Home

of the theory manual including course material

Contents

1. l l l i l ID rule diameter thickness yield str 1 Fa E E 7 poo PETELINI ate 1 680E 01 4 500E 03 2 7808 100 K capacity l i D RRA DOE MESES 1 li 1 Brace Angl Conn Facing Gap i Axial MipB MopE 100 Y capacit ID deg Type brace i Cap Qf Cap Qf Cap Q pacity l 3g l 1 l i 4 6011 97 K 5 016 4 122E 05 2 584E 04 1 973E4 l n Y i 3 877E 05 2 584E 04 973E4 11100 gt i 4 114E 05 2 584E 04 973E4 i 7 93 85 Combined n f n 5 60 K 4 016 i 3 747E 05 2 584E 04 N 1 973E 97 K 3 Y I 3g I ea ER HMM S Re hoe se CN A 1 00 1 00 capacity Q factors Table 5 3 2 Print from the MSL routines on the lt res gt out file SINTEF group 2001 06 10 Load kN USFOS GETTING STARTED PRU M Figure 5 3 1 2D K frame 5 5 Rigid plastic ISO Ultiguide MSL 700 600 ed 2 l 500 400 2 oO 300 200 Fugit joints Rigid plastic 100 F ISO Ultiguide MSL Test 0 0 20 40 60 80 a Figure 5 3 2 Deformation mm 2D K frame Load deformation curves 20 30 40 Deformation mm SINTEF group 2001 06 10 USFOS COURSE MANUAL 6 1 6 Foundation Modelling 6 1 Nonlinear spring A general nonlinear spring element is available in USFOS The spring has 6 uncoupled degrees of freedom The behaviour of each degree of freedom is defined by discrete P 5 points see Hyperelastic materia
2. 8 17 SINTEF group 2001 06 10 USFOS COURSE MANUAL 9 Workshops 9 1 Workshop I USFOS Element formulation Analysis Control The purpose of this workshop is to get familiar with the USFOS program system and to demonstrate the basic features of the program Three simple models are presented where the results can be verified by hand calculations or engineering judgment Theme Relevant input Non linear Material MISOIEP MPLASMON MPLASCYC GBOUND SURF2OFF Load Control CNODES CCOMB CUSFOS CICYFOS Analysis solution control DETEROFF CITER Imperfections CINIDEF GIMPER GELIMP SINTEF group 2001 06 10 USFOS GETTING STARTED 9 1 1 Workshop I a 9 1 1 1 Objective Elasto Plastic Beam Bending The purpose of this case is to demonstrate the membrane action of the USFOS beam element and to investigate the different material models in USFOS A tubular beam is loaded by transverse forces Y Y Beam length L 100m um s 2A Tube diameter D 0 2407 m Tube thickness t 0 005 m Yield stress o 330 MPa Young s modulus E 2 1 10 MPa Reference load qo 10 0 kN m Section area A 370210 m Moment of inertia I 2 572107 m Moment of Wp 2 778 107 resistance m L p Axialyieldload Np 1221 6 kN Plastic moment Mp 91 67 kNm cap Figure 9 1 1 Beam Model 9 1 1 2 Description Plastic hinges form when the bending capacity of the critical sections is exhaus
3. Elevation 50m Elevation 50m m 85 85 80 80 85 80 ma ms pm pm pe pp pn pL 9 2 2 2 Environmental loads Wave loads are calculated using Stokes 5th order theory Wave and current loads are represented as distributed line loads Module loads and wind loads are applied as concentrated forces moments on the deck nodes Wave and current loads are calculated by the SESAM 80 program WAJAC Environmental loads are determined for 8 platform directions Loads are calculated for 100 year wave 100 year wind and 10 year current acting simultaneously SINTEF group 2001 06 10 USFOS GETTING STARTED 9 12 Resulting global forces are listed below Global load envelopes are shown in Figure 9 2 3 Table 9 2 3 Global loads base shear in MN Direction ma OOU m p pe e e e e Ps Wa Wave loading Cu Current loading Wi f Wind loading 9 2 2 3 Static loads Static loads 415 0 MN Buoyancy 77 7 MN Global Loads MN Wa Cu Wi Wa Cu Figure 9 2 3Global loads SINTEF group 2001 06 10 USFOS GETTING STARTED 9 13 9 2 3 FEM Model The structure is modelled by 160 nodal points and 455 elements 20 elements are defined as non structural 19 materials and 119 cross sections are defined Modules or topside facilities are not modelled No joint cans or brace stubs are modelled Piles guides are modelled by equivalent elements The equivalent stiffness of the pile guides and
4. 1 Perform one collision analysis for each accident scenario 9 3 3 Documentation 1 Generate P plots from each analysis 2 Take hardcopies of deformed geometry with member utilization at max impact load 3 Generate plots of axial force vs global displacement N 6 plots for members in the impact area 9 3 4 General comments The ship collision analysis is essentially like any other pushover analysis The major difference is that the user only specifies the general loading situation e g impact energy position and direction and not the actual impact loads Suitable impact load increments are calculated by USFOS But the user still have to define the impact situation as a load case and apply this load case in the USFOS load specification the CUSFOS CICYFOS records like any other load case SINTEF group 2001 06 10 USFOS GETTING STARTED 9 17 USFOS calculates suitable impact load increments as a fraction of the force required to flatten the tube or as a fraction of the mechanism load of the hit member The impact increments are calculated so that initial yielding should occur after some 10 20 load steps When the specified impact energy is dissipated USFOS reverses the impact load and unloads the structure The total energy dissipation is listed in the global history table This include energy dissipated by the ship by the structure and locally by flattening of the tube Energy dissipated by ship and structure
5. SUSFOS HOME bin usfos 15 ENDIN head stru res ENDIN Table 8 3 2 Content of script file go2 with only head and stru input files SINTEF group 2001 06 10 USFOS GETTING STARTED 8 6 It is possible to access files located on other directories than the directory where the script go is located and started from Table 8 3 3 describes the case where some files are located on different directories SUSFOS HOME bin usfos 15 ENDIN head intact nw 100yr model intact stru loads nw 100yr D temp res intact nw 100yr ENDIN Table 8 3 3 Content of script file 603 with input files located on different directories p g p In this case the control file head intact nw 100yr fem is located on same directory as the script file and where the script is started from The structural file stru fem is located in the directory model which is located on same level besides the current directory and the file is named intact stru fem The load file is located on an other directory also on same level as the other two with name loads in a file with name nw 100yr fem The results are saved on the D disc on a directory named temp and file res nw 100 raf The third variant of the fixed name script go3 indicates a first try to organise an analysis series involving several versions of the structural file f ex intact and damaged and several loads f ex nw 100yr nw 1000yr sw 100yr sw 1000
6. backup purpose mv stru fem Case 1 2 mv load fem Case 1 2 4 Table 8 5 1 Content of script file go which assembles input files amp runs USFOS Support Loa go Spring Support 1 Nodel Load go Spring Support 1 Node3 Load go Spring Support 1 Node5 Load go Spring_Support_2 Nodel Load go Spring Support 2 Node3 Load go Spring Support 2 Node5 Load Table 8 5 2 Content of script file run all which executes the script go After the script run all is completed 6 new file folders directories are created see Higure 8 5 2 All directories contain the actual assembled input stru and load the result files res SINTEF group 2001 06 10 USFOS GETTING STARTED 8 10 Contents of Example 3 Assembling Files File Folder File Folder File FEM File File ee Case Spring Support 1 Nodel Load File Folder Case Spring Support 1 Node3 Load File Folder 9 Case Spring Support 1 Node5 Load File Folder Case Spring Support 2 Nadel Load File Folder Case Spring Support 2 Node3 Load File Folder C3 Case Spring Support 2 Node5 Load File Folder Figure 8 5 2 Content of file folder after running script run all 8 6 Example 4 Using the SED editor to modify master input files In the previous example the input to USFOS was composed by some common files special files and in all cases the content of the files were pre defined In the current example another and even more flexible s
7. with no 1382 Local node 1 J X Element no 138 N Local node 2 of original m of original element no 138 element no 138 oe o Figure 5 1 2 Numbering of extra elements generated by USFOS Extra element at end 1 of the actual number gets the member number plus one extra digit with value 1 At end two the extra digit has value 2 Note All elements and nodes generated by USFOS have negative sign The material and geometry numbering starts from the highest user defined material and geometry numbering SINTEF group 2001 06 10 USFOS GETTING STARTED 5 2 Properties of the extra stub elements The material properties are set equal to the properties of the actual chord but hardening is not permitted Fracture is excluded for the joint no limit of the magnitude of the tension strain The cross section parameters Cross sectional area Plastic resistance moment about local Y axis Plastic resistance moment about local Z axis are derived from the API capacity formulas The other cross sectional parameters are set equal to the ones for the actual brace 5 2 Joint deformation control Joint Capacity with Ductility Control 100000 i i Joint Capacity 80000 60000 40000 20000 r 0 Force Transfer 20000 H 40000 60000 F 80000 H 100000 i i i i i 15 E Joint Deformation The joint capacity option is extended to have a control of the joint deformation
8. To answer the question check for any sharp corners on the P curve check the I values during and just after the spring back and check is iterations have converged or performed normally during the spring back SINTEF group 2001 06 10 USFOS GETTING STARTED 9 16 9 3 Workshop III Ship Impact Ship impact analyses are executed for the following accident scenarios 1 Longitudinal impact on leg A4 at elev 1 0 meter 2 Transverse impact on leg A4 at elev 1 0 meter 3 Brace impact row 4 elev 1 0 meter 4 Brace impact row A elev 1 0 meter 5 Transverse impact on leg A3 at elev 1 0 meter Gravity loads buoyancy and operational loads are applied up to unfactored characteristic values Then the impact load is incremented until the specified impact energy is dissipated Finally the impact forces are stepped down Energy dissipation in ship and structure is determined Permanent deformations and residual stresses are recorded and may be used as input to a residual strength analysis restart 9 3 1 Program input Control file head case X gt fem One file for each impact case Structure file stru fem One single file Load file load fem One single file with north wave loading Load case 1 Dead load live loads Listed on the stru file Load case2 Buoyancy Given in the load file Load case 3 Wave and current loads Given in the oad file Load case 4 Ship impact loads Given in the head file 9 3 2 Analyses
9. USFOS checks the following criteria and scales the load step down is necessary Introduction of plastic hinges e Exceedance of the user defined maximum displacement increment e Adjustments during equilibrium iterations If yielding occurs in any cross section the load step is scaled so that the element forces comply exactly with the yield surface scaled increment p 1 unscaled increment 7 p 1 i increment number M M o Figure 1 3 1 Increment scaling due to introduction of plastic hinges By this procedure only one hinge is detected per load increment To avoid unreasonably small step length in case of frequent occurrence of hinges the user may specify a minimum load step for the scaling In this way exact scaling to the yield surface is not always possible and several hinges may be inserted during one load increment In regions where the current stiffness parameter is small very large incremental displacements may result This may reduce the accuracy of the analysis as shown in a2 3 2 To control too large displacements the user may define a global displacement of the structure and set a limit to the size of the displacement increments The global displacement is jpeeued as a weighed sum of some characteristic degrees of freedom supplied by the user DR A3 1 32 A ras 2 Ark k NCNODS x o k Ani displacement increment for control displacement k at step i Ok weight factor associated with
10. as the case of K joints and X joints The tubular joint capacity is evaluated according to the recommended practice of API RP2A API 1994 The conductor guide framing is modelled in a simplified way In the deck only primary elements of the floor system are modelled and some equivalent elements are used instead of secondary elements In order to model the transmission of lateral forces between piles and legs as well as between conductors and guide framing linear dependencies are defined Since the conductors were assumed as structural elements and therefore the soil conductor interaction is also included in the model The appurtenances such as boat landings barge bumpers risers and caissons were modelled as non structural elements in the analyses Figure 9 5 1 Drilling and production platform SINTEF group 2001 06 10 USFOS GETTING STARTED 9 22 9 5 3 Foundation The foundation of the platform consists of eight tubular steel piles each of which is driven into the sea floor and allocated inside each jacket leg The piles are fixed to the corresponding jacket leg and deck column at elevation 7 315 m working point Corner piles were driven at a depth of 98 74 m from the mudline whereas for inner piles this depth is 90 37 m The piles are fabricated with six segments of different wall thickness in a range from 2 54 cm in the pile tip to a thickness of 5 72 cm in the top In all the segments a yield stress of 248 MPa 2530 kg
11. eee sanus BN ERE ERN RN Ein Table 9 4 2 Characteristic load levels members removed Damage condition mcm pce IE p Mp e E pcc a He ee ieee Ee ae a SINTEF group 2001 06 10 USFOS GETTING STARTED 9 19 9 4 1 Program input Control file head lt case X gt fem One file for each damage scenario Structure file stru fem One single file Load file load fem One single file with north wave loading Load case 1 Dead load live loads Listed on the stru file Load case2 Buoyancy Listed on the oad file Loadcase3 Wave Current Wind loads Listed on the load file 9 4 2 Analyses Perform one pushover analysis for each damage scenario specifying an initial damage to the element use GELIMP and e assign the specified out of straightness as initial GIMPER records deformation e switch the local buckling formulation ON 2 Perform one pushover analysis for each damage scenario use the NONSTRU record removing the damaged element 9 4 3 Documentation 1 Generate P plots from the final analysis of each direction 2 Take hardcopies of deformed geometry with member utilization at max load and at the final analysis step For analysis 2 3 Generate plots of axial force vs global displacement N 6 plots for critical members in the failure mechanism buckling members tension failure members or failing leg members For analysis 2 9 4 4 General comments Residual strength analyses are generally a repetition o
12. l u pet ge 2 K El Vix Ely Ww 1 3 where the first integral comes from axial straining and the last integral represents bending The expression in the first parenthesis is the element strain x The total displacement is decomposed into axial displacement u x and lateral deflection v x and w x in three dimensions Torsion is not included in the variational formulation but is added directly into the element stiffness matrix The potential of external loads is written as l l l H Fiu q udx a v dx a wx 1 4 0 0 0 The total potential for an elastic element is now IIZU H 1 5 Total and incremental equilibrium equations are established by taking the first and second variation of the strain energy and the potential of the external work The first variation of internal strain energy comes out of equation as shown in equation 1 6 l l N U EA ux ux dx zr t V xx va TU Vx Ov dx 0 0 EI l j 1 6 N Er C Wax 5 ws g wiwa dy N EAu 8 u dx 0 y 0 The first term of s the linear contribution from axial strain The two next integrals represents bending deformation including the influence of axial forces membrane effects These terms are represented by the Livesly stability functions in the stiffness matrix The SINTEF group 2001 06 10 USFOS GETTING STARTED 1 7 last integral comes from the nonlinear axial strain contribution from lateral deflections and gives a correction to t
13. 06 Q 0 0 Unloading at end 1 Unloading at midspan Yield at end 1 22 OL 29 04 S22 7255 O E STARC qi 29 04 23 06 Qe y Yield at end 1 32 00 33 01 94 02 CaS ee 280 00 Saee oou ud c3 Q Yield at end 2 34 03 els 00 eset S00 0O Unloading at end 2 31 00 a 32 00 VES 02 O 0 64 47 I2 DORIA 74 60 cu PE ae DS 55 64 46 ssbD o 50 Li00 1 00 60 50 78 70 84 78 57 43 78 70 84 78 57 43 B10 35 RPL 60 90 84 SLC 125 PL 60 90 84 GLOBAL TOTAL DL SSE L A C EM ENT S c X dis Y dis Z dis X rot Y rot Z rot 2 872E 02 000E 00 1 772E 03 000E 00 1 210E 03 000E 00 GLOBAL REACTION BSOSRCUBSS 4n se X for Y for Z for X mom Y mom Z mom 1 219E 05 000E 00 6 534E 05 000E 00 1 104E 11 000E 00 1 157E 05 000E 00 6 534E 05 000E 00 1 440E 11 000E 00 2 376E 05 000E 00 2 910E 10 1 19 cpu 1 90 sec Total accumulated cpu 32 68 sec Figure 2 4 4 Load step output SINTEF group 2001 06 10 USFOS GETTING STARTED 2 12 Table 2 4 2Terminology load step output ELEM ES INTERACTION FUNCTION VALUES Yield at Plastic hinge at Unloading at Tension failure mode Fracture at Capacity lim at end 1 end 2 midspan joint 1 joint 2 Q 0 0 0 0 GL
14. 1 2 RUNNING USFOS 2 1 System architecture The USFOS analysis system consists of three main program modules e The USFOS analysis module performs all numerical calculations and generates two or more files of analysis data The analysis print file out is a text file containing general analysis results the Analysis data file raf is a binary file containing structure data and analysis results data This file is as a result database for XFOS and POSTFOS In addition global analysis results are logged on terminal or batch output device e XFOSisan interactive system for visualization and presentation of USFOS analysis results Three dimensional pictures of the analyzed structure may be presented in colours at selected deformation states in order to investigate the collapse process of the structure XFOS also generates XY plots of global structural behaviour as well as element history results Colour pictures plots are generated in PostScript format for plotting or text document inclusion XFOS accesses the USFOS binary result database through POSTFOS e POSTFOS a module designed to extract data from the USFOS binary result database POSTFOS is command oriented with extensive built in HELP functions POSTFOS generates text files of selected analysis results POSTFOS is mainly used through XFOS but can also be used as a standalone program to extract data to tables or for plotting outside XFOS 2 2 Memory allocation Both USFOS and
15. Kf Av 1 31 The considerations above are valid for a plasticity formulation where the cross section is either elastic or fully plastic That is only one yield surface is used SINTEF group 2001 06 10 USFOS GETTING STARTED 1 13 A plasticity model which accounts for partial plastification and strain hardening is formulated according to the bounding surface concept This model employs two interaction surfaces one yield surface and one bounding surface The yield surface bounds the region of elastic cross sectional behaviour when the force state contacts the yield surface this corresponds to initial yielding in the cross section The bounding surface defines the state of full plastification of the cross section This surface has the same shape as the yield surface Figure 1 2 2 illustrates the yield and bounding surfaces for a tubular cross section plotted in the mzn plane When the cross section is loaded the force point travels through the elastic region and contacts the yield surface upper plot This represents first fibre yield in the cross section At this stage a yield hinge is introduced When further loading takes place the yield surface translates such that the force state remains on the yield surface Ty 0 middle plot The bounding surface also translates but at a much smaller rate This is used to model strain hardening according to a kinematic hardening model The translation of the yield surface which approaches
16. LOCAL Z POP 000 707 TOT 000 707 707 000 707 000 1 000 000 000 1 000 000 000 1 000 000 Sh area Sh area y axis z axis 3 552E 03 3 552E 03 8 128bE 02 3 418E 02 Thermal expan 1 400E 05 1 400E 05 Figure 2 4 3 Structural data SINTEF group 2001 06 10 USFOS GETTING STARTED 2 11 The load step output in the analysis control file gives detailed information of the structural response At each load step the load energy and stiffness are listed interaction function values and status of selected elements total displacements of specified nodes and global reaction forces of fixed nodes Formation removal of plastic hinges at each element is commented The accumulated displacements are printed for all nodes included in the global displacement defined by the user Load step Load step yo NN He N eOoOooooocdoco 1 19 INCREMENTAL SOLUTION Z AYA s FRAME U S F O S progressive collapse analysis SINTEF div of Structural Engineering USFOS load combination no 1 Load step no T9 Load increment scaled to minimum step length Load increment 050 New load level 5 939 Current stiffness parameter 409 Solution accuracy parameter 2 800 E 00005 Determinant of tangential matrix 7 980 E 00523 Energy absorbtion 3 699 E 00003 INTERACTION FUNCTION VALUES Fb Fy Nodel Midspan Node2 Yield at end 1 19 10 ebbe 091 19 04 O 0 Lbs 17 18 11 06
17. are incremental displacements and forces Ap is the relative load increment size at each load step The Current Stiffness Parameter is a normalized parameter representing the stiffness of the structure during the deformation It may be regarded as the incremental work carried out in the first load step divided by the incremental work at load step no i Thus the current stiffness parameter will have an initial value of 1 0 For stiffening systems membrane effects it will increase For softening systems it will decrease A small absolute value of the current stiffness parameter will represent an unstable structure the instability point having 0 0 Current Stiffness The Current Stiffness will be negative in the post collapse range SINTEF group 2001 06 10 USFOS GETTING STARTED 1 18 rc Pd i l I l soft 0 Pir T instable instability Load Insta point reversal Figure 1 3 3 Current stiffness parameter 1 3 4 Elastic Spring Back The elastic spring back problem is characterized by an extremely brittle behaviour At a specific level the load drops accompanied by a temporary reduction in displacement The structure may later regain stiffness and the deformations increase Deformation Figure 1 3 4 Elastic spring back problem Along path b c the load should be reduced The Current Stiffness is positive but with a higher value than the initial stiffness 1 The
18. control displacement k The load step will be scaled down if the control displacement increment of the current step exceeds mxpdis times the control displacement of the initial load step of that load combination D If only one degree of freedom is specified the control displacement will not be normalized I e the control d o f will only be multiplied by the weight factor SINTEF group 2001 06 10 USFOS GETTING STARTED 1 17 A r 4 S mxpdis A T slob 1 33 AT gon mxpdis Ar 7 Figure 1 3 2 Scaling by maximum control displacement 1 3 3 Sign of load increment The sign of the load increment is governed by the Current Stiffness Parameter and by the determinant of the tangential stiffness matrix 1 3 3 1 Determinant of Stiffness Matrix The stiffness matrix determinant is the classical stability criterion for nonlinear analyses As long as the determinant is positive the stiffness matrix is positive definite and the structure is stable As the load increases the structural response becomes more and more nonlinear and the determinant will decrease for softening systems Zero determinant signifies a global instability point or a bifurcation point and a negative determinant one or more negative terms on the stiffness matrix diagonal represent an unstable structure 1 3 3 2 Current Stiffness Parameter The Current Stiffness Parameter is defined by Ar Y AR Ap Pd S Ari AR Ap where Ar and AR
19. feature The MSL equations are implemented with ductility limits and post rupture unloading for tension loading but with no ductility limits for compression loading Joint failure in tension invokes the FRACTURE option in USFOS Joint utilisation will be visualised by colour fringes in Xfos The following shows the input required to include MSL joint characteristics in the analysis of a 2D K frame The input is described in more detail below JNT FORM 0 beam stub 1 P delta spring 3 plasticity model JNTCLASS 0 OFF i gt 0 interval for re classification nodex chordi chord2 Can Rule CapLevel GammaQf CHJOINT 7 6 7 0 MSL mean 4x0 Table 5 3 1 USFOS control input activating MSL joint classification Comparison between the USFOS analysis and alternative joint models and tests results are presented in Figure 5 3 2 SINTEF group 2001 06 10 USFOS GETTING STARTED 5 4 Each time joint re classification is performed the following information is printed to the out file Load step 1 60 E JO TN T CTi Av DOn E EAT C ASTEON R 2D K F RAME U S F O S progressive collapse analysis SINTEF div of Structural Engineering Specified capacity USFOS load c amp mbination no 1 Joint ident Load step no z 60 462 683 Load level PSS Sa eS ee mc mi m Se ee ee oe ee Se oe Eee re l 1 l li I NODE f i Capacity i Chord Chord Chord l i I 1 l
20. introduced errors in the solution 12 Decide if you can let the member move free until this load vector or terminate the previous load vector earlier 2 5 4 2 Tension Failure For a member yielding in pure tension increased loading should result in increased axial forces due to strain hardening and reduced or constant bending moments Le cross sectional forces should move in direction of the positive N axis of the yield surface However the plasticity formulation states that the forces should move along the yield surface from one point on the yield surface to another point on the yield surface And at pure tension yielding the yield function is singular The force point may cross from one quadrant of the yield surface over to the other quadrant instead of moving in direction of the positive N axis In the next step it may cross back over again A special membrane element is therefore implemented to model pure tension yielding of a member The membrane element is automatically introduced when the axial force exceeds 98 of the tension yield force AXIAL failure mode The element accounts for the geometric stiffness of the member i e conserves the axial stiffness of the member and allows transformation back to a beam element if the member is unloaded However if the two surface material model is used gradual plastification of the cross section then the over crossing can still occur at the tip of the inner surface the yie
21. is stepped through the structure with time increment 1 s The wave position giving the highest base shear in the interval Time 0 20s is used in the push over analysis NOTE As all hydrodynamic calculations are using SI units the forces are calculated in N Newton If f ex MN is used as force unit the wave forces must be scaled before they are used in the pushover analysis The command WAVMXSCL lt factor gt is used see also User s manual Ch 6 In the current example the wave forces are scaled with a factor 1 3 just for demo purpose a For both the buoyancy forces and the wave forces it is possible to print the calculated forces to separate files but in the example printing is switched off nowrite SINTEF group 2001 06 10 USFOS GETTING STARTED 4 4 Lcomb 1 is gravity loads and static deck loadstcalculated buoyancy Lcomb 2 is Stoke Wave 45 deg diretion nloads npostp mxpstp mxpdis CUSFOS 10 LS 1 00 OOS lcomb lfact mxld nstep minstp 1 Leo 1 0 10 0 05 Dead Buoyancy 2 04 5 3 0 50 0 001 Wave 0 1 6 0 100 Wave Apply automatic out of straightness Use loads from Waves lcase 2 Separate Bouyancy from wave forces Add Buoyancy to load case 1 Option noWrite Define Wave type H Period Direction Phase Surf Lev Depth WAVEDATA Stoke 25 0 16 0 45 0 0 0 0 100 Speed Direction Surf Lev Depth Profile CURRENT 2 45 0 0 100 00 L 20 20 0 T0 0 0 0 0 Identify Worst Phase Max Base Shear an
22. not only the force level to be transferred through the brace chord connection The force displacement characteristics P_d curve of the individual brace chord connections are derived from the actual peak capacities according to f inst API as follows Deformation 0 1 of chord diameter defines yielding Deformation 1 0 of chord diameter defines peak value Deformation 5 0 of chord diameter defines end of peak value Deformation 10 of chord diameter defines joint fracture The generated curves are printed to the out file and in XFOS the peak capacities are printed using the Verify Element Information option The joint behaviour is inspected in XFOS using the Result History plot and by selecting Element displacement vs Element force for end 2 of the joint spring elements the load through the joint is visualized This new option is controlled by the new USFOS command JNT FORM SINTEF group 2001 06 10 USFOS GETTING STARTED 5 3 The peak capacities are easily scaled up and down using the command JSURFSIZ sensitivity studies cracked joints reinforced joints etc In addition a new user defined joint capacity option is implemented This option allows the user to 100 control the P d curves of any brace chord connection S 3 Joint classification MSL joint characteristics This write up is a preliminary description of the implementation of MSL joint formulation in USFOS for use with the B release of the new
23. particular result plot and so on It is highly recommended to not use one input file set which is modified over and over again until all cases are run because Q Possible confusion about input parameters used a Difficult to repeat the analyses after a time a Requires manual editing before each new run impossible to automate It s better to plan and organise the USFOS analysis in a way that makes it possible to ultimately perform hundreds of analyses using only one magic command One solution among several is using UNIX scripts and the following sections will describe this solution USFOS even on Windows NT runs in a UNIX environment and all procedures described in the sections below are running on all computer platforms However some differences may occur f ex C TEMP on PC and tmp on standard UNIX The next sections will deal with use of UNIX commands typed in from the keyboard in the old fashion way It s therefor worth spending some minutes adjusting the UNIX command prompt window 8 1 1 Adjusting the UNIX korn shell window Windows NT 2000 Installations Only Before you start using the UNIX korn shell it s recommended to modify slightly the layout 8 1 1 hows the default window with white text an black background and with size 24 lines 80 columns To modify the window point on the blue top frame of the window and press the right hand button The menu pump imd Figure 8 1 1 The default NutCracke
24. plastic rotations as described by Van Langen 2 as the secant stiffness correction suggested by SNAME RP is not suitable for USFOS we use a tangent stiffness approach A new input card SPUDMAT has been introduced This can be used to give input both for sand and clay The old input card MSPUD is now obsolete but will still be available for compatibility with old models Also new in 2001 is visualization of the spudcan element in XFOS The element needs equilibrium iterations Spudcan elements should not be used without equilibrium iterations switched on Define Spud model 3 Spring 2 ground elements Elements Elem ID Node ID Material ID SPRNG2GR 1 359 SPRNG2GR 2 380 SPRNG2GR 3 Sample Clay parameters Mat No Typ V pre R eff D emb As_lat SPUDMAT 2 Clay 48 9E 06 7 0 i23 102 46 Gv Gh Gr Pois cuo 5 39E 06 5 39E 06 5 39E 06 0 5 26 98E 03 adh fac Suction flag backfill_flag I amp I IC7 0 9 0 0 0 5 T0 Table 6 3 1 Input for SpudCan elements with clay material parmeters SINTEF group 2001 06 10 USFOS GETTING STARTED Define Spud model Elem ID Node ID Material ID 359 380 SPRNG2GR SPRNG2GR SPRNG2GR Sample Sand parameters Mat No Typ SPUDMAT 2 Sand Vpre 10 0E7 R tot Apex ang 10 00 85 0 Fr ang Cohe E UW 35 0 0 0 0 01E 06 3 Spring 2 ground elements 6 6 Elements Gv Gh Gr 50e06 50e06 50e06 Pots 0 25 Table 6 3 2 Input for Spud
25. s Figure 3 2 1 Edit Clip Group Useful USFOS commands for the model repair work Q GROUPDEF Define element groups Q GROUPNOD Add nodes to groups guide loads towards nodes Q NONSTRU Define elements nonstructural Q STRUCTEL Define elements structural override NONSTRU for some elem Q LIN ELEM Define element linear elastic with and without elastic buckling SINTEF group 2001 06 10 USFOS GETTING STARTED 4 4 Wave Loading 4 1 Load Module In connection with dynamic analysis of structures exposed to loads which are dependent on the structural response it is not possible to pre define the load history The loads must be calculated ABERESZSEN during the analysis aa D A new load module has been developed and implemented in USFOS This new module is designed to handle e Hydrodynamic loads e Aerodynamic loads User defined load routines In connection with the hydrodynamic load module the following are implemented p e Hydrodynamic coefficients Ca and Cm The coefficients may be defined by depth profiles and or element by element e Marine Growth The marine growth thickness is defined by a depth profile e Buoyancy The buoyancy is calculated during the analysis which means that elements in the splash zone become buoyant non buoyant as the surface moves up and down e Flooded members Current The current is defined by speed direction and depth profile e Kinemat
26. should be necessary Elastic unloading in yield hinges is not allowed An arc length iteration procedure is implemented with a special algorithm for passing load limit points or bifurcation points Instead of keeping the external load level fixed during iterations the external load and displacement vectors vary according to a prescribed function in the load displacement space In the current formulation the loads and displacements are forced to move along a plane normal to the direction of the original load and displacement increment This is illustrated in Higure 1 3 6 A Rgxr AR LO Ayi 0 Ari Figure 1 3 6 Arc length iterations SINTEF group 2001 06 10 USFOS GETTING STARTED 1 20 Iterations have converged when the change in iterational load and displacement becomes smaller than a specified limit This is expressed by the following parameters S Jar 1 35 Ri TRITT it i it MI where AR and Ar are the load and displacement vectors at iteration j of step i and AR and Ar are the load and displacement increments for step number i The convergence parameters compare the changes in load displacement at each iteration with the corresponding changes during the initial increment iteration zero Thus the test values start from 1 00 and should be steadily reduced as the unbalanced forces vanish The actual convergence criterion is defined by the parameter epsit Iterations are terminated when t
27. spring back behaviour is often characterized by a large Current Stiffness Sp gt 1 in combination with a negative stiffness matrix determinant This is included in the load control algorithm of USFOS The user may change the value of the parameter cmax A Current Stiffness larger than cmax will be interpreted as spring back and the sign of the load increment will be reversed Currently cmax is set to a high value 999 i e only the determinant criterion is active in detecting spring back SINTEF group 2001 06 10 USFOS GETTING STARTED 1 19 1 3 5 Load control algorithm The following load control algorithm implemented in USFOS Positive increment Ap gt 0 IF The tangential stiffness matrix has no negative diagonal terms AND S gt 0 0 Negative increment Ap lt 0 IF The tangential stiffness matrix has one or more negative diagonal terms OR S lt 0 0 OR S gt cmax Figure 1 3 5 Load control algorithm 1 3 6 Equilibrium iterations The pure incremental algorithm generally causes a drift off from the true solution path Each step is a solution of the tangential stiffness matrix each step will move at a tangent to the true curve Corrections for this deviation can be taken care of by specifying equilibrium iterations on the unbalance between external loads and internal forces after each load step In the USFOS the tangent stiffness matrix is updated after each iteration New plastic hinges are inserted if so
28. the bounding surface during the loading process provides for a smooth transition from initial yield to full plastification In the bottom plot the force state has reached the bounding surface the cross section has reached full plastification From this stage the force state remain on the bounding surface and both surfaces will translate in contact Figure 1 2 3IRelates the multidimensional illustration in stress resultant space to a uniaxial stress strain curve SINTEF group 2001 06 10 USFOS GETTING STARTED 1 14 impl Taraa PP Lu Wu z moe Pug bh vu 8 8 D poe a a a a a a a a a aa es ee a an Se a eo ME n 40 man 7 5 m el nears Don Figure 1 2 2 Two surface plasticity model SINTEF group 2001 06 10 USFOS GETTING STARTED 1 15 S fully plastic o s a Figure 1 2 3 Analogy between multidimensional stress space and uniaxial stress strain curve 1 3 Implementation According to the updated Lagrangian formulation the load is applied in steps and the system stiffness equations are solved at every step After each step the structural configuration is updated element forces nodal coordinates etc are updated and plastic hinges are introduced if necessary Thus each step constitutes a full linear analysis based on the updated information from all previous analysis steps A combined incremental iterative loading algorithm is implemented As default
29. to develop 4305 requires additional data storage Figure 2 4 2 Input verification SINTEF group 2001 06 10 USFOS GETTING STARTED 2 10 exe NODAL POINT DATA NPEX NP X Y 1 1 000000 000000 2 2 1 524000 000000 3 3 3 048000 000000 11 H 3 048000 000000 12 12 000000 000000 i se 3 048000 000000 BED ELEMENT DATA DES ELEX ELNO ELTYP GEOM MATER NP1 1 1 BEAM 4 1 6 2 2 BEAM 4 1 7 3 3 BEAM 4 A 4 21 21 BEAM 1 2 8 22 22 BEAM 1 2 10 23 23 BEAM 1 2 11 RT LOCAL COORDINATE ELEX LCNO LOCAL X LO 1 1 707 000 707 000 1 2 2 707 000 707 000 1 3 3 HUI 1 000 707 000 1 21 21 000 000 1 000 1 000 22 22 000 000 1 000 1 000 95 d 000 000 1 000 1 000 vice GEOMETRY PARAMET GEO TYP Area Ixx Iyy Sect Sect mod x mod y 1 PIPE 7 101E 03 1 782E 04 8 908E 05 1 1288 03 7 1608 04 5 I H 1 570E 01 1 449E 04 6 045E 02 3 435E 02 9 372E 02 MATERIAL PARAMET MAT TYP Youngs Poiss Yield modul ratio stress 1 1 2 100EM11 3 000E 01 2 480E 08 3 1 2 100E411 3 000E 01 3 240E408 000 000 000 000 000 3820 3820 3820 CO CO 7620 0000 0000 NP2 5 6 6 LL 12 I3 S UxUS T CAL Y 000 000 000 000 000 000 000 ERS DAZ Sect mod z 908E 05 160E 04 519E 02 761E 02 ERS Density 7 850E 03 7 850E 03 BOUN COND 00 00 00 00 00 XXX xX X 00 XXX X X LCOOR ECCEN1 ECCEN2 1 0 0 2 0 0 3 0 0 2 0 0 22 0 0 23 0 0 EMS
30. to handle other input format is available free of charge on a as is basis no formal support is given STRUMAN is still considered to be a SINTEF in house program Since USFOS requires only one element per physical element a structural model developed for LINEAR analyses may be used more or less directly in the USFOS nonlinear analysis Little extra input is needed SINTEF group 2001 06 10 USFOS GETTING STARTED 2 4 All control parameters and additional input for nonlinear analysis may be specified separately in the Analysis Control File Structure data can also be read from this file but is usually given on one or two separate files The specific content of these files is not important as long as all data are present For convenience these files are labled Structure File and Load file igure 2 3 2 shows interactive program initiation User input is underlined RUN lt USFOS gt USFOS SUS ESO So Progressive Collapse Analysis of Frame Structures Version 5 3 Release 92 02 01 SINTEF div of Structural Engineering Control file prefix ZAYAS FRAME HEAD Structure file prefix ZAYAS FRAME STRU Load file prefix Result files prefix ZAYAS FRAME Figure 2 3 2 Interactive running SINTEF group 2001 06 10 USFOS GETTING STARTED 2 5 2 4 Output The main results of a USFOS analyses are Ultimate collapse load or critical collapse temperature Energy absorption Load displacement rel
31. to which the element are connected to are listed GROUP DEFINITIONS GROUP label Geometry Group no Contains following Geometries 10101 10228 10229 10230 10231 10252 10352 15198 15199 16106 17535 17600 17634 5001 5002 5003 5004 5005 5006 5007 5008 5009 5010 5011 5012 5013 5014 5015 5016 5017 5018 5019 5020 5021 5022 5023 5024 5025 5026 5027 5028 5029 5030 5031 5032 5033 5034 5035 5036 78614 78615 755507 755508 726550 726551 726500 726501 54531 54837 54942 54943 54859 Table 3 2 5 Print of group data geometries elements and nodes on the out file The example shown in Figure 32 1 represents a first stage in a model repair procedure The entire structure is still structural but members are grouped as specified above By using the Edit Clip Group command in xfos it s possible to visualise the different groups include exclude The image to the right shows the full model and by excluding all groups as seen in the Specify Clip Group menu the image to the right appears If the NONSTRU command in Table 3 2 3 s activated note that the passives the command only the elements in the image to the right remains structural but loads are attracted on the full structure image to the left SINTEF group 2001 06 10 USFOS GETTING STARTED 3 5 Specify Clip Group C Include matching element s Exclude matching element s 1000 woup no C Include matching element s Exclude matching element
32. 00 v 2 105 X 6 0000 006 i A 40000006 Fomat Startimes FROM DORN dH Zt o x o d o si o d r i Plot Export Cancel Time ssconds Ok Print Save Cancel Figure 7 3 1 Selecting Dynamic Results from XFOS SINTEF group 2001 06 10 USFOS GETTING STARTED Following results are a NODAL Displacement Velocity Acceleration Relative displacement between two nodes a ELEMENT Displacement Force a GENERAL Internal Energy Plastic Energy Kinetic Energy Total Energy See Table 7 3 1 for example of use Type Node ID Dof RES Node Dis 10 1 RES_Node Dis 130 1 RES_Node Vel 130 1 1 RES_Node Acc 130 Type Node_ID Dof Node ID Dof RES Node RelDis 10 130 Type Elem ID RES Elem Disp 20 RES Elem Force 20 Type RES General Wint RES General Wplast RES General Wkin RES General Wtot Table 7 3 1 Input for Dynamic result saving See also in the example folders dyn drop dyn imp dyn imp2 dyn quak Oooo SINTEF group 2001 06 10 USFOS COURSE MANUAL 8 1 8 Efficient use of USFOS 8 1 General Seldom only one USFOS analysis is performed for a given problem The more typical use is repeated runs due to several load cases parametric sensibility study model change etc In cases where many USFOS analyses should be performed well organising of both input and output files is important There should be no doubt about what was the parameters used for this
33. 17 16517 15108 16216 10123 10126 17617 17616 16116 16218 16119 16318 16118 20085 17619 17618 20122 16418 20112 16518 17620 16120 16320 16620 16220 19101 GroupDef 3000 Geo Pipes 10102 10104 10105 10106 10107 10112 10113 10128 10130 10131 20072 20113 19109 20075 20076 20077 20073 20074 20080 20082 10185 10186 10102 10253 16402 16502 17502 17602 10114 10365 10183 GroupDef 16319 Geo 16319 GroupDef 16219 Geo 16219 Table 3 2 3 Shrinking model using the GROUPDEF and NONSTRU commands SINTEF group 2001 06 10 USFOS GETTING STARTED 3 4 If the definition of the bounding surface the gbound command is left out for general sections default values are used and a warning is printed see able 8 4 2 The default values are shown in the same table Warning GBOUND specified General Default used Warning GBOUND specified General H Default used Warning GBOUND specified General i Default used Warnina GROUND soe ified General Default used GBOUND 10101 0 8 1 0 0 6 1 0 Table 3 2 4 Default Gbound data assigned to general beams When element groups are defined the contents of the different groups are listed in the out file see Table 3 2 5 In the actual example group no 000 is defined through geometry ID s and the specified ID s are listed first similar if the group was defined through material ID s Next the elements which are members of group no 1000 are listed and finally all nodal point
34. 760E 01 1 2 2 000 1 000 8 760E 03 3 504E 02 1 3 3 000 1 000 1 314E 02 7 885E 02 1 4 4 000 1 000 1 752E 02 1 402E 03 J 5 5 000 1 000 2 190E 02 2 190E 03 3 6 5 500 1 000 2 409E 02 2 650E 03 1 7 54923 1 000 2 595E 02 3 074E 03 6 PLAST END1 1 8 6 099 948 2 676E 02 3 270E 03 8 PLAST END2 1 9 6 489 887 2 869E 02 3 754E 03 1 PLAST END2 1 10 6 500 865 2 874E 02 3 769E 03 1 11 6 535 865 2 892E 02 3 815E 03 3 PLAST END1 1 12 6 585 837 2 918E 02 3 884E 03 1 13 6 609 836 2 931E 02 3 917E 03 3 PLAST ID 1 14 6 631 585 2 943E 02 3 949E 03 3 UNLOD END1 1 PLAST END1 1 PLAST ID L AXIAL ID 2 PLAST ID 2 AXIAL ID Figure 2 4 1 Global history output SINTEF group 2001 06 10 USFOS GETTING STARTED 2 7 Table 2 4 1 Terminology global history table USFOS load combination Load combination number or basic load case number Load step Number of times the initial load has been incremented Load level Relative load level of the current load combination and load step The load level is local within each combination starting from zero when a new load combination is specified Current stiffness Structural stiffness The initial stiffness is 1 0 Decreasing value represents a decreasing stiffness in the structure Control displacement Equivalent displacement of the structure The displacement is calculated as a balanced average of selected displacements Energy absorption Accumulated external work absorbed by the structu
35. ARTED 8 3 Figure 8 1 4 Defining window layout Le om Figure 8 1 5 Selecting permanent modification of the short cut The UNIX window will from now on look like the one in Figure 8 1 6 with two scroll bars and it s resizable and a comfortable colour Figure 8 1 6 The modified NutC window with scroll bar SINTEF group 2001 06 10 USFOS GETTING STARTED 8 4 8 2 Some UNIX commands The procedures described in the examples below require that the users knows some UNIX commands and in the following a brief summary of the commands used in the scripts is given Description Copy one file into another Rename a file or directory dump the content of a file to screen dump the content of file 1 into file 2 dump content of filel behind existing content of file 2 append create a directory folder change directory directory path one level up directory path two levels up directory path one level up and one down Environmental variable with name NAME Show me the content of the environmental variable whit name NAME Stream Editor Delete file s Delete directory List files List all files with extension fem Table 8 2 1 UNIX commands overview SINTEF group 2001 06 10 USFOS GETTING STARTED 8 5 8 3 Example 1 Fixed USFOS input file names The simplest example on a UNIX script which saves you for tediously typing is a file with name gol containing following SUSFOS HOME bin us
36. Can elements with sand material parmeters Figure 6 3 1 Simplified Jackup structure with spudcan elements References 1 SNAME 1997 Recommended practice for site specific Assessment of mobile Jack Up units Rev 1 The society of naval architects and marine engineers Jersey City Nj 2 Langen H van Hospers B 1993 OTC 7302 Theoretical model for determining rotational behaviour of spud cans Proc Offshore Technology Conference Houston Texas May 1993 Dynamic Analysis SINTEF group 2001 06 10 USFOS COURSE MANUAL 7 1 7 Dynamic analyses 7 1 Introduction Dynamic analysis can be performed for given load time histories and for ship collision In the latter case the impact velocity of the ship mass is treated as the initial condition for a free vibration problem Two options exist for the mass of the structural element Consistent mass based on interpolation functions for the linear 3D beam Thus it is not truly consistent with the displacement shape function used in USFOS but accurate enough for most practical purposes Lumped mass yielding a diagonal mass matrix In this case the rotational masses are scaled by a factor denoted rotmass The scale factor should be fairly low in order to maintain accuracy for high frequency modes By default this is set equal to 0 01 Linear damping may be given in the form of Rayleigh damping with one term proportional to the system mass and one term proportional to the s
37. IELT SUR ERREUR MEE NE SINTEF group 2001 06 10 USFOS GETTING STARTED 9 14 9 2 4 1 Program input Control flle head lt direction gt fem One file for each loading direction Structure file stru fem One single file Load file load lt direction gt fem One file for each loading direction Load case 1 Dead load live loads Listed on the stru file Load case2 Buoyancy Listed on the oad file Load case2 Wave Current Wind loads Listed on the oad file 9 2 4 2 Analyses l e Perform one pushover analysis for each platform direction with the following input no GELIMP or GIMPER e no initial deformations records e local buckling formulation OFF use DENT OFF 9 2 4 3 Documentation 1 Generate P plots from the final analysis of each direction 2 Take hardcopies of deformed geometry with member utilization at first member failure max load and at the final analysis step 3 Generate plots of axial force vs global displacement N 6 plots for critical members in the failure mechanism buckling members tension failure members or failing leg members The purpose of this is to document the redistribution of forces and the development of the final failure mechanism 9 2 4 4 General comments Now is the time for real life analyses We should utilize any prior knowledge about the structure a If the structure is designed according to NPD regulations it will have elastic response up toaloading 1 6 times the characte
38. KN load PARENTHESIS SINTEF group 2001 06 10 USFOS GETTING STARTED 9 1 4 3 Description One compression brace of the upper panel fails when the external load reach alevelof From this level the load in the other compression brace decreases and the load is transferred to the compression braces of the panel Atloadlevel _ a compression brace of the lower panel fails Again the other compression brace is unloaded and further loading carried by the tension braces Atloadlevel the full plastic capacity of the tension braces is utilized Any further loading is carried entirely by the legs 9 1 4 4 Program input Control file head fem Structure file stru fem Load file load fem Load case 1 Gravity load Load case2 Horizontal load Ho Load case3 Horizontal load Ho 9 1 4 5 Analyses 1 Perform a pushover analysis of the frame with the following input e Load case 2 e no initial deformations e local buckling formulation OFF Perform a pushover analysis of the frame with the following loads e Load case to loadlevel 1 0 e pushover with Load case 2 Create a load combination of loadcase 2 amp 3 named loadcomb 4 e Load case to loadlevel 1 0 e pushover with Loadcomb 4 Use Load case 2 Apply initial deformations Experiment with step size and equilibrium iterations use DENT_OFF modify CUSFOS use CCOMB modify CUSFOS use CINIDEF modify CUSFOS use CITER use DETEROFF SI
39. NTEF group 2001 06 10 USFOS GETTING STARTED 9 6 9 1 4 6 Documentation Use XFOS to generate the following 1 Global load vs global deformation 2 Deformed geometry plots at first member failure collapse and at final analysis step 3 Axial force vs global displacement for members 1 3 5 and 7 9 1 4 7 Comments to the analyses As stated previously the load step size should reflect the nonlinearity of the structure This is not known in advance and some trial analyses should usually be expected Generally the load steps may be relatively large in the linear range and should be reduced as the response becomes nonlinear Particulary in spring back areas both the load step size and the min step should be considered carefully Generally speaking the predicted P 6 curves moves at a tangent to the true P d curve Too large steps in a spring back area could cause considerably drift from the true solution As a general statement the comments from the previous cases still apply a use some 5 10 steps in the linear range to activate geometric nonlinearities b check the min step so that the analysis is not blocked by too many load steps but that the forces still are scaled to the yield surface Too many steps scaled to MIN STEP LENGTH too early in the analysis is usually a sign that the min step may have been set too large Close to the peak load however it is inevitable that many steps are scaled to minimum c check t
40. OBAL TOTAL DISPLACEMENTS GLOBAL REACTION FORCES NODE Element identification number Element status These values represents the accumulated stress level of each element position with the value 1 00 in the initial stress free configuration and the value 0 00 when first fibre yield or the full plastic capacity is reached The primary columns concerns the full plastic capacity Value 0 when the full plastic capacity is reached the secondary column in brackets reaches zero on first fibre yield Short comments when the status of an element is changed Position where the element status changes Representation of the element status corresponding to the value ES Yield hinge inserted Position where element forces are checked for plasticity Plastic hinge removed at element midspan Internally the element is still divided in two sub elements Plastic tension failure The axial tension force has reached the plastic capacity and a membrane element is inserted This table shows the total accumulated displace ments up to and including the current load increment This table show the total accumulated reaction forces up to and including the current load increment Nodal point identification number SINTEF group 2001 06 10 USFOS GETTING STARTED 2 13 2 5 Practical Considerations 2 5 1 Load specification The user supplies the basic oad cases as input to USFOS In the run commands to the program th
41. POSTFOS will allocate a default amount of your computers memory at startup Sometimes this is not enough to read your big model To allocate more or less than default memory at startup simply add a number after the command ex gt usfos 50 this will start USFOS with 50 million words of memory To change the amount of memory POSTFOS allocates when started from XFOS you have to edit the Xfos file in your HOME directory Change the parameter postfos size to suit your needs SINTEF group 2001 06 10 USFOS GETTING STARTED 2 2 Table 2 2 1 System files Type Content USFOS Input ANALYSIS FEM Control parameters for the USFOS nonlinear analysis CONTROL FILE Generated manually STRUCTURE FEM Finite element idealization of the structure Generated manually MODEL FILE or by preprocessor program LOAD FILE FEM Structural loads Generated manually or by load generation program USFOS Output ANALYSIS OUT Print of analysis results Input verification global history output PRINT FILE or output of each load step ANALYSIS RAF Structure data and analysis results of each load step DATA FILE POSTFOS Output POSTFOS PRINT PRI Printed tables of USFOS analysis results FILE S POSTFOS PLOT PLO Plot data of USFOS analysis results FILE S SINTEF group 2001 06 10 USFOS GETTING STARTED 2 3 2 3 Input The USFOS analysis module reads data from one two or three input files With exception of
42. RTED 9 26 Wave current induced forces on the jacket have been estimated with the use of the computer program WAJAC DNV 1982 by applying a Stokes 5 order wave theory Base shear forces obtained for eight wave heights in longitudinal 0 transverse 90 and diagonal 45 wave heading are given rosea Base shear Wave height ton m Longitudinal 0 Transverse 90 Diagonal 45 Table 9 5 5 Wave current induced loads on the jacket 9 5 6 Program input Two different structure models have been prepared One is without piles fixed fem The other one is piled utilizing the USFOS pile and soil modelling piled fem A load file utilizing the USFOS Wave input has been prepared Wave and current data corresponding to 99 year return period and transverse loading is given initially Control file head fem One file Structure file stru fixed fem Two files stru piled fem Load file load fem One file Load case 1 Dead load live loads Listed on the stru file Load case2 Buoyancy None Load case2 Wave Current loads Listed on the oad file SINTEF group 2001 06 10 USFOS GETTING STARTED 9 27 9 5 7 Analyses e Perform one pushover analysis with the given load input and the fixed structure e Perform one pushover analysis with the given load input and the piled structure e Repeat analysis 1 with joint check swiched on use CHJOINT e Repeat analysis 1 and 2 with differen
43. The first yield and fully plastic capacities are represented by yield surfaces based on plastic interaction between element forces e The load is applied incrementally e The load increment is automatically reversed if global instability is detected e The effect of initial deformations and local buckling are included for beam elements e Joint capacity checks and joint behaviour is implemented according to the API rules e Member rupture and redistribution of forces from ruptured element is fully integrated in the analysis procedure e Hydrodynamic loads may be specified directly without using a separate hydrodynamic load generation program e Pile and Soil data may be specified directly without using a separate pile soil interaction program SINTEF group 2001 06 10 USFOS GETTING STARTED 1 6 1 3 Theoretical basis The formulation behind the program is valid for large displacements but restricted to moderate strains USFOS follows an updated Lagrange formulation 1 2 1 Continuum Mechanics The formulation is based on Green strains defined by 1 1 1 57 up E VE Wa 1 1 For moderate element deflection the von Karman approximation applies and x simplifies into 1 1 T Uxt VE Wi 1 2 The stiffness formulation of USFOS is derived from potential energy consideration or the virtual work principle For an elastic beam element the internal strain energy reads i 1 1 1 U EA Pan io 7 A dx mo Zz E xa dx 1 3
44. USFOS GETTING STARTED 1 1 USFOS Getting Started USFOS GETTING STARTED 1 2 Table of Content SINTEF group 2001 06 10 USFOS GETTING STARTED 1 3 SINTEF group 2001 06 10 USFOS GETTING STARTED 1 4 1 UNDERSTANDING USFOS 1 1 Basic features USFOS is a numerical tool for ultimate strength and progressive collapse analysis at space frame structures The formulation includes nonlinear geometry and nonlinear material properties The basic idea of the program is to use only one finite element per physical element of the structure 1 e to use the same finite element discretization as in linear elastic analysis e USFOS operates on element stress resultants i e forces and moments e Material nonlinearities are modelled by plastic hinges at element midspan and at element ends te A I J 1elemen P M cosl P Es P E pee ze te plastic hinge CES 0 Elastic 1 Mo Non linear material Non linear geometry Figure 1 1 1 USFOS basic concepts SINTEF group 2001 06 10 USFOS GETTING STARTED 1 5 e Effects of large displacements and coupling between lateral deflection and axial strain are included by using nonlinear strain relations Green strain This gives a very accurate representation of element behaviour including membrane effects and column buckling e Material models are included both for elastic perfectly plastic behaviour and gradual plastification strain hardening characteristics
45. a pure incremental procedure is adopted Equilibrium iterations may be specified by the user Global instability collapse is detected by a formulation based on the Current Stiffness Parameter in combination with a Determinant Criterion When instability is detected USFOS reverses the sign of the load increment and the analysis proceeds into the post collapse range 1 3 1 Load Specification The user supplies the basic oad cases as input to USFOS In the run commands to the program the load cases are combined into load combinations if necessary Then the user specifies the loading history for the analysis which loads are to be applied in which order and how large increments are to be used The loads are incremented a given number of steps up to a specified load level or until a defined displacement is reached The load is incremented on top of the previous loads i e each load increment is added to the accumulated load of all previous load steps Thus the calculated results at each step are the combined results of the total load history prior to and including that step The results of different load cases may not be superposed since the response of the structure is highly history dependent In nonlinear analyses the actual loads must be combined SINTEF group 2001 06 10 USFOS GETTING STARTED 1 16 1 3 2 Load step scaling In each load increment the program first applies the load increment specified by the user Then
46. ain Strucutre a Main load located in loa Main Load Special information Q Support Structure str Spring Support 1 and 2 a Special Load loa Nodei Load Contents of Example 3 Assembling Files File Folder File Folder a head fem FEM File ge run all File ja go File Figure 8 5 1 Content of file folder before running script run all The idea is as follows Q Use the control file head fem in all cases a Compose a structural file consisting of the common Main Structure and the special support and assemble the complete structural model in the file stru fem a Compose a load file which should consist of the common load file Main Load and the special nodal load and collect all load info in the file oad fem a Create a new unique directory below current directory for each case with informative name reflecting the actual case a Run USFOS an save stru and load files result files on the actual directory Q Create script go for running on case and run all for running all 6 combinations In Table 8 5 T he script with name go is described in detail as it appears in the example folder Lines staring with the sign is comment lines and may appear anywhere in the script file except between lt lt ENDIN and ENDIN It is recommended to use comments both in scripts and in the USFOS input files Firstly the cp command is used to copy the main structure to the file stru fem Next the selected suppo
47. ameter pipes D lt 300mm are grouped because elements referring to those beams are the typical secondary members which should be removed from the analysis model 3 2 Utilizing Group definitions Q Groups are introduced in the latest USFOS version 7 7 A group is identified by its ID which is a number up to 8 digits a Elements become members of groups and the same element may participate in several groups a The nodal points to which the elements are attached becomes members of the actual group D The groups are referred to in connection with assigning properties to elements which will ease the input reduce the amount of input lines In xfos its possible to include exclude groups in the structural image Edit Clip Group Elements are defined members of a group using the GROUPDEF command The element may be identified through Element ID All elements referring to given material ID s All elements referring go given cross section geometry ID s All elements members of existing groups DODDO The actual way of defining the e using the parameters Elem Mat Geo or Group as shown in Table 3 2 1 ID List GroupDef 10 20 30 GroupDef 1 GroupDef 5 GroupDef 88881 88 Table 3 2 1 Defining element groups using of the GROUPDEF command SINTEF group 2001 06 10 USFOS GETTING STARTED 3 3 If wanted extra nodes could be defined members of an actual group and th
48. ations at any nodal point Element forces at all load levels Formation of plastic hinges Redistribution of forces The results are presented in the following way As plots and images presented by the graphical post processor XFOS As printed tables presented by the POSTFOS module As analysis print out on the Analysis Print File generated by USFOS during the analysis As on line print out to terminal or batch output stream SINTEF group 2001 06 10 USFOS GETTING STARTED 2 6 2 4 1 Global history output The global history output gives an overview of the total global behaviour of the structure during loading Load displacement energy and structural stiffness are listed at each loadstep with the formation or removal of element plastic hinges The global history output may be written to the batch out stream during analysis or may be generated by POSTFOS The formation of plastic hinges is listed at each load step This is of particular interest to determine the redistribution of forces throughout the structure and to isolate the elements that trigger the final collapse The global history output is shown in Each term is briefly commented in m U USFOS ANALYSIS RESULTS vb Z AYASs FRAME USFOS progressive collapse analysis SINTEF div of Structural Engineering USFOS load Load Load Current Control Energy Elem Event Event comb step level stiff displ absorb no type pos 1 1 1 000 1 000 4 380E 03 8
49. ble progress in the analysis the minimum step length parameter is introduced However the minimum step size should not be specified too large If more than say 5 hinges are introduced in a MIN STEP LENGTH step this may indicate that the step is too large The true yielding process may not be identified how yielding in one member influence the load redistribution to other members and the correct failure mechanisms may not be initiated As already stated finding the right load specification will often be an iterative process 2 5 3 Analysis verification 2 5 3 1 Iteration Convergence Iterations may be specified to ensure equilibrium between external loads and internal element forces To verify the numerical accuracy of the analysis the iteration test parameters should be checked for each load step The test parameters are printed to the OUT file for each iteration at every load step To ensure the required numerical accuracy the final values should be below the input convergence criterion Note that the program may terminate iterations without having reached convergence This is done to avoid known divergence situations identified by the following criteria e Ifthe Current Stiffness parameter changes sign during iterations e Ifthe Determinant changes sign during iterations SINTEF group 2001 06 10 USFOS GETTING STARTED 2 15 If iterations fail to converge observe the following 1 Verify that the solution conver
50. boundary conditions for the beam The advantage of using the above shape functions is that all integration in the element stiffness expression can be carried out analytically and the element stiffness matrices presented as closed form expressions Furthermore the quality of the shape functions allows for a very simple modelling one element between each joint is normally sufficient to simulate the nonlinear column behaviour with satisfactory accuracy As USFOS employs exact element displacement functions satisfying the governing differential equation USFOS should predict the elastic buckling load for the three cases in Figure 1 2 1 Itis observed that the element equilibrium equations becomes singular at the exact Euler buckling formulas SINTEF group 2001 06 10 USFOS GETTING STARTED 1 10 P M4 P k 2EJ 20 zi L 9 Pon Pe 4EJ Me SSeS det k 0 for 9 0 Pry 2 04 P t x be d P 3e 9 29 3lo 270 det k 0 for 129 9 99 P 0 25 P Figure 1 2 1 Elastic Column Buckling SINTEF group 2001 06 10 USFOS GETTING STARTED 1 11 1 2 4 Plasticity Formulation Material nonlinearities are modelled by yield hinges Plastic hinges may be inserted at element ends or at element midspan In the latter case the original element is divided into two subelements The extra nodal point is introduced automatically and eliminated by static condensation before adding into the global stiffness ma
51. by connecting a loadvector to a time history AT AT AT End Time 1 End Time 2 End Time 3 Time Figure 7 2 2 Specification of time increment to be used with the different time interval 7 3 Dynamic Analysis results Time Series A dynamic analysis may involve a large number of analysis steps 1000 100 000 and saving of analysis results is then a challenge It is then necessary to select a few results which could be saved every analysis step while the rest of the results could be saved more seldom In this way the user obtain following a High density on the time series of the selected most important results a Acceptable density on the results presented in XFOS for inspection of the global behaviour of the structure f ex generation of animation etc The few selected result quantities are stored in a separate file with extension dyn in addition to the usual raf file The dynamic results are accessed from XFOS through the result dynamic result dialogue box see Figure 7 3 1 Plot 9 000e 005 Internal Energy oaj E 6 000e 005 5 000e 00 4 000e 00 3 000e 005 2 000e 005 Vertical axis customization gt E 5 a z E 2 Scientific notation Decimal floating point 3 1 000e 005 Number of decimals 0 000e 000 f EE 1 000e 005 f 0 0 i Export Ok ini ave File prefix t Riserifinallimp Mass 40
52. cm was used except in the upper 11 meter segment of each pile in which a 345 MPa 3520 kg cm yield stress steel was utilized The soil in the Bay of Campeche is mainly comprised of alternating layers of calcareous clays and sands The calcareous sands are of moderate density and the clays are generally stiff and consolidated with exception if the upper layer which is soft A total of ten soil layers are identified in the site and P Y and T Z curves for twenty seven different elevations are given in the geotechnical study of the soil The soil characteristics are summarized in Friction Shear Skin friction Skin friction angle strength Compres tension ton m ton m ton m Unit weight Soil layer ton m 1 2 3 4 5 6 7 8 9 Vertical coordinate of the layer bottom measured from the mudline Table 9 5 1 Soil characteristics SINTEF group 2001 06 10 USFOS GETTING STARTED 9 23 Depth corner piles 98 74 m Figure 9 5 2 Foundation and soil characteristics for the platform SINTEF group 2001 06 10 USFOS GETTING STARTED 9 24 9 5 4 Load modelling for reliability analysis The wave and wind loading on the structure is calculated according to the recommended procedure in the API RP2A API 1994 by using the environmental parameters required for assessment of a fixed platform located in the Bay of Campeche These parameters are sum
53. d do not create a loadfile Criterion EndT Baseshear 1 20 0 noWrite Scale the Wave load This option is required when Force Unit is not N generated wave loads are always using Newton In this demo case scale by 1 3 Table 4 2 1 Input for automatic wave calculations and automatic member imperfections SINTEF group 2001 06 10 USFOS GETTING STARTED 5 S Joint Modelling 5 1 Joint capacity check Depending on the joint geometry the capacity of the connection brace chord is less than the brace capacity This means that the brace can not be utilized 100 96 In convential joint models the limitations in load transfer through the chord surface are neglected The user specifies the nodes where tubular joint capacity should be considered USFOS then calculates the geometry of the tubular joints and introduces extra elements nodal points geometries and materials in the finite element model The capacitites are calculated according to API Figure 5 1 laldescribes the user defined finite element model of a tubular joint and describes the modified input model Beam J Beam P oie Two extra nodes Two extra elements Beam __ Beam a Conventional joint model b E withcapacity check included Figure 5 1 1 Joint capacity modelling The numbering of the extra nodes and elements are as follows see generated by USFOS Extra element Extra element with no 1381
54. dure has been implemented to identify members with repeated on off loading and to prevent them from stopping the analysis e The user can specify an acceptable number of subsequent load steps with plastification elastic unloading for one element e fthis number is exceeded the element is prevented from unloading in the subsequent steps but new hinges may still be introduced e The restriction is removed the first time the element goes through a load step without trying to unload e The restriction is also removed on the first step of every new load vector each new CUSFOS CICYFOS line In particular all elements are free to unload when the external load is reversed In general such a restraint introduces artificial restraints in the solution and should be used with care But the error introduced will often be less than the inconvenience of clogging up the analysis If you still get problems related to repeated on off observe the following steps 8 Check if there is any indication of a bifurcation point in the previous load steps 9 If not set a reasonable number of on off s before locking members 10 Check if they continue on off at the next load vector next CUSFOS CICYFOS line in that case the locking is probably OK SINTEF group 2001 06 10 USFOS GETTING STARTED 2 17 11 If they unload and stay that way after the next load vector this indicates real physical unloading and that the previous locking may have
55. e beam column form a kinematic mechanism For a simply supported column buckling occurs when a plastic hinge is formed at member midspan The load is then reversed and the column is unloaded into the post buckling range If a two surface plasticity model is used buckling takes place some load steps after first fibre yield when the stiffness of the mid cross section has been reduced enough for the column to become unstable 9 1 3 4 Program input Control file head fem Structure file stru fem Load file none Load case 1 Axial load Po SINTEF group 2001 06 10 USFOS GETTING STARTED 9 5 9 1 3 5 Analyses Perform a buckling analysis of the column with the following input e initial deformation 0 001 use GELIMP and GIMPER records e local buckling formulation OFF use DENT_OFF 2 Repeat analysis 2 with initial deformations set to 0 002 use GELIMP and GIMPER records 3 Repeat with initial deformations set to 0 003 4 Turnon the local buckling formulation and repeat analysis Ig e initial deformation 0 001 use GELIMP and GIMPER records local buckling formulation ON use DENT_OFF 5 Define an initial dent of 10 of the tube diameter and use the GIMPER record repeat analysis 5 9 1 3 6 Documentation Use XFOS to generate the following documentation 1 Global load vs global deformation 2 Axial force vs bending moment 3 Axial force vs global displacement and bending moment vs global displacement For analysis 1 on
56. e command groupnod is used for this purpose see Table 3 2 2 This command is used in connection with guiding loads from non structural members towards kept structural nodes i Group ID GROUPNOD 8 Table 3 2 2 Assigning extra nodes to a group using the GROUPNOD command When the groups are defined one single NONSTRU command will remove all the members of the actual groups from the analysis model but loads are kept Specify Groups Which should GroupDef 1000 Geom GenBeams 10101 10228 10229 10230 10231 10352 15198 15199 16106 16129 16193 16194 16195 16196 16198 16199 16206 16229 16293 16294 16295 16296 16297 16299 16306 16329 16393 16394 16395 16396 16397 16398 16406 16429 16493 16494 16495 16496 16497 16498 16499 16529 16593 16594 16595 16596 16597 16598 16599 16606 16693 16694 16695 16696 16697 16698 16699 17529 17592 17594 17597 17598 17606 17629 17693 17694 17695 17696 17698 17535 17600 17634 GroupDef 2000 Geo t Pipes 19107 19108 16202 16302 16102 17502 17602 16602 10102 10104 15110 19106 10106 10107 19105 15106 20110 10113 10360 10111 15185 15107 10112 19103 20096 15114 15113 20097 15191 20099 16607 16407 16307 16107 10365 10243 17607 10367 10118 10116 20111 10122 16213 16214 16109 16110 16209 16114 16112 17509 17510 16614 16612 16613 17512 17612 17613 17610 17514 17609 16314 16410 16313 16309 16310 16312 16512 16514 16510 16414 10121 10120 19102 16617 16616 17517 15115 16516 16417 163
57. e history is a scaling factor time curve as shown in interpolated extrapolated 2 n 3 zr Scaling factor Time t t t 5 4 3 o S D amp i o o 1 2 4 5 Time Figure 7 2 1 Time history examples The upper time history is a typical apply dead loads history The loads connected to this time history are scaled up to the actual level at time t and then be kept constant the rest of the analysis For times greater than tz the extrapolated line through the two last points is used The lower time history example may be an apply impact load history The loads connected to this time history is sleeping up to time t where the loads are scaled to the actual level at time t3 Then the load is reduced causes negative load increments internally in USFOS until time t4 is reached from where the load level is kept equal to zero SINTEF group 2001 06 10 USFOS GETTING STARTED 7 3 A loadvector combined with a time history is called a load history and an unlimited number of load histories may be defined A loadvector may be combined with several time histories and a time history may be combined with several load vectors The records used to define the analysis are DYNAMIC Defines At time increment to be used within a time interval defined by the time terminating the interval see TIMEHIST Defines a time history identified by an ID an described by discrete points LOADHIST Defines a load history
58. e load cases are combined into load combinations if necessary Then the user specifies the loading history for the analysis which loads are to be applied in which order and how large increments are to be used The loads are incremented a given number of steps up to a specified load level or until a defined displacement is reached The load is incremented on top of the previous loads i e each load increment is added to the accumulated load of all previous load steps Thus the calculated results at each step are the combined results of the total load history prior to and including that step The results of different load cases may not be superposed since the response of the structure is highly history dependent In nonlinear analyses the actual loads must be combined 2 5 2 Analysis spesification In nonlinear analyses the accuracy of the results depend on the size of the load steps The load steps may be large as long as the structure behaves linearly The more nonlinearly the structure behaves the smaller the load steps should be That is the optimum load specification is closely linked to the nonlinear characteristics of the structure itself In the USFOS formulation this problem is partly solved by the automatic load scaling used when plastic hinges are introduced but the user still have to supply sensible values for the size of the initial load increment and for the minimum load step in the load scaling algorithm The cor
59. e the structural model is stored in the model folder The content of the script files are described in able 8 7 T TabIe 87 2 and Contents of Example 5 Element Removal Contents of Example 5 Element Removal Elem 01 Elem 02 Elem 03 Elem 04 Elem 5 Elem 05 5 and 12 Elem 6 Elem 07 Contents of etc Contents of model Elem 08 Elem 09 Elem 10 Elem 11 Elem 12 and 13 ai substitute Figure 8 7 1 Files Folders before and after running the script Define varible SCRATCH directory for Raf file storing export SCRATCH tmp scratch Local Dir Element to remove elmdel Elem 01 01 elmdel Elem_02 02 elmdel Elem_03 03 elmdel Elem_04 04 elmdel Elem_05 05 elmdel Elem_06 06 elmdel Elem 07 07 elmdel Elem 08 08 elmdel Elem 09 09 eimdel Elem 10 10 elmdel Elem 11 ld elmdel Elem_12_and_13 I2 elmdel Elem 05 06 and 12 5 End of Run A11 Table 8 7 1 Script file run all The run postfos script runs POSTFOS and creates the default history table using the define history and print history commands Similar scripts could be created for extracting nodal displacements of selected nodes element forces etc SINTEF group 2001 06 10 USFOS GETTING STARTED 8 16 SUSFOS_HOME bin usfos lt lt ENDIN SUSFOS_HOME bin postfos lt lt ENDIN head stru od load SSCRATCH res define hist ENDIN print hist vrs Table 8 7 2 Scrips run_usfos and run_pos
60. each is listed on the analysis print file the OUT file Residual forces and permanent deformations are stored on the RAF file like the results from any load step of a pushover analysis This can be utilized for residual strength analyses by simply specifying a restart from the final step of the ship impact analysis Then the effect of residual forces and permanent deformations will be included in the residual strength analyses Note that the environmental forces then will have to be read into USFOS in the initial analysis read into the RAF file but not applied to the structure It is not possible to read in new load cases in a restart analysis 9 3 5 Longitudinal Impact on Leg A4 An impact energy of 14 0 MJ is applied to corner leg A4 at elev 1 0 meter Impact is specified in the longitudinal direction Impact position end 2 of member 732 The ship is assumed infinitely stiff all energy is absorbed by the structure The applied impact forces can be absorbed by the structure without danger of capsizing Permanent deformations at the point of impact is m Permanent deformations at deck level is m Initial yielding occurs in the hit member at 9o of the impact energy Membrane action is activated in the hit member and forces are redistributed to the surrounding structure Member 7 _ falsat oofthe impact energy followed by member at 96 The global response is fairly linear during the impact Significant local deformation
61. ection Period 20 0 0 16 0 20 p 16 20 16 20 E 16 24 T 20 24 20 Table 8 6 4 Script file run all USFOS Extreme Wave Height 20 0 Dir 00 0 Progressive Collapse Analysis JACKET model SINTEF 2000 Define Wave type H Period Direction Phase Surf Lev Depth WAVEDATA Stoke 20 0 16 0 00 0 0 0 0 0 100 Speed Direction Surf_Lev Depth Profile CURRENT 2 00 0 0 0 100 OO ct iO 20 0 1 0 100 0 0 0 110 0 0 0 Table 8 6 5 USFOS control file modified by the SED editor After all 8 cases are run 8 new directories are created see Figure 8 6 1 containing the modified head fem and the analysis results Figure 8 6 2 hows results from one of the 8 analyses and NOTE that the member imperfections command CINIDEF are applied automatically according to the actual wave load direction which here is 30 SINTEF group 2001 06 10 USFOS GETTING STARTED 8 14 Figure 8 6 2 Case with H 20m Dir 30deg and T 16s SINTEF group 2001 06 10 USFOS GETTING STARTED 8 15 8 7 Example 5 Procedure for element removal redundancy analysis The final example solves following problem a Remove the structural members one by one a Use the same structural file and control file a Save the results from the analyses in separate file folders Figure 8 7 1 shows the content of the example folder before and after running the actual scripts The scripts are organised in the etc folder whil
62. ended when the HHT o method is used The integration may be performed with normal direct integration or with the predictor corrector approach In the latter case the displacement and velocity at the next step are first predicted on the basis of the known displacements velocities and accelerations at the present step assuming implicitly that the acceleration at the next step is equal to zero This is performed without any need for solving system equations Then the accelerations at the next step is solved iteratively by means of the dynamic equilibrium equation and the predictor velocities and displacements are updated accordingly SINTEF group 2001 06 10 USFOS GETTING STARTED 7 2 The predictor corrector approach is convenient because a scaling of the step length may be carried out in the predictor phase At least one equilibrium iteration has to be carried out in order to determine the acceleration at the next step With the direct integration approach a pure incrementation can be carried out However no scaling of the time step is performed With respect to CPU consumption the direct integration with no iteration and the predictor corrector method with one iteration should be comparable because both methods employ one solution of system equation 7 2 Input According to the dynamic input all load control is controlled by a parameter time and the loads to be applied at the different times are specified using time histories A tim
63. eries of spring back failures until the final leg mechanism is activated 9 2 4 5 Comments to the analyses The general comments from the previous workshop still applies and a few trial analyses or restarts should be expected a Check the size of load steps to ensure that the nonlinear effects have been activated For fixed offshore structures the material nonlinearity is by far the dominating one so the main verification lies in checking the I values Check the I values at some characteristic positions along the P 6 curve at max load at the final analysis step etc b Iteration convergence should be checked at every step And since the iterations include a correction to the yield surface convergence imply I 0 0 for all members except AXIAL FAILURE members c AXIAL FAILURE members must be checked separately USFOS scales the load step when ordinary yield hinges are formed But in the present version the load step is not scaled when the element reaches AXIAL FAILURE The I values may jump far off the yield surface Load step scaling will be implemented in the coming version c If the structure shows any sign of spring back behaviour the step size just before and during the spring back must be carefully evaluated Is the analysis detailed enough to capture the redistribution of forces during spring back If not or if in doubt smaller load steps should be specified and the analysis restarted just before the spring back
64. es the user should specify some 5 10 load steps even in the elastic range Try the beam example with large steps and with small steps and compare If you use the one surface model and keep same min step the material nonl inearities should be the same The beam example is a typical stiffening system where the Current Stiffness parameter will increase during loading 9 1 3 Workshop I b Elasto Plastic Column Buckling 9 1 3 1 Objective The purpose of this case is to demonstrate column buckling by the USFOS beam element A tubular beam column is loaded in axial compression The buckling load predicted by USFOS SINTEF group 2001 06 10 USFOS GETTING STARTED 9 4 may be compared to hand calculations according to different codes Further the effect of initial deformations the effect of local buckling and the effect of pre existing dents may be studied 9 1 3 2 Model Column length L 10 0 m Tube diameter D 0 2407 m Tube thickness t 0 005 m Yield stress o 330 MPa Young s modulus E 2 1 10 MPa Reference load Po 1 00 MN Section area A 3 702 10 m Moment of inertia I 2 572107 m Moment of Wp 2 77810 resistance m Axial yield load Np 1221 6 kN Plastic moment cap Mp 91 674 kNm Euler buckling load Pg 533 077 kN L D 0 2407 m 9 1 3 3 Description Due to the initial deformations the column is loaded by combined bending moments and axial forces Under elasto plastic buckling the member buckles when th
65. esults are saved on the actual Case directory and result prefix is res When USFOS is finished the manipulated head fem is moved into the actual Case directory see Tlable 8 6 5 for example on modified head file SINTEF group 2001 06 10 USFOS GETTING STARTED 8 12 USFOS Extreme Wave Height WAVEH Dir DIRECT T PERIOD Progressive Collapse Analysis JACKET model SINTEF 2000 Define Wave type H Period Direction Phase Surf Lev Depth WAVEDATA Stoke WAVEH PERIOD DIRECT 0 0 00 100 Speed Direction Surf Lev Depth Profile CURRENT 2 DIRECT 0 0 100 0 0 20 0 Script for assembling USFOS input and run USFOS Usage go Wave Height Direction Period Create Directory Copy Master control file into the current head file model Master Headfile head fem Substitute the string WAVEH with the first Script parameter 1 substitute WAVEH head fem Similar for par 2 amp 3 substitute DIRECT head fem substitute PERIOD head fem Run USFOS and save results in unique directories SUSFOS HOME bin usfos 15 ENDIN head model stru Case H 1 Dir 2 T 3 res ENDIN Move head fem into actual Case Dir for backup purpose mv head fem Case H 1 Dir 2 T 3 Table 8 6 3 Script file go SINTEF group 2001 06 10 USFOS GETTING STARTED 8 13 Script file run all Table 8 6 4 The script file run all starts go 8 times with different input parameters Wave Height Wave Curr Dir
66. f the earlier pushover analyses The only difference is that we are addressing a new structural configuration And since the structure is damaged we might expect a somewhat more difficult behaviour That is a bit more checking and tuning of load steps should be expected especially if the damage provokes serious spring back behaviour But the basic procedure is the same SINTEF group 2001 06 10 USFOS GETTING STARTED 9 20 9 5 Workshop V Pile and Soil Modelling with USFOS 9 5 1 Description of the structure The structure object of the study is an 8 legged drilling and production jacket platform located in the Bay of Campeche Mexico This platform was designed to operate as a drilling platform and it was installed in 1984 in 46 64 m of water depth The total height of the platform is 69 3 m measured from the upper deck to the bottom of the jacket The two level deck is supported by eight columns arranged so that two longitudinal frames and four transverse frames integrate the structure The upper deck located at the elevation 421 64 m above the mean sea level supports the equipment and supplies required for drilling operations since recently three wells were added to the structure as well as separation tanks Its overall dimensions are 25x45 m The lower deck is located at the elevation 15 85 m and contains 15 christmas trees as well as production equipment The deck is supported by an 8 legged jacket provided with two longitud
67. fos 15 ENDIN head stru load res ENDIN Table 8 3 1 Content of script file gol with 3 fixed USFOS input files Explanation The variable USFOS HOME is set during installation of USFOS on both UNIX and NT computers It contains the file path of the root of the actual USFOS version By prefixing the variable name with the contents of the variable name becomes available for use in connection with any UNIX command USFOS_HOME bin usfos is the address to the USFOS code and by adding 15 after the file name a workspace of 15 mill is required The ENDIN defines that the usual screen input output is given between in the lines between ENDIN and ENDIN The name ENDIN is an arbitrarily chosen name of the label In a usual USFOS run it s first asked for the control file name prefix which here is set to head Further it s asked for the structural and load files which here are stru and load respectively Finally USFOS asks for the result file prefix which is set to res By typing gol USFOS will start use the input files head fem stru fem and load fem and store the results in files with prefix res All input files must be located on the same directory as the script file gol and results are stored in the same directory As USFOS accepts input from one two or 3 files it s possible to leave up to two file names blank as shown in Table 8 3 2 where the load file is left out
68. ges in the next step or that number of steps before next iteration convergence is limited With iterations we can control any deviation from the true solution but any step without iterations or iteration divergence will introduce an error in the solution 2 The solution may fail to converge but still show a steady stable behaviour of the test values The final value after max number of iterations may be close to but not quite below the specified convergence criterion This indicates a stable solution but you should make sure the solution converges after a limited number of steps 3 On the other hand the test values may increase severely or may jump up and down This can be a sign that the solution has broken down and the further results should be regarded with great caution Probably a minor or major discrepancy have been introduced in one of the previous steps leading to uncontrolled behaviour If you have problems due to the iteration divergence modify the control parameters for the steps prior to the divergence Reduce step size Reduce the minimum step size minstp Reduce max displacement increment mxpdis Increase the number of iterations Strengthen reduce the convergence criterion 2 5 3 2 T values Interaction Function Values The I values should be checked at every step I values above 0 00 1 0 in XFOS imply that the solution have deviated from the true solution and may be considered as if that member
69. hat at least 5 10 steps passes from fist yield to full plasticity to activate the elasto plastic transition If the following question can be answered with YES then the load specification is OK IS THE ANALYSIS DETAILED ENOUGH TO CAPTURE THE REDISTRIBUTION OF FORCES AT EVERY STAGE SINTEF group 2001 06 10 USFOS GETTING STARTED 9 9 9 2 Workshop II USFOS Jacket Pushover 9 2 1 Structure Description A realistic North Sea jacket structure is analyzed The structure is shown in Figure 9 2 1 The structure is an 8 leg jacket designed for a water depth of approx 110 meters The legs are arranged in a two by four rectangular grid with the central pair of legs on the platform north side serving as launch runners Overall dimensions a top elevation is 27 x 54 m with launch legs twenty meters 20 m apart Overall dimensions at mudline is 56 x 70 m Total height is 142 m with horizontal bracings at 5 levels The module support frame MSF is a rectangular grid of built up trusswork beams 2 longitudinal and 4 transverse trusses The trusswork is designed of built up box sections Overall dimensions are 27 x 68 meters Trusswork height is 9 75 meters The MSF is shown in Figure 92 2 The jacket foundation is made up of four corner clusters with eight skirt piles in each group no leg piles Longitudinal jacket frames are diagonal braced with X braces between central and corner legs at the bottom bay Transverse frames are K braced w
70. he above expression is satisfied or if the maximum number of iterations is performed In addition iterations are terminated if a load limit point or bifurcation point is detected The Current Stiffness and the stiffness matrix determinant are calculated at each iteration A load limit point or bifurcation point is detected if either the Current Stiffness or the stiffness matrix determinant changes sign from one iteration to the next If a limit point or a bifurcation point is detected the iterations at the current load step are terminated The results from the last iteration are accepted as the results of the load step even if equilibrium has not been obtained 1 3 7 Plastic hinges When a plastic hinge has been introduced the state of forces should move from one plastic state to another plastic state following the yield surface so that I 0 However in each increment the element forces will move at a tangent to the yield surface The state of forces will depart from the yield surface as shown in Figure 1 3 T If the pure incremental solution procedure is used this yield surface departure will lead to I gt 0 Small load steps should be used to keep this drift off small However the iterative procedure include a correction to bring the cross section force state back onto the yield surface As long as the iteration process converge the forces will always remain on the yield surface SINTEF group 2001 06 10 USFOS GETTING STARTED 2
71. he equilibrium axial loads Equation 1 6 is the basis for calculating internal equilibrium forces to be compared with external loads during equilibrium correction Denoting by A the increment between two close configurations the variation of increment in strain energy is given in equation 1 7 l AU EA u u dx 0 N N zr Ay Op Ay Ov dx fer Aw wa Aw Ow EI EI l l T fEA Aux Vix Ov Ay Vix ux dx S zt Aux Wx O wx Aw Wx ux 1 7 0 0 l l EA Av vi 8v dx EAAw wi w dx 0 0 l EA Avavxwx9 wxtAwewevxOv dx higher order terms 0 1 2 2 Finite Element Formulation The incremental stiffness is obtained by introducing interpolation functions shape functions for element displacements u x 9 q V x q 1 8 w x 9 qu The variation of increment in strain energy can now be written as SINTEF group 2001 06 10 USFOS GETTING STARTED 1 8 1 SAU 65u EA dx Au 0 l N dy EL Dy xx 9 EL 9 x Cia dx Av 0 Z 1 N dw rr Dw xx x vk p J dx AW 0 EL l 1 uh T Ji T T dy ps 0 Va 9 dx Aut dy Pt Pax Vx QI dx Av i i 1 50 EA Ao de Au bu EA Bua ul dk Aw 5 0 1 1 dv EA Die Vix dx Av w EA Pun Wx QV dX Aw 0 0 1 1 tw EA pux wx v dx Av dy EA vx wx d AK i 0 o Arranging the parameters in the order u v w the separate terms of the elastic stiffness matrix may be determined K
72. heights whose airgap is less than zero including the wave that corresponds to zero airgap e g the wave height that touches the bottom of the lower deck Haeck which was estimated to be as high as 18 65 m Wave heading Blockage factor Longitudinal 0 7 Transverse 0 8 Diagonal 0 85 Table 9 5 3 Current blockage factors used for estimation of forces in a jacket The API RP2A recommended procedure API 1997 for estimating wave current forces is applied and the provisions indicated in the PEP IMP requalification criteria are considered PEP IMP 1998 API s procedure recognizes the dispersion of the extreme wave by applying a wave kinematics factor of 0 85 PEP IMP 1998 as well as the current blockage effect due to disturbance of the flow when passing through the structure The blockage factors used in this study are shown in Table 9 5 3 Range of water Marine growth depth from MWL thickness m 1 to 20 20 to 40 40 to 80 Table 9 5 4 Marine growth thickness to be considered in requalification of fixed platforms in the BoC PEP IMP 1998 The drag and inertia coefficients in all the elements of the jacket are set to be 1 05 and 1 2 respectively as recommended in API RP2A approach for rough tubular members In addition marine growth is considered in the calculations of wave current forces on the structure with the thickness indicated in SINTEF group 2001 06 10 USFOS GETTING STA
73. ic rotation are given by the surface normal of the yield surface Ag at the curret force state Sj The magnitude of the plastic displacements are given by a scalar factor AA SINTEF group 2001 06 10 USFOS GETTING STARTED 1 12 The hardening rule describes loading from one plastic state to another plastic state When a plastic hinge has been introduced the state of forces should move from one plastic state to another plastic state following the yield surface so that 0 For an elastic perfectly plastic material model this can be expressed as AI a ke Koa AU E np Rape AM N Q 7 JO M 9M 9M gt 1 23 g AS 0 Or G AS 0 1 24 The elastic stiffness expression for the beam element is expressed as AS Kr Ay 1 25 To determine the elasto plastic stiffness expression the total displacement increment is separated into an elastic and a plastic component Av Ay Ay gt 1 26 The stiffness equations can then be expressed as AS Av Av Kr Av Av 1 27 K Av K G Ad when the flow rule is introduced Pre multiplying with G the right hand side takes the form of the hardening rule T AS G Av G G Ad G G Kr AV G Kr 1 28 0 and the plastic increment can be solved AA G Kr G J G Kr Av 1 29 Substituting AA into eq 1 27 an expression for the elasto plastic stiffness of the beam can be determined AS Kr Av Kr G AA Kr Av Kr G G Kr G G Kr Av 1 30 c k G G Kr G re Kr JAv or AS
74. ics Reduction Factor e Hydrodynamic Damping The wave theories implemented in USFOS are e Airy extrapolated e Airy stretched e Stoke s 5 th Skjelbreia Hendrickson e Stream Function Theory Dean Dalrymple Current and wave must be combined in the same loadcase and it is possible to combine several basic waves to an irregular seastate SINTEF group 2001 06 10 USFOS GETTING STARTED 4 2 In XFOS the sea surface elevation is visualized as a carpet with dimensions 2 wave length in X and Y direction The surface elevation which in fact presents the travelling of the wave accounts for the actual current which increases decreases the propagation of the waves An irregular seastate is visualized adding components from each basic wave to the resulting surface elevation The aerodynamics implemented is a part of a Ph D study within dynamic response of slender structures exposed to fluctuating wind The study is not yet completed For more information please contact us The user defined load routines open for a possibility for the users to link their own load routines with USFOS For more details please contact us 4 2 Extreme Wave calculation Automatic member imperfections Modules for calculation of hydrodynamic forces are included in USFOS This means that using separate wave load pre processor is not needed Using the UsFOS hydrodynamic in connection with static push over analysis will typically contain following S
75. in order to repair the linear model A few years ago a typical jacket structural model consisted of 500 1000 members Today the same structure is represented by 5000 10000 members An increasing part of the model is non structural members introduced of different reasons in the linear analysis see or typical example If possible the original structural model should become read only and an intelligent filter should transfer the linear model into a model accepted by the non linear tool see Original Linear Model Intelligent filter 2 Shrinked correct model accepted by read only the non linear tool Figure 3 1 2 Preferred Model Repair solution Often the original linear model will not run at all the analysis fails due to lack of boundary conditions etc To be able to inspect the structure in XFOS the use of the dynamic load procedure is a useful intermediate solution see Table 3 1 1 In an early modelling stage the gravity loading is sufficient load to ensure that all elements are connected boundary conditions correct etc SINTEF group 2001 06 10 USFOS GETTING STARTED 3 2 Dynamic 7 0 025 022 0 1 LoadHist 1 TimeHist 00 11 10001 Table 3 1 1 Using dynamic load procedure Table 3 2 3 jhows the group definition used on a real example and it s here defined 5 groups which all use geometry ID s to identify the elements The general cross sections and the small di
76. inal frames and four transverse battered 1 8 frames with horizontal framing at elevations 6 10 6 10 18 2 32 4 and 46 3 m The transverse frames are K braced whereas the longitudinal ones are a mixture of diagonal bracing and X bracing between central legs Most of the structure was fabricated with mild steel e g ASTM A 36 steel whereas joint cans were constructed with high strength steel The dimensions of the jacket at the mudline are 51X28 m Every leg allocates a pile of 48 inch diameter 1 22 m which are intended to fix the structure to the sea floor The 4 corner piles are driven into the sea floor some 95 m and 4 inner pile tips are 90 m below the mudline Jacket legs deck columns and piles are connected in the working point at elevation 7 62 m This platform is supporting 7 risers whose diameters are 30 5 cm 1 20 3 cm 1 and 50 8 cm 5 as well as 15 conductors The conductors are 30 inch 0 76 m tubular elements driven 60 m below the seabed Every leg is provided with a bumper to protect them against barge collisions There are as well two boat landings located at each longitudinal frame SINTEF group 2001 06 10 USFOS GETTING STARTED 9 2 9 5 2 FEM model The platform was modelled with 1777 beam elements and 1131 nodal points 352 different cross sections were considered and 3 different material types were utilized 39 main tubular joints are checked during the pushover analyses where join cans are considered
77. ing along an erroneous failure path To verify the analyses observe the following steps 1 Ensure that the determinant criterion is active and that the cmax parameter has a large value e g 999 SINTEF group 2001 06 10 USFOS GETTING STARTED 2 18 2 Check the development of the Current Stiffness parameter and the determinant If the sign changes simultaneously the solution is stable and USFOS should follow the correct path 3 If the determinant changes sign but the Current Stiffness parameter remains positive this may signify that the analysis has passed a bifurcation point To improve the bifurcation point traversal the procedure with buckling mode injection may be employed SINTEF group 2001 06 10 USFOS GETTING STARTED 3 1 3 LINEAR MODEL NON LINEAR ANALYSIS Model Repair 3 1 Large Models Creating an accurate structural model is time consuming and costly and it is therefor normal to use existing models rather than create new Existing models in most cases are created for linear design analysis Figure 3 1 1 Large Challenge for Non Linear Analysis Seldom existing models are created with non linear analysis in mind and substantial work has to be done before it s suited for non linear problems As computers are getting faster the model size may increase correspondingly But modification of models means in practice manual work and the bigger models the more man hours have to be spent
78. ith the bottom K inverted to form a double X In the end frame by the conductor area frame 4 the bottom K is substituted with two X bracings The horizontal levels are K braced with X bracing in the conductor area Leg diameters range from 1 6 m at deck level to 3 0 m at elevation 104 m Vertical braces range from 1 1 to 1 6 m diameter horizontals 0 9 1 3 m and horizontal braces 0 8 1 0 m The central legs have a D t of 21 29 from elevation 14 to 43 a D t of 50 67 from elev 43 to 74 and 80 from elev 74 to 104 Corner legs have a D t of 9 12 at the upper bay and 21 29 from elev 14 to 104 Typical D t values for vertical braces are 20 36 with one member at mudline with a D t of 43 D t for the horizontals range from 26 to 48 with two members at mudline with D t of 52 Horizontal braces range from 20 to 33 with 40 as the highest for one member at mudline and one member at elevation 14 SINTEF group 2001 06 10 USFOS GETTING STARTED 9 10 Q 7 Figure 9 2 1 Jacket structure Pile guides not shown SINTEF group 2001 06 10 USFOS GETTING STARTED 9 11 Figure 9 2 2 Module Support Frame 9 2 2 Loads 9 2 2 1 Environmental conditions Table 9 2 1Wave 100 year return period Direction EN LE AME LE PURA pee ww 2 e qwe me ws ws wr ws ur u5 Table 9 2 2Current 10 year return period Direction re pue e e LOC VR NN 1 15 1 10 1 05 1 05 1 10 1 05 1 05 m s
79. l behaviour loading and unloading follows the same curve and an elasto plastic material behaviour with kinematic hardening are available A SIG Linear extrapolation EPS Origo should not be specified Linear extrapolation Figure 6 1 1 Definition of spring properties by discrete points The curve should be straight through origo i e do not break the curve at origo Illegal specification Possible solution Figure 6 1 2 Example of legal and illegal spring definition Both 1 node spring to ground and 2 node spring elements are available The input accounts for the lack of nonlinear preprocessors and therefore the following data handling are performed If the linear spring to ground SESAM element no 18 refers to a nonlinear spring definition MREF the element will be handled by USFOS as a 1 node nonlinear spring to ground If the 2 noded beam element SESAM no 15 refers to the nonlinear spring definition MREF the element will be handled as a 2 node nonlinear spring SINTEF group 2001 06 10 USFOS GETTING STARTED 6 2 6 2 Pile soil interaction In the analysis of fixed offshore structures a proper modelling of the interaction between soil and structure both the static and dynamic case is of major importance The purpose of this activity was to improve the models for soil structure interaction To simplify the user input of the pile geometry and soil properties a specific pre processing functio
80. ld surface In some cases this over crossing may reduce the accuracy of the analysis It is therefore recommended to check any elements with AXIAL failure if any over crossing has occurred If the relative magnitude of the bending moments is small the analysis is probably OK If the over crossing is too large the analysis should be repeated with smaller loadsteps in that area USFOS scales the load step when ordinary yield hinges are formed But in the present version the load step is not scaled when the element reaches AXIAL FAILURE TheI values may jump far off the yield surface Therefore the I values of AXIAL FAILURE members must be checked separately It may be necessary to use extremely small steps to pass a point where an AXIAL FAILURE element has taken off from the yield surface 2 5 4 3 Bifurcation Sometimes the solution fails to detect the correct failure path during traversal of a bifurcation point This is not a general problem but can occur for specific structures In some cases this can be identified by erroneous development of element forces For example compression members get increasing axial forces even after a three hinge mechanism has been formed and buckling is expected In this case the analysis should probably have performed unloading of the structure at some previous load level In other occasions repeated plastification elastic unloading of single members may be an indication that the solution is proceed
81. ly 9 1 3 7 Comments to the analyses USFOS nonlinear effects The same comments apply as for the beam example a To activate the nonlinear geometric terms of the element formulation some 5 10 load steps should be applied before yielding takes place b To activate the elasto plastic transition of the plasticity model there should be 5 10 load steps between first fibre yield and full plastification of the cross section The predicted buckling load will show some variation with the load step size In particular small load steps should be used around the peak load Try with different step sizes and see SINTEF group 2001 06 10 USFOS GETTING STARTED 9 1 4 2D Frame Analysis 9 1 4 1 Objective The purpose of this study is to investigate system effects and redistribution of forces in a relatively simple structure Also this case can be used to study the effect of the different plasticity models on global system behavior The example structure is a two story X braced frame loaded by a horizontal force at the top The model is shown in Figure 9 1 2 9 1 4 2 Model DECK ag le s UPPER o PS PANEL TYPICAL LOAO K LOWER PANEL Q vA A gt o Sv 1920 i D t RATIO IN en ERREUR RE PME EERE 1t 0083 48 zetl Ki Kj gia E 10 0 Figure 9 1 2 Plane frame structure Yield stresses Braces o 248 MPa Legs o 324MPa I beam o 324MPa Reference Ho 40 0
82. marized in table 1 The wave current loads is calculated by means of Stokes 5th order nonlinear wave theory It is assumed that both wave and wind act simultaneously on the platform The loading is determined for the end on broadside and diagonal directions Assessment of existing platforms Elastic Range Ultimate Limit State Exposure Category All categories Moderate Very High 8 1485 3s m 13 70 14 24 92 53 30 72 60 74 22 79 09 51 51 70 16 71 73 76 43 47 65 64 91 66 36 70 71 42 79 58 28 59 58 63 49 35 75 48 70 49 79 53 06 1 1 75 47 09 He eL et Viera ILE FEE E T X AE TTA Os Fi PLEA A T Figure 9 5 3 Loading directions considered in the analyses End on longitudinal broadside transverse and diagonal Plan view of upper deck SINTEF group 2001 06 10 USFOS GETTING STARTED 9 25 9 5 5 Lateral load on the jacket The lateral load effect on the jacket due to wave and current is calculated by means of the response surface function given by equation 5 1 Fixed wave steepness is assumed in all analyzed wave heights H by calculating the corresponding wave period as T 2 95694 H The other environmental parameters e g maximum wind velocity storm tide and maximum current velocity are also estimated as function of the wave height by fitting the values given in corresponding to 99 180 753 892 and 1485 year return periods Then the wave forces are calculated for wave
83. mat Pile_geo lcoor Imper 1 2 3 PILE 4 5 6 i 8 0 PILE PILE PILE PILE ww PILEGEO E ID Type Do T PILEGEO 2 2 1 22 0 05 ID Type Z Mud D ref Ffac API Soil ID SOILCHAR 10 ARI 93 725 Z0 140 101 Clay 201 Clay 301 Clay 401 Sand 501 Sand 501 Sand 501 Sand 601 Sand 0o 120 H2 00 NO ID Type load Gam Plug Su eps50 Tresf QPLim API_SOIL 101 SoftClay Static 9500 l 50E3 013 0 74 0 2E6 API SOIL 201 StifClay Static 9500 1 120E3 012 0 72 1 2E6 API SOIL 301 StifClay Static 9500 1 150E3 010 0 73 1 0E6 API SOIL 401 StifClay Static 9500 1 190E3 2019 0 75 1 9E6 0 0 0 0 0 0 0 0 i ID typ load Gam Plug Phi Delta rNq QPlim API_SOIL Sand Static 8000 0 33 22 22 1 4E7 Table 6 2 1 Input for automatic calculation of piles and soil capacities SINTEF group 2001 06 10 USFOS GETTING STARTED 6 5 6 3 SpudCan Element A Non linear SpudCan element specially designed for Jack Up Structures has been available for some time but until recently only sand models were implemented The implementation was revised in 2001 and the current implementation available in usfos version 7 8 uses capacity and interaction formulas from the SNAME RP 1 e Both sand and clay models are now implemented e Elastic stiffness is taken from SNAME RP but with embedment corrections as described in the commentary to SNAME RP 1 e The nonlinear rotational spring stiffness correction is based on
84. n is implemented in USFOS This saves the user from defining the detailed geometry of the pile spring model The soil pile interface material behaviour i e the spring characteristics is implemented according to a general plasticity formulation The model is 3 dimensional in the sense that both lateral and axial springs are applied to each node The user defines the soil characteristics for each soil layer by P Y T Z and Q Z curves With this information combined with the user s definition of the pile location type single or group and dimensions diameter and thickness USFOS generates finite elements beam and spring The size of this foundation model varies from approx 100 1000 elements depending on number of piles and number of soil layers In XFOS the pile is visualized with discs representing the soil behaviour at the different levels and the size of the discs reflects the relative strength of the soil Both soil deformations and utilization are visualized SINTEF group 2001 06 10 USFOS GETTING STARTED 6 3 6 2 1 Pile Soil Automatic generation of piles and soil capacity The automatic generation of piles and corresponding soil capacities is a powerful option which requires a few input lines only The user s structure ends at mud line and all elements below mud line are generated automatically by USFOS see In Table 6 2 1 overleaf the necessary commands used to produce the foundation model shown in
85. n one member influence the load redistribution to other members and thus the formation of new hinges If too many new hinges are introduced in a MIN STEP LENGTH step the true yielding process may not be identified and the correct failure mechanisms may not be initiated Preferably no more than say 5 hinges should be introduced per step 2 5 4 Potential problem areas 2 5 4 1 Repeated Plastification Elastic Unloading A known problem in USFOS is repeated plastification elastic unloading of specific members This is termed as false on off loading When USFOS detect unloading of a member during the load increment the stiffness matrix for this element is re calculated and the system stiffness matrix reassembled The load step is then repeated and all members are checked for yielding as usual The problem occurs when yielding is detected in the same element that just unloaded The load step is scaled to zero at least to minimum step size and a plastic hinge is introduced In the next step the procedure may repeat itself Specific elements may keep on loading on off for a significant number of load steps clogging up the analysis Of course during significant redistribution of forces within the structure or at bifurcation points the analysis will often need some steps of on off to hit the correct failure path But if one member stops the analysis this can be a significant obstacle To circumvent this problem a proce
86. o the maximum load The collapse load typically varies from around twice the characteristic load for slender 4 leg structures up to about three four times the characteristic load for 8 leg jackets In general the size of the load step should reflect the degree of nonlinearity in the response As the load increases the load increment and the minimum step length should be reduced 2 5 2 1 Iterations Equilibrium iterations may be specified to ensure equilibrium between external loads and internal element forces For pushover analyses the pure incremental formulation will in most cases give satisfactory results and require less computational time Moreover use of iterations introduce additional complexity and new potential error sources iteration divergence etc However the iterative procedure include a correction to the yield surface As long as the iterations converge the forces will always be brought back to the yield surface Thus use of iterations will be a priority between computational time simplicity on the one hand and accuracy easier analysis verification on the other hand In most cases use of equilibrium iterations seem to be beneficial 2 5 2 2 Minimum step length When yielding occurs the load step is scaled so that the state of forces comply exactly with the yield surface In the nonlinear range several cross sections may yield almost simultaneously resulting in far too small load steps To ensure reasona
87. olution is chosen Instead of assembling pieces of input the content of the input file s are modified prior to the analysis As the modification should be performed in a batch run a batch editor is necessary The UNIX shell on both UNIX workstations and the NutCracker UNIX shell on Win NT offers the SED editor the Stream EDitor The operation needed from the stream editor is the REPLACE or SUBSTITUTE command where one character string should be replaced by another The cryptic UNIX command is wrapped into a file which here is named substitute able 8 6 1 and which is used as follows Substitute string 1 string 2 FileName In all connections where string occur on the specified file it s replaced by string 2 The SED editor is case sensitive differs between upper and lower case characters Quotes must be used if blank character s occur in the strings sed 1 8 s 1 82 g 3 gt subst string temp mv subst string temp 3 Table 8 6 1 Script substitute which utilises the SED editor for substituting strings SINTEF group 2001 06 10 USFOS GETTING STARTED 8 11 With the powerful substitute script available following operations should be done a Create only one master USFOS control file which should be used for all cases a Use one structural file a Run USFOS wave analysis for 8 different wave current conditions As indicated in Figure 8 6 1 some files are present before the analyses a
88. p if test count eq S then echo echo fi fi count expr count 1 Assumes structural file on Results are stored on file Usage elmdel Label eleml Creates the directory Label creates a copy of usfos control file and adds necessary NONSTRU commands model stru fem SSCRATCH res elem2 elem3 2re March 2000 Si Heading gt gt gt gt gt gt gt gt gt gt gt nonstru_elem nonstru_elem nonstru_elem nonstru_elem nonstru_elem nonstru_elem Add to file gt gt nonstru elem fem gt gt gt gt nonstru_elem nonstru_elem Update counter Grabbing USFOS master control file from model cp model head fem k echo Adds nonstru commands cat nonstru elem fem gt gt head fem echo Creates Case identifier 1 on head fem etc substitute CASEID 1 cp model stru fem cp model load fem echo and start USFOS etc run usfos gt echo and POSTFOS etc run postfos SCRATCH res gt gt echo echo Saves Global History on current echo echo echo echo head fem echo Grabbing USFOS stru amp load file from run log run log directory mv SCRATCH res pri Global History mv SCRATCH res status text Table 8 7 4 Script file elmdel model t F J xt F ok ke oe ke ok e oe e oe e o e ok e c e oe e oe e ck e oe eo e ck e ce e oe e oe e e e e e e e e e e aa
89. pecify the actual wave type height period direction Specify the corresponding current 1f any Switch on buoyancy optional Specify criterion to be used for selecting worst wave position max base shear or max overturning moment DODO DS d i LS AG S LI im S ARS TUA n 5 ifa Y m A any i LS i the i y yz j jl ow im F Direction of wave Direction of Wave 4 Figure 4 2 1 Automatic member imperfection according to wave force direction SINTEF group 2001 06 10 USFOS GETTING STARTED 4 3 USFOS will then step through the actual wave and identify the worst wave position the position causing the highest base shear or overturning moment The hydrodynamic forces from this wave phase position are saved in memory to be used in the pushover analysis The calculated buoyancy forces are possible to separate from the other hydrodynamic forces and the user may specify how to use the buoyancy forces add to an existing deadweight loadcase etc Applying member imperfections one by one is a time consuming task but by using the new option CINIDEF the correct member imperfection is applied automatically for all beam elements The most common buckling curves are available defining the size of the imperfection see User s manual Ch 6 The direction of the imperfections follow the direction of the loads for a specified load case In Fig
90. piles is computed and used The wave load is modelled by equivalent hydrodynamic diameters Conductor guide frames and the conductor guide arrangement is simplified Wave loads are modelled by equivalent element diameters The launch runners are modelled as an integrated part of the launch legs A2 and A3 Conductors are lumped together as tubular elements with appropriate drag diameter and inertia diameter Risers caissons J tubes boat bumpers walkway and ladder are modelled as equivalent elements These elements are defined as non structural in the USFOS analyses Soil structure interaction is modelled by linear springs 9 2 4 Traditional pushover loads from WAJAC One pushover analysis is carried out for each platform direction Gravity loads buoyancy and operational loads are incremented up to characteristic unfactored value Then environmental forces are applied incrementally until collapse The following events are identified for each loading direction 1 Initial yielding A 2 First member failure member buckling or tension yielding A 3 Ultimate collapse load A T Table 9 2 4 Summarizes the results of the pushover analyses Load factors for characteristic events in each loading direction are listed These are relative load levels referred to the characteristic loading for each direction ref Table 3 2 3 Table 9 2 4Characteristic load levels load factors pushover analyses Direction Pee ne eee El ERH
91. r Window layout SINTEF group 2001 06 10 USFOS GETTING STARTED 8 2 Restore Move Size Minimize Maximize Close Edit gt Properties ce Figure 8 1 2 Menu Select Properties and the select colors menu shown in Figure 8 1 3 hppears Select screen text and screen background among the indicated colours The light grey background together with black text is a good combination UNIX Properties x Options Font Layout Colors C Screen Text Selected Color Values Screen Background Bed b 4 C Popup Text Ceen p 4 Popup Background Blue pa ERRATEA Se HS Selected Screen Colors C WINNT gt dir SVSTEM lt DIR gt 03 01 92 SVSTFM32 lt NTR gt 03 01 92 Selected Popup Colors C WINNT gt dir SVSTEM lt DIR gt 03 01 92 3 10a SVSTFM32 nTR 803 n1i 92 OK Cancel Help Figure 8 1 3 Defining screen and text colour The default window has no screen buffer has no scroll bar but the buffer sizes in vertical number of lines and horizontal number of columns are possible to specify under the layout menu see Figure 8 1 4 Type in or us the arrow the actual sizes which here is set to 132 2048 The window size when it pops up is set to 80 40 When the OK button is pressed the menu shown in Figure 8 1 5 appears Select Modify Shortcut to save the settings permanently SINTEF group 2001 06 10 USFOS GETTING ST
92. re This is the total energy of all load combinations Element number Element identification number Event position Position where a plastic hinge is formed removed ENDI First element end END2 Second element end MID Element midspan JNT1 Joint at first element end JNT2 Joint at second element end Event type Change of element status YIELD The forces has reached first fibre yield of the cross section and a yield hinge is formed PLAST The forces has reached the full plastic capacity of the cross section UNLOD The element has unloaded and the cross section has returned to the elastic state AXIAL The element forces have reached the full plastic tension capacity of the member A membrane element is introduced accounting for geometric stiffness of the member FRACT Fracture is detected in the member JOINT The full capacity of the joint has been reached and the joint is yielding MIN STEP LENGTH Attempt to scale the load step below the minimum size specified by the user MAX DISPL INCR Load step scaled due to large displacement increments 2 4 2 Analysis print file The Analysis Print file is a text file generated during USFOS analysis The file contains input verification data and analysis results at each load step The amount of print is controlled by the user through the parameters inprint and outprint of the CPRINT record governing input verification print and analysis output respectively The amount of input verifica
93. re performed and some are created during the analysis executing the scripts defined in this section Contents of model Case H 20 0 Dir 00 0 T 16 0 jCase H 20 0 Dir 30 0 T 16 0 Case H 20 0 Dir 60 0 T 16 0 jCase H 20 0 Dir 80 0 T 16 0 Case H 24 0 Dir 00 0 T 20 0 a stru fem These files folders are These files folders are present before running present after running the scripts run all Figure 8 6 1 Files Folders before and after running the scripts Master Headfile Table 8 6 2 able 8 6 2 The file is an ordinary control file for USFOS but some parameters are not yet set Instead the parameters are represented by arbitrarily chosen key words In the actual study the wave height direction and period should be varied and the keyword for the wave height is WAVEH the keyword for direction is DIRECT and the keyword for wave period is PERIOD Script file g0 Table 8 6 3 The first operation in the script is creating a directory using the mkdir command and all 3 parameters wave height direction and period are included in the directory name Next the nearly complete USFOS control file named Master Headfile and located in directory model is copied into the file head fem on current directory The script for substituting named substitute is used three times for replacing the keywords with the actual parameter values Then USFOS is run and the same structural file stru fem is used for all cases R
94. rect size for these parameters will be a compromise between accuracy and time cost and the right load specification will often be determined through an iterative process as outlined below 1 Determine an initial load history based on the prior knowledge of the structural behaviour e g linear elastic analyses 2 Check the global behaviour of the structure if significant and sudden redistribution of forces seems to occur at any load level Check if the load increments at this load level are small enough to capture these effects 3 Check the interaction function at the plastic hinges whether the I values are at an acceptable level or if the state of forces show significant exceedance of the yield surface 4 Determine at which load level the analysis accuracy deteriorates 5 If analysis results beyond this load level is required then specify a new load history and restart the analysis REPEAT FROM STEP 2 For offshore structures the global behaviour is normally fairly linear up to and a bit beyond the factored characteristic load design load To ensure that nonlinear effects are properly activated an initial load step of 0 10 0 30 of the unfactored characteristic load is recommended SINTEF group 2001 06 10 USFOS GETTING STARTED 2 14 Depending on the structure 4 leg jacket 8 leg jacket K braces X braces etc the behaviour becomes more and more nonlinear from about 1 5 2 5 of the characteristic load and up t
95. ristic load from the worst direction For the other directions we may assume elastic response up to approximately the same load level in MegaNewtons If the structure is designed according to API then first member failure should coincide with 1 6 times the characteristic loading First fibre yield may then occur at approximately 1 2 1 3 b From first fibre yield to first member failure there would only be a slight reduction in stiffness The response would still be practically linear c For X braced structures or more generally if the loading is carried by a statically indeterminate bracing system we should expect a gradually softening behaviour up to a peak load and then a gradual load reduction in the post collapse range Prior analyses indicate the reserve capacity from first member failure up to the collapse load to be in the order of 30 SINTEF group 2001 06 10 USFOS GETTING STARTED 9 15 d For K or diagonal braced structures if the loading is carried by a statically determinate bracing system the collapse load may be roughly equal to the first member failure load also depending on the strength of the legs of course These bracing systems have very little redundancy and may exhibit an extremely brittle behaviour almost linear up to first member failure system collapse only slightly above first member failure and then a sudden reduction in capacity spring back Several braces may fail in sequence giving a s
96. rt structure is appended to the stru fem using the cat gt gt command Similar is done for the load file assembly SINTEF group 2001 06 10 USFOS GETTING STARTED 8 9 A unique directory for each case is created using the mkdir command and the directory name with prefix Case contains information about both support and load USFOS is started with 15 mill and results are saved in the actual Case directory using the result file prefix res for all cases the directory contains information about the different cases Finally the actual stru fem and load fem are moved into the actual Case directory using the mv command Note that if only directory name is defined in connection with the mv command the file name will be unchanged in the new directory just moved Script for assembling USFOS input and run USFOS m Usage go pari par2 parl Support Structure par2 Load definition eae a a a ee Copy Main Structure into file stru fem and add selected support cp str Main_Structure stru fem cat str 1 gt gt stru fem Copy Main Load into file load fem and add selected load cp loa Main_Load load fem cat loa 2 gt gt load fem Run USFOS and save results in unique directories Create Directory mkdir Case 1 2 SUSFOS_HOME bin usfos 15 lt lt ENDIN head stru load Case 1 2 res ENDIN Move stru fem and load fem into actual Case_Dir for
97. s bending moment 3 Generate plots of axial force vs global displacement and of bending moment vs global displacement for the one surface plasticity model 9 1 2 3 Comments to the analyses USFOS include non linear effects both due to material non linearity yielding and due to geometric non linearity change of global geometry and effect of internal forces in the structure To get a fair representation of the material non linearities the following points should be kept in mind a The minimum load steps should be reasonably small If the minimum step size is too large the element forces will not be scaled back onto the yield surface On the other hand too small min steps may completely block the analysis of large complex structures But all in all a too small min step is better than a too large b With the two surface plasticity model there should be a reasonable number of steps between first yield and full plasticity at one cross section some 5 10 steps should be OK Some load steps are needed to activate the elasto plastic transition of the model If the transition is too abrupt the two surface model will degenerate into the one surface model elastic pure plastic behaviour Try with different load steps and see the difference on the M N interaction The geometric non linearities are a result of the deformations and forces built up during the proceeding analysis steps a To activate the geometric non lineariti
98. s the following script named run all would run through all 16 cases without need for any human interference TWO MIURTE Gee ee OUR See Peer AE Ee See Lee LO LE ee Ee IUE LO LO ETE SO LE Se LO Pe Script for running 2 structural conditions 4 load directions and 2 load conditions E Totally 2x4x2 16 cases Xx ec FE ch eh Oh hs PEE ELLE NA eee Le ee Structure Load go intact nw 100 go intact sw 100 go intact se 100 go intact ne 100 go intact nw_10000 go intact sw_10000 go intact se_10000 go intact ne_10000 go damaged nw_100 go damaged sw_100 go damaged se_100 go damaged ne_100 go damaged nw_10000 go damaged sw_10000 go damaged se_10000 go damaged ne_10000 End of Script File Table 8 4 2 Content of level 2 script file run all which refers to go SINTEF group 2001 06 10 USFOS GETTING STARTED 8 6 8 5 Example 3 Assembling input files before USFOS analysis In the previous examples all input files were complete before the script was executed In may cases only a small fraction of the entire input is different from one case to another Instead of making lots of copies of near 100 equal files the key in this example is to show how the input files could be composed by common information some special information Common information a Control file head fem a Main structure located in str M
99. s develop SINTEF group 2001 06 10 USFOS GETTING STARTED 9 18 9 4 Workshop IV Residual strength analysis The structure is analyzed in different damaged conditions The residual strength of the structure is determined compared to the collapse capacity in intact state Gravity loads buoyancy and operational loads are incremented up to characteristic unfactored value Then environmental forces are applied incrementally until collapse The following damage conditions are analyzed 1 Brace 2261 damaged row 4 Loading from west 2 Brace 363 damaged row 4 Loading from west 3 Brace 463 damaged row 4 Loading from west 4 Brace 355 damaged row A Loading from north 5 Brace 455 damaged row A Loading from north First the members are assigned an out of straightness equal to one tube diameter at member midspan Then the member is removed entirely The following events are identified from each analysis 1 Initial yielding A 2 First member failure member buckling or tension yielding A 3 Ultimate collapse load A 7 Table 9 4 1 and Table 9 4 2 summarizes the results of the pushover analyses Load factors for characteristic events in each loading direction are listed These are relative load levels referred to the characteristic loading for each direction The residual strength ratio RIF is calculated for each damage condition Table9 4 1 Characteristic load levels members damaged Damage condition
100. s load carrying capacity has been overestimated by the same value Generally if iterations are applied successfully the I values should be equal to zero The exception is AXIAL FAILIURE members For pushover analyses some deviation from the yield surface can be tolerated depending on the importance of the actual member in the global load carrying behaviour of the structure As a rule of thumb the I values should usually not exceed 0 05 for primary members e g legs and primary braces of jackets whereas values up to 0 20 might be acceptable for secondary members However this must be considered in each case against the use of the analysis results and the necessary accuracy of the results To verify the analyses observe the following steps 4 Check that the I values are small 5 Ifthere are non zero I values in any step check that are within acceptable limits or that they become zero in the succeeding step s if iterations are specified 6 Determine at which load level the deviation starts Are accurate results beyond this load level required 7 If yes then the analysis has to be resumed with modified control parameters at this load level Reduce step size Reduce the minimum step size minstp Reduce max displacement increment mxpdis SINTEF group 2001 06 10 USFOS GETTING STARTED 2 16 2 5 3 5 Number of Hinges per Load Step Number of yield hinges introduced in one load step should not be too large Yielding i
101. t wave and current directions Edit the load fem file 9 5 8 Documentation 1 Generate P plots from the final analysis of each direction 2 Take hardcopies of deformed geometry with member utilization at first member failure max 3 load and at the final analysis step Generate plots of axial force vs global displacement N plots for critical members in the failure mechanism buckling members tension failure members or failing leg members The purpose of this is to document the redistribution of forces and the development of the final failure mechanism SINTEF group 2001 06 10
102. ted When hinges are formed at both ends and at midspan the beam forms a kinematic mechanism From this stage further loading is carried by axial tension forces as the beam deforms When the mechanism is formed the beam first deforms in a V shape with rotations concentrated to the plastic hinges and each half beam elastic For a further increase in loading the beam deforms into a chain link The beam yields at the quarter lengths and the rotation angle at beam midspan is reduced the midspan area is straightened out The yield hinges at beam midspan unloads A further loading will be carried almost entirely by axial straining and the beam will enter a state of pure membrane action 9 1 2 Program input Control file head fem Structure file stru fem Load file none Load case 1 Transverse load qo SINTEF group 2001 06 10 USFOS GETTING STARTED 9 3 9 1 2 1 Analyses Perform a nonlinear analysis of the beam 2 Repeat analysis 1 but with the One Surface plasticity model use the record SURF20FF 3 Analyse the beam under pure nonlinear elastic bending e Suppress formation of plastic hinges use the CELHINX record OR specify very high yield stress use the MISOIEP record 4 Simulate pure plastic behaviour of the beam e Specify very high Youngs modulus use the MISOIEP record 9 1 2 2 Documentation 1 Take a hardcopy the P 6 plot displayed by XFOS global load global deformation 2 Use XFOS to generate plot of axial force v
103. tfos ee ae the content of one automatically created file folder named Elem_01 which contains the global history created by POSTFOS the log files from the analysis and the different input and output files Table 8 7 3 thows the content of the file nonstru elem fem which is created by the script for two cases To the left the case where element number 1 should become non structural and to the right the case where elements 5 6 and 12 should be removed Contents of Elem_01 Global History is head fem ie load fem p r p i r NONSTRU Element NONSTRU Element r NONSTRU Element NONSTRU Element r Table 8 7 3 Automatically created files containing the NONSTRU comand SINTEF group 2001 06 10 USFOS GETTING STARTED AALL LLLA ELEELE ELEELE LELETET ELETE ETETETT ETETETT ETETETT ETETE Author Tore Holmas SINTEF Group Norway Date 2000 03 18 HHHFHHEA HEHEHE EEE EAE AEA E EAE EEE EH EE HEE EE EH HE E V 4 if test S t 2 then echo OK e soe eoe eoe koe koe koe ke o e o e oe e e e e eoe eoe e e eoe e e e e e e e e e e e e e n n echo E echo echo echo p echo d echo E echo x echo echo ci echo t echo echo i echo e else echo ud echo Creates directory 1 mkdir 2 81 cd 2 81 count 1 for i do if test Scount gt 1 then echo Processing Element if test count eq 2 then echo echo echo echo echo echo fi echo NONSTRU Element o
104. the control parameters for the non linear analysis the user is free to organize the data on these files The specific content of each file is not important as long as all data are present on the files used Only the filenames are input to USFOS The filetypes are predefined by the program system the files MUST have filetype fem User input to USFOS is read from text files Two input formats for structure and load data are currently supported directly e USFOS reads structure and load input prepared for the SESAM program system directly e A simplified USFOS specific format UFO may also be used for structure and load input In the current version of the User s manual one chapter describing the UFO file format is added The UFO file format is used to describe the same type of information which normally is described in SESAM file format and has been used since 1994 by non SESAM users The type of information is Nodal ID s Coordinates and Boundary conditions Element ID s connectivity and properties etc USFOS recognises the file format automatically and the results are unaffected by the structural load file format used However mixing commands from the two input formats are not possible Special USFOS Control Parameters head fem J un General Structure and Load input are specified in SESAM file format or USFOS UFO file format J Figure 2 3 1 Input files to USFOS STRUMAN a converting tool
105. the figure are given See also the example in folder PSI 2 User s Strucutrual Model Generated by USFOS Figure 6 2 1 Automatic generation of piles and soil capacity Comments to the input in Table 6 2 1 a The foundation consists of 4 pile clusters each with 7 piles and 4 single piles This foundation is defined as 8 PILE elements which refer to one of the two PILEGEO records Q PILEGEO number 1 consists of 7 pipes with diameter 1 22m The individual positions are specified through local Y and Z co ordinates referring to the PILE local axis a The PILE local x axis goes downwards from the pile head towards the pile tip Q PILEGEO number 2 is a single pile here defined as a group with only one pipe in the centre of the pile element axis The single pile option could also been used see UM Ch 6 a For all the 8 piles the same soil exists refer all to the same SOILCHAR record a The SOILCHAR is specified with 3 clay layers and 3 sand layers However in order to obtain a reasonable element density in the rather thick sand layer no 2 24 1 to 48 8m the same soil D SINTEF group 2001 06 10 USFOS GETTING STARTED 6 4 property no 501 is referred to three times The soil spring is inserted in the middle of the layers defined under SOILCHAR a The soil strength is calculated according to API 1993 by specifying the geotechnical data in the command API SOIL Elem ID n PILE d PILE PILE pl Soil ID Pile
106. tion data is governed by the input parameter inprint The minimum amount of print is shown in Figure 2 4 2 This is key parameters load control data displacement control data and element imperfection data if any Additional input verification data is e Structural Data Nodal point data Material data SINTEF group 2001 06 10 USFOS GETTING STARTED 2 6 Element data Spring characteristics Local coordinate system data Cross sectional data e Load Data Distributed element loads Nodal point loads Gravity loads Internal F E M parameters Element degrees of freedom Nodal degrees of freedom Nodal point connectivity An example of structural data is shown in Higure 2 4 3 2 4 3 Analysis Status file USFOS creates a text file at the end of the analysis with name f inst Jacket status text if jacket was the result file prefix The status file gives a brief overview of the analysis Time Load level for first yield first plastic hinge first buckling first element exceeding a specified utilization threshold Number of negative pivot elements Number of new hinges in one step Iteration convergence information List of the 10 first elements yielding getting plastic hinges buckling and exceeding the specified utilization threshold SINTEF group 2001 06 10 USFOS GETTING STARTED 2 9 ANALYSIS PARAMETERS ZA YA s FRAME US F O S progressive collapse analysis SINTEF div of Str
107. trix The behaviour of the hinges is governed by plastic flow theory The basic assumptions of plastic flow theory can be summarized as follows 1 There exists a yield condition which can be illustrated by an initial yield surface 2 There exists a flow rule relating plastic strain increments to stress increment 3 There exist a hardening rule relating the extension of the yield surface to the amount of plastic deformation N Q Q Mx My M Ne Qp Qp M xp M yp M p U f 1 0 1 19 In USFOS the yield condition or plastic capacity of a cross section is represented by a plastic interaction function yield surface for stress resultants For a tubular section the plastic interaction function is given as N M M DU f Em RS Im 1 Np Mp Mp 1 20 cos aN jo SEM M 0 2 Np Mp when torsion and shear forces are neglected The function is defined so that 0 for all forces giving full plastification of the cross section T 1 is the initial value of a stress free cross section In principle a state of forces characterized by T gt 0 is illegal The flow rule is given by g O0 AA Ay l a AA 1 21 0 5 AA where r OF T A A OW F A Baa tan 50 90 aA aw ML 1 22 Si IN Q Q Mx My Mz and index i refers to beam end 1 and beam end 2 These equations state that the plastic displacements are normal to the yield surface The direction of the plastic displacements plastic elongation vs plastic shear or plast
108. u Kuv Kus Kr Kvu kw kw 1 10 kwu I kww The following expressions emerge for the sub matrices 1 ku EA Pu x P x dx 0 1 N kl EL ra Pye Px Gy dx 1 11 E l 1 N k fer y xx Xx VES x dx dx z EL l are the diagonal sub matrices that also are present in the secant stiffness matrix ref The subsequent terms comes from large rotations and are nonlinear contributions 1 kw EA 0 vx D ua OX k wy 3 1 12 1 kw 7 EA Oy v dx k 0 These two are coupling matrices between axial and lateral deformation and are linear in rotation SINTEF group 2001 06 10 USFOS GETTING STARTED 1 9 Finally the diagonal sub matrices for deflection kyy and k get additional contributions that are second order in rotation 1 JEA 6 1 13 J 0 The last integral of eq 9 give coupling matrices between the two directions of deflection 1 kw EA 6 Wx Va Py dx 7k 1 14 0 1 2 3 Shape Functions The shape function used for the transverse displacement field w x q USES is taken as the exact solution to the 4th order differential equation of a beam subjected to end forces 9 coshkx sinh kx x L 1 1 16 for positive N compression 9 cos kx sin kx x L 1 1 17 for negative N tension The value of k is given by k ae 1 18 EI l Similar expressions are used for the displacement fields w x and u x The generalized constants q are then determined by the
109. uctural Engineering umber of input lines read 220 umber of nodal points 13 umber of structural elements 23 umber of springs to ground 0 umber of shell property elements 0 umber of overlaps 0 umber of damaged elements 0 umber of materials 3 umber of cross sections 5 umber of spring characteristics 0 umber of linear dependencies 0 umber of element imperfection groups 0 umber of local element coord systems 23 umber of local nodal coord systems 0 umber of eccentricity vectors 0 umber of load cases 1 umber of temperature fields 0 umber of load combinations 0 umber of control nodes 1 umber of steps in post collapse 15 umerical accuracy equation solver 1 00E 20 umerical accuracy interaction surface 1 00E 01 Combined shape function load level 050 ax recalculations due to unloading 5 Elastic spring back introduced at CSTF Local dent Restart dat 1 20E 00 formulation used a stored at intervals 1 Soret LOAD CONTROL DATA eue USFOS Load Max Max Min load Scaling load no of displ comb factor level steps step 1 1 000 5 000 0 010 1 500 6 500 0 010 1 050 000 20 005 1 100 8 000 40 010 i ER DISPLACEMENT External node no T Total Used Max no of plastic hinges at mid span is ALLOCATED CONTROL Global Displacement displacement weight direction factor X 1 000 DATA SPACE Integer Real data data 200000 2500000 10743 19747 elements
110. ure 4 2 1 the jacket to the right is exposed to waves with direction 45 while the jacket to the left is exposed to a wave with opposite direction 225 It is seen that the direction of the imperfections are opposite in the two cases size is scaled All necessary input is shown in able 4 2 1 and it should be noted that these few commands replace 1000 s of input lines and use of separate wave load pre processor load files Comments to the input given in Table 4 2 1 see also example folder wave maxwav Q Load case is used for dead weight and calculated buoyancy Q Load case 2 is used for the extreme wave Load case is not scaled beyond factor 1 0 that s why the calculated buoyancy forces is separated from the other hydro forces and added to this load case Load case 2 forces are scaled to platform collapse a The direction of the member imperfections CINIDEF par no 2 and 3 follows the direction of the member forces defined by load case 2 which is the calculated wave forces a The size of the imperfection CINIDEF par no 1 is calculated according to Chen column curve a A Stoke 5 th wave with height 25m period 16s 45 direction is applied The sea surface is located for global Z coordinate 0 0 Water depth is 100m a A current profile with peak value 2 m s is defined with same direction as the wave From depth 20m Z 20m relative to the sea surface the current is reduces linearly a The actual wave
111. yr etc This leads to the next example which will give an example on how a slight modified go3 could be used for many different analyses 8 4 Example 2 Varying USFOS input file names The fixed name script 203 described above is slight modified Instead of defining the file names 100 some of the file name is substituted by the keywords 1 and 2 It s possible to give input parameters to UNIX scrips and 1 is parameter no 1 2 is parameter no 2 etc SUSFOS HOME bin usfos 15 ENDIN head 1 3S2 model 1 stru loads 2 D temp res 1 2 ENDIN Table 8 4 1 Content of script file go with varying input file names By typing go intact nw l00yr the same analysis as described under example 1 go3 will be performed The 1 variable will be expanded to intact inside the script and 2 will be expanded to nw 100yr which gives the actual file names SINTEF group 2001 06 10 USFOS GETTING STARTED 8 7 Control file head intact nw 100yr Struct file model intact_stru Load file loads nw 100 yr Result file f D temp res intact nw 100yr A script file may not only refer to UNIX commands it s possible to refer to other script files as well This leads to next level in script programming defining a top level script which refers to user defined script s If f ex one analysis series should consist of a number of different structural conditions different load directions and condition
112. ystem stiffness Generally the mass proportional term damps the lower modes of vibration and the stiffness proportional term damps the higher modes of vibration The two proportionality constants can be calibrated such that a desired damping level may be obtained at two frequencies It should be born in mind however that the Rayleigh damping terms will often be of minor importance because since the effective damping will be predominated by hysteretic material behaviour in plastic hinges The numerical integration scheme is based upon the HHT a method which condenses to the Newmark p method for a 0 The property of the o parameter is to introduce artificial damping of the higher order vibration modes which is beneficial for the accuracy of the solution In order to obtain numerical stability during integration the step length has to be adjusted such that it is less than a prescribed fraction of the fundamental eigenperiod of the system For a system with a large number of dofs the highest natural period may become very small This restriction requires many more time steps than needed for accuracy especially when low mode response is governing Hence it is recommended to use an unconditionally stable algorithm For The HHT a unconditional stability is obtained when the following conditions are met pm 3 1 1 2 y ii 1 2 1 a Peat where D and y are the free parameters in the Newmarck D method Generally o 0 3 is recomm

Download Pdf Manuals

Related Search

Related Contents

Window Lock    Becker READY 50 EU 19  Equip Mini DisplayPort Cable, 2m  VDS2700 Installation Manual  One For All SV3250 flat panel wall mount    Caution Warning Caution Warning DATASHEET  Journal Avril 2014_2  

Copyright © All rights reserved. Failed to retrieve file