oktober 2018 10:58:37
>>>>> An: Leonie Brodbek
>>>>> Cc: gate-users at lists.opengatecollaboration.org
>>>>> Betreff: RE: Problems with scoring edep and dose in non rectangular geometries and CAD geometries
>>>>>  
>>>>> Hi Leonie,
>>>>>  
>>>>> thanks for your email. Concerning the first point you spotted could you be more precise? (for example what do you mean with “different results with different order of magnitude”?). With respect to the second point, this is a known behavior of the dose actor. At the moment, it cannot handle correctly non-regular geometries as far as I know. I would recommend you to work with the edep and then post-process the dose. 
>>>>>  
>>>>> Hope it can help!
>>>>>  
>>>>> Best
>>>>> Alessio
>>>>>  
>>>>> From: Gate-users [mailto:gate-users-bounces at lists.opengatecollaboration.org] On Behalf Of Leonie Brodbek
>>>>> Sent: Montag, 08. Oktober 2018 10:31
>>>>> To: Gate-users at lists.opengatecollaboration.org
>>>>> Subject: [Gate-users] Problems with scoring edep and dose in non rectangular geometries and CAD geometries
>>>>>  
>>>>> Dear all,
>>>>>  
>>>>> we try to perform simulations of different detectors with Gate and have some problems when scoring dose and edep in different volumes, perhaps anyone can help us with these issues:
>>>>>  
>>>>> - When we score dose and edep in a gate constructed geometry and in the same CAD geometry we obtain different results with different orders of magnitude (even in rectangular geometries).
>>>>>  
>>>>> - When we score dose and edep in non rectangular geometries we cannot recalculate the dose out of edep and it seems that the dose in Gate is calculated by a geometry surrounding the real geometry but having rectangular corners, e.g. for a sphere of radius 0.5 cm the dose can be recalculated out of edep  by a cube with side length 1 cm.
>>>>>  
>>>>> We would be very thankful for any help!
>>>>>  
>>>>> Kind regards 
>>>>> Leonie Brodbek
>>>>> Disclaimer:
>>>>> Please notice our E-Mail Disclaimer http://www.medaustron.at/email-disclaimer/ 
>>>>> 
>>>>> _______________________________________________
>>>>> Gate-users mailing list
>>>>> Gate-users at lists.opengatecollaboration.org
>>>>> http://lists.opengatecollaboration.org/mailman/listinfo/gate-users
>>>> 
>>>> _______________________________________________
>>>> Gate-users mailing list
>>>> Gate-users at lists.opengatecollaboration.org
>>>> http://lists.opengatecollaboration.org/mailman/listinfo/gate-users
> _______________________________________________
> Gate-users mailing list
> Gate-users at lists.opengatecollaboration.org
> http://lists.opengatecollaboration.org/mailman/listinfo/gate-users
-------------- section suivante --------------
Une pièce jointe HTML a été nettoyée...
URL: <http://lists.opengatecollaboration.org/pipermail/gate-users/attachments/20181010/68c5b643/attachment-0001.html>

More information about the Gate-users mailing list