MSC.Nastran Error List Update February 22, 2006 1-5447411 2005.5 Documentation/QRG - TAble A2 - Element type 38(gap) missing Element type (38) cgap is missing from Table A2 in the MSC.Nastran Quick Reference Guide. 1-12856910 2005.5 Postprocessing - PARAM,OUGCORD,BASIC and PARAM,POST,-1 - Global Displacements in op2 File Global displacements are written to the op2 file even when parameter OUGCORD=BASIC is used for PARAM,POST,-1. In versions 2004 and beyond this error only occurs when PARAM,SPARSEDR,NO is also specified. Avoidance: Specify DISP=ALL and insert the following DMAP alter: compile sedrcvr alter 'ELSE IF ( NOT(post=-1','' ELSE IF ( post>-3 and OUGCORDX='BASIC' and not(exterse and noug<1) ) THEN $ alter 'sdr2.*ougv1pat','' $ delete /ugb,,,, $ drmh1 ougv1,,,/tougv1,mougv1,,,,,,/ $ IF ( NOCSTMS>0 ) THEN $ MPYAD TRANSG,mougv1,/UGB $ ELSE $ EQUIVX mougv1/UGB/-1 $ ENDIF $ SDR2 CASEDR,CSTMS,,,EQEXINS,,,OL1,BGPDTN,,, UGB,,,OINT,PELSETS,VIEWTB,,,,,,/ ,,OUGV1pat,,,,/APP///////ACOUT/PREFDB/-1./-1./ ADPTINDX/ADPTEXIT/FALSE $ endalter 1-13661505 External Superelements - EXTSEOUT - UFM 311, non-unique field 2 on GRID, External SE, .asm file When using the .asm file generated by EXTSEOUT, if the boundary GRID IMAGES don't match, then the following fatal message may be issued: *** USER FATAL MESSAGE 311 (IFPDRV) NON-UNIQUE FIELD 2 ON BULK DATA ENTRY GRID XXX SORTED ENTRY COUNT = XXX The problem is that the IMAGE is searched, not the "equivalent". That is GRID,101,0,1.,1.,1. and GRID,101,,1.,1.,1. are not considered the same, and the fatal message is issued. 1-14144521 Superelements - BNDFIX1 - Unexpected ASET when QSET omitted If BNDFIX1 degrees of freedom are specified, but no QSET is specified, then the BNDFIX1 entry is ignored. Avoidance: Use the ASET entry. 1-14410701 SOL 400 - Solution diverges with SPCD loads and linear rigid elements for LANGLE=1,2 and SOL400. For SOL400, the solution diverges if there are SPCD loads and linear rigid elements in the deck for LANGLE=1, or 2. However, the solution will converge if LANGLE=3. 1-17025310 2005.1 Installation - mscsetup - Linux install fails; cannot determine disk space Nastran 2004 "mscsetup" script fails on Mandrake Linux 9.1/ 10.0 because it cannot determine the amount of free disk space. Suggestions: "mscsetup" uses "df -k" to determine disk space. However, if "df -k" outputs multiple lines, the "awk" procedure cannot determine the actual disk space. Avoidance: Replace the following line in mscsetup linux* ) set -- $(df -k $dir 2>/dev/null | awk 'NR==2{printf("%d %d\n",$4/1024,$2/1024)}') with: linux* ) set -- $(df -k $dir | awk 'NR==2 && NF>1{printf("%d %d\n",$4/1024,$2/1024);exit} NR==3{printf("%d %d\n",$3/1024,$1/1024);exit}') 1-17384835 2005.5 Aeroelastic Analyis - Attempt to trim with no mass matrix. If a solution 144 trim analysis is executed on a model with no mass the following fatal message that does not describe the problem will occur: *** SYSTEM FATAL MESSAGE 6144 (PARTN1) THE SIZES OF THE INPUT MATRICES AND PARTITIONING VECTORS ARE INCOMPATIBLE. SPECIFICALLY: The number of columns in PUXLRPV is not equal to the number of rows in the column partitioning vector PVPR User Information: Size of input matrix and partitioning vectors: PUXLRPV : Rows= 2 by Cols= 9 PVPR : Rows= 11 An attachment contains a version of the AEMR.dmap that traps the missing mass matrix. 1-17505001 2005.5 External Superelements - EXTSEOUT, different ID - missing connection For external superelement using the EXTSEOUT command, the boundary points for the assembly structure at which the components will be attached are generated in the ASMBULK file. It is also assumed that the boundary IDs for the component and the assembly structure are the same. If the boundary IDs are not the same, then the structure may not be connected as intented. Avoidance: If the boundary IDs for the component and assembly structure are different, do not use the ASMBULK file, but create a SEBULK entry using the automatic feature instead. This problem has been resolved in Version 2005R3 by the incorporation of a new option called ASMBULK=AUTO for the EXTSEOUT Case Control command. 1-18450701 2005.1 Nonlinear Analysis - CBAR, CBEAM with offsets and param,lgdisp fails with UFM 6174 USER FATAL MESSAGE 6174 is incorrectly issued when a nonlinear model contains bar or beam elements with offsets and param,lgdisp,-1. When param,lgdisp,-1 is specified the differential stiffness is not required, thus the fatal should not be issued. Avoidance: There is no avoidance. 1-18602293 2005.5 Eigenvalue Analysis - EIGRL - USER FATAL MESSAGE 4405 (PRTPRN) For most real world problems the CRAY X1 Unicos system's acrhitecture is such that it realizes a performance gain when MAXSET on the EIGRL entry is set to 12 rather than to default value of 7 used on most other machines. One small test problem (s6821) diverged when trying to solve the eigenvalue problem. This small test completes when MAXSET is changed to 7. 1-18612550 Nonlinear Transient Analysis - TSTEPNL - UFM 5423, TSTEPNL with NO = 2 The entry: TSTEPNL 2 2000 .0005 2 resulted in the following fatal message: *** USER FATAL MESSAGE 5423 (SADD5) ATTEMPT TO ADD INCOMPATIBLE MATRICES, (SADD5) COLS ROWS FORM TYPE NZWD DENS TRAILER FOR PST1 IS 1002 6 2 2 2 58 TRAILER FOR IFS IS 1001 6 2 2 0 0 Avoidance: Use either: TSTEPNL 2 2000 .0005 1 or TSTEPNL 2 2001 .0005 2 1-18678552 2005.1 Optimization/Design Sensitivity - Incorrect updated PBEAM entries written to the punch file In design optimization Incorrect updated PBEAM entries are written to the punch file. 1-18679258 Nonlinear Analysis - GPFORCE, TEMP(LOAD) - GPFORCE incorrect in nonlinear with thermal load The GPFORCE balance is incorrect in SOL 106 when there is thermal load because the themal load contribution is included twice. SOL 400 is correct. Avoidance: (1) Replace "SOL 106" by "SOL 400" (2) Add "ANALYSIS=NLSTAT" in the case control deck (you really need not do it in this deck since the default of ANALYSIS is NLSTAT in SOL 400.), (3) Replace "SUBCASE" by "STEP" in the case control deck (you can also skip this change since this test deck is single SUBCASE only.). 1-18723414 2005.5 Documentation/QRG -MODALKE and MODALSE case command descriptions for SORT1 & SORT2 are interchanged The descriptions of the SORT1 and SORT2 output formats for the MODALKE and MODALSE case control commands are interchanged. The MODLAKE SORT1 description actually shows SORT2 output and the MODLASE SORT2 output actually shows SORT1 output. 1-18809018 2005.5 Optimization/Design Sensitivity - MCID - Incorrect punch file In an optimization run, if the MCID is used in a plate element, the punch output will contain an incorrect MCID. Avoidance: Define the angle using the theta field instead of MCID. 1-18890681 2005.5 Elements - CQUADR and TEMPP1 - Wrong Answers (STRESS) A simple plate, modeled with with CQUADR elements, that is subjected to a linearly varying temperature gradient through the thickness using TBAR and TPRIME on the TEMPP1 entry, will not result in the correct zero stress field. Avoidance: Use the CQUAD4 element. 1-18935320 Documentation/QRG - GPRSORT - Missing Documentation GPRSORT Case Control command is undocumented Avoidance: See Nastran V2005 Release notes for an example. For the QRG additional examples of non-default values need to be added. 1-18998425 2005.1 Rigid Elements - RBE, TEMP(INIT), TEMP(LOAD), RIGID = LAGRAN Models with temperature loads on rigid elements, rbes, with thermal coefficient of expansion defined, may fail in the routine getemp on Solaris and possibly other machines. Avoidance: There is no avoidance. 1-19006951 2005.5 Nonlinear Analysis - ESE, SET - Missing Output for ESE referencing SET Nonlinear analysis, SOL 106, may ignore an ESE request that references a SET. Avoidance: Request ESE=ALL 1-19047395 Optimization/Design Sensitvity - AUXMODEL - Wrong stresses When an Analytic Boundary Shape optimization job also includes SPCD entries in its primary Bulk Data section, it may produce wrong element responses such as stress for the final analysis at the last design cycle. Avoidance: Apply the following dmap alter compile nddl=nddl include 'nddl.dat' Notice the nddl.dat above is a modified complete listing of the nddl statements. The modification is made to the following one DATABLK statement where the Path has been changed fro NOPATH to YS DATABLK YSX PATH=YS LOCATION=EQVSCR, 1-19104750 2005.1 Composites - STRAIN(BILIN), PCOMP - UFM 3005 (SDR2), Attemp to Operate on Singular Matrix A model with composites that requests strain(bilin) or strain(corner) but does not request stress will fail with the message: USER FATAL MESSAGE 3005 ATTEMPT TO OPERATE ON THE SINGULAR MATRX NONE IS SUBROUTINE SQD43D 1-19124784 2005.5 Documentation/QRG - CWELD - Missing documentation on new V2005 capability Beginning in V2005, CWELD is able to connect elements that share a common node. 1-19192460 2005.5 Postprocessing - RANDOM, SORT1 for DISP and ACCE - SORT1 Request Ignored in Random Analysis The following request in Nastran 2004 and 2005 will generate OP2 results in the incorrect format. OUGPSD2 is written to the op2 file instead of OUGPSD1. DISPLACEMENT(SORT1,PLOT,PHASE,PSDF) = ALL ACCELERATION(SORT1,PLOT,PHASE,PSDF) = ALL This is a limitation in releases prior to Version 2005R3. Nastran only generates SORT2 random results because the original design required all results to be output though the XY-PLOT module which only supports SORT2. 1-19203339 2005.1 Dynamic Analysis - GPFORCE=set yields wrong answers in dynamic analysis Although the Version 2005 QRG Case Control table 4-5 shows that the GPFORCE request is not supported in SOL 108,109, 111, and 112, including it in the Case Control produces GPFORCE output. If the GPFORCE request references a SET then the resulting numbers are incorrect. If the GPFORCE references "ALL" then the results are correct. 1-19209144 2005.5 Heat Transfer Analysis - View factor calculation with CHBDYG,AREA8 - Wrong Answers Coarse mesh models with curved surfaces will produce incorrect view factor calculations with CHBDYG elements using the AREA8 option. Avoidance: Refine the mesh using CHBDYG,AREA8 with CQUAD8 elements or use CHBDYG,AREA4 with cquad4 elements. To address this problem the code is modified to use corner grids to compute view factors for warped elements. 1-19210511 2005.1 MDACMS - DOMAINSOLVER, CWELD, PWELD with mset=on - Poor Accuracy. A SOL 200 analysis using mdacms gave incorrect answers with cwelds and mset=on unless upfact was turned up to 3.0 from the default value of 2.0 on the DOMAINSOLVER entry. . 1-19325591 2005.1 Nonlinear Heat Transfer Analysis - CHBYDE, CHBDYG - Coarse Mesh, poor results or non-covergence Heat transfer analysis models with CHBDYE and CHBDYG elements with coarse meshes as can be found in tpl problem ht15306 may fail to converge or produce poor results. Avoidance: 1) Model using CHBDYP with view view factor instead of CHBDYE or CHBDYG elements. 2) Refine the mesh. 3) Insert the DMAP alter COMPILE NLKR $ ALTER 'RMG2.*RGG' $ $ MODIFY THE RADIATION TO AVOID TEMPERATURE OVERSHOOT WITH COARSE MESH $ VERSION 2005 FEBRUARY 25, 2005 BY BOB HARDER ECHOOFF $RMG2 EST,MPOOL,MUGNI,KGGNL, MPTS, DIT, BGPDTS,SILS,USET1/ $ RDEST,RECM,RGG,KGGPNL/TABS/SIGMA/S,N,NORADMAT/LUSETS $ $ MOVE POSITIVE OFF-DIAGONAL MATRIX TERMS TO DIAGONAL $ PARAML RGG//'TRAILER'/1/S,N,GSIZE $ DIAGONAL RGG/DRGG/'SQUARE'/1. $ DIAGONAL TERMS ADD RGG,DRGG/ORGG/(1.,0.)/(-1.,0.) $ OFF-DIAGONAL TERMS DIAGONAL ORGG/ARGG/'WHOLE'/1. $ ABS(OFF-DIAGONAL TERMS) ADD ORGG,ARGG/XRGG/(.5,0.)/(.5,0.) $ OFF-DIAGONAL TERMS .GT. 0 MATGEN, /ONES/6/V,N,GSIZE/0/V,N,GSIZE $ COLUMN OF ONES MPYAD XRGG,ONES,/SRGG $ ROW SUM(OFF-DIAGONAL TERMS .GT. 0) MATMOD SRGG,,,,,/IRGG,/28 $ VECTOR TO DIAGONAL MATRIX ADD5 RGG,XRGG,IRGG,,/NRGG/(1.,0.)/(-1.,0.)/(1.,0.) $ EQUIVX NRGG/RGG/ALWAYS $ MESSAGE //'^^^ RADIATION MATRIX HAS BEEN MODIFIED' $=================== ADD KGGPNL,KGGNL/RGGNL/(1.,0.)/(-1.,0.) $ TRNSP RGGNL/TGG $ DIAGONAL TGG/DTGG/'SQUARE'/1. $ ADD TGG,DTGG/OTGG/(1.,0.)/(-1.,0.) $ DIAGONAL OTGG/ATGG/'WHOLE'/1. $ ADD OTGG,ATGG/XTGG/(.5,0.)/(.5,0.) $ MPYAD XTGG,ONES,/STGG $ MATMOD STGG,,,,,/ITGG,/28 $ ADD5 TGG,XTGG,ITGG,,/NTGG/(1.,0.)/(-1.,0.)/(1.,0.) $ DELETE /RGGNL,KGGPNL,,, $ TRNSP NTGG/RGGNL $ ADD KGGNL,RGGNL/KGGPNL/(1.,0.)/(1.,0.) $ MESSAGE //'^^^ NL-CONDUCTION MATRIX HAS BEEN MODIFIED' $=============== ECHOON 1-19361520 2005 DMAP - INPUTT4 - SFM 6445 when reading unformatted file. INPUTT4 may fail with the following message when attempting to read unformatted files: *** SYSTEM FATAL MESSAGE 6445 (INPTT4) READ REQUESTED RETURNED INCORECT DATA LENGTH WHILE PROCESSING UNIT = 11 FOR LOGICAL NAME INPUTT4 EXPECTED 4 WORDS, ACTUALLY READ 5 WORDS USER ACTION: VERIFY THE FILE ASSIGNMENT AND CONTENTS OR CONTACT MSC.SOFTWARE CLIENT SUPPORT. In a test case which attempts to read in the GM, GOT, KFS and KSS matrices, the fatal error was encountered on both a SGI and PC machine. Tests on 2004.0.6 and 2005 were successful. Avoidance: Possible avoidance is to try using a formatted fortran file or use version 2005.4 or later systems where this problem has been addressed. op4util -fc inputfilename outputfilename 1-19369064 2005.5 Frequency Response - STRESS=set id, OUTPUT(XYPLOT) - Extra Printed Output If a model contains a STRESS=set id request but also contains an XYPLOT type request for different element ids, then printed output may be generated for the elements with an XYPLOT request. Avoidance: There is no avoidance. 1-19372191 2005.5 DMAP - MATPRN - Matrices The MATRPN module can only print 6 digits for rows and columns, ****** will be printed for matrices greater than one million rows and columns. 1-19381619 2005.1 Nonlinear Heat Transfer - TEMPD - Traceback, Overflow The solution for nonlinear steady state thermal analysis with radiation enclosures is depedent on the initial guess for TEMPD. If the initial guess is incorrect by a large amount it may cause the solution to have trouble converging as evidenced by extremely large convergence factors and the job may end in a floating point exception. . . MAIN: The FPU has been reset after floating point exception. MAIN: For reference, A(MAINAL) = 4012FE, A(/SYSTEM/) = 1FD0A00 MAIN: A(/XNSTRN/) = 3CF0020 MAIN: "Floating point overflow" (C0000091) exception encountered. MAIN: Exception occurred at address 00524C15. 1-19390023 2005.5 Documentation/QRG - CQUAD4 differential stiffness remark obsolete Remark 6 for CQUAD4 should now read: ..."The specification of offset vectors gives wrong results in solution sequences with geometric nonlinear analysis (SOL 106, 129, 153, 159 with PARAM LGDISP>0) The differential stiffness is computed correctly in V2004 and higher (SOLs 5, 16, 105 and 200 and SOL 106 with PARAM LGDISP <=0) 1-19394948 2005.5 CDITER - Infinite CPU Loop, No FATAL diagnostics for an ill conditioned model On rare occasions for ill-conditioned models, the round off error in caluculating the QR1 matrix used for testing convergence in CDITER will make it difficult to find a new random vector and an infinite loop may be encountered. The problem can be diagnosed by setting PARAM,CDPRT,YES which will print the QRI terms. Tthe symptom is a very small negative (tensile) force on the order of E-4 or smaller. 1-19406406 2005.5 Documentation/QRG - params CWDIAGP and CWRANDEL - Descriptions have been swapped The text describing the PARAMs CWDIAGP and CWRANDEL have been swapped. Current Version 2005 MSC.Nastran QRG: CWDIAGP Default = NO For CWELD element: if not zero, then it specifies as a decimal percent the number of fasteners to randomly delete. CWRANDEL Default = 0 For CWELD element: prints elements randomly deleted if set to YES. Should be: CWDIAGP Default = NO For CWELD element: prints elements randomly deleted if set to YES. CWRANDEL Default = 0.0 For CWELD element: if not zero, then it specifies as a decimal percent the number of fasteners to randomly delete. 1-19415074 2005.1 SOL 600 - Incorrectly translated DIST LOAD entry in Parameter section of marc input - EXIT 13 DIST LOAD entry in Parameter section of the marc input deck is not written correctly. The 4th field in the DIST LOAD entry in the PARAMETER section is the Maximum number of nodes with point loads applied. When Gravity load is applied in a model with Point Mass elements, the Marc input deck is created with POINT LOAD entries in the model and/or history sections. But the 4th field in the DIST LOAD entry is not given a large enough number. The Marc run exits with the EXIT 13 indicating the following error: *** error - attempt to define too many point loads 0.000E+00 0.000E+00-0.208E-04 0.000E+00 0.000E+00 0.000E+00 a list of nodes given below 3749 1-19418098 2005.1 SOL 600 - Gravity Loads are not converted correctly into DIST LOAD Gravity Loads are not converted correctly in to DIST LOAD. For example, the following entries in the .bdf are converted to .marc.dat incorrectly: $ Loads for Load Case : GRAV_0.10_RELIEF.SC2 LOAD 51 1. .1 46 1. 72 1. 73 $ Loads for Load Case : GRAV_0.20.SC3 LOAD 55 1. .2 46 1. 72 1. 73 Converted gravity loads may have an incorrect sign. Only one load case has the correct sign (-ve) and value of gravity. The other load cases have an incorrect sign (+ve) and wrong values for gravity. 1-19429671 2005.5 Elements - CBAR - UFM 2007 (GP2), OFFT references "B" and either "E" or "O" with vector V specified In v2005.0.0, bars and/or beams will issue UFM 2007 complaining that an undefined grid point 0 has been referenced if all of the following conditions are met: - The OFFT field starts with a "B" to designate an orientation vector in the BASIC system - The orientation Vector option (e.g. 0. 1. 0. ) is specified - Either "E" or "O" is used to designate one or both of the shear center offset coordinate systems 1-19456411 2005.1 Elements - CHEXA, CTETRA, CPENTA, TEMP(LOAD) - Poor Answers If a solid element model has a temperature rise which is linear in (x,y,z), then there should be zero thermal stress if the constraints are nonredundant. Due to limitations imposed by the isoparametric shape functions, MSC.Nastran's elements give poor results for displacements and stresses. This occurs for HEXA, PENTA, and TETRA elements, especially if there are no midside nodes. A new theory is now available that lets us eliminate this limitation. This implementation gives exact displacements and zero stresses for bodies modelled with HEXA, PENTA, and TETRA elements if the temperature is linear and the constraints are nonredundant. This is even true for a TETRA-4 element which is notoriously inaccurate. This revised theory was released in version 2005r2 (2005.1). If you want to compare with the old method, you can use NASTRAN,SYSTEM(428)=1 to get the old answers. 1-19491251 2005.1 Distributed Memory Parallel - Small jobs eigenvalue jobs may fail with SFM 9160 (MDACMS) Very small eigenvalue problems that specify Lanczos, automatically revert to the AHOU eigenvalue method. If this occurs during a DMP execution the model will fail with the message: ^^^ SYSTEM FATAL MESSAGE 9160 (MDACMS) ^^^ PROBLEM ENCOUNTERED IN CALCULATING MDACMS DOMAIN MODES. ^^^ USER ACTION: IF LARGE MASS IS BEING USED FOR ENFORCED MOTION THEN REPLACE LARGE MASS METHOD WITH DIRECT METHOD USING THE SPCD BULK DATA ENTRY. ^^^ IF THE LARGE MASS METHOD IS NOT BEING USED THEN MASSLESS MECHANISMS MAY EXIST. PARAM,MECHFIX,YES MAY BE USED TO IDENTIFY MASSLESS MECHANISMS. 1-19528421 2005.1 Elements - SYSTEM(370)=5 - Wrong Answers for STRESS(punch,center) If nastran system cell(370)=5 is used to convert CQAUD4/CTRIA3 to CQAUDR/CTRIAR, and (STRESS(punch,print,SORT1,REAL,VONMISES,CENTER) is selected the punch output will be incorrect. 1-19529350 2005.1 Optimization/Design Sensitivity - PARAM,DESPCH - Incorrect values in punched bulk data - HP only Optimization, SOL 200, may produce incorrected punched bulk data on HPUX. AIX, Sun also produce different punched bulk data. Problems might be in PBAR elements. Avoidance: Setting param,despch,-1 outputs the bulk data in small field format and avoids the problem. 1-19535231 2005.5 Optimization/Design Sensitivity - PCOMP - SFM 4276, EC =3061, UFM 2042 (DSABO) When an optimization job specifies identical PID on a designed PCOMP entry and another designed property entry such as PBEAML, it may fail with a misleading USER FATAL MESSAGE 2042 (DSABO): Missing Material Table ***** for Element 1. Avoidance: Use unique PIDs among all designed property entries. 1-19557001 2005.1 Frequency Response - CQUAD4 - Corner Qx and Qy for plate elements are wrong. Frequency response analysis produces incorrect QX and QY CQUAD4 complex corner stresses. Avoidance: Other than using the results from the center of the element, there is no avoidance. 1-19579411 2005.1 Case Control - K2GG, M2GG, B2GG, K42GG, A2GG (etc.) - Continuation Lines lead to UFM 614 (IFP1D) If the case control command K2GG (or any of its variants) calls up more matrices than can be specified on one line, and a continuation line is added, the solution may fail with the following message: *** USER FATAL MESSAGE 614 (IFP1D) THE ABOVE SET IS BADLY SPECIFIED. Avoidance: If the list of matrices does NOT contain scale factors, place the list of matrices on a SET entry and call the SET from K2GG. For example: SET 992 = KX1, KX2, KX3, KX4, KX5, KX6, KX7, KX8, KX9, KX10, KX11, KX12, KX13, KX14, KX15, KX16, KX17 K2GG = 992 If the list has scale factors, there is no avoidance. 1-19600121 2005.1 SOL 600 - CBUSH/PBUSH elements are not translated correctly - Wrong Answers The CBUSH element property orientation is not translated correctly in SOL 600. This may lead to wrong answers. 1-19603931 2005.1 Optimization/Design Sensitivity - SPCD, MODESELECT scr=yes - Wrong Answers, Poor Convergence SOL 200 with scr=yes for ANALYSIS=MTRAN,MFREQ,MCEIG, or FLUTTER may differ from a scr=no run in the following areas: 1. Wrong answers in enforced motion 2. Different convergence behavior or no convergence 3. SFM 3200 in GKAM1C if MODESELECT based on effective mass fractions is requested. Avoidance: Run with scr=no or insert the following DMAP alter: compile fea alter 'call feamode','' call feamode, $..Input Datablocks: adrdug,ag0,ajp,bgpdtvu,CASECC,caseds,casem,CASERS, CASES,CASESX,CASEXX,cstms,DEQATN,DEQIND,dit,drg, DYNAMIC,DYNAMICS,ectas,edom,EDT,eht,ehta,elemvol, EMAP,EPT,epta,EPTS,EQEXINS,ETT,extdb,FORCE,gdntab, GEOM1,geom1qx,geom1s,GEOM2,GEOM2S,GEOM3,GEOM3S,GEOM4, GEOM4S,GM,GPDTS,GPLS,gpsnts,ibulk,INDTA,MAPS,matpool, matpools,medge,mgg,mmcdb,MPT,MPTS,oint,PCDB,PCDBS, pcompts,pelsets,POSTCDB,pval,PVT,pvtbc,PVTS,R1TABR, rmg,rvg,scstm,SILS,SLIST,SLT,tabevp,tabevs,tempall, USET0,VGFS,vgq,viewtb,XYCDB,YG,xsop2dir,pvt0, asf,aay,aaa,day,kdmg,kdss,gei,pmyg,pbyg,pkyg, mss,bss,cee,mee,gge,bec,mec,meh,cmlama1,epsseo,mphvz,bk4zz,phqmg, kgg,kay,may,bay,k4ay,k4ss,phba/ $..Output Datablocks: AGG,BAA,bdict,belm,BGG,BGPDTS,bgpdt,BJJ,BMG,BSF, CMLAMA,CMPHA,CMPHO,Dar,DFFDNF,ditid,ects,error, errorg,EST,frl,GOQ,GOT,GPAK,GPECT,K4AA,K4JJ,K4MG, K4SF,kaa,KDICT,KELM,KFS,KJJ,KLL,KMG,KSF,KSS,L2RR, LAMA,LAMAF,LAMAS,LLL,LOO,LRSEQ,MAR,MDICT,MDSF,MEA, MELM,mjj,MLL,MLR,MMG,MR,MRR,MSF,MTRAK,orseq,PG,PHA, pharef,PHFA,phg,PHGref,PHQG,PHSA,PJ,PSS,PTELEM,QR, USET,VAFS,VELEM,vgfd,YS,maa,mdphiq,mdgot,mdgoq,mdvft, mdmap,mdgrdrm,adsf,lammats,lammatf,mis,mif,mechg, bdsf,kdsf,k4dsf,efmfacf,efmfacs/ $..Input Parameters: AERTYP/ALTRED/APP/AUTOSPC/cnvflg/conv/DELTA/DESCYCLE/ DESCYCL1/deseig/desmax/DFREQ/DFRQCC/DMRESD/DROPT/ERROR/ EXTERSE/FEAOPT/FIXEDB/fluidse0/FSCOUP/INREL/LUSETS/ MODACC/NASOUT/nk/NOPH2/NSKIP/PEID/pexist/ploop/pmode/ PSTAT/RSONLY/SEKR/SELG/SELR/SEMG/SEMR/SOLTYP/desiterp/ $..I/O PARAMETERS: ADPTEXIT/ADPTINDX/CARDNO/freqdep/FS/luset/METHF/ MDTRKFLG/s,modetrak/NORSET/NOSASET/PFILE/PVALID/ SLAVEJOB/SUPORT $ compile feamode alter 1,1 SUBDMAP FEAMODE, $..INPUT DATABLOCKS: adrdug,ag0,ajp,bgpdtvu,CASECC,caseds,casem,CASERS,CASES,CASESX, CASEXX,cstms,DEQATN,DEQIND,dit,drg,DYNAMIC,DYNAMICS,ectas,edom, EDT,eht,ehta,elemvol,EMAP,EPT,epta,EPTS,EQEXINS,ETT,extdb,FORCE, gdntab,GEOM1,geom1qx,geom1s,GEOM2,GEOM2S,GEOM3,GEOM3S,GEOM4, GEOM4S,GM,GPDTS,GPLS,gpsnts,ibulk,INDTA,MAPS,matpool,matpools, medge,mgg,mmcdb,MPT,MPTS,oint,PCDB,PCDBS,pcompts,pelsets,POSTCDB, pval,PVT,pvtbc,PVTS,R1TABR,RMG,rvg,scstm,SILS,SLIST,SLT, tabevp,tabevs,tempall,USET0,VGFS,vgq,viewtb,XYCDB, YG,xsop2dir,pvt0,asf,aay,aaa,day,kdmg,kdss,gei,pmyg,pbyg,pkyg, mss,bss,cee,mee,gge,bec,mec,meh,cmlama1,epsseo,mphvz,bk4zz,phqmg, kgg,kay,may,bay,k4ay,k4ss,phba/ $..OUTPUT DATABLOCKS: AGG,BAA,bdict,belm,BGG,BGPDTS,bgpdt,BJJ,BMG,BSF, CMLAMA,CMPHA,CMPHO,Dar,DFFDNF,ditid,ects, error,errorg,EST,frl,GOQ,GOT,GPAK,GPECT, K4AA,K4JJ,K4MG,K4SF,kaa,KDICT,KELM, KFS,KJJ,KLL,KMG,KSF,KSS,L2RR,LAMA,LAMAF,LAMAS,LLL,LOO, LRSEQ,MAR,MDICT,mdsf,MEA,MELM,mjj,MLL,MLR,MMG,MR,MRR,MSF,MTRAK, orseq,PG,PHA,pharef,PHFA,PHG,phgref,PHQG,PHSA,PJ,PSS,PTELEM, QR,USET,VAFS,VELEM,vgfd,YS,maa,mdphiq,mdgot,mdgoq,mdvft,mdmap, mdgrdrm,adsf,lammats,lammatf,mis,mif,mechg,bdsf,kdsf,k4dsf, efmfacf,efmfacs/ $..INPUT PARAMETERS: AERTYP/ALTRED/APP/AUTOSPC/cnvflg/conv/DELTA/DESCYCLE/DESCYCL1/ deseig/desmax/DFREQ/DFRQCC/DMRESD/DROPT/ERROR/EXTERSE/FEAOPT/ FIXEDB/fluidse0/FSCOUP/INREL/LUSETS/MODACC/NASOUT/nk/NOPH2/NSKIP/ PEID/pexist/ploop/PMODE/PSTAT/RSONLY/SEKR/SELG/SELR/SEMG/SEMR/ SOLTYP/desiterp/ $..I/O PARAMETERS: ADPTEXIT/ADPTINDX/CARDNO/freqdep/FS/luset/ METHF/MDTRKFLG/modetrak/NORSET/NOSASET/PFILE/PVALID/slavejob/suport $ alter 'type db',''(,3) TYPE DB ophfa,ophsa,rph,rpd,rsgg,rmgg,gpke,msf1,bsf1,ksf1, vgdr,vspcg,vmpcg,MMG1,BMG1,KMG1 $ endalter 1-19623877 2005.5 Documentation/QRG - MFLUID, ELIST - Only CQUAD4 and CTRIA3 Supported The QRG should note that the ELIST entry only apply to the ctria3 and the cquad4 elements. 1-19644046 2005.5 Superelements - Missing OLOAD RESULTANTS The OLOAD RESULTANTS output for the upstream SE is equal to zero for all components. Avoidance: Insert the following DMAP alter: compile phase1dr,nolist alter 'type.*parm.*nddl' TYPE PARM,NDDL,I,Y,RESLTOPT $ TYPE PARM,,I,Y,ALTSHAPE $ alter 'ENDIF.*SETYPE=.*EXTRNA' IF (SETYPE = 'EXTRNA') THEN $ expand PA to Jsize UMERGE USET0,PA,/PJext/'G'/'A'/'O' $ message //'OLOADS from external se follow' $ VECPLOT PJext,BGPDTS,SCSTM,CSTMS,CASESX,,MEDGE,,/ ,PJRESext,PJHDext,,/ GRDPNT/S,N,ZERO/RESLTOPT/'OLOAD'///ALTSHAPE//SEID $ ENDIF $ 1-19650401 2005.1 DMAP - INPUTT2 - INPUTT2 won't rewind a file that has been read to its end. INPUTT2 won't rewind a file that has been read to its end, if attempted the job will fail in the INPUTT2 module and the only message will be in the log file: forrtl: severe (24): end-of-file during read 1-19654201 2005.5 Dynamics - Rayleigh Damping Incorrected implemented The current Rayleigh damping is not implemented correctly in GMA. Only the structural matrices should be scaled, but the final matrices are scaled (after param, G, k4, tf, *2pp are added). Beginning in 2005r3 - Alpha1 and Alpha2 will affect only the structure - New Alpha1fl and new Alpha2fl will affect the fluid. 1-19680181 2005.1 Aeroelastic Analysis - MONPNTt3, SET1 - UFM 2258 If a monpnt3 entry references a non-existant set1 id, USER FATAL MESSAGE 2258 is issued, however the monpnt3 id is in the wrong slot of the error message and the set1 id may or may not be the id of the bad set1. In addition, this type error should not be a fatal, but a warning message because the checking is done late in the analysis. 1-19689679 2005.5 Case Control - Scale Factor applied to P2G lead to UFM 300 Attempting to apply a scale fact to P2G, (i.e. P2G = 1.25*PAX) may lead to user fatal message 300 in the case control section. Tested 2005t1 and 2006t1, both run with NO error. - SJO 1-19722224 2005.1 External Superelements - SPOINTS - SFM 3001 in the BDRYINFO module. An external SE creation run employing only SPOINTS fails with SYSTEM FATAL MESSAGE 3001 in the BDRYINFO module. 1-19744067 2005.1 SOL 600 - TITLE, SUBTITLE, LABEL - Missing Title and Label information In SOL 600 titling information is not carried through in the f06 output. Avoidance: Add param,mop2titl,-1. 1-19747531 2005.1 Elements - CHEXA8 with bubble functions and bad geometry - Missing Fatal Message, Poor Answers A CHEXA8 element should have a Jacobian determinant with the same sign at all integration points. Failure to do so indicates an invalid mapping to parametric space caused by an extremely skewed element, which could yield bad results. This should but does not cause a fatal message. Avoidance: There is no affect for well shaped elements. Check the element geometry using IN 2 or 3 on the PSOLID entry. 1-19749179 2005.1 Optimization/Design Sensitivity - PBAR, PBEAML - UFM 7190 (DOPR1I) When an optimization job designs both properties on PBAR and PBEAML entries and no PBEAM entry is present, the job fails in module DOPR1 with the following error message. *** USER FATAL MESSAGE 7190 (DOPR1I) PTYPE PBEAM DEFINED ON A DVPREL3 0 ENTRY CAN NOT BE FOUND ON A PROPERTY ENTRY. USER ACTION: DELETE THE DVPREL3 ENTRY OR INCLUDE AN ADDITIONAL PROPERTY ENTRY. Avoidance: Add a dummy PBEAM entry. 1-19769081 2005.5 Miscellaneous - INCLUDE - Missing Input If an INCLUDE statement is missing and the statement is near the end of the input stream, for example, INCLUDE 'platest.dat ENDDATA the include file is skipped and only a warning about a missing ENDDATA statement is produced. 1-19780791 2005.1 Optimization/Design Sensitivity - SFM 3003 (DSADJ) ATTEMPT TO READ PAST THE END OF FILE in DSDJ2D An optimization job may fail in DSDJ2D with SFM 3003 when it meets the following three conditions: 1) GE is present on a MATi entry associated with the designed element group A 2) GE is not present on a MAT entry associated with designed element group B where both group A and B belong to the same element type. 3) One design element is touched by more than one design variable. For example, one design variable designs one PBARL or PBEAML dimension,.it will spawn multiple non-linear design properties that are also treated as design variables. So one element is touched by more than one design variable. Avoidance: Use Param,cdif,yes. 1-19780814 2005.5 Optimization/Design Sensitivity - PBUSH - SFM 3001 in DSDJ1D An optimization job may fail in DSDJ1D (or DSDJ3D or DSVG1D) with SYSTEM FATAL MESSAGE 3001 when the input contains one or more PBUSH entries that have non-zero values in the GE's 2-6 fields and none of these PBUSH entries are designed. Avoidance: Create a dummy desvar and dvprel1 entry to design one PBUSH property where the dummy desvar has the same lower and upper bounds. 1-19791494 2005.1 Elements - PBUSH, TABLED1 - Wrong Answers Version 2004 and 2005 do not provide the correct results for a bush element if pbusht/tabled1 is used. The stiffness in the pbush entry is used instead of the force-displacement curve defined in the pbusht/tabled1 entry. 1-19795091 2005.1 Superelements - SEP1X, CSUPEXT, THRU - Wrong Answers Superelements using CSUPEXT entries containing THRU ranges may produce incorrect results. Avoidance: Use delineated grids instead of THRU range. 1-19822213 2005 Frequency Response - FRLG - Wrong Answers when Large Memory Requested Frequency response analysis that request more than 2 gig of memory may yield incorrect results. In the cases that have been observed, the response results contained spikes. Avoidance: Either always run with less than 2 gig of memory or plot the response for all frequency steps and do not trust the results if it contains unexplained spikes in the response. This problem is fixed in Version 2005. 1-19823963 2005.1 Aeroelastic Analysis - FA1, FLFACT, PKS, PKNLS - Wrong Answers Wrong flutter results are computed by the PKS and PKNLS flutter analysis methods if MAX or POINT normalization is specified on the eigr or eigrl entries. Futhermore, if flutter eigenvectors are requested for the PKS or PKNLS methods using negative velocities on the FLFACT entries, the eigenvetors are not computed or printed. Avoidance: To obtain correct flutter results, use the MASS normalization on the EIGR or EIGRL entry. If flutter eigenvectors are required, the PK or PKNL method should be chosen. 1-19859061 2005.5 Acoustics - RESTART from SOL106 into SOL111 - Prestress affects not included. If a modal frequency response acoustic analysis restarts from a geometric nonlinear analysis, the prestress affects are not included on the fluid. 1-19903501 2005.5 SOL600 - Cross section orientation of CBEAM/PBEAML is not correctly translated In SOL600, the cross section orientation of the CBEAM/PBEAML entry is not translated correctly and this leads to wrong answers. Please also see related quality 1-19558101. 1-19903564 2005.1 SOL600 - Torsional constant J of PBEAML entry is not tranlated correctly - Wring Answers In SOL600, the torsional Constant J of the PBEAML entry is not tranlated correctly and this will lead to wrong answers. 1-19914786 2005.1 Miscellaneous - INPUTT4 Fails to Read a Matrix written to a Binary Fortran File A binary fortran file that is written with the utility subroutines to produce and OUTPUT4 file cannot be read into MSC.Nastran with INPUTT4. The unformatted writes are writing more information than INPUTT4 will read. When attempting to read the file with INPUTT4, a message such as follows may be encountered: *** SYSTEM FATAL MESSAGE 6445 (INPTT4) READ REQUESTED RETURNED INCORECT DATA LENGTH WHILE PROCESSING UNIT = 23 FOR LOGICAL NAME INPUTT4 EXPECTED 1263 WORDS, ACTUALLY READ 2523 WORDS USER ACTION: VERIFY THE FILE ASSIGNMENT AND CONTENTS OR CONTACT MSC.SOFTWARE CLIENT SUPPORT. In Version 2005.1 the op4util has been extended to ignore the extra data (generating a warning) and a new option, fconvert, has been added that allows the conversion of a file, correcting the long record problem, while preserving the file's endian format. 1-19919211 2005.5 Connector Elements - CWELD - SFM 7558 Nastran CWELD elements may have a problem projecting the GS node on SHELL elements. This may lead to SYSTEM FATAL MESSAGE 7588. 1-19947997 2005.1 SOL 600 - CQUADX - Wrong Answers SOL 600 with axisymmetric CQUADXC elements does not correctly translate the constraints. This lead to failure or incorrect answers. 1-19948040 2005.5 Elements - PSHEAR with MAT1 and Nu=-1 gives traceback If a PSHEAR entry references a MAT1 entry with a value nu=-1, the job will fail with a traceback since formulation of the element contains the term (1+nu). 1-19948321 2005.5 External Superelements - CSUPER - Inefficient Restart If CSUPER elements are attached in an assembly restart run, then the downstream modes maybe un-necessarily recomputed. Avoidance: Insert the following DMAP alter in the assembly run cold start: compile extin alter 'putsys'(,-1) endif $ if ( setype='extrna' or exter ) then $ endalter 1-19969821 2005.1 DMAP - INPUTT2 and INPUTT4 generate scrambled erorr messages when an I/O error is encountered The INPUTT2 and INPUTT4 modules generate error messages using MSGPOP where the arguments do not match the text in ANALYSIS.TXT. Also, the order of the message lines in ANALYSIS.TXT may generate incomplete messages. 1-19978074 2005.1 Optimization/Design Sensitivity - p2rset on doptprm may fail in dom10 module with illegal input. SOL 200 job with p2rset specified on the doptprm entry may fail in the dom10 module with a fatal message that indicates illegal input to the domcb subroutine. Avoidance: Remove p2rset from the doptprm entry. 1-19980761 2005.1 Aeroelastic Postprocessing - PARAM,POST,0 and CAERO5 - SWM 6065 If PARAM, POST,0 is specified along with CAERO5 entries, then the following warning message may be issued and the xdb file corrupted. UNDEFORMED GEOMETRY RECOVERY FOR SUPERELEMENT 0 GRIDX 1 0 2 0 0 0 0 0 0 0 CAQ4 1 0 2 0 0 0 0 0 0 0 *** SYSTEM WARNING MESSAGE 6065 (DBC) *** DBCMSG ERROR MESSAGE 4 FROM SUBROUTINE CAQ4 ELEMENT 406 GRID POSITION 3 CONTAINS AN INVALID INTERNAL INDEX. LIMITS 1 0 2 0 0 0 0 0 0 0 EQEXING 1 0 2 0 0 0 0 0 0 0 EQEXINE 1 0 2 0 0 0 0 0 0 0 Avoidance: There is no avoidance. 1-19981679 2005.1 Plotting - OUTPUT(PLOT) - PLOT Module, Access Violation A large static analysis input file encounters an access violation in the PLOT module in version 2005.0.0. Job aborting... Signal caught, signal=11 job aborting User signal caught, errno=0 Command exited with non-zero status 1 Avoidance: Remove the OUTPUT(PLOT) request. 1-19990290 2005.5 SOL 600 - RBE3 Elements Incorrectly Translated SOL 600 may incorrectly translate RBE3 elements and this leads to wrong answers. 1-19992961 2005.1 Optimization/Design Sensitivity - PARAM,AUTOMSET - UFM 2101 (GP4) If PARAM automset is used in sol 200, the job will execute one design cycle and then fail on the update step before the next design cycle with a message similar to: *** USER FATAL MESSAGE 2101 (GP4) GRID POINT 23 COMPONENT 1 ILLEGALLY DEFINED IN SETS UM US User Avoidance: apply the following dmap alter: compile update alter 'gp4','' TYPE PARM,NDDL,CHAR8,Y,automset $ GP4 CASESX,GEOM4S,EQEXINS,SILS,GPDTS,BGPDTS,CSTMS, MEDGE,MFACE,,geom2s,gdntab,,gpsnts,geom1s/ RMG,YG1,USET0,gpsntn/ LUSETS/S,N,NOMSET/0/S,N,NOSSET/0/S,N,NORSET/ NSKIPX/0/0/0/0/0/ALTSHAPE/sebulk/automset/snormprt/ s,n,newgpsn $ 1-20009232 2005.5 Rotordynamics - UNBALNC - Null Load If the vector components (X1, X2, X3) on the UNBALNC entry aligns with the GRIDA and GRIDB direction defined on the RSPINT entry, then a null solution is obtained. This should result in a fatal error rather than continuing the execution. Avoidance: The avoidance is to correct the UNBALNC X1, X2 and X3 components. 1-20032521 2005.5 Distributed Memory Parallel - EIGRL, DOMAINSOLVER, STATSUB - UFM 9139 in PHASE0 DMAP checks for fmax on the eigr and/or eigrl entries do not take a static subcase (statsub) into account and the job may incorrectly fail with USER FATAL MESSAGE 9139 in PHASE0. Avoidance: Apply the following dmap avoidance. 2005r1: compile phase0 alter 34,49 $ 2005r2: compile phase0 alter 'IF ( (NFOUND\<0 OR FMAX', '' $ 1-20049736 2005.5 Buckling, Superelements - TEMP(LOAD), multiple subcases - Wrong answers If superelements exist in SOL 105 and multiple load cases are analyzed, the results may be incorrect. The results may vary depending on the order of the subcases. The use of superelements is not recommened in SOL 105 because the upstream shapes may not be correct. However, if superelements are used and multiple subcases exist the results should not depend on the ordering of the subcases. Avoidance: Apply the following dmap alter: compile sekdr alter 'setq','' case cases,/case1/'tran'/colno $ call setq case1//seid/peid/s,mtemp/s,k2gg/s,m2gg/ s,b2gg/s,mpc/s,spc/s,load/s,deform/s,templd/ s,p2g/s,dyrd/s,meth/s,mfluid $ 1-20058831 Thermal Analysis - Large Model, SFM 7796 (DFMSA) A large MSC.Nastran 2005 thermal analsyis encountered the following fatal message: *** SYSTEM FATAL MESSAGE 7796 (DFMSA) THE MAXIMUM INTEGER VALUE IS EXCEEDED. USER INFORMATION: THE MAXIMUM INTEGER NUMBER ALLOWED ON YOUR MACHINE IS 1073741823. HOWEVER, THE NUMBER OF REAL STACK ELEMENTS WAS COMPUTED TO BE 1084008517. USER ACTION: USE A DIFFERENT RE-ORDERING SCHEME VIA SYSTEM(206), OR REDUCE THE SIZE OF YOUR MODEL. Avoidance: Specify the iterative solver by setting NASTRAN ITER=1 in the first line of the input. 1-20066405 2005.5 Optimization/Design Sensitivity - MODE TRACKING and SCR=NO - SFM 4276, EC 3001 GREAD A SOL 200 job that includes mode tracking and eigenvector responses may fail in the DOM10 module with a message similar to: BIOMSG: ERROR 3001 HAS OCCURRED IN ROUTINE GREAD , FILE INDEX = 0. STATUS = 0 This error only occurs if SCR=NO is used. Avoidance: Run with SCR=YES or apply the following DMAP alter: compile desopt alter 'RESP3rg,rr2idrg',' ' DELETE /RESP3RG,rr2idrg,wgtmg,RQATABRG,modrprg $ 1-20081221 2005.5 Postprocessing - DBC, PARAM,POST,0 - XDB CBUSH object orientation vector is incorrectly translated The items within the CBUSH MSC.Access object contain incorrect values for the orientation information. 1-20088574 2005.1 Heat Transfer - View Factor Output truncated View factor output may be incorrectly truncated. The output below demonstrates the problem where the exponent in the last line of the FIJ column is incomplete. *** VIEW FACTOR MODULE *** OUTPUT DATA *** CAVITY ID = 1 *** ELEMENT TO ELEMENT VIEW FACTORS C* PARTIAL SURF-I SURF-J AREA-I AI*FIJ FIJ ERROR SHADING ERROR SCALE 100001 - 100002 6.5403E-03 1.36031E-05 2.07989E-03 2.1633E-03 NO YES 100001 100003 6.5403E-03 7.94674E-06 1.21504E-03 1.3356E-03 NO YES 100001 100002 6.5403E-03 1.32132E-05 2.02027E-0 240/256 1-20107532 2005.5 Postprocessing - XDB file missing when job is submitted to AIX 5 using "node" option from Windows MSC.Nastran 2005 Local Computer: Win2000 Remote Computer: AIX 5.1 Hostname = dcadsi03 When user submits job to remote computer: nastran test1.bdf node=dcadsi03 all Nastran output files are copied (rcp) back to the local WinPC except for the xdb file. The mscrsh.exe is being used on the local WinPC when the Nastran job communciates with the remote computer. This was verified by looking at the Task Manager on the local WinPC, and seeing the mscrsh.exe process. When sumbitting from Windows to a unix machine using the "node" option, the XDB file is not copied back to the Windows machine. The following warning message appears after submital: MSC.Nastran beginning job test1.T2368_12. The filename, directory name, or volume label syntax is incorrect. MSC.Nastran job test1@dcadsi03.vought.com completed. 1-20135551 Miscellaneous - param usetprt integer overflow ****** with more than 1 million DOFs An integer will occur in USETPRT output for inputs greater than one million DOFs. 1-20135658 2005.5 Dynamic Response - PBUSH, PSHELL - Incorrect structural damping GE Variable GE damping on the PBUSH or PSHELL entries may lead to incorrect damping and thus wrong answers in Version 2004 and Version 2005. Avoidance: Although the error is only likely to occur for large models, there is no avoidance other than not using this capability. The error has been fixed in version 2005.1.2. Requested platforms have been put to the Training and Support - Nastran Product Update web site. 1-20143203 2005.5 Frequency Response Analysis - PARTN - UFM 2170 A frequency response analysis, SOL 108 , with no null columns in the damping matrix (BJJX) may fail with USER FATAL MESSAGE 2170 in the PARTN module in the PHASE1A subDMAP. Avoidance: Insert the following DMAP alter: compile phase1a alter 'if ( nonull\<','' dbstatus vb//s,n,nullvb $ if ( nonull<>-1 and nullvb>-1 ) then $ alter 'if ( nonull\<'(2),'' if ( nonull<>-1 and nullvb>-1 ) then $ endalter 1-20160333 Optimization/Design Sensitivity - DOM9 - SFM 4276 (GREAD) Optimization jobs, SOL 200, with a large number of design variables and a large number of active/violated constrains may fail in the DOM9 module and SYSTEM FATAL MESSAGE 4276 may be issued. This error is the result of an incorrect memory allocation in DOT. Avoidance: There is no avoidance. 1-20164298 2005.5 Optimization/Design Sensitivity - PBEAM - Incorrect or Missing Punch Output When a SOL 200 input file includes PBEAM entries with the NO option, the updated PBEAM entries are not all written correctly to the .PCH file. Avodiance: There is no avoidance other than manually creating the PBEAM entries. 1-20166404 Transient Response - SOL 146, OLOAD - Missing Output OLOAD output is not supported in SOL 146 for transient reponse. Avoidance: To request OLOAD output for transient response in SOL 146, add the following DMAP alter: compile seaero alter 'delete \/ppf,psf','' delete /psf,,,, $ IFT ppf,CASES,TRL,FOL/ pxf,fol3/IFTM $ equivx pxf/ppf/-1 $ 1-20171152 2005.5 Aeroelastic Analysis - PERQ field on the AEFORCE bulk data entry is not supported The 2005 Quick Reference Guide indicates that field 9 of the AEFORCE entry can be used to specify that the associated data is PERQ (that is, per unit of dynamic pressure). This feature is not supported until Version 2005R3 (2005.5). Avoidance: Include the dynamic pressure factor in the loads and leave the PERQ attribute blank The following alter can be applied in the tier1 version of 2005 or 2006. I have not checked 2005r2 compile aestatrs list alter 'pavst'(2),'' type parm,,i,n,nopvperq $ dbstatus pvperq//s,n,nopvperq $ if ( nopvperq > 0 ) then $ $ partition the v-set into those to be scaled by Q and those $ that are absolute. Then, scale and remerge partn pavst,pvperq,/paast,,paqst,/1 ADD paqst,/qpaqst/CQ $ scale by q merge paast,,qpaqst,,pvperq,/pavstf/1 $ $ partn pgvst,pvperq,/pgast,,pgqst,/1 ADD pgqst,/qpgqst/CQ $ scale by q merge pgast,,qpgqst,,pvperq,/pgvstf/1 $ else $ equivx pavst/pavstf/always $ equivx pgvst/pgvstf/always $ endif $ call appndadb aedbstdx,uxvst,,,,pavstf,pgvstf/ mrgindx,mrguxv,mrgwjv,mrgfjv,mrgqkv,mrgqav,mrgqgv $ PARAML mrgindx//'TRAILER'/1/S,N,NvTERMS $ delete /pavstf,pgvstf,,, $ 1-20172731 2005.5 Complex Eigenvalue Analysis - CHACBR, EIGC - Wrong Answers, Primepower Only Models with acoustic barrier CHACBR elements may yield incorrect complex eigenvalues In MSC.Nastran 2005.1 on Fujitsu Primepower machines. 1-20184226 2005.5 Connector Elements - CWELD, TEMP(LOAD) - UFM 4016 A model which has cweld elements and temperature loading terminates with USER FATAL MESSAGE 4016 (GETEMP) Avoidance: A TEMPD must be defined. Temperature loading is not supported for cweld, cfast, and cwseam elements. The problems has been addressed by skipping the nodal temperature check for these elements. 1-20194247 2005.5 Optimization/Design Sensitivity - PBRSECT - Missing Punch Output An optimization analysis, SOL 200, with designed PBRSECT entries may fail to punch the complete PBRSECT entry. For some entries, only the outer loop definition was punched, the inner loop definition was missing. Furthermore, the outer loop definition was that of the original definition and not the optimized definition. Avoidance: There is no avoidance other than manually creating the entries. 1-20218686 2005.5 Loads - Combining load cases with references to non-existant loads - Wrong answers If a reference is made to a non-existant load the following warning message is issued. *** USER WARNING MESSAGE 3204 (SSG1) UNABLE TO FIND SELECTED SET ( 16064) IN TABLE (SLT ) IN SUBROUTINE (SSG1). However, if an attempt is made to combine non-existant load cases with existing load cases, wrong answer result. Avoidance: Remove the reference to the non-existant load cases. 1-20220161 2005.5 Frequency Response - DPHASE - Single Precision Loads A SOL 108 run with DPHASE data generates loads in single precision on short word machines. Loads should be generated with double precision accuracy. This may lead to results that do not contain the fidelity expected by the user. Avoidance: There is no avoidance. 1-20221811 2005.5 Distributed Memory Parallel Processing - Acoustic with Akusmod input, DMP=1 - SFM 3007 When an Acoustics job (with Akusmod input) is submitted with dmp=1 the job fails with the message: SYSTEM FATAL MESSAGE 3007 (DISUTIL) ILLEGAL INPUT TO SUBROUTINE DISUTIL Avoidance: DMAP avoidance for SOL 111 version 2005. compile read70 $ alter 'if ( nproc','endif $ nproc' $ if ( nproc>1 ) then $ if ( procid=1 ) then $ disutil agg0,,,////3 $ else if ( procid>1 ) then $ delete /agg0,,,, $ disutil ,,,,/agg0///3 $ endif $ endif $ endalter $ 1-20221988 2005.5 Nonlinear Analysis - CGAP, PUNCH, NLSTRESS - Missing Output in Punch File. If an NLSTRESS and STRESS output request is made in the same run, GAP FORCES are not written to the PUNCH file. Avoidance: Remove the NLSTRESS request form the output. 1-20222042 2005.5 Dynamic Response - PARAM,LFREQ, PARAM,SCRSPEC - Response spectrum SPCFORCES are incorrect Response Spectra analysis produces incorrect SPC forces when param,lfreq or the MODESELECT command is used. All other results are correct. Avoidance: Apply the following DMAP alter: compile postreig alter 'sdr1' $ equivx qgh/phqg/-1 $ 1-20240253 2005.5 Elements - CHEXA, TEMP(INIT) - Wrong Answers for Temperature Dependent Materials In a linear static thermal stress analysis with a temperature dependent material wrong results may be generated if - HEXA elements are present and a - TEMP(INIT) Case Control Command is used Avoidance: There are two avoidances: 1. Use a TEMP(MATERIAL) or TEMP(BOTH) Case Control Command instead of TEMP(INIT) and specify TREF on the material entry. 2. Setting NASTRAN SYSTEM(309)=1 also avoids the error and still allows the usage of TEMP(INIT). But this second avoidonce should not be used for highly skewed HEXA8. 1-20240331 2005.5 Optimization/Design Sensitivity - DRESP1 - Poor Message, SFM 3062 A SOL 200 optimization job with DRESP1 that references a missing GRID may fail with SYSTEM FATAL MESSAGE 3062 instead of a USER FATAL MESSAGE that identifies the missing GRID. Avodiance: Use PARAM,SOLADJC,-1 1-20258966 2005.5 Cyclic Symmetry - UFM 4319 on AIX 5.1 only SOL 114, cyclic symmetry, jobs may fail in Version 2005 on IBM AIX 5.1 with USER FATAL MESSAGE 4319. Avoidance: Run on a different platform or use an earlier version. 1-20275644 2005.5 Nonlinear Analysis - PCOMP, PARAM,COMPMATT - SFM 6352 (TDCINI) If a nonlinear analysis, SOL 106, contains a PCOMP entry with LAM="SYM" specified in field 9 and parameter COMPMATT is set to "YES" or "NONSMEAR" the analysis may fail with the message: *** SYSTEM FATAL MESSAGE 6352 (TDCINI) LOGIC ERROR: KERROR= 350. USER ACTION: CONTACT MSC.SOFTWARE CLIENT SUPPORT. PROGRAMMER INFORMATION: IPTR= 10916, ENDREC= 10900 PROGRAMMER INFORMATION: THE PARENT SUBROUTINE IS NLINIT OR MAIN: "Access violation" (C0000005) exception encountered. MAIN: For reference, A(MAINAL) = 4012FE, A(/SYSTEM/) = 1FD0A00 MAIN: A(/XNSTRN/) = 3DF0020 MAIN: Exception occurred at address 00C35C86. MAIN: Program attempted to access data at location E346AA1C. Avoidance: Other than removing PARAM,COMPMATT, no avoidance exists. 1-20277571 2005.5 Frequency Response - NLRGAP - SFM 3001 in the DPD module in XRANDOM subDMAP A SOL 108 deck with an unused NLRGAP entry in the input fails with SYSTEM FATAL MESSAGE 3001 in the DPD module in the XRANDOM subDMAP. Avoidance: Remove the unused NLRGAP entry. 1-20290197 2005.5 Enforced Motion - CBUSH - UFM 2170 The enforced motion example attached contains two CBUSH and two CONM2 elements. It used to run in nastran 2005r1 but is failling during restart in 2005r2 with the following message: USER INFORMATION MESSAGE 7559 (FRRD1A) THIS FREQUENCY RESPONSE ANALYSIS INVOLVES THE FOLLOWING TYPES OF EXCITATION -- ENFORCED DISPLACEMENT USING SPC / SPCD DATA USER FATAL MESSAGE 2170 (PARTN2) BOTH THE ROW AND COLUMN PARTITIONING VECTORS ARE PURGED AND ONLY ONE MAY BE. Avoidance: Insert the following DMAP alter into the restart run: compile semfreq alter 'freqdep='(,-1) putsys(0,109) $ alter 'freqdep=' putsys(1,109) $ 1-20375480 2005.5 Acoustic Analysis - MFLUID and param,vmopt,2 - Wrong Answers Starting in V2004r3, PARAM,VMOPT,2 gives wrong answers. Avoidance: Insert the following DMAP alter: compile mgen2a alter 'mpyad phsa,phh' equivx eh/lammat/-1 $ endalter 1-20375725 2005.5 Optimization/Design Sensitivity - DVPREL2 - SFM 3007 (DOM9) If a DVPREL2 (or DVCREL2 or DVMREL2) references more desvar id than there are desvar entries and optimization is performed, a non-useful fatal message is produced. This can also happen when the user calls out the same desvar id on a dvprel2 entry more than once. On windows, the job ends with: MAIN: "Access violation" (C0000005) exception encountered. MAIN: For reference, A(MAINAL) = 4012FE, A(/SYSTEM/) = 1FD0A00 MAIN: A(/XNSTRN/) = 3BF0020 While on AIX the job ends with the message: *** SYSTEM FATAL MESSAGE 3007 (DOM9) ILLEGAL INPUT TO SUBROUTINE DOM63 User Avoidance: Remove the duplicate reference to the same ID on the DVPREL2 entry or add dummy desvar entries to make the number of desvar's greater than the maximum number of DESVAR's called out on a DVxREL2. 1-20500621 2005.5 Elements - CPENTA, SDR2 - Poor Performance Large problems with large number of PENTA elements could experience significant performance issues in SDR2 element data recovery. Avoidance: The only avoidance is to use other elements or do not include the PENTA's in the output set unless necessary. 1-20523074 2005.5 Elements - CGAP, TEMP(LOAD) - UFM 4016 Temperature loading is not supported for CGAP elements. If temperature loading exists in a model with CGAP elements, NASTRAN incorrectly performs a check of the nodal temperatures of the grids attached to the gap elements too. If there is no temperature specified, Nastran terminates with USER FATAL MESSAGE 4016 (TA1ETD) indicating that there is no temperature data for the gap element. 1-20580861 2005.5 Large Models - SSG1 - UFM 6827 (PLOAD4D), Incorrect Memory Management If there is a large model with many grid points and elements, the job may fail with an inappropriate USER FATAL MESSAGE 6827 indicating a problem with geometry. Avoidance: Increase 'mem' on the Nastran command. 1-20609309 2005.5 Elements - PSHELL referencing MAT10 - No connection between structure and fluid If a PSHELL entry references a MAT10 entry, Nastran should issue a fatal message that indicates that a shell element cannot reference fluid materials. 1-20622051 2005.5 Composites - PCOMP - PRTPCOMP=0 does not work if LAM="SMEAR", Extra Output PRTPCOMP=0 [same as SYSTEM(361)=0] should supress the printout of equivalent PSHELLs and MAT2s for PCOMP entries. However, it does not work when LAM="SMEAR" on the PCOMP record. 1-20632561 2005.5 Acoustics - RBE1, RBE2, RBAR, RBE3 etc. - Wrong Answer if rigid elements attached to fluid dof If a rigid element (RBE1, RBE2, RBAR, RTRPLT, RSPLINE, RBR3, etc) is connected to a fluid degree of freedom, no message is given, and wrong answers occur. The rigid element family was not intended for fluid degrees of freedom; a fatal message should have been issued. Avoidance: Use MPC equations for fluid constraints. Always refer to component "1". 1-20647414 2005.5 Nonlinear Analysis - TEMPD - Floating Point Exception A nonlinear analysis, which includes nonlinear heat transfer, may fail with a floating point exception. The error has only been encountered on Windows and was the result of a poor guess for TEMPD. MAIN: The FPU has been reset after floating point exception. MAIN: For reference, A(MAINAL) = 401304, A(/SYSTEM/) = 21A4920 MAIN: A(/XNSTRN/) = 3A00000 MAIN: "Floating point overflow" (C0000091) exception encountered. Avoidance: Make a better guess on TEMPD for the initial temperature. 1-20752361 Optimization/Design Sensitivity - Poor Message, DRESP2 references a custom DRESP1 (MAX or RSS) If a DRESP2 entry uses Flag DRESP1 to reference a custom DRESP1 such as MAX or RSS DRESP1, then the job will fail with USER FATAL MESSAGE 7147. Avoidance: The message should be improved to direct the user to use Flag DRESP2 instead of DRESP1. 1-20780764 2005.5 External Superelements - EXTSEOUT - Excessive Disk Space Solution 101 requires excessive disk space when reducing an external superelement model to aset. The excessive disk space is required because the RFA matrix is unnecessarily created. Avoidance: Apply the folloiwng DMAP avoidance: compile extout alter 'call genrga'(,-1) if ( sdrs<>1 ) then $ alter 'call genrga' endif $ sdrs<>1 1-20798001 2005.5 Elements - PBRSECT, PBMSECT, EMG - Fatal Message If on the PBRSECT or PBMSECT entry a straignt segments, the job may fail in the element generation routines in the EMG module. Avodiance: Do not split a straight segement into multiple l segments. 1-20805090 2005.5 ACMS - CAABSF - Wrong Answers Nastran 2005r2 gives wrong answers when absorbers (CAABSF) are used together with MDACMS. This error only occurs for models which contain frequency dependent elements attached to fluid points. At this time only CAABSF elements have this feature Avoidance: Apply the following DMAP avoidance: compile fdrmgen alter 'partn mdotm,,vgfd','' add5 rph,mdotm,,,/mdotm1 $ partn mdotm1,,vgfd/,rgxfd,,/1 $ delete /mdotm1,,,, $ endalter 1-20868511 Documentation/IOG, QRG - SYSTEM(319) - Messages Documentation System cell 319 provides a mechanism to allow users to print additional information associated with warning, information and fatal message. The V2005 QRG contains a description of this system cell but it should also be included in Section 6.6 of the IOG. system(319) keyword=XMSG type=Integer This is used by MSGAPIC.C to switch on/off additional user/system information for messages output to the f06. SYSTEM(319) = 0 (Default) Do not show extra information. = 1 Show extra information. (Default=0) 1-20899028 Acoustic Analysis - ACMODL, ALLSET=yes - possible wrong answers In an acoustic analysis, if ALLSET is set to yes on the ACMODL entry, then the results may be incorrect. Avoidance: Remove ALLSET and use the default value. 1-20900111 2005.5 Nonlinear Analysis - MATS1, TABLES1 - UFM 2076 User fatal message 2076 may be encountered incorrectly if a model contains a very large number of material entries that reference temperature dependent tables. *** USER FATAL MESSAGE 2076 (TA1) TABLES1 BULK DATA ENTRY ID 52020, WHICH IS REFERENCED BY MATS1 ENTRY ID48490980, IS INCORRECTLY DEFINED FOR NONLINEAR ANALYSIS 1-20944681 2005.5 Access - DEMO1 component fails on some operating systems The MSC.Access DEMO1 program has shown a FATAL traceback on some platforms and operating systems. This error was due to libdbio dependencies not being in the makefile. The Platforms affected are Linux64 and IRIX64. 1-20947181 Job may fail when using K2PP with namelist and factors SOL 109 job may fail in multiple SUBCASE's model when there are more than one K2PP defined in the Case Control Deck and at lease one K2PP is using the format of namelist and factors. (M2PP and B2PP should have the same problem.) [P.S.] See the test decks and output in /nast/app/etl/q1-20947181/xk2pp.zip - 3 test decks included. 1-20954421 2005.5 MDACMS - F1 specified on EIGR/EIGRL may lead to Wrong Answers For MDACMS, if F1 is specified on the selected EIGR or EIGRL entry, and if F1 is less than 1.0, and if there are eigenvalues between F1 and 1.0 Hertz, then the eigenvalues in [F1,1.0] will be recovered twice. Avoidance: Do not specify F1 less than 1.0 for ACMS. 1-20979791 2005.5 Optimization/Design Sensitivity - DOPTPRM, PENAL- Incorrect dependent design variable constraints The DOPTPRM parameter PENAL can be used to transform an infeasible design task to a feasible one. If this is used with the DOT optimizer and there are dependent design variables, the constraints on the first dependent design variable as seen by DOT are incorrect. This error should not affect the results unless the dependent design variable is close to or exceeding its bounds. Avoidance: Use the MSCADS optimization algorithm in place of DOT 1-21038768 2005.5 Nonlinear Analysis - RIGID=Lagrange Fails in SOL 106 with SFM 4017 from NLCOMB Lagrange rigid element RBE2 is not available for SOL106. If used the following fatal message is encountered in the NLCOMB module *** USER FATAL MESSAGE 4017 (NLCKDT) ESTNL AND FILE (103) ARE NOT PROPERLY ALIGNED *** SYSTEM FATAL MESSAGE 3007 (NLCOMB) ILLEGAL INPUT TO SUBROUTINE NLCKDT In Version 2005R3 as documented in the QRG, the RIGID=LAGRANGE command will be ignored in SOL 106 and RIGID=LINEAR will be used. Avoidance: A possible avoidance is to replace the RBE2 elements with stiff CBEAM elements. Else use SOL 400 in MD Natran. 1-21115634 2005.5 Postprocessing - ASSIGN FORM=BIGENDIAN or LITTLEENDIAN - UFM 659 (IFP1PO) If a run has multiple subcases and multiple assign statements that specify a non-native form of the output file, i.e. bigendian on a little endian machine or littleendian on a bigendian machine, the job may fail. For example the following input leads to the fatal message below: ASSIGN output2 = '12_little.op2', UNIT = 12,form=bigendian ASSIGN output2 = '61_little.op2', UNIT = 61,form=bigendian ASSIGN output2 = '62_little.op2', UNIT = 62,form=bigendian SOL 103 CEND post patran tofile 12 SUBCASE 1 METHOD = 1 $ modes=5 disp(plot)=all post displace tofile 61 SUBCASE 6 METHOD = 2 $ modes=10 stress(plot)=all post stress tofile 62 *** USER FATAL MESSAGE 659 (IFP1PO) THE FORTRAN LOGICAL UNIT REFERENCE NUMBER 61 ON THE ABOVE POST COMMAND REFERENCES A FILE THAT HAS A FORM=FORMATTED ATTRIBUTE. A PREVIOUS POST COMMAND SPECIFIES FILE UNIT 12 FOR FORMATTED OUTPUT. THE POST COMMAND DOES NOT ALLOW A FILE UNIT CHANGE FOR FORM=FORMATTED FILES. 1-21221351 2005.5 Optimization/Design Sensitivity - temp(load) and buckling response may have incorrect sensitivity Sensitivity coefficenet for a SOL 200 job may be in error if the following conditions are met. a. has buckling eigenvalue as constraint b. has temp(load) in static subcase c. has 'PARAM,DSNOKD,1.0' Aovoidance : attached following DMAP alter compile response $ alter ' bdiag,budg,dkdgg,budg'(,-2),''(,-2) $ EMG EST,CSTMS,MPTX,DIT,,ugb,ETT, EDT,,,BGPDTS,GPSNTS,,,,,/ KDELMDg,KDDTDg,,,,/ 1/0/0/-1/-1/-1/ 0 /-1//////1 $ 1-21243388 2005.5 Optimization/Design Sensitivity - DESVAR - SFM 3001 A SOL 200 job will fail with system fatal message 3001 if it has a DESVAR command in the Case Control section referencing a SET ID to select design variables and that SET covers all DESVAR entries in the input deck. Avoidance : Either use DESVAR = ALL or remove DESVAR in case control 1-21249491 2005.5 Executive System - RESTART - SFM 1013 SYSTEM FATAL MESSAGE 1013 may occur in a restart if the solution sequence is large enough to exceed an array size. For example, the index to FREDBD in common block DBFREX is out of bound. That array is sized to 300, but in OPNDBS, the FREPTR is calculated to be 303 (outside the array bounds). Avoidance: Execute the job as a cold start. 1-21251041 2005.5 Transient Analysis - DDRMM, SORT2, SPARSEDR=NO - Missing Output for multiple subcases If PARAM,SPARSDR,NO is set, the second and subsequent subcases in a transient analysis will not produce SORT2 output. 1-21289481 2005.5 installation - linux - Missing Libraries - LD_PRELOAD LD_PRELOAD was set on Linux systems in 2005.1. This setting may cause a message in the log file about missing libraries. Avoidance: Replace the following line in $MSC_BASE/bin/nast20051.lcl and $MSC_BASE/bin/nast20051.dmp: %MSC_ARCH=linux:% export LD_PRELOAD with %MSC_ARCH=linux:%# export LD_PRELOAD 1-21298401 2005.5 Postprocessing - CQUADR/CTRIAR center format not supported by DBC The CQUADR/CTRIAR center format was not supported by DBC 1-21304451 2005.5 MPYAD - Arithmetic Exception - machine dependent failure The MPYAD module may fail for large problems in which the first matrix has a large number of rows and a small number of columns. This error only occurs when a very large memory is specified. Avoidance: Reduce the memory request. 1-21317981 2005.5 AUTOMSET - UFM 9163 (SEKR0) Models that contain redundant, i.e over constrained, rigid elements will fail with USER FATAL MESSAGE 9163 if PARAM,AUTOMSET,YES is specified. An avoidance is available in the following DMAP alter: compile sekr0 alter '9163', 'RETURN' $ message //' user'/wrn/'message 9163 (sekr0)' $ message //' The following mpc and/or rigid element '/ ' equations are redundant and will be discarded ' $ matgpr bgpdts,uset1,,rax//'h'/'a' $ endalter 1-21318011 2005.5 AUTOMSET - Wrong Answers for Multiple BC If PARAM,AUTOMSET,YES is specified for any job with multiple boundary conditions (BC), results for the second and subsequent BCs will be wrong. An avoidance is available via DMAP alter: compile sekr0 alter 'DBSTATUS knn'(,-1) delete /knn,,,, $ 1-21345191 2005.5 Optimization/Design Sensitivity - FLUTTER, Superelements - EOF on vtqu in dsfltf. SOL 200 job with flutter response and design model covers more than one upstream SE will fail in DSFLTF module with EOF on vtqu. This problem does not affect SOL 200 job with residual only or design models that span only one upstream SE. Avodiance : use following alter to avoid the problem ompile respsen,list $ alter 'lpflg=0' $ type parm,,i,n,lpflgcu $ lpflgcu=0 $ alter 'call flutsens','' $ IF ( DOFLUTDG>0 ) then $ lpflgcu = lpflgcu + 1 $ call flutsens, KDICTDSF,KELMDSF,BGPDTS,SILS,CSTMS,KDICT, KELM,DSPT1,MDICTDSF,MELMDSF,MDICT,MELM,EST, MPTS,DIT,ett,GPSNTS,BDICTDSF,BELMDSF, KDICTDSB,KELMDSB,MDICTDSB,MELMDSB,BDICTDSB, BELMDSB,CASEM,EDT,PHDH,EMAP,CASEXX,SLT, MAPS,USET,GOAT,GOAQ,GM,CASES,KHH,BHH,MHH, QHHL,FLUTAB,R1TABR/ DELFL,VDKU,cdelk,cdelm,cdelb,dug,dvg/ LPFLGcu/DMRESD/SEID/RSONLY/CDIFX/NOPSLG/ IMAGE/shapeopt $ endif $ 1-21368531 DMP - Installation - Scyld Clusters Supporting Scyld Linux clusters requires minor changes to the installation and submittal script. The necessary information is documented below. Preparing the Computer Nodes Nastran needs to have a small set of binaries available on the compute nodes. The small script shown below will handle copying the files to the compute nodes. Place this script into /etc/beowulf/init.d and reboot the compute nodes. You can also run this script by hand, as root, and passing it a single integer arugmentas the compute node you want to prepare. #!/bin/bash [ "$NODE" == "" ] && NODE=$1 if [ "$NODE" == "" ]; then echo "No node specified" exit 1 fi # Copy binaries out to node for i in /bin/date /bin/cat /bin/grep /usr/bin/time /bin/sed /bin/uname /bin/ksh do bpcp ${i} ${NODE}:$i done Configuring Nastran Several things must be done to allow Nastran to transparently run serial jobs on the compute nodes. First we work around a bug in KSH shell. Change the file $NAS_HOME/bin/nast2005 near the top From: # Pick up the command name now. # cmdname=${0##*/} arg0=$0 To: # Pick up the command name now. # #cmdname=${0##*/} cmdname=nast2005 arg0=$0 Now do the same thing for $MSC_HOME/bin/msc2005 and $MSC_HOME/bin/nastran substituting when needed. From: # Pick up the command name now. # cmdname=${0##*/} arg0=$0 To: # Pick up the command name now. # #cmdname=${0##*/} cmdname=msc2005 arg0=$0 Nastran now must be told how to send jobs off to the compute nodes. Since Scyld's Bproc system handles mapping and migrating processes to compute nodes, we must tell Nastran to use the beorun facility. Edit the global Nastran configuration file in $MSC_HOME/conf/nast2005rc and add that begins with tcmd, as shown below: auth=1700@target sdirectory=/tmp buffsize=8193 memory=estimate tcmd="beorun -no-local" Now jobs will be automatically migrated out to the least busy compute node. The -no-local command to beorun tell the system to never attempt to run any job on the master node. In this configuration it is NOT necessary to have the RSH commands running on the compute nodes. The beorun command handles starting the process remotely rather then RSH. Distributed serial jobs should now work as expected. Configuring Nastran for DMP parallel jobs 1) LAM versions are not binary compatible. Scyld ships with version 7.1.2 while Nastran expects version 7.0.4. In order to use LAM under Scyld you will need to apply the Scyld patch to the LAM source tree. Both the source for LAM 7.0.4 and the patch file are available from Scyld Support. Build and install LAM $>cd /usr/src 4>tar -zxvf lam-7.0.4.tar.gz $>cp lam-bproc-vexecmove.patch /usr/src/lam-7.0.4 $>patch -p1 < lam-bproc-vexecmove.patch $>./configure --prefix=/usr/local/lam/7.0.4 $>make; make install The version information should read: LAM 7.0.4/MPI 2 C++/ROMIO/bproc - Indiana University 2) Nastran uses a templating engine to determine how to start itself up. You'll need to alter this template in order to correctly adjust for your installation of LAM. Edit $MSC_HOME/msc2005/linux8664/nastran.dmp and about line number 229, change the setting of LAMHOME to reflect that PATH to LAM. The line should read: %MSC_ARCH=linux:% export LAMHOME="/usr/local/lam-7.0.4" and slightly below that at about line 235 make sure the PATH is properly set. The line should read: %MSC_ARCH=linux:% export PATH=${PATH}:/usr/local/lam-7.0.4/bin the last change is to make sure that if localhost is not in the hostfile for LAM, that Nastran properly adds "localhost" rather then the output of "hostname" to the hostfile. So at about line 224 Change: %MSC_ARCH=linux:% hostname >> $hostfile To: %MSC_ARCH=linux:% echo "localhost" >> $hostfile 3) Setup your environment. You'll want to set MSC_BASE, and LAMHOME appropreatlye and add $MSC_HOME:/bin to your $PATH. It is a good idea to add these to your .bashrc or something similar. $> export MSC_BASE=/mnt/raid/apps/nastran $> export LAMHOME=/usr/local/lam/7.0.4 $> export PATH=$PATH:/mnt/raid/apps/nastran 4) Start Nastran $>msc2005 nastran -n dmp=4 hosts=.0:.1:.2:.3 5) Run the resulting script. $>./lgqdf.T22479_38 Running Distributed Serial Jobs If you have configured Nastran to run serial jobs using the section above, then Nastran will run as expected. It is of importance to note that there is no need to specifiy a node= to Nastran because Scyld will automatically run the job on whatever node is least free. An example of a valid command is: $> msc2005 nastran vcdp.dat dbs=/tmp Running DMP Parallel Jobs In order to run Nastran DMP jobs, you must have carefully followed the section on configuring Nastran for DMP jobs in the above section. Unlike serial jobs, at this time you still must tell Nastran what nodes you wish to run your job on. An example of a vaild command for DMP jobs is: $>msc2005 nastran -n lgqdf.dat dmp=4 hosts=.0:.1:.2:.3 Known Issues and Limitations It is not possible to run Nastran in DMP mode without the -n option. A wait_pid error is generated. Automatic selection of nodes is not currently available with Nastran's DMP mode 1-21377192 2005.5 Optimization/Design Sensitivity - DOM9 - SFM 4276 EC 3021 When an optimization job includes DVPREL1 entries with a single DESVAR entry and PMIN and PMAX is specified on some but not all of the DVPREL1s than the job may fail with SYSTEM FATAL MESSAGE 4276 EC 3021 in the DOM9 module. Avoidance: 1) Use ADSCOD = 1 on the DOPTPRM or 2) Either leave all PMIN/PMAX fields blank on the DVxREL1 entry and reference a single DESVAR or add a non-zero number to all PMIN/PMAX fields of DVxREL1 entries. 1-21389158 2005.5 Postprocessing - RTRPLT, PARAM,POST,0 - RTRPLT missing from xdb RTRPLT information is missing from the XDB. 1-21431967 2005.5 Model Checkout - Singularities not identified, UFM 9050 Not Issued on Windows PC MSC.Nastran 2005r2 (version 2005.1.0) on Windows may fail to identify singularities and not issue USER FATAL MESSAGE 9050. The analysis may procede and issue very bad answers that contain rigid body motion. Avoidance: The error has been corrected in Version 2005.1.2 (a.k.a 2005r2a). Either run the problem in Version 2005.0.0 or get the latest release. The updated 2005.1.2 Windows analysis.exe executable is available online at: http://www.mscsoftware.com/support/software_updates/index.cfm#nastran 1-21462191 2005.5 SOL 600/Marc Executables - 2005R2 Linux/Itanium CD - RHEL 3.0 - Memory Error The SOL 600 marc executables on the MSC.Nastran 2005r2 Linux/ Itanium CD will not work on Red Hat Enterprise Linux (RHEL) 3.0. They will work with RHEL 2.1 A Marc memory allocation error will appear in the log files when trying to run a Nastran SOL 600 analysis using RHEL 3.0. Avoidance: Download the file linux64_rhel3.tar.gz from MSC ftp site. This zip file contains the SOL 600 marc binaries for the RHEL 3.0 / Itanium platform. 1. unzip and untar the file in the /msc/nastran/msc20051/marc directory. It will create a linux_rhel3 directory. 2. Rename the current linux64 directory to linux64_rhel2. The Marc binaries in the current linux64 were valid for RHEL 2.1 3. Rename linux64_rhel3 to linux64. Now, the default Marc binaries are for RHEL 3.0 The problem with the wrong Marc executables for RHEL 3.0 / Itanium also applies to the Nastran 2005 installation CDs. This zip file will be posted at the following site. http://www.mscsoftware.com/support/software_updates/index.cfm#nastran 1-21475421 2005.5 Optimization/External Superelements - EXTSEOUT - SFM 4276 (PACK), EC 105 When a shape optimization job includes Part External SE, it will fail in DOPR2 at 2nd design cycle with the following error message: SYSTEM FATAL MESSAGE 4276 (PACK) ERROR CODE 105 PID= 0 Avoidance: Apply the following dmap alter. compile desopt alter 'desiter=desiter+1' luset = luset0 1-21570101 2005.5 Optimization/Design Sensitivty - EXTSEOUT - SFM 6551in EXTOTMIN module A SOL 200 job employing an external superelement created by using EXTSEOUT with a DISP request fails with SYSTEM FATAL MESSAGE 6551 while reading the OP2 file in the EXTOTMIN module. Avoidance: The same job runs successfully to completion if the external SE is created without the DISP request. 1-21572991 2005.5 Superelements - EXTSEOUT for fluid - SFM 5423 If the EXTSESOUT command is used to create an external superelement with fluid elements and fluid grid points exist on the boundary the assembly may terminate in the MPYAD module with SFM 5423. Avoidance: There is no avoidance. 1-21602236 2005.5 Aeroelastic Analysis - Flutter Analysis - UFM 3005 (FA1) The PK and PKE flutter methods use an eigensolution which requires an invertible mass matrix. This may lead to the following fatal message: *** USER FATAL MESSAGE 3005 (FA1) ATTEMPT TO OPERATE ON THE SINGULAR MATRIX MHH IN SUBROUTINE PKINIT 1-21699761 External Superelements/Rigid Elements - RBE3, RSSCON, AUTOMSET - Wrong Answers An external superelement model with RBE3 and RSSCON elements that specifies PARAM,AUTOMSET,YES will produce incorrect results. Avoidance: Remove PARAM,AUTOMSET,YES. 1-21724661 2005.5 Enforced Motion - SPCD - UWM 4152, Wrong Answers If more than one degree of freedom is specified on a SPCD entry and USER WARNING MESSAGE 4152 (G4SPCDR) is issued, wrong answers may result. Avoidance: Only specify one dof on the SPCD entry. 1-21737131 2005.5 Distributed Memory Parallel/ACMS - FLSPOUT - Wrong Answers / Missing Output Incomplete FLSPOUT (Fluid/Structure Panel Participation) output is produced when MDACMS is used with Distributed Memory Parallel (DMP). There are two avoidances: 1. Run in serial; do not use DMP for MDACMS with FLSPOUT. 2. Employ the following DMAP alter: compile freqrs alter 'if ( acms and fdfreqr','' if ( (acms or mdacms) and fdfreqr and nproc>1 and nfreqt>1 ) then $ endalter 1-21764817 2005.5 Optimization/Design Sensitivity - SFM 3007 (DSABO) ILLEGAL INPUT TO SUBROUTINE DSAB1I When an optimization job designs a temperature independent material property while specifying temperature dependent material property entrIes, MATT9 or MATT2, it will fail with the following error message: *** SYSTEM FATAL MESSAGE 3007 (DSABO) Avoidance: Do not include MATT9 or MATT2 entry in the job. Use MATT1, MATT3 or MATT5 instead. 1-21857794 2005.5 Acoustic Analysis - CHEXA, MAT10 - Incorrect complex acceleration and pressure Acoustic analysis produces incorrect complex acceleration velocities and pressure levels. 1-21857803 2005.5 Topology Optimization - PCOMP - UFM 2025 TA1 A topology optimization job with composite elements encounters a USER FATAL MESSAGE 2025 in the TA1 module, stating that there is an UNDEFINED COORDINATE SYSTEM ID = 3000002 even though a coordinate system with that id is present in the input file. Avoidance: Other than removing the composite elements, there is no avoidance. 1-21860501 2005.5 Restarts - PSD - Wrong Answers The psd response calculated in a RESTART may be incorrect in Version 2005.1. Avoidance: In the RESTART run apply the following DMAP alter: ECHOOFF $ COMPILE SEMFREQ NOLIST $ ALTER 'IF ( NOUHF\<1','' $ IF ( NOT(NOPH2) AND EXTDR='NO' ) THEN $ ENDALTER $ ECHOON $ 1-21869596 2005.5 Topology Optimization - Multiple Subcases - UFM 2025 (TA1), UWM If there are CQUADx/CTRIAx referencing MCID the job may fail with one of the messages below: *** USER FATAL MESSAGE 2025 (TA1) UNDEFINED COORDINATE SYSTEM ID = 3000002 *** SYSTEM WARNING MESSAGE 2072 (SEP1X) CARD TYPE 14101 NOT FOUND ON DATA BLOCK. *** USER FATAL MESSAGE 2010 (TA1) ELEMENT 201 REFERENCES UNDEFINED PROPERTY ******** Avoidance: Either use a single subcase or apply the following dmap alter: compile predom alter 'ta1 ,,ec', '' ta1 ,,ect,ept,bgpdt,sil,,CSTM,,,/ estx,,,gpect,,,,,/ luset/-1/0/1/0/0 $ 1-21892107 2005.5 Case Control - GPFORCE SFM 2362 (SETA2L) SET Spill AREA HAS BEEN EXCEEDED. The following Error message was issued in V2005.0 in a run with more than 1200 subcases and GPFORCE requests in each subcase. *** SYSTEM FATAL MESSAGE 2362 (SETA2L) SET Spill AREA HAS BEEN EXCEEDED. USER ACTION: CALL MSC CLIENT SUPPORT. *** USER WARNING MESSAGE 2354 (GPFDR) MODULE IS UNABLE TO CONTINUE AND HAS BEEN TERMINATED DUE TO ERROR MESSAGE PRINTED ABOVE OR BELOW THIS MESSAGE. THIS ERROR OCCURRED IN CODE WHERE THE VARIABLE -NERROR- WAS SET = -99 Avoidance: Add a SUBCASE with a different set request for any GPF output. 1-21895337 2005.5 Optimization/Design Sensitivity - MAT9 - Incorrect design sensitivity coefficients Design sensitivity coefficients for MAT9 entries are incorrect. 1-21901601 2005.5 Aeroelastic Analysis - DMIG, AEPRESS - SFM 4276 (FREAD) If an aeroelastic analysis has an AEPRESS entry that references an undefined DMIG entry than the job will fail with SFM 4276 in the FREAD module. Avoidance: Correct the input. 1-21957121 2005.5 Adams Integration - ADAMSMNF, AUTOQSET - UFM 1126 (FNFIST) A SOL 103 run that combines param,autoqset with a ADAMSMNF FLEXBODY request will fail with the message: *** USER FATAL MESSAGE 1126 (GNFIST) DMAP MODULE GP4 IS ATTEMPTING TO OUTPUT DATABLOCK 202 NAME = YG , WHICH ALREADY EXISTS. 1-22012914 2005.1 Installation - nastran command - analysis.8664.exe Not Found The nastran command delivered with the MSC.Nastran 2005r2a Windows cd will attempt to find the non-existant analysis.8664.exe on em64t and opteron systems running Windows XP64. When MSC.Nastran version 2005.1.2 is executed on a Window XP64 system it fails to find the file, analysis.8664.exe, and the run terminates. An updated delivery, MSC.Nastran 2005r2b, is available on the internal ESD site and a copy will be made and sent on request. Avoidance: Although the executable for the XP64 system is not currently available, the 32 bit executable should run by using the EXE keyword to point to the analysis.exe executable or renaming the analysis.exe to analysis.8664.exe. 1-22013693 2005.5 Case Control - CMSENERGY - UFM 601 The CMSENERGY command must be abbreviated with 4 charachters, i.e. CMSE, else the job will fail with USER FATAL MESSAGE 601. 1-22050590 2005.5 Optimization/Design Sensitivity - SFM 3002 (DOM9) An optimization job may fail in DOM9 with SYSTEM FATAL MESSAGE 3002 with and "END OF FILE" error message if the following conditions are met: 1) Either DVCREL1 or DVMREL1 is present and it references more than one design variable or it specifies non-zero PMIN/PMAX; 2) DVxREL2 is present (or is internally spawned by designed BEAM properties), Avoidance: 1) Replace DVCREL1 or DVMREL1 with DVCREL2 or DVMREL2 or 2) Remove PMIN/PMAX from DVxREL1 Shenghua, It would be nice if we could give better message. - Raj optimisation run wuth dvcrel entries gives *** SYSTEM FATAL MESSAGE 3002 (DOM9) EOF ENCOUNTERED WHILE READING DATA BLOCK SCRATCH (FILE 330) IN SUBROUTINE SKPREC If you replace the DVCREL1 entries with DVCREL2 entries and a linear equation the job runs ok. If there is only one design variable on the DVCREL entries it also runs ok. Example file is attached. 1-22053652 Transient Analysis - OTIME, GPFORCE - Missing Output, Extra Output OTIME in transient (direct or modal) does not work correctly for GPFORCE output request. When fewer time steps than exist are specified, all timesteps are output. If more time steps are specified than exist, gpforce results are output for the first time step only. 1-22059227 2005.5 Connector Elements - CWELD - Missing element force output If CWELD elements connect points defined using a local coordinate system, then element force requests for the element will be ignored. 1-22099431 2005.5 Dynamics - PARAM,AUTOQSET,YES and PARAM,POST,0 - Missing data in xdb When PARAM,AUTOQSET,YES is used then any grid/scalar point results will be incomplete on the xdb file. The following warning message will be issued. *** SYSTEM WARNING MESSAGE 6065 (DBC) *** DBCMSG ERROR MESSAGE 22 FROM SUBROUTINE GRIDBK DATA BASE READ RELATION PROBLEM FOR EQEXING RETURN CODE = 0. Avoidance: Insert the following DMAP alter: compile extduset alter 'call augqset' TYPE PARM,NDDL,I,N,CP,DBCPATH $ TYPE PARM,NDDL,CHAR8,Y,DBCCONV,DBCOVWRT $ TYPE PARM,NDDL,I,Y,GEOMU,LOADU,POSTU,DBCDIAG,POST $ if ( post=0 ) DBC BGPDTS,GPLS,GEOM1S,EQEXINS,,,,,,,,,,,,,,// 'BGPDT'/'GPL'/'GEOM1'/'EQEXIN'///////////////// -1/DBCPATH/S,N,CP/'reig'/0/GEOMU/LOADU/ POSTU/DBCDIAG/DBCCONV/DBCOVWRT/0 $ endalter 1-22164401 2005.5 Connector Elements - CWELD elements with PARTPAT option connect shell elements with wrong PSHELL IDs If the shell element that includes the projection point GA or GB is connected to several shell elements with different PSHELL IDs, the CWELD elements with PARTPAT option may connect shell element with wrong PSHELL IDs. 1-22166731 2005.5 Optimization/Design Sensitvity - SFM 3008, Insuffucient Memory in DSAB1J A large optimization job that has many designed elements may fail in DSAB1J with a fatal error indicating insufficient memory. Avodiance: The amount of memory required has likely been overestimated but the only avoidance is to increases the memory. 1-22194181 2005.5 Optimization/Design Sensitvity - Static Superelements may not utilize adjoint load method Adjoint load method may not be utilized in Statics SOL 200 job with SEs. The jobs will still run to the completion with correct results. The impact of this problem is, if any, in the performace. Avoidance: insert following alter $ compile descon,list $ alter ' if ( doeigv\>0 )'(2,-1) $ IF ( APP='STATICS' ) THEN $ TCOLADJS = TCOLADJS + COLADJ $ IF ( TADJC<0 ) THEN $ GADJFLGS = 0 $ ELSE IF ( ADJFLG>0 AND GADJFLGS=0 ) THEN $ GADJFLGS = ADJFLG $ ENDIF $ TADJC<0 COLADJ = 0 $ TADJCS = TADJC $ ENDIF $ APP='STATICS' $ 1-22195429 2005.5 Multipoint Constraints & Rigd Elements - RSSCON, MPC - SMF 2104 (GP4) A SOL101 model with MPCs/SPOINTs and RSSCONs may fail with the message: *** SYSTEM FATAL MESSAGE 2101 (GP4) GRID OR SCALAR POINT ID= 300740 COMPONENT = 0 FOUND ON THE MPC ENTRY IS ILLEGAL. USER ACTION: CHECK MPC ENTRIES. 1-22266681 Distributed Memory Parallel - GDMODES, MDMODES - SFM 6144 (MERGE1), Job Hangs GDMODES is not designed to support component mode reduction with external superelements. If attempted the job will fail with SYTEM FATAL MESSAGE 6144 (MERGE1). MDMODES should work but may hang. Avoidance: Run without DMP. 1-22331531 2005.5 External Superelements - SFM 3200 in BDRYINFO module An external SE creation run with no T-set DOFs, but only Q-set DOFs, terminates with SYSTEM FATAL MESSAGE 3200 (Logic Error 11) in the BDRYINFO module. 1-22398781 2005.5 DMAP - MATMOD option 47 - SFM 4276 If the input matrix to MATMOD option 47 is null then the module will fail with SYSTEM FATAL MESSAGE 4276. Avoidance: Check for a null matrix using the PARAML module before executing this option. 1-22405031 Different from CBAR versus beam orientation I use the workshop provided in the nas101 course file (Workshop number 2). In this whorshop only element number 12 is a CBAR. In order to orient the CBAR we have physically two options (1,0,0) or (0,0,1). These 2 orientations should give the same result BUT I get different result for displacement. The PBARL is used to give the bar property. So I checked the 2 pbar card in the F06 file, and there are the same. For me the 2 files should give the same result. See soln2_X.bdf and soln2_z.bdf file in the attachement of this quality. 1-22407911 2005.5 SOL 600 - NSTATE of NLAUTO entry is not translated to .marc.dat file correctly with DYNAMIC The value of the NSTATE field of NLAUTO is not translated to the .marc.dat file in Transient Dynamic Analysis. 1-22408855 Optimization/Design Sensitivity - PBARL, PBEAML, DVPREL1 - SFM 3007 (DOPR1) When an optimization job designs either PBARL or PBEAML and the DVPREL1 references a non-existing PBARL or PBEAML entry, the job fails in DOPR1 with SYSTEM FATAL MESSAGE 3007 with the following message: "Illegal Input to Subroutine DOPR1C" Avoidance: Do not reference non-existent entries. 1-22416502 Transient Analysis - Enforced Motion, SPCD, PARAM,MODACC,0 - SFM 3001 in MODACC A SOL 112 deck with SPCD enforced motion and PARAM,MODACC,0 fails with SFM 3001 in MODACC module as indicated below. *** SYSTEM FATAL MESSAGE 3001 (MODACC) THE INPUT DATA BLOCK NAMED PDF0 AT POSITION 4 DOES NOT EXIST. USER INFORMATION: THIS ERROR IS CAUSED BY ONE OF THE FOLLOWING: 1. THE DATA BLOCK WAS NEVER CREATED. 2. THE DATA BLOCK WAS DELETED. 3. THE DATA BLOCK IS SPECIFIED ON A TYPE STATEMENT IN THE CURRENT SUBDMAP OR A HIGHER SUBDMAP BUT THE CURRENT QUALIFIER VALUES DO NOT MATCH THE QUALIFIER VALUES OF THE DATA BLOCK(S) ON THE DATABASE. USER ACTION: 1. IF YOU ARE EXECUTING AN MSC.SOFTWARE SUPPLIED SOLUTION SEQUENCE AND NOT USING THE ALTER EXECUTIVE CONTROL STATEMENT, THEN CHECK FOR BULK DATA AND/OR CASE CONTROL INPUT ERRORS. IF NO ERRORS CAN BE FOUND, THEN CONTACT MSC.SOFTWARE CLIENT SUPPORT. 2. IF YOU ARE EXECUTING A DMAP PROGRAM NOT SUPPLIED BY MSC.SOFTWARE, THEN FOR DEBUGGING PURPOSES INSERT STATEMENT DIAGON(20) BEFORE THE MODULE SHOWN ABOVE. PROGRAMMER INFORMATION: THE FIST NUMBER IS 104 AND SUBROUTINE IS MODACC 0FATAL ERROR 1-22445937 2005.5 Connector Elements - PFAST - UFM 2025 (MODGM2), Undefined Coordinate Cystem If a large number is used for the MCID that is referenced on the PFAST entry the job may with USER FATAL MESSAGE 2025 being issued from the MODGM2 module indicating an undefined coordinate system. Avoidance: User smaller ids for the coordinate system. 1-22461444 2005.5 Eigenvalue Analysis - MEFFMASS output request - UFM 5423 An MEFFMASS request for an analysis that produces very large, very sparse matrices by fail with USER FATAL MESSAGE 5423. The model that experienced this fatal message had a density on the order of 1.0e-6. 1-22502406 2005.5 MCFRACTION - Fluids - SFM 3007 - ILLEGAL INPUT TO SUBROUTINE MCFRAC If MCFRACTION is used for a model that includes fluid nodes, the modal contribution output is missing, and the following error message is issued: *** SYSTEM FATAL MESSAGE 3007 (MCFRAC) ILLEGAL INPUT TO SUBROUTINE MCFRAC Avoidance: Remove the MCFRACTION request. 1-22539420 2005.5 Postprocessing - PARAM,POST,0, DBC - Missing Output in xdb, SWM 6065 DBCMSG 14 SOL106 restarts using 'PARAM,POST,0' produces an .xdb file that is missing output. The f6 output contains the following warning messages: In the run that produces .xdb nastran gives this SYSTM WARNING MESSAGE, *** SYSTEM WARNING MESSAGE 6065 (DBC) *** DBCMSG ERROR MESSAGE 14 FROM SUBROUTINE DBCROL UNABLE TO OPEN INPUT RELATION NLLF RETURN CODE = 1. *** SYSTEM WARNING MESSAGE 6065 (DBC) *** DBCMSG ERROR MESSAGE 14 FROM SUBROUTINE DBCROL UNABLE TO OPEN INPUT RELATION NLLF RETURN CODE = 1 Avoidance: There is no avoidance for the xdb file other than running a cold start. Restarts with PARAM,POST,-1 produce a complete op2 file. 1-22555556 2005.5 Optimization/Design Sensitivity - Designing ZOFFS on CQUADR and CTRIAR - UFM 7193 If an optimization job designs ZOFFS of a CQUADR or CTRIAR, the job fails with USER FATAL MESSAGE 7193. Avodiance: Use CQUAD4 or CTRIA3 to design ZOFFS. 1-22611621 2005.5 External Superelements - AUTOQSET, DMIGPCH, K4AAX - UFM 4353 (MTRXIN) Creating an external superelement (SOL 103, DMIGPCH option) using 'PARAM, AUTOQSET, YES' produces an unsymmetric K4AAX matrix. The following assembly (SOL 111) run then fails with the message: *** USER FATAL MESSAGE 4353 (MTRXIN) DMIG MATRICES SELECTED BY K2GG, M2GG, B2GG OR K42GG DO NOT HAVE SYMMETRIC FORM (IFO=6). Running the job with SPOINT and QSET1 entries instead of PARAM AUTOQSET produces a symmetric K4AAX matrix and the assembly run works fine. Avoidance: Insert the following DMAP alter: compile phase1b alter 'UMERGE1 uset,k4QQ,k4TQ' modtrl k4aa////6 $ alter 'UMERGE1 uset,bQQ,bTQ' modtrl baa////6 $ endalter 1-22618721 2005.5 Distrubuted Memory Parallel - DMP, PARAM OUGCORD BASIC - Missing Results in op2 file Distributed memory parallel jobs that use PARAM,OUGCORD,BASIC will only have partial results output to the op2 file. Avoidance: Insert the following DMAP alter: compile sedrcvr alter 'OUTPUT2.*OUGV1PAT'(,-1) if ( disofp and not(acms) ) then $ $ distributed and high level parallel IF ( procid<>1 ) THEN $ disofps ougv1pat,,,,,,, $ ELSE $ disofpm ougv1pat,,,,,,,/ ougv1ptt,,,,,,,/ mpibuff $ equivx ougv1ptt/ougv1pat/-1 $ ENDIF $ procid<>1 ndisofp=ndisofp+1 $ endif $ disofp 1-22653481 2005.5 Adams Integration - param,cmsortho not valid in Version 2004 & 2005 The nastran/adams interface in Version 70.7 and 2001 had a param,cmsorthono option. This option was removed in Version 2004. 1-22716859 2005.5 MDACMS - EIGRL, ND - Extra Eigenvalues MDACMS ignores ND on the EIGRL entry and will output more eigenvalues than specified by ND. Version 2005R3 will issue a warning message that ND is ignored. 1-22732561 2005.5 DMAP - PARAML Options 'VERS___' or 'COMP____' - Access violation The VER__ and COMP___ options of the PARAML module may fail with an access violation on some operating systems; e.g., Linux RedHat 9 and Windows. Avoidance: There is no avoidance. 1-22832921 2005.5 Aeroelastic Analysis - TRIM entry with AEQR set is not parsed correctly If a TRIM entry contains an AEQR value that is a real number 0 < AEQR <= 1.0 and one or both of the label fields on the parent TRIM entry are blank, the AEQR value will be reset to 0.0 with the result that there is no aeroelastic feedback. Avoidance: If AEQR=1.0 is desired, leaving the AEQR field blank will result in the default of AEQR=1.0 being used. It some value other than 1.0 is desired, the only avoidance is to create dummy aerodynamic extra points (AESTAT or AEPARM) and constrain them to 0.0 on the TRIM entry so that there are no blank label fields on the parent. 1-22840061 2005.5 External Superelements - EXTSEOUT - OUTPUT2, SFM 6551 (INPTX2), AN END-OF-FILE WAS DETECTED If element forces are requested in the external component reduction run using EXTSEOUT and the DMIGOP2 option, the datablock IEF1O may be erroneously identified (in the XSOP2DIR record) as one of the datablocks that are to be written to the op2 file. This will cause SFM 6551 in the assembly run that tries to read the op2 file when the datablock MEF1O was not written to the op2 file. SYSTEM FATAL MESSAGE 6551 (INPTX2), AN END-OF-FILE WAS DETECTED Avoidance: Insert the following DMAP alter: ECHOOFF $ COMPILE EXTOUT NOLIST $ ALTER 'XNAME = \'MEF1O'(,-1),'' $ IF ( NOIEF1o>0 ) THEN $ ENDALTER $ ECHOON $ 1-22850801 2005.5 Topology Optimization - TOPVAR - Traceback in dop1x6 Topology optimization jobs with symmetric constraints specified on the TOPVAR entry may fail with a traceback in dop1x6. 1-22912575 2005.5 Distributed Memory Parallel - Large models run with MDACMS and DMP may hang Large models that are executed with a large memory request that use MDACMS and DMP may hang. The matrix partition process (module PRESOL) may attempt to acquire extra (dynamic) memory it needs to partition the model. If there is insufficient memory available on the system to complete the memory allocation successfully, the DMP Slave process may hang. Avoidance: There are two possible workarounds: 1. Reduce the amount of memory given to Nastran with the "mem=" command line option. For most models, "mem=1gb" should be sufficient. 2. Increase the amount of memory given to Nastran. Option (1) is preferable, since the amount of extra memory may be much more than is truly needed to solve the problem. For option (1), there must be sufficient memory available on the system to satisfy the request. Exercise this option on a lightly loaded system. 1-22916281 2005.5 Aeroelastic Analysis/Rotordynamics - MATREDU- SFM 3007, MATREDU Rotordynamics problems may fail in SOL 146 (aeroelastic analysis) in the GENGYRO subDMAP with SFM 3007 from the MATREDU module. Avoidance: There is no avoidance. 1-22955943 2005.5 Nonlinear Analysis - NLITER, CGAP - Divide by Zero, UFM 4658 (NHEXD) A nonlinear analysis with CGAP entries and large displacement may encounter a divide by zero or *** USER FATAL MESSAGE 4658 (NHEXD) DIFFICULT GEOMETRY PREVENTS FURTHER COMPUTATIONS FOR ELEMENT WITH ID =916 User information: Difficult geometry for solid elements occurs when faces have zero areas, or when Jacobians at integration points do not all have the same sign, which signifies that the edges are crossed. In geometric nonlinear analysis, tests are made for initial and deformed geometry. Elements should be modeled such that their faces are as nearly equilateral as possible and midside nodes should be positioned close to the center of their edges. 1-23000557 2005.5 Acoustic Analysis - PARAM,AKMPOST and GRIDMP keyword on FLSPOUT command - Wrong Answers If PARAM,AKMPOST,YES is specified and PANELs are present and the GRIDMP keyword is specified on the FLSPOUT command then the results written to Fortran unit 90 will be incorrect. Avoidance: Remove the GRIDMP keyword from the FLSPOUT command or insert the following DMAP alter: compile akmpost alter 'numfgrid'(2) PARAML mpfmap//'TRAILER'/6/S,N,numgrdmp $ alter 'rowmap = rowmap + 1'(4),'' rowmap = rowmap + 1 + numgrdmp $ endalter 1-23040291 2005.5 Initial Stress Entry Can Not be Translated Correctly for solid elements The 2005 QRG does notnot have a complete explanation concerning this. We have added the the statement that LAY1 and LAYN should be set to zero or left blank if a model has no beam or shell elements. However in checking this it was discovered that some of the possible 7 stress values did not make it to the Marc deck. This problem has now been fixed. Ed Dickerson 1/23/06 Dear sir The initial stress entry "ISTRESS" of slo600 will be translated to "INIT STRESS" of Marc. Unfortunately, it can not be translated correctly. Please check the attached files. Supplements by Charles Chen, 16/Jan/2006, "ISTRESS" would be translated in "INIT STRESS" card in Marc. Regarding "INIT STRESS" card, for shell element, Marc needs 6th datablock to define "layers"; for solid element, 6th datablock should be skipped. However, SOL600 would always generate 6th datablock for "INIT STRESS" card, which will make Marc fail when executing analysis for solid elements. 1-23061191 2005.5 SOL 600 - Stress free initial contact of BCTABLE not translated correctly - Wrong Answers Stress free initial contact of BCTABLE is not translated correctly into marc.dat file in Nastran 2005r2 SOL 600. If multiple contact pairs are defined and stress free initial contact definition is not same for them in BCTABLE, this happens. For example: Nastran input deck: BCTABLE 0 2 SLAVE 2 0. 0. 0. 0. 1 0. 0 1 0 MASTERS 1 SLAVE 3 0. 0. 0. 0. 2 0. 0 0 0 MASTERS 2 Marc input deck translated by SOL 600: CONTACT TABLE 2, 2, 0.000000E+00, 0.000000E+00, 0.000000E+00, 0.000000E+00, 0.000000E+00, 1, 0, 1, 0, 1, 3, 0.000000E+00, 0.000000E+00, 0.000000E+00, 0.000000E+00, 0.000000E+00, 2, 0, 1, 0, 2, Marc correct input deck CONTACT TABLE 2, 2, 0.000000E+00, 0.000000E+00, 0.000000E+00, 0.000000E+00, 0.000000E+00, 1, 0, 1, 0, 1, 3, 0.000000E+00, 0.000000E+00, 0.000000E+00, 0.000000E+00, 0.000000E+00, 2, 0, 0, 0, 2, 1-23086666 2005.5 Transient Analysis - IC(STATSUB) with PARAM,INREL,-2 - SFM 3001 In SOL 109, direct transient analysis, if IC(STATSUB) references a subcase with PARAM,INREL,-2 , the job will fail with the message: *** SYSTEM FATAL MESSAGE 3001 (TRNSP) THE OUTPUT DATA BLOCK IN POSITION 1 DOES NOT EXIST. Avoidance: This error does not occur in SOL 112. Apply the following dmap: compile sedtran alter 'call super1','' CALL SUPER1 /CASECC ,USET ,GOT ,KAA ,DYNAMICS, dynamic ,GPLS ,SILS ,SLT ,dar , CASES ,EQEXINS ,MATPOOL ,GM ,DIT , BGPDTS ,CSTMS ,EST ,MPTS ,XYCDBS , lll ,kll ,ull , ,PCDB , EMAP ,XYCDB ,MAPS ,ETT ,PVTS , KFS ,KSS ,KSF ,EDT ,EDOM , GEOM2S ,GEOM3S ,POSTCDB ,ECTS ,EPTS , INDTA ,KELM ,KDICT ,GPECT ,VELEM , FORCE ,casers ,SCSTM ,matpools,ditid , SPSEL , ,PJ ,PTELEM ,PG , PA ,YS ,qr ,UOO ,PSS , PL , ,gpsnts , ,KJJ , KGG ,edts ,LOO ,BJJ ,BGG , K4GG ,MJJ , ,MLR ,MLL , MRR ,CMLAMA ,CMPHA ,CMPHO , , MR ,GOQ , , ,MAA , MAR ,MEA ,BAA ,K4AA ,MGG , , , , ,GEOM4S , VGFS ,PCDBS ,ESTL ,ESTNL ,AGG , DMIs ,DMINDXs ,DTIs ,DTINDXs ,CONTACT , btopo ,GPDTS ,MELM ,MDICT ,GEOM1 , GEOM2 ,GEOM3 ,GEOM4 ,EPT ,MPT , SLIST ,KMG ,MMG ,MSF ,BMG , BSF ,K4MG ,K4SF , ,pvt , oint ,deqatn ,deqind ,medge ,gdntab , pelsets ,viewtb ,bgpdtvu ,orseq , , vgfd ,eht ,geom1s ,pvtbc , , , , , ,vgq , , , ,mmcdb , , , , , ,axic , , , , , , epsseo ,drg ,mphvz ,bk4zz ,pcompts , pval ,elemvol ,pmyg ,pbyg ,pkyg , mss ,bss ,k4ss ,belm ,bdict , , , , , , , , , , , , , , , , , , ,xsop2dir,rvg , geom1qx ,cee ,mee ,gge ,bec , mec ,cmlama1 ,ajp ,gei ,pvt0 , kay ,bay ,may ,k4ay ,goy , yg ,phba ,aaa ,aay ,asf , pggrv , , , , , ,ym , , , , , , , , , , , , , , , / s,APP /DMAPNO /S,RSONLY/SOLTYP /FALSE / FALSE /FALSE /FALSE /APRCH /S,SEID / S,PEID /S,MTEMP /S,LOAD /S,TEMPLD/S,DEFORM/ S,MPC /S,SPC /S,SUPORT/S,K2GG /S,M2GG / S,B2GG /S,P2G /S,DYRD /S,METH /S,MFLUID/ s,nloop /S,LASTSEID/S,PFILE/S,LUSETS/S,FSCOUP/ enfmeth /AUTOSPC /S,NOPH2 /FALSE /s,pvalid/ s,pexist/false /s,kdiff /' ' /0 / 0 /0 /0.0 /s,noph1 /s,adptindx / false /'no ' /'no ' $ 1-23097181 2005.5 Optimization/Design Sensitivity - DESSUB, DFREQ, MFREQ, DSAD - Fatal Message, Wrong Answers If a frequency response (ANALYSIS=DFREQ or MFREQ) with multiple subcases is submitted with the first frequency response subcase lacking a dessub request, the job can fail with system fatal messages in the dsad module or it can complete with wrong answers for the design responses and sensitivities. Avoidance: Remove the subcase that lacks a DESSUB command or place this subcase after a frequency response subcase that does have a dessub request. 1-23113802 2005.5 Buckling - multiple superelements and multiple subcases - SFM 6912 A linear buckling solution with more than one superelement and more than one buckling subcase may fail with the following message issued by the SEDR module: *** SYSTEM FATAL MESSAGE 6912 (SEDRML) THE BASIC GRID POINT DEFINITION TABLE NAMED BGPDTS AND THE DISPLACEMENT MATRIX NAMED BUG FOR SUPERELEMENT 20 ARE INCONSISTENT. USER ACTION: IF THIS IS AN EXTERNAL SUPERELEMENT DATA RECOVERY, ENSURE THAT THE BGPDTS TABLE AND NOT THE EQEXINS TABLE IS DBLOCATE'D. PROGRAMMER INFORMATION: THE NUMBER OF DEGREES OF FREEDOM IN THE BASIC GRID POINT DEFINITION TABLE IS 36. THE NUMBER OF DEGREES OF FREEDOM IN THE DISPLACEMENT MATRIX IS 60. Avoidance: Use the following DMAP alter. compile super3 alter 'ugb','' DBVIEW UGb=UG WHERE (( (seid=np2 and seid>0 and peid>0 and spc=* and mpc=* and suport=* and bmeth=* and mtemp=* and p2g=* and k2gg=* and m2gg=*) or (seid=np2 and seid=0 and peid=0 and spc=cspc and mpc=cmpc and suport=csuport and bmeth=cmeth1 and p2g=cp2g and k2gg=ck2gg and m2gg=cm2gg and mtemp=cmtemp) )) $ 1-23128147 2005.5 Optimization/Design Sensitivity - Buckling with external superelements - UWM 6911, Wrong Answers If buckling sensitivity analysis with superelements is performed with multiple subcases UWM 6911 (as shown below) is issued indicating there are multiple copies of a requested datablock. This may lead to wrong answers. *** SYSTEM WARNING MESSAGE 6911 (QCBFIN) THE DBVIEW HAS YIELDED MORE THAN ONE DATA BLOCK BUG . THE PROGRAM WILL SELECT THE "FIRST" DATA BLOCK FOUND. PROGRAMMER ACTION: TO FIND THE DUPLICATES, SPECIFY THE WHERE CLAUSE FROM THE DBVIEW STATEMENT ON THE DBDICT STATEMENT AND THEN MODIFY THE WHERE CLAUSE ON THE DBVIEW STATEMENT TO REMOVE THE DUPLICATES. Avoidance: External superelements can be replaced with DMIGPCH and K2GG instead of DMIGOP2. This avoids the SE processing. Else the following dmap alter may be used: : compile respsen alter 'dbview bugx=bug','' dbview bugx=bug where (seid=temp and wildcard) $ 1-23241901 2005.5 Large Models - GP4 - SFM 2101 pointing to an illegal dof A large model, more than 3 million dof, failed in GP4 with SFM 2101 indicating that an illegal degree of freedom existed on an mpc entry. . *** SYSTEM FATAL MESSAGE 2101 (GP4) GRID OR SCALAR POINT ID= 16700047 COMPONENT = 2 FOUND ON THE MPC ENTRY IS ILLEGAL. USER ACTION: CHECK MPC ENTRIES. 0FATAL ERROR Avoidance: There is no known avoidance. 1-23267428 2005.5 Aeroelastic Analysis - GI, SPLINEi - Floating Point Overflow A floating point exception can occur due to an attempt to divide by zero when processing aerodynamic spline input. Avoidance: The problem may be avoided by running on a different platform. The only other avoidance is to use a release prior to Version 2005.1. There is no avoidancethere is not avoidance other than to use a different platform (linux seems to work okay) or to use a version of Nastran earlier than 2005.1 1-23354553 2005.5 QRG/Aeroelastic Analysis - Plotting with SET 2 = AERO1 Fails In aeroelastic analysis, the plotting of the AERO1 elements are no longer available from the PLOT module using SOL 145 or 146. The Quick Reference Guide SET description Remark 1 needs to be updated to reflect this If AERO1 is used then the program issues the message: ERROR MESSAGES FROM THE PLOT SET DEFINITION MODULE (PLTSET) AERO1 ELEMENTS CANNOT BE PLOTTED BY THE STRUCTURAL PLOTTER Avoidance: Use Patran with PARAM,OPPHIPA,0 and PARAM,POST,0 and request displacement output. Create a new patran db and access both (model and results) from the xdb file. 1-23474784 Optimization/Design Sensitivity: Multiple types of MATTi entries leads to SFM 3003 (EMG) When an optimization job includes more than one type of MATTi entries, or temperature Dependent Material propertiy entries, the job may fail in EMG at design sensitivity phase as follow: *** SYSTEM FATAL MESSAGE 3003 (EMG) ATTEMPT TO READ PAST THE END OF A LOGICAL RECORD IN DATA BLOCK MPTX (FILE 103) IN SUBROUTINE PREMAT Avoidance: Include only a single type of MATTi entry in an optimization job. 1-23481533 Optimization/Design Sensitivity - DESVAR - SFM 4276, EC 3001 (gread) - DOPR1 When an optimization job includes a DESVAR command that selects all the DESVAR entries and also includes DLINK entries, the job may fail in DOPR1 with SFM4276/3001. Avoidance: Use DESVAR = ALL or remove DESVAR command. 1-23482561 2005.5 File Management Section - ASSIGN, OUGPSD1 - Incorrect LTLEND/BIGEND format If the LTLEND describer is used on the ASSIGN statement to write a little-endian format op2 file on a big-endian machine in a random response analysis, the OUGPSD1 datablock will have the first 3 words of the case control record written in big-endian order. The error is caused because OUTPX2 did not properly detect that there is no valid NDDL description for OUGPSD1 and other datablocks for the same OUTPUT2 DMAP statement. Just as for FORM=FORMATTED (neutral format), OUTPX2 shoud issue a UIM and skip the data block. The same problem would affect INPUTT2. 1-23494381 2005.5 Matrix Methods - ADD5 - BIOMSG: ERROR 999 HAS OCCURRED IN ROUTINE alxa2 The following error occurs in the ADD5 module when scaling a matrix when the input datablock is from a database with a smaller buffsize. BIOMSG: ERROR 999 HAS OCCURRED IN ROUTINE alxa2 1-23548203 2005.5 SOL 600 - CQUAD4, CHEXA, CBEAM, RBE2, RBE3 - Contact Model, Access Violation A contact model with CQUAD4, CHEXA, CBEAM, RBE2 and RBE3 elements fails during processing the input file with an Access Violation. 1-23573231 2005.5 Optimization/Design Sensitivity - Uniqueness on DRESP2 IDs in SOL 200 is not checked Uniqueness of DRESP2 IDs in SOL 200 jobs is not checked for MSC.Nastran v2005 onwards. The last verion that the check is performed is v2004r2. Users must manualy make sure there is no duplication on DRESP2 IDs. If duplication on DRESP2 IDs exists and the ID is selected as a constraint or objective, the optimization results may be based on the unintended constraints. Avoidance - There is no avoidance. 1-23623711 2005.5 Optimization/Design Sensitivity - SFM 3007 (DSFLTF), Misleading message If a design response is associated with a complex eigenvalue (RTYPE = FLUTTER or CEIG) and the eigenvalue is 0.0,0.0, the following message is produced: *** SYSTEM FATAL MESSAGE 3007 (DSFLTF) ILLEGAL INPUT TO SUBROUTINE DSFLTF Avoidance: Determine and remove the cause of the zero eigenvalue. A UFM message such as the following needs to be added: The sensitivity of a complex eigenvalue that is exactly zero (0.0,0.0) is being requested. The sensitivity algorithm does not permit this and indicates a problem with the underlying analysis or the prescribed design task. User Avoidance: Determine the cause of the zero eigenvalue and remove it. 1-24081681 2005.5 External Superelements - Duplicate GRID and Element IDs - Wrong Answers If one or more grid point IDs and element IDs are the same in an external superelement, then output requests for displacements and stresses in the creation run for grid points and elements with the same IDs may lead to erroneous results in a subsequent assembly run. Specifically, displacement output requests for grid points in the creation run may lead to erroneous element stress output for elements with the same IDs in a subsequent assembly run. Similarly, element stress output requests for elements in the creation run may lead to erroneous displacement output for grid points with the same IDs in a subsequent assembly run. Avoidance: Make the following two changes to the input data of the external SE creation: (1) If there are any VELO or ACCE requests in the Case Control, be certain to consolidate those sets with a DISP request. (2) Employ the following DMAP alter: compile extout alter 'sdr2','' $ SDR2 CASEx,CSTMS,MPTS,DIT,EQEXINs,,ETT,nulllama,BGPDTS,,qg, rga,EST,XYCDBs,,,,,,,,,,,,,,/ ,iqg1,IUG1,IES1,IEF1,,,,,,,/ 'reig'/S,N,NOS2A/S,N,NCOMPS $ 1-24090721 2005.5 Optimization/Design Sensitivity - PBMSECT - UFM 3007 in dompftn. A SOL 200 job with pbmsect entries may fail with USER FATAL MESSAGE 3007 in dompftn if some pbmsect are not included in the design model. Avoidance: Include all PBMSECT in the design model. If some pbmsect should not be altered buring the optimization process, use the XINIT field of the DESVAR entry to specify an appropriate XLB and XUB.