Using FLUKA to Study Radiation Fields in ERL Components
Contents
1. l S 9 01 nyu 2 H a 5 0 0001 EL E izl z H H z v g E o eel 1e 006 j ll ad ec ae S i a Me TS 5 belt for H H D _ gt 1e 008 it 2 Z O gS Oo E a 1e 010 1e 012 100 o 100 200 400 x cm FIG 17 Photon fluence particles per cm along the z axis at y z 0 FLUKA 0 0001 MCNPX rag 1e 006 P a i serie tf EE H Fa a ii T E 1e 008 v g Z 1e 010 w 2 N S 5 1e 012 E aa Qa 1e 014 if 100 o 100 200 300 400 x cm FIG 18 Photon fluence particles per cm along the z axis at y 0 one meter upstream from the back face of the collimator 15 FLUKA 0 0001 MCNPX ay oO E 1e 006 H HT 2 T a ml z 1e 008 i V i 5 L gt a 1e 010 p a L Nn 5 6 1e 012 lt a h 1e 014 H l 100 400 x cm FIG 19 Photon fluence particles per cm along the x axis at y 0 one meter downstream from the fron2. cutoff is set at 1MeV 2 5 10 primaries are run per cycle Both LPB and lambda biasing are used with the lambda factor set at 0 01 The 3D modeling software SimpleGeo 10 is used to create the simulation geometry A diagram of the geometry is shown in Figure 5 The beam particles travel in the positive z direction The collimator is centered at the origin of the coordinate system and a cutaway of the collimator geometry is shown in Figure 6 FIG 6 Cutaway of the collimator geometry C Results 1 FLUKA The calculation took 118 minutes to complete and consisted of 7 cycles Neutron fluences in the xz plane at y 0 are shown in Figure 7 and in the xy plane at z 0 in Figure 8 Fluences are reported in units of neutrons per cm per primary Photon fluences in the xz plane at y 0 are shown in Figure 9 and in the xy plane at z 0 in Figure 10 Fluences are reported in units of photons per cm per primary The simulation was originally run for 5 cycles but the photon fluences outside of the collimator volume were excessively noisy The number of cycles was increased to 7 which improved the data somewhat However as is apparent in Figures 9 and 10 there is still significant variance in these regions It is worth noting that data of spatially distributed quantities can be imported into SimpleGeo and visualized with the simulation geometry Examples of this are illustrated in Figures 11 and 12 Planar and linear projections of
3. and save Note that with the Extract button near the bottom of the DaVis 3D window data can be extracted by a number of methods A very useful extraction is Plane which takes the slice of data being displayed by any of the projection planes and produces an ASCII The data is extracted in a minimal column format so it can immediately be used by gnuplot or other such programs 4 USRBIN Data in Flair Flair can be used to project any range of USRBIN data onto the xy yz or xz planes as well as geometry cross sections When using Flair for spatial plots like this we must keep in mind that FLUKA like MCNPX uses a left handed coordinate system In geometries such as the present one where asymmetries make this relevant we can circumvent this by simply browsing for the data we want using Flair then re plotting it with gnuplot applying an extra factor of 1 to the appropriate coordinate This is the ideal approach because SimpleGeo is right handed If we were to manually mirror the geometry once it s imported into Flair as well as any right handed external data such as the source distribution then the FLUKA data would not display correctly with the SimpleGeo geometry 34 An elegant alternative approach would be to build a left handed geometry in Simple Geo and modify any SimpleGeo screenshots by reflecting them across the vertical Such screenshots don t have labels to act as fingerprints of the image mirroring operation Note tha
4. LMB release the LMB and finally release Shift This creates a reference node of twall1 in TUNNEL which might not appear until the display is refreshed by selecting a different node and then reselecting TUNNEL With TUNNEL selected create a difference node and with this node selected create a cylinder e Name twall2 22 e X Rel 145 e Z Rel 310 e Radius 213 e Height 620 Observe that this inner cylinder is longer than twall1 This is because bodies used in boolean algebra statements must not have coplanar surfaces For instance if twall2 were the exact length as twalll then it would be ambiguous whether the points at the ends of twalll are part of TUNNEL or not Create a plane Ctrl Shift P as a sibling of twall2 e Name tfloor e Y Rel 140 e Rot X 270 Try temporarily changing Rot X to 90 and thus observe that points on the green side of a plane are considered to be external to the plane body while points on the red side are internal to it Note that planes are automatically hidden when they are not selected With the root node selected create a difference node name it TNLAIR and change the material to vacuum Create a reference node in TNLAIR to the difference node in TUNNEL which contains twall2 and tfloor That is Shift drag the difference node in TUNNEL to TNLAIR Next create a reference node in TNLAIR to BLKHOLE Observe that the inclusion of BLKHOLE as a younger sibling of twall2 tfloor c
5. back to the list of output files and notice the four usrbin_ files listed there These are still binaries but we can convert any of them to ASCII for viewing in SimpleGeo We could have simply asked for ASCIIs when creating the score cards but then we would not be able to take advantage of the data merging program which significantly impacts the statistics of the simulation To convert a binary to ASCII in the Output Files screen simply select it hold the RMB and then release it over the option To ascii that appears in the list After a moment the list of files will refresh and a _usrbin_ lis ASCII file will appear Open SimpleGeo and load the DaVis 3D plugin by selecting Macros Load plugins Check the box next to DaVis 3D click Load selected Plug Ins and close the window The plugins 33 tend to hide themselves in the toolbar next to the automatic update icon Open the plugin by clicking DaVis3D or if this button is not visible on the toolbar open it by clicking the Toolbar Options button which appears as a vertical bar with two tiny arrows at the top very well disguised See section 5 3 in the FAQ section of the SimpleGeo manual http theis web cern ch theis simplegeo manual manual2 htm FAQ if this is still not clear Load the data by clicking the points of ellipsis button next to the File name field Select All Files in the Files of type pull down menu and then open the appropriate lis file Click Load data a
6. but if it is not select it in the left frame or press F2 We can open a template to make a new input or we can open an existing input Open the input just exported from SimpleGeo by clicking on the open icon next to the Input field Change the Title field to Monster75 and click yes on the window that pops up asking to update the T TLE card Go ahead and save the project as Monster75 01 flair by selecting the save icon next to the Project field or on the toolbar Press F3 or select Input in the left frame A list of all the input cards is displayed in the right frame and the text of the currently selected card is displayed at the bottom as it actually appears in the input file Selecting different fields of the card will highlight the corresponding WHAT at the bottom The blue lines of text at the top of some cards are the commented lines preceding the card in the input file Pressing F1 while a card is selected will open the FLUKA manual in Flair to the description of that card Note that the 24 Flair manual is integrated into this FLUKA manual and that pressing F1 when a different frame is selected e g Compile or Run opens the Flair manual to the relevant entry I find Flair to be well documented in this manual but it does assume complete knowledge of FLUKA C Geometry Debugging Before making any changes to the input it s a good idea to debug the geometry to make sure that we have not assigned points to more than one
7. data can be extracted during the visualization and multiple data sets can be loaded simultaneously 0 1 0 001 0 0001 1e 05 rimary 1e 06 1e 07 1e 08 1 cm2 p 1e 09 1e 10 1e 11 1e 12 100 50 0 50 100 150 200 250 300 350 x cm FIG 7 FLUKA result for neutron fluence neutrons per cm per primary in the xz plane at y 0 This is the plane of the electron beam 10 x cm FIG 8 FLUKA result for neutron fluence neutrons per cm per primary in the ry plane at z 0 1e 12 1e 14 1e 16 100 50 0 50 100 150 200 250 300 350 x cm FIG 9 FLUKA result for photon fluence photons per cm per primary in the xz plane at y 0 This is the plane of the electron beam 200 100 y cm 100 200 0 01 0 0001 1e 06 1e 08 1 cm2 primary 1e 10 1e 12 1e 14 100 0 100 200 300 x cm FIG 10 FLUKA result for photon fluence photons per cm per primary in the zy plane at z 0 toor a J TAR RINGO BQ Intersection 3 Z Difference shell perso ps0 shel RINGOZA RINGOZB RINGOZC Nane Type Conmert Importance Magnetic eld Fe Y Fel Ze Rot x Rot Y Rot Z Cortours Contour color Contour style Contour width BLKHOLE bifference no comment defaut Ho Yes Black Solid Hormal YeRay mode Off Edge Ray mode Off Exploded view Off Exosion factor 1 00 Vishie
8. implying boolean statements from the tree hierarchy with bodies and boolean operators appearing as nodes in the tree The visibility of regions can be toggled with the check box of the corresponding node in the boolean tree The middle mouse button MMB toggles between move object mode and rotate camera mode Minimal coordinate axes are always visible in rotate camera mode and I never use move object mode so I often select View Coordinate axes to turn off the main coordinate axes The scroll wheel zooms the camera and holding the RMB pans the camera Changing the visibility of regions will cause the camera to automatically zoom and pan when Automatic camera is turned on under View Camera settings on by default Ctrl a resets the camera I prefer to darken the working area for the sake of my eyes by navigating to the View Settings window and selecting Background color Black gradient and Background type Full 2 Creating Geometry The following instructions recreate part of the Monster75 01 dat file One could simply open that file and skim this section to see how it was created or the instructions can be followed verbatim using Monster75 01 dat only as a reference Upon opening the file a message will most likely appear indicating that the auto update has been turned off It can be turned back on by selecting it at the bottom of the View menu Select the tab Simplel if it s not already selected This
9. incorporates the basic FLUKA materials for regions Materials can be assigned with the drop down list at the bottom of the right frame when a valid node is selected Material BLACKHOLE is the default thus we have now completely defined the geometry of the blackhole region surrounding our primary geometry For the first generation nodes of the root the name listed in the right frame becomes the name of the region Change the name of the difference node to BLKHOLE Note that regions cannot share the names of materials in a FLUKA input Before continuing we must address reference nodes The use of a boolean tree intro duces a complication that does not exist when one simply lists bodies and declares regions by naming the bodies in boolean algebra statements We ve created bodies as children of a first generation node a region but we d like to use the same bodies in other first generation nodes This is accomplished with reference nodes Any change to a node that is referenced will propagate down to all reference nodes A reference node cannot itself be referenced but any other node can Hide BLKHOLE by un checking the box next to its name in the tree With the root node selected create a difference node and name it TUNNEL We want to use concrete as the material but this is not a default material of FLUKA We will have to modify this later in Flair Expand the BLKHOLE node select twalll hold down Shift drag twall1 to TUNNEL using the
10. is the new project that is started by default when SimpleGeo is launched Select the root node in the boolean tree then create a difference operation node by pressing Ctrl Shift D or by selecting its icon from either the toolbar or the Objects menu The children of a difference node define the volume that is the first child minus all its younger siblings This is the only case where the order of nodes of the same generation has meaning This order can be changed by selecting the node to move then right clicking it and selecting CSG tree tools Reorder nodes Hierarchy changes can be made by selecting a node then dragging it onto the node which will become its parent With the difference node selected create a cylinder Ctrl Shift C and use the right frame to set the following properties e Name blkhole this becomes the name of the body in the input file 21 e X Rel 145 FLUKA uses the unit cm for distance Z Rel 320 e Radius 260 e Height 640 The entry in the Comment field is NOT exported as a comment in the input file Also note that the number of faces of bodies is only for rendering in SimpleGeo See section 8 2 4 of the FLUKA manual for a description of bodies and their parameters Create another cylinder again with the difference node selected e Name twalll e X Rel 145 e Z Rel 300 e Radius 245 e Height 600 First generation children of the root node are interpreted as regions and SimpleGeo
11. 001 0 01 GeV 1e 06 F 1e 07 1e 08 1e 09 F 1e 10 0 001 0 0001 1e 05 1e 06 1e 07 F 1e 08 1e 09 1e 10 T 1 T T f P 0 0001 F 4 0 0001 F a 1e 05 1 fi f j 1 1 1 1e 07 1e 06 1e 05 0 0001 0 001 0 01 0 1 LPB and lambda biasing 7 25 minutes 1e 07 1e 06 1e 05 0 0001 0 001 0 01 0 1 GeV FIG 2 Total neutron fluence out of the target particles per cm per primary as a function of energy GeV for different biasing combinations Statistical errors are plotted in red Computation times are listed in the titles mary particle s energy contributes cascades the number of secondaries tracked will increase exponentially as will the computation time This is an important consideration for any sim ulation not using LPB but especially for those using lambda biasing without LPB because the number photoneutrons tracked are further artificially increased Leading particle biasing significantly increases the statistical errors for most of the spec trum but LPB together with lambda biasing is an effective combination Statistics are improved over the un biased calculation for the upper end of the spectrum while becoming worse only for neutrons below about 5keV The large reduction in computing time makes it clear that the combination of LPB and lambda biasing will be the efficient choice for more complex calculations involving photoneutron spectra A comparison of the MCN
12. 5 e Xmax 105 e NX 100 e Part NEUTRON e Ymin 250 e Ymax 250 e NY 100 e Zmin 310 e Zmax 310 e NZ 124 Press Esc to finish editing the card press Ctrl c and then Ctrl v to create a duplicate of the card Make the following changes e Unit 51 BIN e Name phFluence e Part PHOTON 28 Notice that when the name is changed Flair asks if we want to update anything in the input that has referenced this name This is useful in some cases but nothing references the detector we ve just created Finish editing the card paste the first USRBIN again and make the following changes e Unit 52 BIN e Name nDose e Part DOSE EQ Copy this third USRBIN and paste a fourth making the following changes e Unit 53 BIN e Name phDose These last two USRBINs currently score equivalent dose for all particles so we supplement them with AUXSCORE cards to filter just neutrons and photons Press Insert select Auzscore and enter the following values e Type USRBIN e Part NEUTRON e Set AMB74 e Det nDose The default values of the last two options are fine so leave them blank Set is this card s SDUM Press F1 to see the available sets of conversion coefficients Create a new AUXS CORE or copy and paste this one using PHOTON as the particle and phDose as the detector 2 Physics Many physics options are set in FLUKA by using the DEFAULTS card When the DEFAULTS card is not presen
13. A and MCNPX the fully un biased or analogue functionality is available to the user as well as the methods used to bias these analogue simulations 6 5 III SIMULATION OF AN ELECTRON INDUCED CASCADE IN LEAD A Introduction A basic problem with a minimal geometry is chosen to serve as an initial comparison of FLUKA and MCNPX simulations The FLUKA input is a recreation of one written by Vaclav Kostroun Calculations are performed with and without biasing to see firsthand how different biasing methods affect computation performance of the simulations as well as the statistics FRONT SIDE point of beam incidence DRY AIR 100 cm 6cm FIG 1 Diagram of the simulation geometry B Simulation Setup A 2GeV electron beam is incident on a cylindrical lead target centered coaxially in a cylindrical volume of dry air Fig 1 Leptons and photons of energy less than 6 737MeV are not transported 10 primaries are run per cycle Neutron fluences out of the column of air referred to as the lab and out of the Pb target are calculated and compared with MCNPX calculation data Spatial distributions of neutron fluence and energy deposition are also calculated These are measured in multiple axial and radial intervals and in only one angular interval due to the symmetry of the problem Two methods of biasing are used Leading particle biasing LPB is designed to reduce the mean CPU time per primary t by reducing the
14. PX calculation and the FLUKA calculation using both LPB and lambda biasing is shown in Figure 3 Total neutron fluences particles per cm per primary out of both the lead target and the lab air column are plotted as a function of energy GeV The FLUKA results are plotted in red and the MCNPX results are plotted in black Both codes ran the same number of primaries and the computation time of 7 25 minutes for FLUKA is to be compared with the 27 minute computation time for MCNPX FLUKA e MCNPX 0 001 0 0001 1e 05 1e 06 1e 07 1e 08 Neutrons per square cm per primary 1e 09 1e 10 1e 08 1e 07 1e 06 1e 05 0 0001 0 001 0 01 0 1 1 10 GeV FIG 3 Total neutron fluence particles per cm per primary as a function of energy GeV out of the target and out of the lab as computed by FLUKA and by MCNPX The FLUKA result took 7 25 minutes to compute and the MCNPX result took 27 minutes The two codes give plots which are roughly the same shape while the MCNPX results for neutron energies below 10keV tend to be cleaner It may be possible to improve FLUKA s low energy neutron statistics by using a biasing method called weight windows where particle creation and transport in user defined energy ranges are given preference 6 IV SIMULATION OF THE COLLIMATION OF TOUSCHEK PARTICLES A Introduction Touschek scattering is the collision of b
15. Using FLUKA to Study Radiation Fields in ERL Components Jason E Andrews Department of Physics and Astronomy University of Washington Seattle WA 98195 Dated August 11 2010 Various components of particle accelerators are sources of ionizing radiation These radiation fields must be considered during the design of the Energy Recovery Linac ERL at Cornell University so that the radiation can be shielded against and so that activation in the shielding components is understood To model the fields the use of the Monte Carlo code FLUKA I1 will be explored This will be the first time that FLUKA has been implemented to evaluate the ERL designs at Cornell Therefore instructions on the use of FLUKA and its accompanying software will be created as an appendix to supplement the existing documentation Radiation calculations have previously been performed with the Monte Carlo code MCNPX 2 Several of these problems will be recalculated with FLUKA and the results of the two codes will be compared I INTRODUCTION A Radiation Fields Ionizing radiation poses a hazard to both personnel and to electronics Beta particles being charged are directly ionizing Gamma radiation ionizes indirectly via the photoelectric effect and Compton scattering Free neutrons are also indirectly ionizing there are a number of neutron nucleus reactions which result in radioactive nuclei The degree to which these particles are ionizing depends on the energy o
16. ansport code CERN 2005 10 2005 INFN TC_05 11 SLAC R 773 Statistics and Sampling presentation from the 7th Official FLUKA Course NEA Paris 2008 H Bichsel D E Groom and S R Klein Passage of Particles Through Matter chapter 26 in Review of Particle Physics Particle Data Group 2002 A Xiao and M Borland Phys Rev Special Topics Accelerators and Beams 13 074201 2010 C Theis and K H Buchegger SimpleGeo version 4 2 0 3 V Vlachoudis et al Flair version 0 8 2 R1 100 Appendix FLUKA By Example and Notes or How I Learned to Stop Worrying and Roll the Dice I INTRODUCTION This informal document is intended to function as an introduction by example to using FLUKA and supplementary software and to provide a medium for me to record additional things I ve learned from the fluka discuss archive The work flow illustrated here uses Flair and SimpleGeo Flair is the FLUKA GUI Several important processes such as the imple mentation of user routines and the merging of multiple data runs are automated by Flair If gnuplot is installed Flair can also be used for plotting data though I often find the need to use the gnuplot command line directly SimpleGeo is a 3D modeler and data visualization tool One can export geometries created in SimpleGeo for use with FLUKA as well as import certain data for visualization with the geometry Radiation about the M7 ERL collimator is the project used in the example The following
17. ction 13 1 of the FLUKA manual we must compile the source mod f file and link the resulting source mod o to the FLUKA routines by producing a new FLUKA executable These processes are automated by Flair With source mod f in the working directory of the project press F5 or select Compile under Process in the left frame Click the button in the upper right and select source mod f and then click Open in the pop up Click the Eze field and another pop up will appear Enter M75fluka as the file name and click Save Select source mod f under the File list and then click the Build button Flair will both compile the source routine and produce the new FLUKA executable using the name entered in the Eze field 31 5 Final Details Flair likes to interject the unnecessary STOP card at the end of an input when a project is saved This can effectively break an input if one saves an incomplete project and then later inserts some cards at the end of the input FLUKA will ignore cards that appear after STOP and indeed it is this functionality that makes STOP a useful card for debugging since one can effectively comment off the bottom of an input But we don t presently need the STOP card so either delete it or move it to the very end of the input When STOP is present the last three cards in the input should be RANDOMIZ START and STOP in that order Once this is verified the input is complete Finally place a copy of M75primaries txt in t
18. e extent that complex phenomena such as cascades emerge extemporaneously from the physics of the simulation 6 Il MONTE CARLO CALCULATIONS AND FLUKA Using a Monte Carlo code to perform a calculation can be as simple as defining an appropriate input and running the simulation The quality of such a calculation however is not to be taken for granted We can think of Monte Carlo calculations as characterizing physical systems by exploring momentum space for each given real space region and for each particle species of interest With this in mind a sense of completeness can be established for a Monte Carlo calculation a more complete calculation being one that thoroughly probes the relevant regions of a generalized phase space and thus well characterizes the system at hand The completeness of a calculation is the degree to which the resulting data has converged to the average behavior of the system In practice this convergence is impractically slow There is usually a large variance in the histories of the primary particles requiring a large number of primaries to be tracked before the average behavior emerges Furthermore tracking a single primary can take a significant amount of processing time given the complexity of cascades A useful quantity to consider is thus the computer cost computer cost o t where o is the variance and t is the mean computing time per primary particle 7 If a problem is to be calculated in a prac
19. eam particles with one another and the momen tum transfer of such collisions can lead to particle loss In linear accelerators Touschek scattering is often ignored 9 However in ERL designs the more compact particle bunches cause a higher Touschek scattering rate so this mechanism of particle loss cannot be ignored Rather than allowing the stray particles to collide with the beam pipe they are occluded from the beam by strategically placed collimators such as the type M collimator The collimation of Touschek particles in collimator M7 of the Cornell ERL is simulated The FLUKA input is again a recreation of that originally written by Vaclav Kostroun The probability distribution of Touschek particles shown in Fig 4 is provided by Chris Mayes B Simulation Setup The beam particles are initialized with an energy of 5GeV and their Touschek distri bution is implemented with particle weights Primaries are initialized at random locations within the distribution and the normalized probability of that location becomes the statisti cal weight of the particle The transport cutoff for leptons is set at 4 67MeV and the photon 0 14 m T T T T T T T T T T AE a Probability T 0 06 4 FIG 4 Probability distribution of Touschek particles incident on collimator M7 beam direction y TEES Z 1m 1m X x FIG 5 Geometry used in the simulation input
20. ery LEY ey E e S E ion E T A H L TD 1e 006 ni FEE Q wn Z 9 5 a 1e 007 p Zz F F 1e 008 100 o 100 200 300 400 x cm FIG 13 Neutron fluence particles per cm along the z axis at y z 0 FLUKA is plotted in red and MCNPX in black 13 0 0001 T T T T T T FLUKA MCNPX LI H FE cot 1e 005 4 oy eft H i Ap 7 a lij H i Al n S 2 5 3 i it H H HH HE Laas E jji Munns Ke 5 1e 006 fA iji Het EE 5 ii H rpn H gt E tH D A a 2 1e 007 v z 1e 008 z 100 o 100 200 300 400 x cm FIG 14 Neutron fluence particles per cm along the z axis at y 0 one meter upstream from the back face of the collimator 0 0001 T T T T T T FLUKA MCNPX gt 1e 005 4 a 9 D r a 2 T To v d 5 1e 006 a i 4 it fe a a i A A H Fry i ji iH ET Hipy a oO 5 1e 007 gt w a 1e 008 1 1 100 o 100 200 300 400 x cm FIG 15 Neutron fluence particles pe
21. f FLUKA lectures http fluka course web cern ch FLUKA Course Lectures_pdf has a good intro to FLUKA s scoring cO8_Scoring pdf and the description of FLUKA input sections in the manual are also valuable Many 27 questions and answers are in the mailing list archive so it s a good idea to search the name of a card there to learn more about it This example uses USRBIN and the supplementary AUXSCORE card USRBIN is the FLUKA equivalent of MCNP X s mesh tally It is a grid of bins in dependent of the geometry and it can be used to score energy deposition or fluence The AUXSCORE card can accompany a USRBIN and convert fluence to dose during the run though it can be used to modify scoring other than just USRBIN Score cards output either binary or ASCII files FLUKA comes with many useful routines for processing the binaries and Flair automates this process Flair can also automate the merging of multiple cycles of the same run for the purpose of calculating statistical errors Thus we usually ask FLUKA to output binaries and generate ASCII files afterwords Multiple instances of the same type of score card can output to the same file but I find it convenient to output to separate files when there will not be very many scoring cards Select Scoring under Input in the left frame Press Insert select Usrbin and enter the following values e Unit 50 BIN BIN indicates binary e Name nFluence e Type X Y Z e Xmin 39
22. f the particle and this dependence is different for each type of particle The energies and fluxes of betas gammas and neutrons thus well characterize the radiation field Various doses are obtained from flux using energy dependent conversion tables It is the quantity equivalent dose which characterizes the biological hazard of a radiation field B Sources of Flux In an accelerator environment there are many causes for beam particles to strike the surrounding structure These may be unwanted causes such as beam scraping or deliberate causes such as collimation When a high energy electron or positron is incident on a dense medium an electromagnetic cascade occurs The incident or primary lepton produces high energy photons via bremsstrahlung when the primary is deflected by the electric field of an atomic nucleus When such a high energy photon is incident on a nucleus conservation of momentum allows for pair production to occur The resulting electron and positron can then be energetic enough to lose energy via bremsstrahlung Hence in a dense medium the bremsstrahlung and pair production processes feed one another in a chain reaction produc ing many secondary particles which may escape the medium and constitute a radiation field 3 A neutron flux is also produced by the bremsstrahlung photons of the cascade In addition to pair production a photon with energy greater than the nucleon binding energies may cause nucleons t
23. files _usr 74 sum lis and _usr _tab lis are created the sum file contains a somewhat expository summary of what was scored while the tab file is the raw data formatted for use with software such as gnuplot o When Flair uses gnuplot the commands issued to gnuplot appear in the Flair out put terminal This can be a valuable way to learn gnuplot commands by example even though Flair s commands to gnuplot can be redundant These output terminal commands can also give insight into how the raw data is formatted
24. files are meant to accompany this document e Monster75 01 dat e Monster75 01 inp e Monster75 01 flair e M75primaries txt e source mod f II GENERAL BIBLIOGRAPHY Official FLUKA manual FM pdf http www fluka org fluka php id man_onl The FLUKA manual is a valuable reference and its complete text is integrated into Flair Pressing F1 within Flair opens the FLUKA manual to the entry corresponding to the currently selected card Flair quick start Flair pdf http www fluka org flair doc html As of this writing Flair has advanced beyond some details in this quick start but Flair pdf remains a good introduction to the use of Flair Presentations from recent FLUKA courses http fluka course web cern ch FLUKA Course nea2008 index php id c_program and http fluka course web cern ch FLUKA Course Lectures_pdf These presentations tend to provide a more complete picture of using FLUKA and Flair together Much of the manual is simply reprinted though sometimes in a more expository manner These links change over time so I have archived the contents of the second link above 19 The fluka discuss mailing list archive http www fluka org web_archive earchive new fluka discuss index html This is a very valuable resource once one is familiar with the basics of FLUKA Many common problems and some not so common are documented here usually with usefully described solutions Users are encouraged to post their input files for e
25. for new FLUKA users but the experienced user may come to view this control as an asset Flair the FLUKA GUI 11 does a great deal to organize input creation running the code and the analysis of data and to some extent it automates these processes SimpleGeo greatly facilitates the creation of complex geometries for use with FLUKA inputs Its data visualization capabilities are also useful FLUKA s method of calculating statistical errors involves running multiple cycles of the same calculation Every cycle saves a random number seed to be used to initialize the next cycle if one is run This means that additional cycles can be appended to a data set without recomputing the entire calculation This was taken advantage of for the M7 collimator calculation and saved about 84 minutes of computing time The target application of FLUKA as a multi purpose particle transport code means that some specialized options are not accounted for by the FLUKA developers so such options must be implemented by the user through custom routines However Fortran templates are provided for most of the common routines a user might need For instance the source particle probability distribution of the M7 collimation simulation was implemented with modifications to the source f template included with the FLUKA distribution Flair au tomates the integration of such routines into the FLUKA code VI ACKNOWLEDGMENTS I would like to thank Vaclav Kostroun of Corne
26. he directory of the project and we will be ready to run FLUKA F Running FLUKA Press F6 to display the Run screen Here we can set the number of cycles in a run monitor its progress and cancel a run if needed Notice M75fluka is listed as the FLUKA executable When FLUKA is executed a directory named fluka_ is created where temporary files are stored is just the number of the process and this directory will be deleted upon completion of the run Flair displays the progress of the run and of the cycle and it displays the cycle progress by reading the 00 out file every 15 seconds where is the title of the input and is the number of the cycle The time Flair waits to check this file can be changed in the Tools Preferences window under Project Also notice the Attach timeout listed there Flair will stop checking for updates and list Timed out as the status after this time has passed with no change to the output file Once FLUKA has made some progress we can ask Flair to resume checking on the status by clicking the Attach button If a run needs to be aborted for any reason perhaps a biasing tweak has caused it to become unreasonably languid then one can select Stop Run This creates a rfluka stop file in the luka_ temporary directory The file is empty and FLUKA will check for the existence of it every 50 primary particles or so This means it might take FLUKA a while to finish with the current set
27. hotonuclear characteristics of the problem Such interactions are much less probable than photon interactions with atoms and electrons so running a simulation using 2 5 10 primaries and no biasing would provide very rough and sketchy USRBIN data for the neutron fluences with many bins going completely untested LAM BIAS certainly increases the computing time of the simulation but increasing the number of primaries to achieve the same degree of convergence to the solution would correspond to a much more dramatic increase in computing time Another form of biasing we will implement is called leading particle biasing LPB which is designed to reduce the number of particles tracked in an EM cascade Its use increases the variance in some regions of phase space which I have seen in neutron fluences below 0 1MeV However it drastically reduces the computing time for each primary particle Indeed in this example a cycle of 2 5 10 primary particles would take 66 times longer without it LPB is more thoroughly explained in Note 9 of section 7 2 EMF BIAS in the FLUKA manual LPB is implemented with either the EMFCUT card or the EMF BIAS card EMFCUT allows us to turn on LPB as well as to set energy cutoffs for particle creation and transport EMF BIAS allows us to turn on LPB with more specifications about when and where it is applied but presently we simply want it across the board so we will just use EMFCUT since we would have created
28. imply increase the radius of the RCC card for the blkhole body to R 360 and the debugger will now run without error Also run the collimator debugger for a higher resolution check of this geometry With these verifications complete we can expect to be done with SimpleGeo and can proceed with editing the input in Flair D The Minimal Input Press F3 select the GLOBAL card and change the Input field to Names The mouse behavior can sometimes be problematic so I often Tab through the fields of cards and select pull down options with the arrow keys Notice the brighter highlighted text at the bottom showing when changes are made One can press Esc to stop editing the fields of a card Select GEOBEGIN change the title to Monster75 and change Fmt to COMBNAME Notice that when the Fmt list is selected the right most field of the card is highlighted at the bottom indicating that this is the SDUM Press F1 to find out what using COMBNAME does and notice that the SDUM default for GEOBEGIN is COMBINAT WHAT s and SDUMs left blank often take on such default values Flair introduces explicit REGION cards which are the region declarations in the input It is instructive to examine the REGION cards to see how the regions made in SimpleGeo are expressed purely as boolean statements ASSIGNMA cards assign materials to regions and SimpleGeo creates an ASSIGNMA card for each region For several consecutive regions with the same material one ASSIGNMA ca
29. ional things that are very useful to know about 35 o Multiple projects can be opened simultaneously in Flair by selecting File New Win dow and opening the project there Note that one can copy and paste cards from one project to another o Cards can be temporarily disabled in Flair by right clicking a card and selecting Disable Card s in the pop up menu This is very useful for debugging or for experimenting with biasing o Additional cycles can be run and merged with data from a run that has already completed We might want to do this if uncertainties are too large overall or if the simulation does not appear complete enough To do this in Flair go to the Run Fluka screen set the Previous field to the number of the last completed cycle then set No Cycles to the number of cycles being added Once these cycles complete simply re process the data in the Data Merging screen Note that if USRBIN plots have already been generated then we will have to make whatever change is necessary in the USRBIN Plot screen to cause Flair to re generate the plot dat and plot_geo dat files from the new binary data If we were to make a plot re process the binaries then make another plot without any changes Flair would just use the previously existing ASCII files created when the original plot was made o USRBDX USRTRACK and similar scoring cards should be requested to output bina ries When Flair processes and merges the cycles the ASCII
30. ions 2 3 2 6 and 2 7 of the manual the FLUKA input is an ASCII file with the inp extension Commands are given as single or double line cards each line consisting of a card name followed by six numbered fields named WHAT and a text field named SDUM As will be familiar to one au fait with MCNPX or a similar MC code the FLUKA input essentially consists of physics settings geometry and scoring with the geometry consisting of definitions of primitive geometric bodies followed by boolean statements defining regions with the bodies followed finally by material assignments for the regions Every geometry is surrounded by a finite blackhole region where particles are killed upon entering Every point of the geometry ie within the blackhole shell must belong to exactly one region For simple geometries we can open a template and complete the input in Flair as per 20 the Flair pdf quick start or we could just create the ASCII input directly For the present example we assume that the geometry is complex enough that we want to use SimpleGeo so that we have visual feedback while we create the geometry A SimpleGeo 1 Quick Start and Settings The SimpleGeo interface is divided into three frames The left frame is the boolean algebra tree the center frame is the working area and the right frame displays the details of the currently selected node in the boolean tree SimpleGeo defines regions by
31. l move stray cards like BEAM out of the sequence of geometry cards but I am in the habit of putting beam related cards at the top of the input Set e Beam Energy e E 5 0 FLUKA uses the unit GeV for energy e Part ELECTRON The default values are suitable for the remaining WHAT s The default beam starts at the origin and points in the z direction and we override these defaults with the BEAMPOS card Press Insert select Primary Beampos and set z 70 1 The manual indicates that a beam should not start on the boundary between regions hence the extra millimeter Next select the last ASSTGNMA card press Insert and select General Randomize Press Insert again and select Primary Start Enter 2 5e5 in the field No this is the number of primaries used in the calculation In terms of required cards we now have a complete FLUKA input even though it s not very useful since we haven t requested any scoring The point however is that the example so far has followed a completely general procedure for creating a FLUKA input that starts with SimpleGeo I have simply assumed a particular geometry to illustrate this procedure with E Completing the Input We now continue by adding the more specialized elements of the input What remains is to specify scoring activate physics relevant to that scoring optimize the running time and statistics with biasing and to specify a custom beam profile 1 Scoring The repository o
32. ll University for guiding me through the research teaching me a great deal about nuclear physics providing me with MCNPX input and data and for letting me take my final exams in his office when I first arrived here at Cornell ld also like to thank Georg Hoffstaetter Ivan Bazarov Lora Hine Monica Wesley and everyone else at Cornell involved in this year s CLASSE REU program for making this REU possible and for making it a very great summer Special thanks to Georg Chris Mayes and Don Hartill for the multiple times we went sailing on the lake This research was supported by the National Science Foundation NSF REU grant PHY 0849885 and by 17 the NSF ERL grant DMR 0937466 1 G Battistoni S Muraro P R Sala F Cerutti A Ferrari S Roesler A Fass and J Ranft The FLUKA code Description and benchmarking Proceedings of the Hadronic Shower Simulation Workshop 2006 Fermilab 6 8 September 2006 M Albrow R Raja eds AIP Conference Proceeding 896 31 49 2007 Denise B Pelowitz ed MCNPX User s Manual Version 2 6 0 LA CP 07 1473 April 2008 B Rossi Cosmic Rays 1964 H DeStaebler T M Jenkins and W R Nelson Shielding and Radiation chapter 26 in The Stanford Two Mile Accelerator p 10380 p 1038 Benjamin 1968 J F Briesmeister MCNP A General Monte Carlo N Particle Transport Code Version 4C LA 13709 M 2000 A Fasso A Ferrari J Ranft and P R Sala FLUKA a multi particle tr
33. nd just use unity for the normalization factor in the window that pops up After clicking Ok the data will be processed and another window will pop up asking for min and max data values The min value will have to be increased to something of order 10712 otherwise most of the relevant data will be compressed into a narrow spectrum of red colors Projection planes are turned on by checking the boxes next to the names of the planes Use the scroll bars next to these names to browse the data Multiple data sets can be loaded by creating a copy of the DaVis 3D plugin file VDaVis3D plx in the SimpleGeo directory then loading this re named file as an additional plugin Data can be viewed together with geometry simply by opening a SimpleGeo dat file after loading data Several options that can be useful for visualization are X Ray mode for regions displayed under Vis Attributes in the right frame when the region is selected clipping planes in the View menu as well as a very well hidden option named Enforce overlay This last option can be found by opening the DaVis 3D window right clicking the title bar then selecting Settings from the list that appears X Ray mode can be too obfuscating so whenever I use it I also select View Contouring Render hard contours To save a screenshot for use in a presentation or paper use the usual Windows Ctrl Print Scrn shortcut to copy the screen to the clipboard then paste it into an image editor to crop
34. number of secondaries produced in a cascade This is accomplished by transporting only one of the two particles produced in an electromagnetic interaction involving leptons and photons Which particle is transported is chosen at random with probabilities proportional to the particle energies The weight of the surviving particle is then multiplied by the inverse of its probability of being selected LPB thus greatly simplifies the cascade event with an obvious sacrifice of an increase in variance By using particle energies as selection probabilities the particles presumably most relevant to the calculation are given preference in an attempt to keep the variance increase at a minimum Photonuclear cross sections are much smaller than those for electromagnetic interactions of photons so photoneutron production is relatively rare Lambda biasing is designed to increase the number of photonuclear interactions by reducing the inelastic nuclear mean free path MFP of photons and thereby reduce the computer cost for photoneutron spec trum calculations With lambda biasing secondaries are always created at photonuclear interaction points with weights determined by both the actual creation probability of the secondaries and the factor by which the photon MFP is reduced This increases t but there is not a direct feedback mechanism between secondaries produced and the photons produc ing them so there is not a geometric increase in the number of particles
35. nuplot which should pop up without error if gnuplot is installed Labels can be added to this plot with the Axes Labels section CB stands for color bar and gnuplot commands for these labels can be added in the relevant Opt fields A plot produced in this screen can be saved by clicking in the File field at the top of the screen I often find it necessary to work with gnuplot directly For instance as mentioned above Flair will naively plot the present projection as though the coordinates were right handed Thus I want to apply a factor of 1 to the x coordinates but there is no way to add this with the Flair interface Flair is still of use though because in such cases we can use it to generate the plot dat and plot_geo dat files from the binary data then work with these in gnuplot Of course we could save some time and just choose to live with left handed plots but this method illustrates how we can take advantage of Flair s methods to extract ASCII data from the binaries without working with the large ASCIIs that we send to SimpleGeo To get a view from the top toggle y to be the projection axis in the Projection amp Limits section set the y limits at 0 to 5 and reset the z limits to blank values or the full z domain of course To plot this rotated 90 check the box next to swap and set the Geometry Axes pull down to X Z V VARIOUS NOTES Not everything got integrated into the walk through so here are some addit
36. o be ejected when the photon is incident on a nucleus There are three primary mechanisms by which a nucleus may be stimulated by a photon to emit a neutron in this context referred to as a photoneutron These are the giant nuclear dipole resonance in the 5 15MeV photon energy range the pseudodeuteron reaction in the 50 150 MeV range and the intranuclear cascade at photon energies above 140MeV 4 The relevant flux and energy calculations are thus calculations of the electromagnetic cascade and of photoneutron production C Calculation Methods There are special cases where it may be efficient to analyze the radiation field analytically but the general case of calculating cascades in a realistic environment is prohibitively complex for an analytical solution to be obtained 4 Such complex systems are readily modeled by the Monte Carlo method and it is the use of the Monte Carlo simulation software FLUKA which is presently explored A calculation using the Monte Carlo method probes a physical system by tracking the history of a single primary particle and all generated secondaries using probability density functions to randomly determine the outcome of particle matter interactions This process is repeated for many primaries and from the statistics of this repeated sampling the simulation results converge to the average behavior of the system 5 FLUKA and similar Monte Carlo codes employ fundamental models and enforce conservation laws to th
37. of primaries and stop but it has always eventually stopped for me The temporary directory will remain when FLUKA is stopped in this manner so it must be removed manually if its contents are of no use When running an input for the first time I usually run only one cycle and ignore un certainties in the data so that I can just get a feel for how biasing is affecting the running time and the completeness of the simulation Here we assume that our input is suitable so go ahead and set No Cycles to 7 and click Run It will take FLUKA a few moments to initialize and then the Primaries progress bar should begin to update On the machine I used this run of 7 cycles with 2 5 10 particles each took about 110 minutes For the sake of continuing with the example the number of primaries could be reduced to 2 5 103 and the run should finish in about a minute Either way the run should complete without timing out and the status message Finished OK should appear 32 G Data 1 Output Files Press F7 to view the Output Files screen The Fortran output units as well as the out files for each cycle are listed and one can view one of the output files by double clinking on it One should always examine one of the output files when a new simulation has been run for the first time because the input is echoed there and one can see how it is interpreted by FLUKA The output files contain much useful information in addition to this and they are
38. r cm along the z axis at y 0 one meter downstream from the front face of the collimator 0 0001 T T T T T T T FLUKA MCNPX 1e 005 xt hain hen at teal i 1e 006 Neutrons per square cm per primary 200 nil ji 1e 007 pi 1e 008 300 100 FIG 16 Neutron fluence particles per the collimator z aR 100 200 cm along the z axis at y 0 one meter from the side of 14 Neutron fluence along the z axis at y 0 and x 1 5m is shown in Figure 16 which puts the plot one meter from the side of the collimator the center of the tunnel The comparisons again show FLUKA reporting greater fluence overall as well as greater statistical errors The curves are again similar with FLUKA reporting a few spikes in the downstream plot especially near z 1 2m Photon fluences are plotted at the same locations as with neutron fluences The plots along the x axis are shown in Figures 17 18 and 19 which are at the origin one meter upstream and one meter downstream respectively The plot along the z axis is shown in Figure 20 In some locations where the FLUKA data has converged reasonably well FLUKA again reports greater fluence than MCNPX The FLUKA data is much noisier than the MCNPX data and the uncertainties in the FLUKA data are significantly greater overall 1 T T i FLUKA MCNPX C
39. rd can be used for the whole sequence of regions to clean up the input if one wishes Notice that the ASSIGNMA for TUNNEL is material BLCKHOLE which we now update Flair contains a database of materials and the author of Flair emphasizes that this is only a reference database and that the material properties should be verified Select Material in the left frame to display the database Enter concrete in the Search field and select Concrete portland Either hold the RMB and release it over Insert to Input or double click on Concrete portland causing a pop up to appear Check MAT PROP and then click Selected to insert the material into the input Press F3 to return to the input 26 and notice the COMPOUND MAT PROP and portland concrete MATERIAL cards As stated in the manual a MATERIAL card must be present to use any material that is not listed in section 5 2 of the FLUKA manual Notice that Flair automatically created the MATERIAL card for potassium which is used in the COMPOUND card for the portland concrete Finally select the ASSIGNMAT card which has the Reg TUNNEL entry and select PORTLAND in the Mat list Note that in these pull down lists one can press the first letter of the desired entry to jump to that section of the list We now define the beam Select the GLOBAL card then select Card from the menu bar or press Insert Select Primary Beam and a BEAM card will be inserted after the GLOBAL card When a flair is opened Flair wil
40. region or to no region at all The debugger is explained in the Flair manual F1 under the entry Debug Geometry as well as in the FLUKA manual under the GEOEND card entry Select Debug in the left frame then click the icon in the upper right of the Geometry Debugger frame twice to add two test regions Enter e Name tunnel e Xmin 395 e Xmax 105 e NX 101 this is the number of mesh points in the x direction e Ymin 250 e Ymax 250 e NY 101 e Zmin 305 e Zmax 305 e NZ 123 for one region and e Name collimator e Xmin 55 e Xmax 55 e NX 111 e Ymin 55 e Ymax 55 NY 111 25 e Zmin 95 e Zmax 95 e NZ 191 for the other In some geometries the odd number of mesh points helps to keep the points being sampled from coinciding with region boundaries Select the tunnel region and select the Debug button in the lower right After a few seconds the debugging will stop with errors found Double click the out file to open the standard out so we can see what FLUKA has to say Select ERROR on the left and you can see a large number of point not contained in any region messages together with the coordinates of the point One can usually gain insight into the problem by examining the ranges of these coordinates In this case there is nothing wrong with the geometry we have simply defined our cuboid debugging region such that it contains points outside of our BLKHOLE region S
41. t as in the current input the setting NE W DEFA is assumed NEW DEFA is suitable for many cases but we must turn on photonuclear interactions manually by creating a PHOTONUC card Otherwise our neutron detectors won t register anything Press F3 press Insert select Physics Photonuc and set the following values e Delta resonance on e Quasi D on e Giant Dipole on e Mat BLCKHOLE 29 e to Mat QLASTMAT We could simply set All E on but for a 5GeV electron beam this slows the calculation with no significant effect on statistics The ZASTMAT specifies the last material in the list of materials used effectively selecting the entire list 3 Biasing There are several ways to implement biasing in FLUKA and much time can be spent tailoring the biasing techniques to a given input Ideally one would strike a balance between time spent tweaking the biasing and computing time saved There are a couple of biasing cards that have a significant impact on performance which we shall use in the present example but there are also several others described in the manual as well as in the course materials on line In the Input frame press Insert and select Biasing Lam Bias then enter the following settings e Part PHOTON e X inelastic 0 01 With these settings LAM BIAS reduces the mean free path for inelastic nuclear interactions of photons which effectively gives the simulation many more opportunities to probe the p
42. t face of the collimator FLUKA 0 0001 MCNPX LT gt L E 1e 006 F a TE 5 L bin 2 I 1e 008 U wv li L 5 a 1e 010 F J a H Nn 5 5 1e 012 F 2 a H 1e 014 F 300 100 0 100 200 300 z cm FIG 20 Photon fluence particles per cm along the z axis at y 0 one meter from the side of the collimator 16 V CONCLUSIONS The neutron fluences reported by the two codes agree in Figure 3 but this is calculated from a very basic geometry The calculations with the more complex geometry of the M7 collimator show FLUKA reporting fluences substantially greater than MCNPX but the curves are nearly identical The codes agree qualitatively but there is further investigation to do to determine the nature of the difference in neutron fluences There are several models in use for neutron transport at energies above 100MeV so investigating how each code transports neutrons above this energy would be a good place to start Making calculations with FLUKA requires a substantial familiarity with how the code performs simulations The relevant physics and all biasing is fully user specified so an effec tive FLUKA calculation depends on an effective input The degree of control available to the user means there is a significant learning curve
43. t this would still require that we mirror any right handed external data which could be easily done with a 1 factor in the source mod f routine Anyways press F9 to display the Plot List screen Profiles of plots can be generated automatically by clicking the Scan input file for possible plots button on the right which has an icon that looks like a gear Click this and notice that plot nodes appear for geometry and for each of the USRBIN units in the left frame Select the node corresponding to unit 50 to display the USRBIN Plot screen Enter 1 in the aspect field to set the x and y axes to the same scale In the Projection amp Limits section set the z limits at 0 to 5 Z should already be toggled as the axis to project Leaving the x and y limits blank defaults them to their min and max values The Norm field takes a numerical factor or expression to multiply the data by Notice that the Geometry Use pull down is set to Auto meaning Flair will generate and overlay a geometry plot This will be a cross section of the geometry occurring on the projection axis at the coordinate given in the Pos field When this field is blank the cross section occurs at the middle value of the projection limits 2 5cm in this case Click Plot in the lower right Flair will generate a _usrbin_50_plot dat file and a usrbin_50_plot_geo dat file which are ASCIIs of the projected data and the geometry Flair then sends plot commands for these to g
44. that card for the cutoffs regardless With the Input frame selected press Insert and select Transport Emfcut then enter the following settings 30 e e e 0 00467 e 0 001 e Bremsstrahlung On e Pair Prod On e e ann Q rest On e Compton On e Bhabha amp Moller On e Photo electric On e e ann Oflight On e Reg BLKHOLE to Reg LASTREG The field Type is this card s SDUM and should be left blank for the card to invoke LPB 4 A User Routine for the Beam We want to simulate a beam with an z distribution given by the discrete locations and corresponding probabilities listed in M 75primaries txt This is accomplished with the user routine source mod f A number of user routine templates are included in FLUPRO usrmvaz and it is instructive to compare source mod f to the file source f found there Source mod f will be called every time a primary particle is initialized in the simulation It will choose a line number from M75primaries txt at random initialize the primary at the location in the line number and apply a weight to the particle using the probability of that line number The source routine is called by adding the SOURCE card into our input Insert this card at the top of the input it s located in Primary Source and leave all the options blank Note that BEAM is still a required card In general the source routine simply overrides the specifications of the BEAM card As explained in se
45. tical amount of time then this quantity must be minimized FLUKA is a general purpose particle transport code and consequentially it must be tailored to suit the problem at hand This means that it is up to the user to minimize computer cost because the method of its minimization is subject to the details of a given problem A Biasing in FLUKA Biasing is the means by which computer cost is reduced in FLUKA An un biased FLUKA simulation samples from true distributions of particle matter interactions and particle transport resulting in histories which are meant to accurately represent the histories of real particles However such realistic histories may be dominated by physics which leave out regions of phase space that are of interest relegating these regions to be sampled only in rare events ie a large computer cost for a quantity of interest A biased simulation samples from distributions which are biased either in favor of aforementioned rare events or in some other way which aims to reduce computer cost Every particle that is tracked is given a statistical weight and these weights are adjusted to compensate for whatever biasing is applied In this way biasing does not alter the solution but how the calculation converges to the solution Most biasing techniques reduce one factor of computer cost while increasing the other and it is up to the user to apply biasing such that the net computer cost is reduced With both FLUK
46. to Gl Yes Vishieto AOC Yes Material No item selected Selec an tem for info FIG 11 Visualizing data in SimpleGeo with projection planes 12 FIG 12 Using a surface extrusion to visualize neutron fluence 2 Comparison with MCNPX data The MCNPX calculation took 590 minutes to complete The FLUKA and MCNPX data for neutron fluence along the x axis at y z 0 are shown in Figure 13 along with the statistical errors The FLUKA results are plotted in red and the MCNPX in black The units are in neutrons per cm per primary The two codes give curves of roughly the same shape but FLUKA reports a significantly greater fluence The statistical errors reported by FLUKA are also greater than those of MCNPX Neutron fluence along the x axis at y 0 is shown at z 1 9m in Figure 14 and at z 1 9m in Figure 15 These coordinates give plots at one meter upstream from the back collimator face and at one meter downstream from the front collimator face respectively 0 001 p T T T T T FLUKA F of MCNPX S A s S 0 0001 a 5 amp j l E Mhp inh ee T ie rt 1e 005 miir ie alem pa _ g a cat gi Li ii Hip Hi at cat he h 4 Pe
47. to view the collimator regions Take note of RINGO6A and the body taper01 therein One must increase the value of View Settings Nr of property digits for greater accuracy of Radius Low and Radius High to be exported Note that these radii cannot correspond the the radii of the infinite cylinders they intersect because of ambiguity this would create in the boolean statements involving taper01 the cylinders and cutting planes Observing the hierarchy of RINGO6A B and BEAMAIRB may be instructive as well As another detail examine shell08 in RINGO7A Try changing the number of faces to see the display anomalies on the tapering part of RINGO7A that sometimes occur in SimpleGeo As evidenced by shell08 changing the number of faces can fix the display However such anomalies are certainly not features of the geometry exported to FLUKA so there is no harm in ignoring them We are ready to export the SimpleGeo dat model as a FLUKA input Select File Export Preferences and check Export material names This is not necessary but it makes the input more readable Export it using a suitable name by selecting File Export Fluka and under Save as type select the parentheses input option This creates an incomplete ASCII input for FLUKA which we shall complete using Flair B Flair Quick Start The left frame of Flair is used to select what frame is displayed on the right The root node Fluka should already be displayed upon launching Flair
48. tracked as there is in a cascade Therefore the reduction of t by LPB is much greater than the increase of t by lambda biasing Where lambda biasing is implemented in the present calculations the factor used to reduce the photon MFP is 0 02 C Results Neutron fluences particles per cm per primary as a function of energy GeV for dif ferent biasing combinations are shown in Figure 2 The quantity plotted is the total fluence out of the lead target Statistical errors are plotted in red and computation times are listed in the titles of the individual plots Lambda biasing alone gives exceptional neutron data but the computation time is as should be expected the greatest Individual cycles averaged a computation time of 47 8 minutes but some random number seeds gave cycles that did not complete even after 14 hours Furthermore the target dimensions are small considering the radiation length and Moli re radius of lead 8 In calculations where a more substantial fraction of the pri No biasing 155 minutes Lambda biasing 239 minutes 0 001 0 001 1e 05 1e 06 F if work ort 1e 08 Neutrons per square cm per primary 1e 09 F 1e 10 ni f L f f 1 1e 07 1e 06 1e 05 0 0001 0 001 0 01 Leading particle biasing 6 69 minutes 0 001 F 0 0001 1e 05 f 1e 06 F 1e 07 1e 08 F Neutrons per square cm per primary 1e 09 F 1e 10 f 1 1 1 1 1 1e 07 1e 06 1e 05 0 0001 0
49. uts off the ends of the region which is necessary because these points were already part of the region BLKHOLE and points must belong to only one region We can skip this action of cutting off extra volumes by using infinite cylinders and cut ting planes to create cylindrical regions instead of using finite cylindrical bodies With TNLAIR selected create an intersection operation Ctrl Shift I The intersection op erator defines the volume of the points common to all children of the operator With this node selected create a difference node and with that selected create a cylinder e Name shell01 e Radius 50 e Infinite Z axis and a plane e Name pzn90 e Z Rel 90 Pzn90 should be the younger sibling of shell01 With the intersection selected create a plane 23 e Name pz90 e Z Rel 90 This coordinate based naming convention makes it easier to locate bodies which will be referenced often such as cutting planes The rest of the geometry is created primarily by the methods just described It should be noted that there are different techniques which achieve the same geometry and that this has only been an illustration of one possibility Open Monster75 01 dat if it is not already open to continue with the example In the tree observe the large fraction of reference nodes used illustrating the utility of a good naming convention for bodies With the auto camera on hiding TUNNEL and TNLAIR makes it easier
50. well explained by c09 _Output pdf in the repository of FLUKA lectures http fluka course web cern ch FLUKA Course Lectures_pdf 2 Data Post Processing and Merging The binaries created by the scoring cards need to be processed by the various programs that ship with FLUKA or by user routines The appropriate FLUKA programs are usually described in the FLUKA manual entry for the score card in question and these programs are found in FLUPRO flutil It s important to understand what those do to the data For instance the USRBDX card can be used to score fluence and one can request only one angular bin In such a case the data scored is still differential in angle so each bin must be multiplied by 27 to get the integrated data However the program usrsuw used to merge multiple USRBDX binaries and compute standard deviations performs this multiplication thus one must not include an extra factor of 27 when plotting the data The point is that one must keep post processing in mind when learning about how the various scoring cards record data Flair automates post processing and merging Press F8 to display the Data Merging screen The scoring units are listed in the Usrxzx frame Select all four of them then click Process in the lower right Flair will run the appropriate program s which is usbsuw in this case We can work with the USRBIN data in Flair or in SimpleGeo 3 USRBIN Data in SimpleGeo Once this has finished press F7 to go
51. xamination so this also serves as a resource for seeing how others build their sometimes broken inputs The SimpleGeo manual http theis web cern ch theis simplegeo manual manual2 htm This is the only documentation of SimpleGeo and its plugins that I am aware of so I have made special effort to clearly describe SimpleGeo here II INSTALLATION FLUKA runs only on Unix while SimpleGeo runs only on Windows Flair is built to run on Unix and Windows though I have only tested it on Linux Installation of FLUKA is fully automated by downloading and executing the current rpm file on the FLUKA website www fluka org For distros that are not RPM based Installation pdf from the current course materials provides the most thorough description of the install process http fluka course web cern ch FLUKA Course nea2008 Lectures_pdf Installation pdf An RPM is also available for Flair at http www fluka org flair download html or for a manual install a brief description is given in the README contained in the tar package for Flair SimpleGeo is a straightforward Windows install but the plugins need to be manually downloaded and unzipped into the SimpleGeo install directory Both SimpleGeo and the plugins are available at http theis web cern ch theis simplegeo IV WALKTHROUGH FLUKA Flair and SimpleGeo all create additional files as they run so it s a good idea to create a dedicated directory for each project As described in sect
Download Pdf Manuals
Related Search