User's Guide to fete: From ENDF/B-VI To ENDL
Contents
1. Tz B7 as Eem 0 which is an integrable singularity The Jaco bian also has a zero for Ecm Eo and Hem 1 This is because the corresponding energy in the laboratory frame is Fia 0 and the direction is arbitrary 1 lt Hiab lt 1 APPENDIX C LLNL DISCLAIMER This document was prepared as an account of work sponsored by an agency of the United States Govern ment Neither the United States Government nor the University of California nor any of their employees makes any warranty express or implied or assumes any legal li 1 B Beck T Hill D McNabb ASCII format specifica tions for the Evaluated Nuclear Data Libraries ENDL in preparation R J Howerton R E Dye P C Giles J R Kimlinger S T Perkins and E F Plechaty Omega A Cray 1 executive code for LLNL nuclear data libraries Lawrence Livermore National Laboratory Re port UCRL 50400 Vol 25 1983 2 Evaluated Nuclear Data File http www nndc bnl gov nndc endf 3 Cross Section Evaluation Working Group ENDF 102 Data Formats and Procedures for the Evaluated Nu clear Data File ENDF 6 Ed V McLane C L Dun ford P F Rose Brookhaven National Laboratory Report BNL NCS 44945 1997 4 fete Code Reference part of the fete code release in preparation 5 LLNL Computational Nuclear Physics Group page http nuclear 11nl gov 6 G W Hedstrom An explanation of the ENDEP code Lawrence Livermore Nat2. 2 Resonance Data 3 MF 3 Cross Section Data Data on massive particles ordered by MF number 1 MF 4 Outgoing Angular Distributions 2 MF 5 Outgoing Energy Distributions 3 MF 6 Outgoing Energy Angle Distributions Gamma data 1 MF 12 Multiplicity and MF 13 Cross Section Data 2 MF 14 Angular Distributions 3 MF 15 Energy Distributions Delayed neutrons IV Conclusions m CDIOTOTWNWNNNNNM WY NI NNN O AAAADAD D CO CO CO 10 10 10 10 Acknowledgments 11 A Preprocessing ENDF B VI Data 11 1 Split the ENDF data 11 2 Official IAEA Preprocessing codes 11 3 Preprocessing code linear 11 4 Preprocessing code recent 12 5 Preprocessing code sigma1 12 B Double differential data in ENDF B VI 12 1 The domain in the laboratory frame 13 2 Center of mass probability density 13 3 The Jacobian 13 C LLNL Disclaimer 14 References 14 I INTRODUCTION This guide is a description of the LLNL Computational Nuclear Physics CNP Groups s From ENDF B VI To ENDL a k a fete code fete translates nuclear re action data from the ENDF B VI to Livermore s inter nal ENDL 1 format ENDF B VI is both an evaluated nuclear data library 2 and an international standard data format 3 for evaluated nuclear data Because of the significant difference in both format and content the translation process is a complicated multi step process All of ENDL data is presented in point wise l
3. ad 23 z n3a n a 452 z f D total prompt plus delayed fission 15 24 z 2na 33 454 z f Independent fission product yields 15 25 z 3na 16 455 z f D for delayed fission 15 27 n abs n a 456 z f D for prompt fission 15 28 z np 20 214 457 z f Radioactive decay data n a 29 z n2a 27 458 z f Energy release in fission for incident n s 15 30 z 2n2a n a 459 z f Cumulative fission product yields 15 32 z nd 22 500 Total charged particle stopping power n a 33 z nt 24 501 Total photon interaction o 70 34 z n He 25 502 Photon coherent scattering 71 35 z nd2a n a 504 Photon incoherent scattering 72 36 z nt2a n a 505 Imaginary scattering factor n a 37 z 4n 4 506 Real scattering factor n a 38 n 3nf 4 chance fission n a 515 Pair production electron field n a 41 z 2np 29 516 Pair production 74 42 z 3np 6 517 Pair production nuclear field n a 44 n n2p 7 522 Photoelectric absorption 73 45 n npa 34 534 572 Various subshell photoelectric o s n a 50 z n 0 excitation of g s 11 9 600 z p 0 40 51 90 z n n excitation of nt excited state 19 601 648 z p n 40 91 z n c excitation to continuum 1 9 649 z p c 40 01 n disappearance n a 650 z d 0 41 02 z 7 46 651 698 z d n 41 03 z p 40 699 z d c 41 04 z d 41 700 z t 0 42 05 z t 42 701 748 z t n 42 06 z He 44 749 z t c 42 07 z c 45 750 z He 0 44 08 z 2a 3
4. database will appear as my_db yi01 za001001 endf orig endf prepro za001002 endf orig endf prepro 4 The mini db utility The mini db utility is a simple script that copies out a subset of targets from one database to another This is extremely useful for troubleshooting problem nuclides The utility is invoked with mini db lt isotope file gt lt output db gt master copy Here master copy is an optional argument and de faults to NDSPATH databases ENDFB6 prepro The required argument lt isotope file gt is a plain text file containing a list of za numbers The other argument lt output db gt is the uppermost directory of the new miniature database you are creating A subdirectory of lt output db gt called yi01 containing the za files will be created by the script C Translating the database amp computing energy deposition Translation is very simple go to the yi01 direc tory containing all of the za subdirectories and execute translate_all Provided everything is working you will have a fully translated database in those za directories complete with energy depositions Your database direc tory should look something like this my_db yi01 za001001 endf orig endf prepro i endep_csh log i endep log translate_csh log u translation log w yo00c01i000s000 yo00c10i000s000 za001002 endf orig i endf prepro The next few subsections explain the translate and translate_al
5. gammas in the individual reactions Otherwise we con vert the discrete lines to a continuum distribution with approximate 6 functions at the lines This scheme also handles the case where an ENDF B VI evaluator repre sents continuum gamma distributions as a set of discrete lines with frequencies dependent upon the energy of the incident neutron We comment that the ENDL unit base interpolation never made sense for gamma distributions represented as approximate d functions and fortunately there were are few targets in ENDL99 which have such data It does make sense to interpolate equally probable bins for this data and that is how the translation code is written 2 MF 14 Angular Distributions The MF 14 files contain information on angular dis tributions for gammas These are usually isotropic and in ENDL the convention is that a distribution is isotropic if no angular information is given There are a few reac tions in ENDF B VI for which the MF 14 data is Leg endre coefficients We convert this data to ENDL I 4 format with Legendre expansions 3 MF 15 Energy Distributions The MF 15 files are for energy distributions of con tinuum gammas The corresponding continuum gamma multiplicity is given as the last table in the MF 12 file or the gamma production cross section is the last table in the MF 13 file D Delayed neutrons For delayed fission neutrons in ENDF B VI some of the multiplicity data is in the MF 1 MT 455 f
6. output not just the reconstructed MF 3 cross sections e g an gular and energy distributions are also included The default input parameter file is called RECENT INP and sets the output tolerance to 0 1 for the entire incident neutron energy range 5 Preprocessing code sigmai The purpose of this program is to Doppler broaden neutron induced cross sections Each section of cross sec tions MF 3 is read from the ENDF B VI format The data is Doppler broadened thinned and output in the ENDF B VI format All cross sections that are used by this program must be tabulated and linearly interpolable in energy and cross section In the unresolved resonance 12 region it is not possible to exactly define the energy de pendence of the cross sections The average widths and spacings given in ENDF B VI are only adequate to de fine average values of the cross sections Therefore all cross sections in the ENDF B VI format for the unre solved region are really average values which cannot be Doppler broadened using the sigma1 method which re quires tabulated linearly interpolable energy dependent cross sections Therefore e All tabulated points within the unresolved res onance region are copied without modification or broadening Adoption of this convention al lows subsequent programs to properly define self shielded Doppler broadened cross sections in the unresolved resonance region e Cross sections are extended as 1 v belo
7. 0 all messages reported if lt 1 Info and Unimplemented messages are not reported but Warning and SevereError messages are reported if gt 1 only SevereError messages are reported endl_datafield_precision 7 Precision saved in ENDL files max_energy 20 0 Maximum energy for ENDL files MeV max_list_size 500 Maximum length of a 1 d list For cross section data this has units of barns for multiplicities and probabilities this has no units gt Esnift F Ethresh 1 F mf5 shift The accompanying y data is lumped into the C 55 file TABLE IV Default parameters for the fete code The parameter names are not case sensitive A General information on the reaction data in ENDF B VI 1 MF 1 Documentation The MF 1 information describes the contents of the file and aside from the MT 455 file we don t translate it The treatment of delayed fission neutron information is detailed in Subsection II D 2 MF 2 Resonance Data The MF 2 data gives the resonance parameters We preprocess this data with Red Cullen s recent code to convert this information to tabular cross sections 10 This table is combined with the MF 3 cross section data to form the actual cross section for the reaction 3 MF 8 Cross Section Data The MF 3 file gives tabular cross section data As a second preprocessing step we use Red Cullen s sigma1 code to Doppler broaden the data to room temperature The corresponding ENDL file is flagged b
8. 7 751 798 z He n 44 09 z 3a n a 799 z He c 44 11 z 2p 18 800 z a 0 45 12 z pa 48 801 848 z a n 45t 13 z t2a 42 849 z a c 45t 14 z d2a n a 851 Lumped reaction covariances n a stored with individual reactions T he C 55 file Only total fission data is stored in Purely a derived file so it is not needed ENDL Below Eine 20 MeV only Li has this reaction and Li n 3np a Li n 3na p ENDF B VI does not distinguish whether the proton or neutron is emitted first Level excitation functions stored in S 1 type files in ENDL f MF 12 15 then the outgoing y s are lumped in the C 55 file otherwise the data is discarded since it should be he non y data is discarded since it is purely derived data If the sum_inelastic option is set then use the MT 4 data otherwise use the MT 51 91 data lumping all y data into h Below Eine 20 MeV only 10B has this reaction and Bin t 8Be t 2a A slot exists in ENDL but this data is not translated ENDL does not yet have a slot for documentation k Delayed fission information is stored in S 7 type files in ENDL TABLE II ENDFB VI MT numbers and their rough ENDL equivalent Preprocessed Database Name CVS Project Name version in CVS ENDF B V ENDF _tapes ENDF B V N ENDF B VI ENDF_tapes ENDF B VI N JEFF 3 0 JEFF 3 0 Y JENDL 3 3 JENDL 3 3 Y TABLE III Databases in ndg 11nl gov cvsroot grees Kelvin so care must be take to downlo
9. NDF B VI data is closely related For a given reaction an MF 12 file gives the multiplicity of emitted gammas while an MF 13 file gives the product of the multiplicity with the reaction cross section the gamma production cross section In ENDL the C 55 file with I 0 gives the gamma produc tion cross section while for individual reactions only the gamma multiplicity may be given an I 9 file Thus if ENDF B VI has MF 13 data for a given reaction we have to divide by the cross section to produce and ENDL I 9 file while if ENDF B VI has MF 12 for a combina tion of reactions we have to multiply by the cross section to produce an ENDL C 55 file There are two wrinkles we encounter when process ing the MF 12 and MF 13 data First the MF 12 or MF 13 files may give detailed information on the y decay chains For the translation to ENDL we have to accumulate this information into files of energy distri butions and multiplicities Second ENDL s mechanism for handling discrete and continuum gammas S 3 files requires more detail than is present in the ENDF B VI data files The problem with the ENDL S 3 option is that it requires those channels with continuum gammas and those with different discrete excitation levels each be treated as separate reactions having their own cross section and outgoing energy distributions To deal with this problem we introduced a split_gammas option If set we split the discrete lines out from the continuum
10. P u P E u where P is the distribution in the I file The individual P and P files are normalized so that de P3 E dE 1 for each value of u and if P ys du 1 Again we use the convention that P 0 and P 0 outside the domain of validity of the data C Gamma data In ENDF B VI the information on emitted gammas is given in files with MF numbers 12 through 15 This data has been the most difficult to put into ENDL format because the ENDL format is more constraining than the ENDF B VI format In ENDL it was customary to lump the gamma infor mation into two categories n y and n Xy the C 46 and C 55 files In fact there are only 7 isotopes in ENDL that have gamma files other than C 55 and C 46 and all of them are light isotopes The majority of these isotopes have gamma data only for inelastic scattering In ENDF B VI1 it is customary to leave the gamma data with the associated reaction We have chosen not to com bine all of the ENDF B VI gamma data into C 55 files but to keep it with the reactions For some targets and certain combinations of reactions ENDF B VI does have total gamma data For gammas from inelastic scattering the combined ENDF B VI data is given in MT 4 files ENDF B VI may also combine gamma data from any reactions having MT 3 files In our translation we join such gamma data and write it to C 55 ENDL files 1 MF 12 Multiplicity and MF 13 Cross Section Data The MF 12 and MF 13 E
11. PATH Src PYTHONPATH set path NDSPATH bin path B Getting and preparing the database 1 Getting a database from the web There are several modern databases that use the ENDF B VI format ENDF B VI JENDL 3 3 JEFF 3 0 BROND 2 2 and CENDL 2 1 ENDF B VI 2 JENDL 3 3 8 and JEFF 3 0 9 are the most up to date and are the main targets of our translation effort All three of these databases are available on the web both in raw and point wise i e preprocessed form Since the preprocessing stage can take quite a bit of time several hours to a day or more depending on your hardware we recommend downloading the preprocessed data ENDL requires that the data be Doppler broadened to 300 de o e d e f g MT Description Q MT Description C 1 n total 1 115 z pd 19 2 z elastic 10 116 z pt 39 3 z non elastic 55 151 n resonance n a 4 z n 11 201 z Xn 56 5 z anything n a 202 z Xy 55 10 z continuum n a 203 z Xp 50 11 n 2nd 32 204 z Xd 51 16 z 2n 12 205 z Xt 52 17 z 3n 13 206 z X He 53 18 z f 15 207 z Xa 54t 19 n f 1 chance fission n a 208 218 Various meson and antiparticle production o s n a 20 n nf 2 chance fission n a 251 253 Various elastic neutrons scattering parameters n a 21 n 2nf 3 chance fission n a 301 Energy release for total and partial o s n a 22 z na 26 451 Heading or title information MF 1 only n
12. UCRL SM 206606 User s Guide to fete From ENDF B VI To ENDL David Brown and Gerald Hedstrom Lawrence Livermore National Laboratory Livermore CA 94550 USA Tony Hill Los Alamos National Laboratory Los Alamos NM USA Dated September 17 2004 This guide describes how to run the fete ENDF B VI to ENDL translation code and outlines some of the general features of the translation process In particular this guide details how to install the code and supporting scripts how to prepare an ENDF B VI formatted database for translation how to translate the database and how to check the translated database This guide also explains how fete treats each ENDF B VI reaction and data type Contents I Introduction II Steps in Translating IIT A Installing the translation code amp scripts 1 Getting the translation code 2 Building the codes 3 The bdf1s file 4 Setting up the environment Getting and preparing the database 1 Getting a database from the web 2 Getting a database from the CVS 3 Database preparation 4 The mini db utility Translating the database amp computing energy deposition 1 translate vs translate_all 2 Invoking fete from the command line Checking the translation 1 The build_errlist py utility 2 The view_all py utility 3 The za_comp and yo_comp utilities Translation Details A D General information on the reaction data in ENDF B VI 1 MF 1 Documentation 2 MF
13. ad the cor rect database 2 Getting a database from the CVS Locally we store preprocessed and unprocessed data in our CVS repository At the time of this writing the repository is located at ndg 11nl gov cvsroot Ta ble III lists the databases and whether a preprocessed version of the data is checked into our CVS 3 Database preparation The ENDF B VI format specification makes no re quirement as to how the data is arranged in the library so each nuclear data library is shipped in a different form In the case of the ENDF B VI database all of the data is lumped into tapes hence the need for splitting the database with the endfsplit code discussed in Appendix A1 and then preprocessing JEFF 3 0 and JENDL 3 3 require unpacking and directory renaming In the end we want the data in something like the ENDL standard directory layout 1 my_db yi01 za001001 endf orig za001002 endf orig Here the endf orig file is the raw unprocessed evalua tion for one nuclide Once this is done we must preprocess each orig file to produce a prepro file The prepro file differs from the original file in that e All resonances are expanded into a linear linear in terpolable form e All cross sections are heated to 300 K This is mainly an automated process using the prepro2000 csh script which controls Red Cullen s linear recent and sigmai codes 10 This step is de tailed in Appendix A 2 After this step the
14. al ENDF B VI data to ENDL format difficult is the interpolation with respect to the energy of the incident neutron The usual prescription in ENDF B V1 is to derive such energy distributions by linear interpolation of the values of Some ENDL processing codes e g ndfgen code 11 use unit base interpolation while others e g mcapm 12 use equally probable energy bins a mathematically similar procedure We have found that the unit base interpolation scheme is usually not compatible with the linear interpolation of the function parameters e g used by ENDF B VI We therefore chose to ensure that linear interpolation of the equally probable bin boundaries in mcapm is consis tent with linear interpolation of the function parameters in ENDF B VI Another complication of translating ENDF B VI en ergy distributions comes from the ENDF option of rep resenting the data as a weighted average of some number of functional formulas p wifi E i 1 For example for an n 2n reaction with different for mulas for the energy distributions of the two outgoing neutrons So far this option is implemented only for models in which the functions f E have the same func tional form such as the evaporation model with different effective temperatures It is not yet possible to handle such cases as when f is an evaporation model and fo E is Watt model 3 MF 6 Outgoing Energy Angle Distributions The MF 6 files con
15. epro2000 Inside the prepro2000 directory are machine specific BAT files to build simply execute the appropriate one Currently only the SUN BAT and LINUX BAT systems has been tested by the group Once the executables are built you can launch the prepro2000 script The script re quires one argument the directory containing the split ENDF data prepro2000 lt split db gt The script launches three codes sequentially for each evaluation file linear recent and sigmai There are three input files used by these codes called LINEAR INP RECENT INP and SIGMA1 INP The explicit directions and descriptions for these codes are in the DOCUMENT direc tory The default input files have been set to linearize the data to a tolerance of 0 1 expand resonance data to 0 1 tolerance and then heated to 300 degrees Kelvin with a final tolerance of 0 5 The heating is applied only to the cross sections not to the angular distributions or energy distributions The final preprocessed output file for each isotope is renamed to endf prepro and these are the files read in by the fete There are several common features to the preprocess ing codes These programs only use the ENDF B VI for mat and can handle data in any version of the ENDF B VI format They assume that the data is correctly coded in the ENDF B VI format and perform no error check ing In particular they assume that the redundant in formation on each line is correct Sequence number
16. for reaction cross sections n a 34 Data covariances for angular distributions n a 35 Data covariances for energy distributions n a 39 Data covariances for radionuclide production yields n a 40 Data covariances for radionuclide production cross sections n a aMT 451 delayed neutron data is handled separately Translated into pointwise data using Red Cullen s recent code Not implemented in ENDL yet 4Use ACTL for this data type Only if not isotropic if isotropic this file is ignored since the ENDL default is isotropic TABLE I ENDFB VI MF numbers and their rough ENDL equivalent Information on where and how to obtain these codes can be found on the Computational Nuclear Physics Group s web page 5 Getting these codes installed and working is your own responsibility however we provide details on the splitting and preprocessing in Appendix A The various control scripts require that you set sev eral environment variables BDFLSPATH ENDLPATH FUDGEPATH add the path to the fudge source to your PYTHONPATH and add the path to all of the executables produced above to your PATH If you are using csh you may accomplish all of this by adding the following lines appropriately modified of course to your cshrc file Nuclear Data Processing System setup setenv NDSPATH home joeuser NDS setenv BDFLSPATH NDSPATH bdf1s setenv ENDLPATH NDSPATH databases setenv FUDGEPATH NDSPATH fudge setenv PYTHONPATH FUDGE
17. ile and some in the MF 5 MT 455 file so we need to handle them together The MF 1 MT 455 file starts with a list of the time constants for delayed fission neutrons In ENDL format the delayed neutron data is identified by C 15 fission along with S 7 delayed neutrons The data corre sponding to the different time constants is put into dif ferent ENDL blocks with the time constant located in the x 1 slot This is an extension of the ENDL format which originally had only one variety of delayed fission neutron For C 15 and S 7 we usually produce an I 7 de layed neutron multiplicity file with separate blocks cor responding to the different time constants Note that some fissionable targets in ENDF B VI lack this data ENDF B VI may also give energy distributions for de layed fission neutrons In that case we write a corre sponding C 15 S 7 I 4 ENDL energy distribution file IV CONCLUSIONS Throughout this manual we have dealt with the de tailed differences between ENDF B VI and ENDL As a result of these differences we have had to add several new reactions to ENDL and these are summarized in Ta ble V Interestingly there is one case where ENDL has more information than ENF B VI In ENDL n np and 10 C Reaction 16 n 3n p 17 n n 2p 18 n 2p 19 n p TABLE V Reactions added to ENDL n pn reactions are treated as separate reactions if en ergy distributions are given only for the outgoing neutr
18. inearly interpolated tables while ENDF B VI data is presented in a variety of manners depending on the reaction and quantity in question The two formats also differ in the arrangement of the data in the databases Data in the ENDF B VI format are divided into materials abbre viated MAT which are further divided into Material Files abbreviated MF An MF is one type of reac tion data e g MF 3 is tabular cross section data These MF s are subdivided into Material Tapes ab breviated MT which contain the data for individual re actions Thus MT 16 MF 3 contains n 2n reaction cross section data while MT 18 MF 3 contains the total n f reaction cross section By contrast ENDL requires that reactions and nuclei be sorted into nesting directo ries first according to projectile e g yi01 means incident neutron then nuclide e g za001002 means Z 1 and A 2 so we have deuterium then reaction The data for each reaction has several numbers associated with it namely yo to denote the outgoing particles C to denote reaction type and I to denote reaction property The ENDL C and I numbers are analogous to the ENDF B VI MF and MT numbers respectively However we cau tion that the mapping from MF MT to C I is not this straightforward as we see in Tables I and II In this note we will first cover the installation and use of the translation codes and supporting scripts Second we will detail the translation process a
19. interpolable form add in any linearly interpolable background cross section and output the result in the ENDF B VI format The cross sections output by this program will be linearly interpolable over the entire en ergy range The resonance contribution is calculated for total MT 1 elastic MT 2 capture MT 102 and fission MT 18 added to the background if any and output In addition if there is a first chance fission MT 19 background present the resonance contribution of fission will be added to the background and output If there is no first chance fission background present the program will not output MT 19 The formats and conventions for reading and interpret ing the data varies from one version of ENDF B to the next However if the section MF 1 MT 451 describ ing the data is present it is possible for this program to distinguish between data in the ENDF B IV V and VI formats and to use the appropriate conventions for each ENDF B version see subroutine file1 for a descrip tion of how this is done If the data description section is not present the program will assume the data is in the ENDF B VI format and use all conventions appropriate to ENDF B V Users are encouraged to insure that the data description section MF 1 MT 451 is present in all evaluations All energies are read in double precision by special FORTRAN I O routines and are treated in double pre cision in all calculations Entire evaluations are
20. ional Laboratory Report UCRL ID 151609 1999 14 ability or responsibility for the accuracy completeness or usefulness of any information apparatus product or process disclosed or represents that its use would not infringe privately owned rights Reference herein to any specific commercial product process or service by trade name trademark manufacturer or otherwise does not necessarily constitute or imply its endorsement recom mendation or favoring by the United States Government or the University of California The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or the Uni versity of California and shall not be used for advertising or product endorsement purposes 7 B Beck For UpDating ENDL fudge manual in preparation 8 Japanese Evaluated Nuclear Data Library http wwwndc tokai jaeri go jp jend1 j33 j33 html 9 Joint European File http www nea fr html dbdata data evaluated htm 10 D Cullen prepro2000 Computer Code available from http www nds iaea org ndspub endf prepro 11 Beck B LLNL s Deterministic Transport Access Rou tines and Data Documentation for the Nuclear Data Files ndf and the libndf a access routines Lawrence Livermore National Laboratory Report UCRL MA 147647 2002 12 Hedstrom G Cox L and Perkins S The Content and Structure of MCF Files Lawrence Livermore Na
21. l scripts and how to invoke the fete code directly The translation may be tweaked using the fete_options inp file and this mentioned below and is discussed in more detail in Section III 1 translate vs translate_all The translate_all loops through a set of nuclei and translates them all using translate It may be invoked using translate_all directory where directory is an optional argument that speci fies the directory containing the za subdirectories If this argument is missing the script defaults to processing the current directory By contrast translate script automates the transla tion of a single nuclide and is invoked via translate lt za number gt This script first invokes fete to translate the nuclide s data into ENDL format then it sets up and invokes the endep csh script The endep csh is the piece of code that actually calls endep to compute energy depositions 2 Invoking fete from the command line Although we usually invoke fete through either the translate or translate_all scripts you may wish to invoke fete directly on the command line via fete f options_file options fete expects that you invoke it in a za subdirec tory and that an endf prepro file exists in this direc tory There are various options that control how the translation proceeds and they can be overridden in the fete_options inp file These options are described in Section III and listed in Table IV The f option allows you
22. lished with the endfsplit code It is currently in ndg 1lnl gov cvsroot endfsplit The endfsplit project contains two FORTRAN files a shell script and a Makefile Typing make will build the two executables summary and endfsplit needed by the splitter script This splitter script requires two arguments when launched splitter lt original db gt lt final db gt The first argument is the name of the directory con taining the ENDF B VI tapes the second argument is the name of the new directory to contain the new file structure The script first runs the summary code which reads in the ENDF summary file and produces a new file neutron tapes which contains the names of the tapes that have incident neutron reaction data in them The script then reads in the neutron tapes file and runs the endfsplit code on each of the tapes listed The endfsplit code breaks the data tapes into single isotope files and places those files in directories named by their za number The endfsplit code will create the new di rectories if necessary and overwrite without hesitation insuring the desired evaluation precedence For other databases such as JEFF or JENDL you may have to split and arrange the databases yourself as the 11 splitter code will not work with them 2 Official IAEA Preprocessing codes The next step in the processing requires the use of official IAEA codes We store them locally in cvs as prepro2000 10 ndg 11nl gov cvsroot pr
23. n code fete is available either as a tarball on Liver more s CNP Group web page 5 or internally on the CNP Group s CVS The CVS is located at ndg 1llnl gov cvsroot and this project s name is fete The tarball unpacks to a directory called fete 2 Building the codes To build the package enter the main project directory fete and type make If compilation went well an executable file called fete will be left in the src direc tory We recommend moving this executable file as well as the contents of the scripts directory to some place in your environment s path Our preference is to put all of the executable files in NDSPATH bin directory Here NDSPATH is an environment variable that points to a di rectory containing this project as well as the other code projects needed for nuclear data processing 3 The bdfls file The bdfls file contains a variety of data required by several of the codes maintained by the CNP group In particular the bdf1s file contains group boundaries fluxes nuclear masses half lives various physical and conversion constants and atomic excitation levels fete endep 6 and fudge 7 all require masses A copy of the bdfls file is included in the src di rectory of the fete package You will need to set the BDFLSPATH environment variable to point to the bdfls file If you do not set this variable fete will attempt to load the bdf1s file from the directory Furthermore m
24. nd highlight issues of which a user should be aware Finally we will discuss some of the outstanding issues in the translation the ex tensions that ENDL will need in order to support the rest of the ENDF B VI format and some of the bugs found both in the format specifications and in the databases This note also contains two appendices covering some of the issues in greater detail The first appendix details the preparation of the databases for translation and the var ious preprocessing codes This includes a discussion of the Doppler broadening performed at this step The sec ond appendix details some of the mathematical issues en countered in translating ENDF B VI double differential data This note does not go into the detailed workings of the code This is covered in the fete Code Reference 4 documentation that accompanies the source code distri bution II STEPS IN TRANSLATING The entire ENDF B VI to ENDL translation process is summarized 1 Install the translation code and scripts 2 Acquire and prepare a database 3 Translate the database 4 Compute energy depositions 5 Check results In the following subsections we will explore each step We assume some familiarity with unix or related systems This code system is known to work on Redhat Linux versions 8 0 and higher MacOS X and higher both using gcc3 3 and on Solaris using gcc2 95 3 A Installing the translation code amp scripts 1 Getting the translatio
25. neutrons You are encouraged to consult Tables I and II while reading this section Key Value Interpolation Options Description tol_id 0 005 Relative accuracy of 1 d list interpolation cut_off_id 0 001 Suspend tol_id below this absolute value arb units tol_2d 0 015 Relative accuracy of 2 d list interpolation cut_off_2d 0 003 Suspend tol_2d below this absolute value arb units dE_tol 0 01 Maximum incident energy difference for 2 d lists MeV log_log_tol 0 005 Relative error in converting log log to lin lin 3d List Options num_mu_3d 12 2 nummu_3d 1 number of u bins in 3d expansion num_E_in_3d 12 Number of incident energies for 3d expansion MF 5 Options see Section II B 2 num_mf5_ bins 32 Number of equiprobable energy bins mf5_shift 0 00002 Amount by which the first distribution is above threshold mf5_tol 0 00001 Energy tolerance in ENDF B VI files MeV Other Options For Specific Types Of Data sum_inelastic 0 If 0 use the MT 51 91 individual inelastic data otherwise use the MT 4 total inelastic data Kalbach_i10 0 If gt 0 calculate the I 10 average energy see Section IIIB 3 split_gammas If gt 0 split discrete gammas from continuum using S 3 see Section III C fission_Q If gt 0 use ENDF fission Q value otherwise subtract Q for delayed fission decays see Section III D General Processing Options skip_date If gt 0 don t bother processing evaluation s date field message_level 0 If
26. on and proton This is done because the average energy for the gammas and the residual differs slightly depending on the order of particle emission ENDF B VI has just one reaction which we treat as n np In addition to these problems the databases them selves often have out of date or sometimes even wrong data Fortunately the data is updated regularly and problems are corrected There are several sources for bug lists e NNDC s Known Errors and Deficiencies in ENDF B VI page at http www nndc bnl gov csewg errors html e NJOY s NJOY 99 Issue Tracker page at http t2 lanl gov codes njoy99 Issues html e CEA s nuclear data listserver archive at http www nea fr listsmh e D A Brown and G Hedstrom Possible problems in ENDF B VI r8 LLNL Report UCRL CONF 200686 2003 Nevertheless we would like to establish something more comprehensive and centralized such as a bugzilla bug tracking page Finally we mention that work that we have yet to do Most serious issues that remain are 1 Documentation for evaluations in ENDL is com pletely decoupled from the actual evaluations 2 Implement translation of outgoing energy distribu tions when the distribution is weighted sum of mod els as discussed in Subsection IIIB 2 3 The F n 2n reaction in ENDF B VI has two separate MF 6 double differential data blocks cor responding to the two outgoing neutrons The tra ditional ENDL forma
27. ost of the supporting scripts will fail without this envi ronment variable 4 Setting up the environment You will need several supporting codes in order to com plete the translation 1 prepro2000 if preprocessing a database yourself 2 endfsplit if preprocessing the ENDF B VI database yourself 3 endep to compute energy depositions 4 fudge for checking the translation results 5 PERL Python csh a c compiler make MF Description I 1 General information n a 2 Resonance parameter data 0 3 Reaction cross sections 0 4 Angular distributions for massive emitted particles 1 5 Energy distributions for massive emitted particles 4 6 Energy angle distributions for emitted particles y s or massive particles 3 or 4 7 Thermal neutron scattering law data n a 8 Radioactivity and fission product yield data n af 9 Multiplicities for radioactive nuclide production n af 10 Cross sections for radioactive nuclide production n a 12 Multiplicities for photon production 9 13 Cross sections for photon production 0 14 Angular distributions for photon production 4 15 Energy distributions for photon production 4 23 Photo atomic interaction cross sections 0 27 Atomic form factors or scattering functions for photo atomic interactions 941 30 Data Covariances obtained from parameter covariances and sensitivities n a 31 Data covariances for nubar n a 32 Data covariances for resonance parameters n a 33 Data covariances
28. quently if the masses of the target projectile and ejected particle are denoted by respectively Mtarg Mproj and Meject then a stationary ejected particle in the center of mass frame Eem 0 has energy M i Meject Eo proj Mrarg Mpro3 Ein B2 in the laboratory frame if the incident particle has lab frame energy Fin The direction of such a particle is forward Hiab 1 For particles emitted with positive energy the mapping from center of mass to laboratory coordinates is given by Eiab Eo ae Ecm as 2 fcm V Ego Ecm B3 Eom Eo ab Heng 4 B4 Pab s Fis Era Some curves of Fab Hiab for constant Eo and Eem for 1 lt uom lt 1 are shown in Figure 1 Note that the direction of emission in the laboratory frame is always forward Hia gt 0 if Eem lt Eo In the case that the ENDF B VI data is for 0 lt Ba lt Emax there are 3 possibilities for the domain Qia de pending on the size of Emax relative to Eo If Emax lt Eo then we have only forward emission in the lab frame and Qia is a disk shaped region bounded by the line pap 1 and a curve as in Figure 1 with 41a gt 0 In this case 1 4m If Emax Eo the region Qia is bounded below by the curve starting at the origin in Figure 1 This is because the point with Eem Eo and u 1 maps to Fia 0 Finally if Emax gt Eo then backward emission is possible in the lab frame and Qia is bounded by a curve as in Figu
29. ranslation_summary html file contains two sec tions The first is a hyper linked table of nuclides and problems encountered during their processing This ta ble is linked to a page for each nucleus containing copies of the log files containing details of the problem The second section is a bulleted list of errors encountered in the translation stage each with a list of the nuclei that encountered that error 2 The view_all py utility view_all py is a tool that uses fudge to plot every dataset in an ENDL formatted database To use it type view_all py lt za gt database where lt za gt is the za number of interest and database is an optional argument for the database containing lt za gt If this argument is skipped view_all py will attempt to look for lt za gt in the current directory 3 The za_comp and yo_comp utilities These two utilities allow you to compare the contents of two databases or two individual za subdirectories To compare two databases do za_comp lt dbi gt lt db2 gt To compare the contents of two za subdirectories do yo_comp lt zal gt lt za2 gt Ill TRANSLATION DETAILS In this section we discuss what is involved in the ac tual translation We start with the treatment of reac tion cross sections Then we discuss angle energy and joint energy angle distributions for massive particles not gammas Following this we discuss gamma data and we conclude with a section on delayed fission
30. re 1 with 1 lt pia lt 1 the minimum value of Hiab iS MIN fap 13 Uns VS E Contours of Constant Ecu Lia A neee Ecu 1 0 E DE sek Ea 1 2 E ji Se Ecu 2 0 E af ii 0 Ew FIG 1 Mapping center of mass energy and angles into the laboratory frame Here the center of mass energy Fem is given in units of the kinetic energy of the center of mass sys tem Eo If we find that Emin gt 0 and Emin lt Eon lt Emax for the ENDF B VI data then Qia is the difference between two of the domains described in the previous paragraph 2 Center of mass probability density For given Fab Hiab in Qian we may invert Eqs B3 B4 to get Ecm Eo Erab 2hiab Eo Fiav B5 Ezab Eo cm al Z B6 H H b4 En En B6 We insert these values of E m Hem into the formulas ap propriate to the ENDF B VI data 3 Chap 6 3 The Jacobian The Jacobian of the mapping from center of mass to laboratory coordinates can be derived several different ways and the result is Era J Eon Hcn zz Note that this Jacobian has a singularity due to the fact that we have Fia Eo and Hiap 1 for Eem 0 and all directions 1 lt Men lt 1 Small perturbations cover so much more area in the center of mass frame than in the laboratory frame that we have J 0 pcm for 1 lt pen lt 1 This is not a serious difficulty because dif ferentiation of B5 B6 shows that J Fon Hem
31. ry coordinates In Eq B1 the dependence of pen and the Jacobian J on Ecm and Hem is parametric so we must compute Eem and Len from Fab and Hia in practice The Jacobian must be included because the probability of finding emitted particles within a region of energies and cosines is the integral of p over that region The translation code assumes that the ENDF B VI dou ble differential is given for particles emitted in all direc tions 1 lt Hem lt 1 for a range of energies Emin lt Ecm lt Emax With Emin gt 0 That is in center of mass coordi nates we have a rectangle with Qem Eon Hem Emin lt Ecm lt Enaz 1 lt Hem lt 1 In practice we always have Enin 0 The steps involved in finding the probability density in laboratory coordinates are as follows 1 Find the domain Qia in laboratory coordinates corresponding to the region Qem where we have data 2 For a point Frab Hiab in Qia find its image Eon Wem and evalu ate the probability density Pem Ecm Hem 3 Compute an approximate value of J Eem 4cm and use B1 to cal culate the probability density in the laboratory frame Let us expand on these concepts 1 The domain in the laboratory frame When ENDF B VI data is given in center or mass co ordinates it is assumed that the mapping to laboratory coordinates is to be done according to Newtonian me chanics 3 p 6 5 It is also assumed that the target nucleus is at rest Conse
32. s are ignored on input but will be correctly output on all lines The format of section MF 1 MT 451 and all sections of MF 3 must be correct The program copies all other sections of data as plain text and as such is insensitive to the correctness or incorrectness In the current ver sions of these codes all MF 3 energies will be output in f instead of e FORTRAN format in order to allow energies to be written with up to 9 digits of accuracy In previous versions this was an output option However use of this option to compare the results of energies writ ten in the normal ENDF B VI convention of 6 digits to the 9 digit output from this program demonstrated that failure to use the 9 digit output can lead to large errors in the data just due to translation of the energies to the ENDF B VI format 3 Preprocessing code linear The purpose of this program is to convert ENDF B VI MF 8 23 and 27 data to linear linear interpolable form Any section that is already linear linear interpolable will be thinned Entire evaluations are output not just the linearized data cross sections e g angular and energy distributions are also passed to the output file The de fault input parameter file LINEAR INP sets the output tolerance of 0 1 for the entire incident neutron energy range 4 Preprocessing code recent The purpose of this this program is to reconstruct the resonance contribution to the cross section in lin early
33. s for ENDF B VI are supposed to provide double differential joint energy angle distributions 2 MF 5 Outgoing Energy Distributions The MF 5 file contains energy distributions of the out going particles This data is often in laboratory coordi nates as is the ENDL equivalent data I 4 When the data is presented as an interpolatable table the transla tion to ENDL format is trivial Unfortunately it is more common for ENDF B VI MF 5 data to be given as a table of parameters to use in some formula Thus the energy distribution may be given by a single temperature evaporation model CE exp E 9 for 0 lt E lt Elax Here E is the energy of the out going particle and values of the effective evaporation temperature in MeV are given in the MF 5 file for var ious values of the energy of the incident neutron The value of E is determined by energy conservation and the constant C is chosen to make the total probability be one We expand this a formula into a piecewise linear interpolation table accurate to an amount specified by mf5_tol The translation code has a cutoff cut_off_1d so that we do not impose accurate piecewise linear interpolation in the exponentially small tail Specifically if the energy distribution is denoted by p f E and if pmax is the maximum value of p then accurate interpolation is not enforced when f E lt cut off_1d Pmax What makes translation of this function
34. t permits only average distri butions for the two neutrons 4 While we have made several changes to the ENDL format not all downstream codes have been up dated to handle our changes In addition to the 19F n 2n problem just mentioned there is also problems with discrete gamma data and delated fission neutron data 5 ENDF B VI has uncertainties and covariances for several evaluations now but ENDL has no way to represent either Acknowledgments This work was performed under the auspices of the U S Department of Energy by University of California Lawrence Livermore National Laboratory under Con tract W 7405 Eng 48 APPENDIX A PREPROCESSING ENDF B VI DATA There are several steps to convert the ENDF B VI database into a form suitable for translation a split the database into individual nuclei b preprocess the database to get it into a pointwise form and c Doppler broaden the data to 300 degrees Kelvin 1 Split the ENDF data The ENDF B VI database ships in a collections of files with the tape extension each containing several mate rials First we break the raw ENDF B VI tapes into smaller files each file containing a complete evaluation for a single isotope The individual ENDF B VI files are placed in parallel directories with names derived from the za number of the isotope This marks the begin ning of the ENDL standardization The splitting of the ENDF B VI data and subsequent file organization is ac comp
35. tain double differential joint energy angle distributions ENDL has two options for representing such data I 4 files giving the joint distri butions as Legendre expansions in the laboratory frame P E u X cn E Pa ts and I 3 files containing tables of joint distributions again in the laboratory frame ENDF B VI allows both of these options and adds the ability to represent data by a formula Several formulas are allowed but only two are used in practice the Kalbach Mann model 13 and a uniform distribution in phase space When an ENDF B VI evaluator uses Legendre expan sions we can copy it directly to an ENDL I 4 file as this data is always in laboratory coordinates When an eval uator uses a table of double differential data we must first boost to the laboratory frame if necessary and then transpose the data We must transpose the data be fore copying it into an I 3 file because the ENDF B VI column ordering is E E then u whereas ENDL s is F u then F The most common MF 6 model used in ENDF B VI is the Kalbach Mann formula We evaluate the formula for the number of outgoing angles and energies specified by the num mu_3d option In the future we hope to au tomate these parameter choices One area of difficulty with Kalbach Mann data is that it is in center of mass coordinates Because the double differential distributions are probability densities the conversion from center of mass to laboratory coordina
36. tes includes multiplication by the Jacobian of the transformation We currently use a finite difference approximation to this Jacobian This is detailed in Appendix B Another trouble with Kalbach Mann data is that the function parameters are given as histogram data in the center of mass frame There is no way to make an exactly faithful representation in the laboratory frame As a check on the accuracy of the translation of the Kalbach Mann data we have implemented a calculation of the average energy of the outgoing particles The stan dard approach within ENDL is to use the endep code to integrate the double differential data in laboratory coor dinates The new method of calculating the average en ergy uses integration of the ENDF B VI data in center of mass coordinates You can use the new method by set ting the Kalbach_i10 option With the current default values of the number of energies and angles to use in the translation the disagreement in the two approaches is typically about 2 We expect that this figure would be improved with a better method for selecting the ENDL data points Finally we need to mention that double differential probability distributions are normalized differently in ENDF B VI and ENDL ENDF B VI double differential data is normalized such that ies p E u dudE 1 as suming p E p 0 whenever E m is outside the range of validity of the data ENDL double differential data is normalized such that p E u
37. tional Laboratory Report UCRL ID 127438 Rev 1 July 1997 13 C Kalbach Systematics of continuum angular distribu tions Extensions to higher energies Phys Rev C 37 1988 2350 2369
38. to override the name of the fete_options inp file The options in the fete_options inp file may also be overridden on the command line by prepending a to the option e g fete tol_id 0 001 mf5_tol 0 000025 D Checking the translation We provide some tools for checking and examining the translation The first build_errlist py builds a web page summary of the translation errors The second view_all py allows you to plot all of the quantities in the newly created ENDL formatted database The last two utilities za_comp and yo comp are useful for com paring the contents of ENDL databases and individual za directories 1 The build_errlist py utility The Python script build_errlist py is a tool to build an HTML formatted report of all errors encountered in translating a database It requires the fudge package 7 be installed to function and it is invoked with build_errlist py lt database gt where lt database gt is the name of the directory contain ing the database to be checked This script combs the log files produced during the translation for errors in the translation stage and the energy deposition stage This script also runs the fudge script check py to check for proper ENDL formatting When finished an HTML formatted report will be generated in the lt database gt directory in the same level as the yi01 directory To browse this report point your web browser to the translation_summary html file The t
39. w the up per energy limit of the resolved resonance region and above the lower energy limit of the continuum region i e into the unresolved resonance region This convention guarantees a smooth behavior close to the unresolved resonance region boundaries Entire eval uations are output not just the broadened MF 3 cross sections e g angular and energy distributions are also included The default input parameter file SIGMA1 INP is set up to heat the isotopes to 300 degrees Kelvin and output the cross sections using a 0 5 tolerance APPENDIX B DOUBLE DIFFERENTIAL DATA IN ENDF B VI For joint probability distributions of the energy and direction cosine of emitted particles double differential data it is common in ENDF B VI 3 Chap 6 for the data to be given in terms of the Kalbach 13 parameter ization A complicating factor is that this representation is in the center or mass coordinate system while dou ble differential data in ENDL is in the laboratory frame Because we are dealing with a probability density the transformation formula is Piad Era Hiab Pem Ecm Hom J Eom Hcn B1 Here piapb and Pem are the probability density in respec tively laboratory and center of mass coordinates Ezab and Eem are the energies of the emitted particle tia and Lem are the direction cosines of the emitted particle rela tive to the incident particle and J is the Jacobian of the mapping from center of mass to laborato
40. y I 0 B Data on massive particles ordered by MF number 1 MF 4 Outgoing Angular Distributions The MF 4 file gives angular distributions for outgoing particles This is most commonly used for discrete 2 body reactions and it consists of coefficients for Legendre expansions in the center of mass frame ENDL also uses the center of mass frame for such data but it only has a piecewise linear representation We approximate the Legendre expansions by piecewise linear functions The corresponding ENDL file is flagged by I 1 Note that the sum of the Legendre expansion is sometimes negative for certain outgoing angles and when this happens we replace these unphysical values by zero Interpolation of this data with respect to energy of the incident neutron is usually straightforward since linear interpolation of Legendre coefficients is equivalent to lin ear interpolation of the sum of the Legendre series For some targets e g 7 Pu za094238 ENDF B VI uses log linear interpolationfor the Legendre coefficients We expand the data to linear linear interpolation In some cases the MF 4 data consists of tabular angu lar distributions in the laboratory frame In these cases energy distributions are also given and the two distribu tions are regarded as uncorrelated The translation of this data to ENDL is no problem because ENDL uses the same conventions Such tabular data in MF 4 is archival data in the sense that new evaluation
Download Pdf Manuals
Related Search