MCNP - Science topic

Hi Aline,

First of all be careful when using DE/DFs for calculating dose. Those conversion factors are for calculating effective dose and therefore are not applicable to medical dosimetry where the interest is in dose (energy deposited per mass).

I am assuming that the brachytherapy source you are simulating emits both photons and electrons. Why don't you run an MCNP case with just photons and one with just electrons? You can then sum the results to get a dose estimate. With that being said a +F6 can be more appropriate because depending on the geometry and the material an F6:p, e may double count the electrons coming from secondary reactions.

My suggestion would be to calculate the dose for cells (rather than splitting the geometry in thinner pieces) using the different tallies compare the results and if they make sense continue with the splitting. Tally segmentation can also assist although it is kind of tricky to make sense of how to use them in 2-D slices. They worth a try.

I hope I helped with some inspiration.

Good luck,

dimitris

There is a software monte carlo Universal by Kurchatov institute. It has "delayed neutron" calculation as an option. It shows delayed neutron fraction and something else that I have not checked yet

Hi Sabrina,

I am not sure if this input in this state was running the previous day and now it does not (unless there were changes). For the input you provided the problem is the format of material 1 (m1); please remember that a continuation line must begin with five blanks otherwise MCNP assumes lines "2.5161e-3" and "92235.66c 1.1760e-3 92238.66c 8.2051e-5" do not belong to any card. Once you fix this formatting error this input runs normally.

PS: When asking about code error it always useful to also include either the output or the fatal error messages.

Good luck,

dimitris

Dimitrios Kontogeorgakos I am using version 6.2

Here is my input file! I have yet to make any changes to it that you recommended but this is what I have so far.

Hi Otto Haubrich ,

Maybe you can try DBCN J 26R x , where x is the bank size. From the MCNP manual (either 6.3 or 6.2) there is this note "DEFAULTs vary by application: x28=2048 for most fixed-source problems, x28=128 for criticality problems, x28=16384 for high-energy problems)".

Do some testing with a larger bank size and with a smaller NPS.

Regards,

dimitris

Hi Sohail Ahmad Raza,

you can also try replacing the material number with zero; i.e. FM14(1 0 -6). By doing so, MCNP will use the reaction cross

sections for the material in which the particle is traveling.

You will also need to add -8 in the FM card; this will give you the equivalent of tally F7. See "2.5.4.1 Equivalence of F4, F6, and F7 Tallies" in the MCNP6.3 manual that is available online.

Regards,

dimitris

read this you may find the solutions and try the help from icrp official help communit.

Hi Dimitris,

Thank you so much for your answer. I have done the calculation using the method you mentioned. Reaction rates are coming nicely.

your welcome.

Hi Sara, although I have not completely understood what the attached picture shows here are my thoughts. The "output (pSv/NPS) are the tally results with the DF/DE multiplier (to get the ambient dose rate from ICRP-74). The H*(10) results are the tally results after normalization. If that's true, then to get 1.5e+9 pSv/h from 3.41e-5 pSv/NPS, you have multiplied by 4.4e+13, which is the disintegrations per hour and not the photons per hour (unless, I am not reading the data on the png file correctly). Basically, 1.5e+9 pSv/h needs to be multiplied by 2 (2 photons emitted per Co-60 disintegration). Is the "Taxa de Equivalente de Dose Ambiente)" the measured values? If so, then by multiplying those by 2 will get you closer to them. What was the MCNP to measured ratio for Cs-137?

Hi. I'm not sure. I think that you can find on this website:

or

or

Is there something you'd like to keep constant? For example, if you want the total density or total atom density to stay the same, then you can re-normalize the atom density for the tissue constituents and the added boron, by dividing each value with their sum. The result will be a material definition with the sum of atom densities equal to 1, and then assign the desired density in your cell definition. I hope this makes sense.

Hi Otto, I am not sure I fully understand the question. Do you need to increase the value of the tally in terms of statistics? i.e. decreasing the error? Or you only want to re-normalize the result from being "per source particle" to something else? e.g. per Bq? As you've stated this is done with the FM card.

However, if you are looking to decrease the error without increasing the number of histories, then you may try tally F5 or take a look at variance reduction techniques.

I hope this helps.

Hi, Khadijeh Kanani

as another method you can use variance reduction specially cut card.

by this, the random number that you need will decrease and a warning not show (most probably)

Dear Dimitrios,

thank you very much for your reply!

You are right, I derive the factor in MCNP as you have said. One has to divide the result of the F4 tally in the sample volume filled with the material and its nominal physical density by the same F4 tally result but infinitely diluted (that means the sample is now filled with void, ~ 0 g cm^-3 density). Your definition with the surface flux is equivalent; for a void sample of constant geometric cross section and a parallel beam of particles the surface flux is also constant for every thickness (particle track length equals the thickness of the sample in this case, divided by the volume yields the surface flux).

The integral statistics (60 keV to 20 MeV) are totally sufficient, something like 0.01% statistical error only.

Basically four our fast neutrons the total macroscopic cross sections equal the scattering cross section, i.e. they only scatter. In that case it is logical that 1 primary neutron can make more than 1 score in the F4 tally, if its energy after interaction still is in the range between 60 keV and 20 MeV and if it is still inside the sample. Therefore, meanwhile I would assume MCNP is correct and. By the way, the shielding factors are not much above 1, mostly something below 1.05.

Iam analyzing different materials, severals metals like aluminum and iron or some compounds like calcium carbonate (CaCO3) or cerium chloride (CeCl3).

Kind regards,

Niklas

MCNP can be restarted but you must have runtpe file (file with "r" ending). Then you can use command

"

mcnp6 c tasks 16 outp= output_file_name runtpe= runtpe_file_name

"

More details can be found in MCNP manual (see "continue run")

Hi there, the SD is used to input the volume for those cells that MCNP cannot calculate the volume. For tallies that are expressed in /gr or /cm3 this is required, and one workaround is to use SD=1 as you mention. Of course, the latter depends on what you are trying to calculate and therefore, you may want to include the actual volume (in the MCNP manual there is a method to assist in calculating volumes; it is really helpful). For DPA calculations, the tally result, using the F4 and the FM card, is per cm3; if MCNP can calculate the volume of the cell, then you do not need the SD card. But if you need DPA/s then you need to multiply the tally result by the volume of the cell you are analyzing. To my understanding by setting SD=1 you assume that the volume of the cell is 1cm3 or 1 gr per tally used. I hope this helps.

Hi,

In any set-up, there are deep-seated areas where large relative errors are recorded after simulation; perhaps, this account for the observed inhomogeneity in your result.

Endeavour to reposition the source, redefine the source location in terms of (x, y, z) with the target distribution size or area in mind. Take note that homogenous distribution is short range, so locate your source close to your target area of interest.

I hope you find this suggestion useful.

Best regards,

Olaseni M. Bello.

Hi,

you must write

MODE N

SDEF POS=0 0 0 PAR=1 erg=d1

SP1 -2 $For Maxwellian spectrum

don't forget the syntax is SP1 -2 a

If you are using Kcode calculation while normalization divide the tally by keff.

I think you can get the pulse height spectrum with Tally f 8

And extract other relevant information from the shape of the spectrum.

You can use the WGT in the SDEF definition as Vadim Talanov said. With that being said, the FM card does multiply the fluxes in every energy bin. If you need to be convinced, you can always try this in a short/test calculation to confirm. I have learned a lot by trial and error.

FMESH defines the mesh superimposed over the geometry, that means independent from the geometry definition, and so provides no access to the properties of the underlying geometry like the material number.

Heating is in units of MeV/collision, so it would be the MeV that resulted from a specific type of reaction.

MCNP uses the default free gas thermal treatment for low energy neutrons. For higher accuracy (e.g., consideration of chemical binding effects and crystal structure), one would need to declare the S(α,β) in the Materials Card. Unfortunately, the S(α,β) for sapphire may not be available in MCNP installed libraries. But S(α,β) of O in sapphire is available for download in JEFF-3.3 in ACE format, in the website of OECD-NEA. To add downloaded cross sections, you would need to insert them into the MCNP library locations and edit the XSDIR.

You cannot use fractions on cell card. Materials should be defined in the data card and you can only use the material number on the cell card. In the material card you can define a single material containing several different elements with different weight fractions. Even if you want to define a mixture of different alloys you can define it by computing the weight fraction of each element in each alloy.

For example, you know that there are 0.12 'O' and 0.88 'Lu' in Lu2o3 and 0.12 'O' and 0.88 'Yb' in Yb2O3. now you want to produce a material containing 1/3 weight fraction Lu2o3 and 2/3 weight fraction Yb2O3. in the material card you should enter ZAID of 'O' with weight fraction of 0.12, ZAID of Lu with a fraction of 0.293, and ZAID of Yb with weight fraction of 0.586.

During my summer holidays I took up again this issue and I am very happy to inform that it worked. I have now the MCNP4C version up and running on Ubuntu 18.4 (I should also upgrade this one, without using Wine!!!). Next I will install the code under CentOS also. I started from the source code (Fortran and C++). This means that, if needed, the source code can be modified or additional functions can be added. One of the first things I will do is compiling with PVM or more modern equivalents for parallel computing on many machines (I was doing this in 1998 on 12 SunOS boxes speeding up calculations a factor 10).

See page 91:

You have not properly defined the rectangular box. You have it as the intersection of the "negative" side of planes 1, 2, 3, and 4. The negative is to the "left" of an x plane, and "below" a y plane. So this cell extends to infinity in the third quadrant, effectively. What you want is:

501 5 -0.740582 10 -1 2 -3 4 imp:n=1 tmp= 4.94E-8

As Norbert suggested, you should also add pz planes; their height is irrelevant, but lets say you have the following:

*5 pz 10.0

*6 pz -10.0

Then, you should add "-5 6" to the definition of each cell. You box will become:

501 5 -0.740582 10 -1 2 -3 4 -5 6 imp:n=1 tmp= 4.94E-8

Also, I'm confused about your cell 601. It should probably be:

601 0 1:-2:3:-4:5:-6 imp:n=0

Finally, your ksrc has to contain source locations that are within a fissionable material; otherwise the fission sites will all be rejected when they are sampled (this is your main error, I believe). Your fissionable material is between surfaces 7 and 8, so try:

ksrc 0.3 0.0 0.0

In addition to being inside of a fissionable material, try not to put the ksrc directly on any cell boundaries, as it could get confused about which cell it is in.

Hi~ Ilker Meric

I had used DICOM to MCNP parser to turn Rando Phantom Dicom image into MCNP inputfile. It is really useful! But, the number of voxel (size: 1 mm3) is too much. I can't run and plot the phantom. Do you have better way to solve the problem?

Thanks

Hao-Ting Chang

check this out.

I think the uncertainty function can do so, to give you a wide range of parameter to check and realize which better fit for you

But as Geometry to make a change like a for loop or if function. unfortunately not available

You can read that article for more about uncertainty in MCNP https://mcnp.lanl.gov/pdf_files/la-ur-16-23533.pdf

Augusto Di Chicco I am glad you could find an answer.

Regards,

Modeste

Thank you Abdelhai Ben Ali.

Do you know where can I buy Scan2mcnp ?

I really appreciate your help very much.

In similar situations, I always used SI1 L and SP1 D with normalized probabilities, but as far as I know the probabilities do not need to be normalized.

If the simulation is a detection problem, the remark of Kengo Shibuya is important when the photons are emitted in cascade and when the detection efficiency is high. E.g. if the probability of detection of each of the two gammas 1.17 and 1.33 is about 10%, then the probability of simultaneous detection is about 1%, which is not small with respect to 10%.

Modeste Tchakoua Tchouaso

you can convert graph to spectrum using xyextract software

@yuda

Dear Mohsen Ali

in the following and the attached article you will fiend what you need

please don't hesitate to contact me if you have any problem with the input file!

My e mail prof_fayiz@yahoo.co.uk

A detailed procedure to simulate an HPGe detector with MCNP5 C.C. Conti*, I.C.P. Salinas, H. Zylberberg Institute for Radioprotection and Dosimetry e IRD/CNEN, Av. Salvador Allende s/no, P.O. Box 37750, 22783-127 Barra da Tijuca, Rio de Janeiro, Brazil

This Link Maybe Useful:

I attach two figures.

1° step. You need to use F4:n 2 tally card... (neutrons/cm2) in the cell 2.

2° step. To do a manual calculation to determine the normalization constant "C". After that, in your input to type this value in the FM4 C 2 (it means the value C in the cell 2)

The results in your output file is the reaction rate in this cell.

Regards.

Agree with

For research and education purpose it’s free

I guess "have" and "willing to share" are two totally different topics. We generate and use those models as the basis for our extremely fast shielding calculation process. If you like to find out more, I just uploaded my presentation from PTCOG 57 to Research Gate. Contact me via my website/email if you like more information. www.meissner-consulting.com

Megan, here you have a MCNP primer with examples of basic concepts for source term problems.

Thanks.

Dose MCNP6 work at nano scales? is Kerma approximation valid ? How about electron cross sections in MCNP at nano metric scales? could a user use MCNP tallies for nano dimension particles? If you have access to related references please introduce here.

Dear friends,

1- There are many papers show that MCNPX can be used for nanoparticle simulations in diameter higher than 30 nm. please see my publications in this regard.

2- MCNPX and other codes use atomic cross section of materials for calculations so the cluster of atoms in a nanoparticles behave like other composite materials as far as we know.

3- The *f8 and F4 tallies can be utililzed in 30 nm scale. electron cutoff is about 1 KeV and its range in water is 10 nm. thus with voxel size of several times higher than this range (10nm) the results can be reliable.

4 -F6 tally which shows energy deposition can not used in place where electronic equilibrium does not exist.

I'm not sure if such an option exists in MCNP but I know of an alternative. There are some modules in PyNE ( https://pyne.io/index.html ) that you can use to create a custom material libraries, mix those materials as you please, export the materials to mcnp input format.

Material: https://pyne.io/usersguide/material.html

mcnp: https://pyne.io/pyapi/mcnp.html#module-pyne.mcnp

you will need to install PyNE on your computer and other dependencies. https://pyne.io/install/index.html

To use PyNE for example:

> Creating material library

------------------------------------- start of creatematlib.py ----------------------------------------------------

#! /usr/bin/python import os from pyne import material from pyne.material import Material, MaterialLibrary, MultiMaterial def He_mat(): nucvec = {20000000: 100} He = Material(nucvec) He.density = 0.0001786 return He #.expand_elements() def W_mat() : nucvec = {740000000: 100.0} W = Material(nucvec=nucvec) W.density = 19.35 return W #.expand_elements()

def main(): # create material library object mat_lib = MaterialLibrary() # get material definition mat_lib['He']= He_mat() mat_lib['W']= W_mat() # remove lib if os.path.exists('my_mat_lib.h5'): os.remove("my_mat_lib.h5") # write fnsf material library mat_lib.write_hdf5("my_mat_lib.h5", datapath='/material_library/materials', nucpath='/material_library/nucid') if __name__ == "__main__": main()

------------------------------------- end of creatematlib.py ----------------------------------------------------

> mix materials

------------------------------------------- start of mixmat.py ----------------------------------------------------

#! /usr/bin/python import os from pyne import material from pyne.material import Material, MaterialLibrary, MultiMaterial # Example: mix water and steel by mass # mix=MultiMaterial({mat_lib['SS_316']:0.8,mat_lib['Water']:0.2}) # new_mat=mix.mix_by_mass() # Example: mix water and steel by volume # mix=MultiMaterial({mat_lib['SS_316']:0.8,mat_lib['Water']:0.2}) # new_mat=mix.mix_by_volume() """ Load material library (created using creatematlib.py) """ def load_matlib(): mat_lib=MaterialLibrary() mat_lib.from_hdf5("my_mat_lib.h5", datapath='/material_library/materials', nucpath='/material_library/nucid') return mat_lib def mix_mat(material_library): mix=MultiMaterial({material_library['He']:0.20, material_library['W']: 0.8}) mixed_mat=mix.mix_by_volume() mixed_mat.metadata['mat_number']=1 mixed_mat.write_mcnp('mcnp_text') return mixed_mat #.expand_elements() def main(): # remove old mcnp_text try: os.remove("mcnp_text") except: pass # remove old mixmat_lib try: os.remove("my_mixmatlib.h5") except: pass # create material library object mixmat_lib = MaterialLibrary() # Load material library mat_lib=load_matlib() # mix OB SR mixed_mat = mix_mat(mat_lib) mixmat_lib['mixed_mat'] = mixed_mat # write material library mixmat_lib.write_hdf5("my_mixmatlib.h5",datapath="/material_library/materials",nucpath="/material_library/nucid") if __name__ == "__main__": main()

--------------------------------------------- end of mixmat.py ----------------------------------------------------

Massimiliano Alvioli Sanjay Kharche

Hi dear

I use win7 64 bit.

And I found it in output that multi-time MCNP is running

I follow.......

You must using tally to get it sir

You can use surface of cell flagging functionality (MCNP 6 manual 3.3.5.12 - 3.3.5.13).

At the basic level, the WGT entry on the source card is a tally multiplier. You can use the default (WGT=1.0), and then multiply your tally cards by your source strength (either particles/second or particles) to match the physical situation of your simulation. Another way to go is to set up the source with a WGT equal to the source in your physical geometry. For example, if you were interested in a dose rate at a location inside a Co-60 irradiator with a 1.0 Ci Co-60 source, then you would provide a WGT like the following:

WGT = 7.40e10

[source strength = (3.7e10 dis/s) * (~2 gammas/dis) = 7.4e10 gammas/s]

Your dose rate tally would already be scaled to the source. You could also use and integral number of gammas (say 100 second irradiation, WGT = 7.40e12) and get the total integrated dose. I tend to use 1 for the weight in my calculations and scale the tallies by the source strength, but you may have a different preference.

With regard to how MCNP uses weight during its calculations, the math behind the scenes with regard to Russian roulette, splitting, implicit capture, etc. is designed to provide a "fair game" - or adjusting the weight to make sure that a complete analog Monte Carlo simulation for the same problem would give you the same estimate of the radiation field quantities. The weight games are played to reduce the variance in the MC estimate and converge the estimate more efficiently. In any case, the F4 tally (track length estimate of cell fluence [or fluence rate]) uses this equation to provide its estimate:

F4 = (Weight [particles or particle/s] * Track Length [cm]) / Volume [cm**3]

F4 = particles / cm**2 (or particles / cm**2-s)

The other tallies use the weight in a similar manner.

Can anybody answer this Q

Dear Muhammad

You need to define particle energy as distribution for 2 gamma rays (1.17 and 1.33 MeV).

MCNP input:

SDEF POS 0 0 0 ERG=d1 PAR=p

SI1 L 1.17 1.33

SP1 1 1

I recommend you to read this manual. There are examples for many situation.

Please read this pdf.

Both the F6 and *F8 tallies of MCNP code can be utilized for calculation of absorbed dose by MCNP.

The tally F6 is total energy deposition per mass in a cell, given in MeV/g

The *F8 tallies is the total energy deposition in a cell given in MeV

The calculated values were converted to Gy by dividing MeV by the mass within the cell and multiplying by 1.602×10-8 to convert the units from MeV g-1 to J.kg-1 (Gy).

I suggest using *F8 with the Mode p e to obtain the deposited energy in the interesting volume or cell

see our paper in JINST maybe will be helpful

A comparative evaluation of luminescence detectors: RPL-GD-301, TLD-100 and OSL-AL 2 O 3 :C, using Monte Carlo simulations

July 2017Journal of Instrumentation 12(07):P07017-P07017

DOI: 10.1088/1748-0221/12/07/P07017

check the definition of your cells, try to zoom in the view of the cells in visual editor in all angles, a red dotted line indicates that cell is not defined well, hence the leakage

Hello

Please where do i locate temperature of the reactor from MCNP input deck??

There is no way. Because of the keff is calculated in criticality tasks only. For the tasks with external source one can calculate number of generated neutrons.

Be very cautious here: KCODE problems are not fixed source (i.e., normalized per source neutron) even if the initial source distribution is given by an SDEF code rather than KSRC. In KCODE, SDEF only defines the initial source distribution. After the first cycle, the fission source points generated in the previous cycle are used.

As for normalization of tallies, the tallies in KCODE are normalized PER FISSION NEUTRON. So, you must normalize to the value related to the reactor power, nu-bar, keff, and energy released per fission.

I need to use the Dicom images in MCNP. Did you find the solution Isaac ?

Hi !

I'm not expert on MCNP but try to put a blank line between "surface card" and "Data card".

MCNP need three blank line to read input file :

- Between the "Cell card" and the "Surface card"

- Between the "Surface card" and the "Data card"

- After the "Data card"

Excuse me for my poor english.

yes

MT 107 is allowed for all types of incident particles, your tally should therefore also include (p,a) reactions. If you do not have incident neutrons then your tally should reflect the desired value. You should be able to verify by just adding a MT 1 tally for all neutron interactions.

See here for further details:

e0 1e-3 599ilog 1e3 (MCNP6 Manual, Section 2.8.1)

Is this what you are looking for?

Actually there is. Look in the history for the 1000 identifier in position 3, indicating that this is the source (src) of the history. Position 6 should be the cell where the event started. The X Y and X are the first 3 entries in the line that follows. Of course you have to include the src otion in the EVENTS= statement on your PTRAC card.

BURN card is very suitable to activate material. But it can be used only in kcode calculation.

With help from a colleague we have found a solution, included below. It involves including an FCEL dependancy on each distribution, as well as calling the cells directly within the CEL distribution without adding another D-card. Then ALL of the cells must be included in the following distributions.

SDEF PAR=SF

CEL=D9

X=FCEL D11

Y=FCEL D12

Z=FCEL D13

c

c

SI9 L (3<2[-5:5 -5:5 -10:10]<5) 25

c Cells 3 and 25 are called (3 is in the lattice notation)

SP9 1 2540r 2079

c assigning importance to each cell: lattice ~55%, PuO ball ~45%

DS11 S 1 2540r 14

c First 2541 cells have D1 attached, the last cell is attached to D14

DS12 S 2 2540r 15

c First 2541 cells have D2 attached, the last cell is attached to D15

DS13 S 3 2540r 16

c First 2541 cells have D3 attached, the last cell is attached to D16

c

c ---- Lattice cell source-----

SI1 -0.12 0.12

SP1 0 1

SI2 -0.12 0.12

SP2 0 1

SI3 -0.12 0.12

SP3 0 1

c

c ---- Separate cell source ---

SI14 22.8 27.2

SP14 0 1

SI15 -2.2 2.2

SP15 0 1

SI16 -2.2 2.2

SP16 0 1

Thank you for reading and trying to answer my questions. It's always tricky trying to do so without much context.

Helen