Home

T07-08-02_Reliability_Analysis_D7_1_Appendix

1. TU7 OB 2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505 20 Distribution Distribution Fxaniple ud ues Distribution n hen arameter arameter Parameter Unique fortran name Description LSE mapping M z 5 PRA p parameter 3 distribution Variation Standard name coefficient Deviation Ca2 3 Cc1 2c Cc2 2a Cc2 2b Da2 5 Spectral peak period inverse of Tp Tp S Ba2 4c normal 0 2 peak frequency Tr Tr Root tensile strength kN m Ba2 4b Aal 1 Bal 1 Bal 5d Duration of wave record test or Bal 5d Aa2 1b Ba2 1a TR StormD 0 lognormal 1 sea state Ba2 1b Bc2 1b Bc2 1h Aa2 4 Ba2 4i Ba2 4d 0 1 u VesVel Vessel velocity m s Da4 1 Horizontal depth mean current U HDMCV m s Bc3 1b Bc3 1d velocity ub Ub Near bed velocity m s Bc3 1b Bc3 1d ul Ul Velocity of the storm debris m s Da4 3 v SFVel Seepage flow velocity m s Bal 5d v IceVel velocity of the ice feature m s Da4 2b V Vol Volume m Da4 2b vs Vs Ship s speed m s Bc2 1db Bc2 1d TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505 20 Parameter Unique fortran name Description Fall velocity of the sand LSE mapping Example distribution Distribution Distribution SES Distribution parameter 1 parameter 2
2. Sliding inner slope Figure A 2 Slope instability Technical Advisory Committee on Water Defences 1998 The following variables above the geometry variables apply to the mechanism slope instability see Table A 3 For more information about this mechanism is referred to Steenbergen and Vrouwenvelder 2003B Table A 3 Variables for slope instability Steenbergen and Vrouwenvelder 2003B Variable nr symbol description 20 Au Deviation water levels 21 c cohesion per layer 22 tan friction angle per layer 23 q Model uncertainty Bishop 1 1 1 5 Heave piping In case of the mechanism heave piping the dike fails because sand under the dike is flushed away see Figure A 3 Two mechanisms are involved First the impermeable layer will heave Second pipes will develop due to the hydraulic gradient and sand from below the dike will be washed away n 9 Piping Figure A 3 Heave piping Technical Advisory Committee on Water Defences 1998 The following variables above the geometry variables apply to the mechanism heave piping see Table A 4 and Figure A 4 For more information about this mechanism is referred to Steenbergen and Vrouwenvelder 2003B T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 68 Task 7 Deliverable D7 1 Appendices 1to 5 PRAN cf Contract No GOCE CT 2004 505420 FLOODS ite Table A 4 Variables for heave piping Steenbergen and Vrouwen
3. Cc1 5 Cc2 2a Cc22b Aa2 4 Bal 1 Bal 5bii Ba2 5 Bc2 1a Bc2 1b h WaterD Water depth m lognormal 0 15 Bc2 1d Bc3 1a Ca2 1b Ca2 2b H IncRWaveH Incident regular wave height m Ca2 1a Aa2 1b Aa2 4 Bal 5bii Bal 5dii Water depth at the toe including n s h WDToe m Ba2 1bii Ba2 1biii the coverlayer Ba2 4d Bb1 2 Ca2 1b Ca2 2a Ca2 2b Water level in the floodplain Bal 5aiii Ca2 2a hb Hb PE m normal 1 0 1 dike ring Ca2 2b Bal 5ai hc Hc Water depth above structure crest m Bcl 1 Ba2 1bii Ba2 1biii Ba2 4b Ba2 4c Ba2 4i Berest m Ba2 4iii Ba2 5 Ca2 3 normal 1 0 1 Cb1 2c Cc1 2aii Cc1 2b Cc1 2c Cc1 2d T07_08 2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505 20 Parameter Unique fortran name Description LSE mapping Example distribution Distribution Distribution K Distribution parameter 1 parameter 2 Standard Deviation ey parameter 3 Variation i name coefficient Hcrest Crest height above SWL Cc2 2a Cc2 2b Da2 5 hf Hf0 Bc2 1j Bc2 3a normal 1 0 1 Crest height of the dune which hgp Hgp Aa2 la respects to SWL Significant wave height Hm0 Hm0 calculated from the spectrum Ca2 1a Hmo 4Amo hs HMobSoil Height of the mobilised soil Ccl1 2aii Ccl 2b Aa2 1a Aa2 1b Aa2 4 Ab2 1a Ab2 1b Bal 1 Bal 5bii Bal 5dii B
4. Reduction factor g gf gb Ba Aiii g RedFG taking into account the effects of oblique wave attack Ba2 4c Cb1 2a Cb1 2c Cb1 2d gd SoilDryWeight Volumetric weight of the dry soil kN m3 normal Cc1 2aii Cc1 2b 0 05 gf SlopeRough Roughness of the seaward slope Ba2 4iii Ba2 4c lognormal 1 0 25 gf c Gfc Roughness at the crest Ba2 4iii1 Ba2 4c Unit weight of the fine grained gfg GammaFg natural soil beneath the kg m3 Bbl 2 normal 0 2 embankment saturated gG Gg Velocity coefficient Ba2 1a Ba2 1biii TU7 OB 2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505 20 Distribution Distribution enii uri ps Distribution arameter arameter Parameter Unique fortran name Description LSE mapping xu E PRA p parameter 3 distribution Variation Standard name coefficient Deviation Cb1 2a Cb1 2c Cb1 2d Volumetric weight of saturated Cc1 2aii Cc1 2b gs GammaS kN m3 e normal 0 2 soil Cc1 2d Bal 5aiii Bc2 3b Ba2 4b Unit weight of the saturated part gsat GammaSat kg m3 Bb1 2 normal 0 2 of the embankment Unit weight of the unsaturated gunsat GammaUnsat kg m3 Bb1 2 normal 0 2 part of the embankment Cb1 2a Cb1 2c Cb1 2d gw GwaterL Groundwater level m Cc1 2aii Cc1 2b normal 0 1 Bal 5aii Cc1 5 Bb1 2 Cb1 2a Cb1 2c Cb1 2d Cc1 2aii gw GammaW Volumetric weight of w
5. 7002 7009 7023 7024 7025 7028 7038 7042 7047 7053 7071 7074 7075 7094 7109 85 2 82 4 71 7 71 6 71 2 70 1 65 1 64 1 63 6 61 9 57 6 569 55 7 51 7 47 4 7111 7116 7124 7129 7136 7139 7152 7159 7163 7167 7185 7202 7211 46 4 45 7 39 367 33 3 32 282 27 1 25 6 24 2 18 8 14 1 12 6 7220 7233 7249 7258 7271 11 5 88 6 4 3 9 0 9 T07_08_02_Reliability Analysis D7 1 Appendix 10 April 2008 18 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505420 5 Onsite Adjustments of profiles DHV has made several adjustments to the PC Ring database during the calculations Apart from the adjustment of the dike section selection as discussed in the previous section the dike profiles are adjusted to recently measured cross sections of the water board The adjustments of the profiles is further commented on in appendix A Schematization of coverings Often more than one type of covering on a dike section is present in dike ring 32 PC Ring is unable to perform calculations for more than type of covering for 1 dike section In case more than one type of covering is present VNK calculates all types individually and determines which one is governing also in relation to concurrent design points This governing covering is consequently accounted for when calculating the probability of flooding Only 1 type of covering per section is calcu
6. Dike with grass covering Dike with asphalt covering TO07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 15 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505420 Gasite Dune Sea walls RWS Public Works and Water Management Engineering structure The following division can be made The division and selection of dike and dune section is looked further into in section 2 3 0 0 8 4 3 20 1 22 0 40 2 44 7 76 0 68 2 69 7 70 1 71 2 76 3 11 3 78 8 79 8 82 7 82 9 84 3 84 6 85 1 0 8 km 4 3 km dike with stone covering dike with grass covering 20 1 km 22 0 km 40 2 km 44 7km 67 0 km 68 2 km 69 7 km 70 1 km 71 2 km 76 3 km 77 3 km 78 8 km 79 8 km 82 7 km 82 9 km 84 3 km 84 6 km 85 1 km 85 7 km dike with stone covering sea wall RWS dike with stone covering sea wall RWS dike with stone covering dune sea wall RWS dike with stone covering dune dike with stone covering dune dike with grass covering dike with stone covering dune dike with stone covering dune dike with stone covering dune grass 14 Structures are present in dike ring area 32 An overview of these structures is given in table 2 1 1 Pumping station Cadzand 2 Pumping station Campen 3 Pumping station Nieuwe Sluis 4 Pumping station Nummer Een 5 Pumping station Othene 6 Pumping
7. Example arameter 1 parameter 2 Parameter Unique fortran name Description LSE mapping i E d parameter 3 distribution Variation Standard S name coefficient Deviation Ccl 2d model uncertainty factor mR MCCI 2DR Cc1 2d E strength Ccl 2d model uncertainty factor mS MCCI 2DS ia Cc1 2d 5 loading Cc1 5 model uncertainty factor mR MCCI 5R Cc1 5 strength Ccl 5 model uncertainty factor mS MCCI 5S Ccl1 5 E loading Cc2 2a model uncertainty factor mR MCC2 2AR Cc2 2a i strength Cc2 2a model uncertainty factor mS MCC2 2AS i Cc2 2a B loading Cc2 2b model uncertainty factor mR MCC2 2BR Cc2 2b i strength Cc2 2b model uncertainty factor mS MCC2 2BS Cc2 2b 7 loading Da2 5 model uncertainty factor mR MDA2 5R Da2 5 B strength Da2 5 model uncertainty factor mS MDA2 5S Da2 5 2 loading TU7 OB 2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 mn m e oy fLOOP Sic Contract No GOCE CT 2004 505 20 Distribution Distribution SET Distribution 9 eee Example parameter 1 parameter 2 Parameter Unique fortran name Description LSE mapping parameter 3 distribution Variation Standard S name coefficient Deviation Da4 1 model uncertainty factor mR MDAA IR Da4 1 a strength Da4 1 model uncertainty factor ms
8. Ed Meyer H A John Wiley and Sons New York 146 190 4 Khuri A I and Cornell J A 1987 Response surfaces design and analyses Marcel and Dekker New York 5 Pandey M D Van Gelder P H A J M and Vrijling J K 2003 Dutch Case Studies of the estimation of extreme quantiles and associated uncertainty by bootstrap simulation Environmetrics DOI 10 1002 env 656 6 PC RING Manual 4 3 QQQ Delft and Demis bv September 2004 7 Schueller G I and Stix R 1987 A critical appraisal of methods to determine failure probabilities Structural Safety 4 239 309 8 Shinozuka M 1983 Basic analysis of structural safety Journal of Structural Engineering ASCE 109 3 721 740 9 VNK Report Safety in the Netherlands mapped Flood risks in dike ring area 32 Zeeuws Vlaanderen December 2005 T07_08_02_Reliability Analysis D7 1 Appendix 10 April 2008 48 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505420 5 Onsite APPENDIX II A SCHEMATIZATIONS AND ADJUSTMENTS BY DHV Selection dike sections In consultation with the water board two weak links in the dikes are added to the selection of DHV It concerns weak links near Hm 72 000 this section was already in the original schematization section 7023 Hm 83 000 this section has eventually been added as section 7009 dike section 7008 was chosen at first This has been changed because the choice between 7008 and 7009 didn t matte
9. Standard Deviation UN parameter 3 Variation name coefficient Ww SPartVel m s Aa2 la particles Width of shingle beach Ww ShingleBeachWidth determined as narrow wide and m Ab2 la lognormal 0 5 condition grade W WSuW Water surface width m Ba3 1 wa Wa Distance between two tie rods m Cb1 2a x coordinate of leaking point at Xu m Bal 5b Bal 5bii Xu the inner berm x coordinate of intersection XW Xw point of still water level and m Bal 5b Bal 5bii outer slope importance of structure factor u y StructImp S Ba2 4iii gt 1 engineering judgement factor Eccentricity of the center of Y e Da4 2b Y gravity from the point of impact y Eccent Eccentricity ship in canal m Bc2 1b Bc2 1d Initial unscoured bed level Zo Z0 m Ba3 1 adjacent to toe of protection TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 ELAARN T reuunsife Contract No GOCE CT 2004 505 20 Distribution Distribution Pease ul a5 Distribution arameter arameter Parameter Unique fortran name Description LSE mapping A z E NS p parameter 3 distribution Variation Standard name coefficient Deviation a TieRAng Angle of inclination of the tie rod P Cb1 2a a Sang Angle of the slope Ab2 1b Ba2 1b Cofficients for determination of Cc1 2c Cc2 2a Cc2 2b a p ABeta i horizontal wave load Ca2 3 Aal 1
10. loading Bc3 1b model uncertainty factor mR MBC3_1BR T Bc3 1b strength Bc3 1b model uncertainty factor ms MBC3 IBS T Bc3 1b gt loading Bc3 1c model uncertainty factor mR MBC3_1CR T Bc3 1c strength Bc3 1c model uncertainty factor ms MBC3 1CS T Bc3 1c loading Bc3_1d model uncertainty factor mR MBC3_1DR Bc3 1d B strength Bc3 1d model uncertainty factor ms MBC3 IDS 3 Bc3 1d loading Ca2 1a model uncertainty factor mR MCA2 1AR Ca2 1a x strength Ca2 1a model uncertainty factor ms MCA2 1AS Ca2 1a loading T07_08 2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Pann crta Contract No GOCE CT 2004 505 20 FLOGS Distribution Distribution SET Distribution Example arameter 1 parameter 2 Parameter Unique fortran name Description LSE mapping i E d parameter 3 distribution Variation Standard S name coefficient Deviation Ca2 1b model uncertainty factor mR MCA2 IBR Ca2 1b E strength Ca2 1b model uncertainty factor mS MCA2 IBS 1 Ca2 1b x loading Ca2 2a model uncertainty factor mR MCA2 2AR Ca2 2a strength Ca2 2a model uncertainty factor mS MCA2 2AS Ca2 2a loading Ca2 2b model uncertainty factor mR MCA2 2BR Ca2 2b i strength Ca2 2b model uncertainty factor mS MCA2 2BS Ca2 2b i loading Ca2 3 model uncertainty factor mR MCA2 3R Ca2 3 i strength Ca2 3 model uncertainty
11. Cc1 2d TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505 20 Parameter Unique fortran name Description Coefficients for consideration of LSE mapping Example distribution Distribution Distribution EN Distribution parameter 1 parameter 2 Standard Deviation PRA parameter 3 Variation name coefficient kd kf KdKf the crest width Bk and Aal 1 Bal 1 Ba2 5 sharpcrestedness of the weir Rk kh VelProF velocity profile factor Bc3 1b Bc3 1d Cb1 2a Cb1 2c Cb1 2d K PassGrainFCoeff Coefficient for passive horizontal i Tn assGrainFCoe T ognorma P grain force Cc1 2aii Ccl 2b 8 Cc1 2d 0 1 Slope reduction factor for critical Bc3 1b ksl Ksl bed shear stress ksl kl kd Bc3 1d is Ki Turbulence amplification factor Bc3 1b for current velocity Bc3 1d kr KLamda Coefficients Cc1 2d kn Kn Coefficients Ccl1 2d Length of horizontal sliding l Hslid Bb1 2 surface beneath embankment Effective span distance between L Espan Ca2 3 the supports L Slab Length of the concrete slab Cc1 2c Cc1 2d Cc2 2b TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to5 mn m e 7 fLOOPS Ac Contract No GOCE CT 2004 505 20 Distribution Distribution K Distribution Example parameter 1 parameter
12. Converged Yes or No Convergence The proximity of the last 2 failure rates checked If Converged is Factor Yes the value will be less than the Convergence Factor in CONVERGENCE CONTROL Samples The number of Samples checked If Converged is No this will be the same as Max Samples Time Taken s Time taken by the Calculator Log messages In order to assist the investigation of the results produced by the Calculator a log file is produced with messages as follows Currently the log file is produced in the folder C TEMP which must exist The name of the file is pcname_nnnn log where pcname is the name of your PC and nnnn is a number derived from the process id of the EXCEL instance you are using This normally ensures that you can accumulate log files for several runs and delete them at your leisure Log files are simple text files viewable by tools such as Notepad 1 Q 3 4 At the start of each Calculate run 2 messages show the fixed values and the distributions to be used Each time convergence is checked a message shows the Sample and Failure counts the current and previous rate of failure and an indication of the proximity of the two The line includes a message if full convergence was detected Two additional messages are produced for each Sample if you have chosen Yes for Log each Sample One shows the specific values used for each parameter The other shows the resulting failure indicatio
13. Deviation Parameter Unique fortran name Description LSE mapping parameter 3 distribution Variation A name coefficient pressure Maximum uplift pressure lower than average outside water level en Hgws 0 higher than average outside Bc2 1j Bc2 3a normal 1 0 5 switc water level 1 or highest of the two 2 hGWS FMGWS Average water level Bc2 1j Bc2 3a lognormal 1 0 5 0 13 Factor to derive the design water fMGWS DeqReqHs Bc2 1k level hMGWS Required equivalent thickness of Deq req Hs StaticDynamic Bc2 1m top layer as a function of Hs Static stability or dynamic switch FCoeffFric stability limit state equation for Bc2 3a Bc2 1m f PilEscMay coefficient for friction Bc3 1b Bc3 1d Choice for Pilarezyk UM switch NormIncReglIrreg i Ca2 la deterministic Escarameia and May models L SmReq incident regular wavelength Ca2 la Normally incident nonbreaking Aa2 4 Bal 1 switch TopSill regular 0 or irregular 1 waves Bal 5bii Bal 5dii TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505 20 Parameter Unique fortran name Description LSE mapping Ba2 5 Ca2 2a Distribution Distribution K Distribution Example parameter 1 parameter 2 Standard Deviation BAG PUE parameter 3 distribution Variation i name coeffic
14. F Il 4 PROBABILITY OF FLOODING CALCULATION DIKE RING AREA 32 This section describes the approach and results of the performed calculations for determining the probability of flooding With the presentation of the results a distinction is made between contributions to the probability of flooding of dunes dike sections and structures and of different failure mechanisms within them The calculated results are compared with the judgment of the water board The computer model used to calculate the probabilities of flooding for dike ring 32 is PC Ring version 4 3 February 2005 Calculations have been made by DHV with checks by VNK and assessments by WZE It proved to be difficult to perform good calculations of the probability of flooding due to the variation in loads and the complexity of the dike profiles Approach and assumptions of the calculations 1 1 1 General The calculations of the probability of flooding of the dike ring and the probability of failure of dike section and dunes have been performed using the computer program PC Ring version 4 3 Input for this program are the schematization and the data as discussed in chapter 2 The program calculates a probability of failure for each dike section based on the contributions of each separate failure mechanism and eventually the total probability of flooding for the entire dike ring Additionally the program provides insight in to what amount the various variables e g the length
15. Variation a name coefficient potential head induced in the pb PotHead Bcl 5 filter or a gabion i idem critical value hydraulic cr PsiCR v Bc3 1c stability mobility parameter of protection Cr ProElem Bc3 1b Bc3 1d element Friction angle between two d TwoMat Ba2 3 lognormal 5 materials Stability correction factor for fsc PhiSC Bc3 1b Bc3 1d current exposed stones Stability correction factor for fsw PhiSW Bc2 1g Bc2 1m wave exposed stones Stability upgrading factor fu PhiU y EE 5 Bc2 1g Bc2 1m depending on system Dn50 of the structure core in Dn50 core Dn50Core Ab2 1b Bc2 1c Van Gent et al Coefficient in Bc1 5 in erosion of c geotextile cGeotex subsoil through revetment or Bcl 5 geotextile TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 muousife Contract No GOCE CT 2004 505 20 Distribution Distribution Distribution A NS Example parameter 1 parameter 2 Parameter Unique fortran name Description LSE mapping A itei ie PS parameter 3 distribution Variation Standard name coefficient Deviation Breaker index in Bc2 1b erosion Be2 1b k si Ksi i g of revetment by shipwaves Rete crest freeboard relative to the Dade Re rear water level at rear side of the crest Area root ratio in Ba2 4b erosion Ar A ARootRatio clay inner slope by wa
16. supporting LSE functions T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 85 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505420 FailureModeParam csv A CSV file defining the Parameter values required Parameter csv Structure csv FailureMode ped SheetPileWall fta Excel exe config by the LSE for each Failure Mode A CSV file defining the names of all the current Parameters A CSV file defining each structure and its associated Fault Tree file A Primary Events database for use with OpenFTA when defining further fault trees A Fault Tree file for a Sheet Pile Wall structure EXCEL configuration file required for PCs with NET Framework 3 0 or later Table 4 Reliability Calculator files T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 86 Task 7 Deliverable D7 1 Appendices 1to 5 mn 7 Contract No GO CE CT 2004 505420 FLOORS Ze APPENDIX V 2 EXAMPLE LSE FROM FLOODSITE TASK 4 Bal aiii Uplifting of impermeable layers behind earth embankment Summary Uplifting behind embankments occurs if the difference between the local water level h and the water level inside h is larger than the critical water level h Reliability equation The reliability function is expressed by z m h m Ah where h critical water level m Ah difference between local water depth in front of dike and water level in the flood
17. 149 Qiobith Discharge Lobith Rijn 150 hoz Water level Delfzijl 151 hos Water level OS11 152 Qvecht Water level Dalfsen Vecht 153 Quissel Discharge Olst IJssel 154 Quith Discharge Lith Maas 155 Ahwk Prediction error water level Maeslantkering 156 Aijsselmeer Water level IJsselmeer 157 itarkermeer__ Water level Markermeer 158 hava Water level Hoek van Holland 159 hoH Water level Den Helder 160 hyiis Water level Vlissingen 161 hir Water level Harlingen 162 hio Water level Lauwersoog 163 Vsp Wind speed Schiphol Deelen 164 Vic Wind speed ligth island Goeree 165 Vak Wind speed de Kooy 166 Wiis Wind speed Vlissingen 167 Vw Wind speed Terschelling West 168 Ahok Prediction error water level Oosterscheldekering 169 two Duration wind setup 170 Atos Phase difference T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 75 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505420 IV C SENSITIVITY COEFFICIENTS DIKE RING 7 32 AND 36 The sensitivity coefficients of dike rings 7 32 and 36 are shown in Table C 1 Table C 2 and Table C 3 Table C 1 Sensitivity coefficients dike ring 7 Noordoostpolder Variable 1 O do oo NO oO A O N XS cw OLX N 13 14 15 16 17 18 19 20 21 22 23 24 25 Description Dike height h d Berm height h B Berm width B Toe height h t Slope outer slope top Slope outer slope bottom Slope outer slope Mode factor critical overflow discharge m qc
18. 6First results per dike section The preliminary results per dike section in beta are provided in table 4 1 These results are discussed with the water board see section 4 3 4 As a result of this discussion it was concluded that a number of sections can be left out of consideration for now These are results that are unidentifiable for the water board and have to be analysed further or weak spots that are nominated to be improved These sections are shaded grey in the table T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 36 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505420 7002 7009 7023 7024 7025 7028 7038 7042 7047 7053 7071 7074 7075 7094 7109 50 66 57 56 60 74 52 58 58 157 54 48 49 50 5 55 6 7 6 5 11 3 11 11 73 16 7 63 98 10 63 6 1 62 70 64 34 137 50 46 9 2 4 62 9 3 7 6 5 1 70 7 8 7 8 6 0 7111 7116 7124 7129 7136 7139 7152 7159 7163 7167 7185 7202 7211 5 2 4 9 3 9 5 0 5 6 4 9 4 9 4 4 4 8 3 0 4 1 4 8 3 9 6 7 6 6 7 1 6 3 8 0 6 1 5 3 5 2 4 5 7 5 4 1 4 8 4 5 5 4 6 5 8 5 5 3 5T 37 6 8 5 2 14 13 36 6 1 14 7220 7233 7249 7258 7271 4 5 4 8 4 5 4 5 4 6 2 7 4 6 5 2 6 0 6 4 37 8 9 37 1 8 2 2 Table 4 1 Reliability indices preliminary per section in
19. A VariousEmpiricalCoeff e Bc3 la empirical formulae Empirical factor according to Aa2 1b Ba2 1bii A LarsonCoeffA Larson et al 2004 A 1 34 10 2 Ba2 biii A IceCrushingArea Area of ice crushing m Da4 2b Ac CanalArea Area of canal s cross section n Bc2 1b Bc2 1d Ba3 1 Ae RockErosionArea Erosion area on rock profile m Ab2 1b Bc2 1c Ba2 4c as FlowDir Flow direction Bal 4 Bb1 4 As ShipArea Area of ship s cross section m Bc2 1b Bc2 1d l ApronW Width of apron m Bc3 la Empirical factor according to Aa2 1b Ba2 1bii b LarsonCoeffB Larson et al 2004 b 3 19 10 4 Ba2 lbiii Ratio of local geometry b 0 for b LocGeometryRatio Da4 2b head on impact TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 muousife Contract No GOCE CT 2004 505 20 Distribution Distribution SET Distribution Example parameter 1 parameter 2 Standard Deviation Parameter Unique fortran name Description LSE mapping Ee parameter 3 distribution Variation A name coefficient Width of segment element or b SegmentWidth i m Bc2 3b slice B ChannelWidth Channel width m Bc2 1b Bc2 1d Aa2 1b Aa2 4 Ba2 4i Ba2 4c Ba2 4d Ba2 5 B CrestWidth Crest width m Bal 5bii Bal 1 lognormal 1 0 20 Bal 5bii Ba2 1bii Ba2 1bii Ca2 1b Aa2 4 Bal 1 BB BermWidth Berm width m Bal 5bii Bal 5dii normal 0 15 Ba2 5
20. Bal 1 Aa2 4 Ba2 41 Ba2 4b Ba2 4c au InsSlopeAng Angle of the inner slope T normal 1 0 05 Ba2 4d Bal 5bii Bal 5dii Aa2 1b Aa2 4 Bal 1 Bal 4 Ba2 3 Ba2 4c Bal 5bii Bal 5dii Ba2 1bii Ba2 1biii ao OutSlopeAng Angle of the outer slope H Ba2 4d Ba2 41 Ba2 5 normal 1 0 05 Ba2 Aiii Bb1 4 Bc1 4 Bc1 5 Bc2 1b Bc2 1c Bc2 1d Bc2 1g Bc2 1h Bc2 1m Bc2 3b Aa2 1a Aa2 1b Aa2 4 SE Angle of wave attack with P WaveObliquity a Ab2 1a Ab2 1b Bal 1 normal 1 15 respect to the structure Bal 5bii Bal 5dii T07_08 2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 muousife Contract No GOCE CT 2004 505 20 Distribution Distribution K Example ei as Distribution arameter arameter Parameter Unique fortran name Description LSE mapping Acc o E PRA p parameter 3 distribution Variation Standard name coefficient Deviation Ba2 1a Ba2 1b Ba2 3 Ba2 1bii Ba2 1biii Ba2 4111 Ba2 5 Bc2 1a Bc2 1b Bc2 1c Bc2 1d Bc2 1g Bc2 1h Bc2 1m Bc2 3b Ca2 1a Ca2 1b Ca2 2a Ca2 2b Ca2 3 Cc2 2a Cc2 2b Da2 5 Aa2 4 Ba2 4d Bl Betal Internal friction angle of sand 2 Bal 5bii Bal 5dii Reduction factor for slope If RFSIoR roughness wave run up wave Ba2 4b normal 0 1 overtopping Bal 4 Bbl 4 Bcl 4 yw WaUWei water unit weight kN m normal 1 Bc2 1j Bc2 3a 0 01 Relativ
21. Contract No GOCE CT 2004 505420 IV B OVERVIEW OF RANDOM VARIABLES IN PC RING Computer program PC ring is used in the Netherlands to calculate failure probabilities of dike rings The different failure modes are described Appendix A an overview of all random variables is provided in table B 1 Table B 1 Overview of random variables in PC Ring Steenbergen and Vrouwenvelder 2003A pp 48 50 and 2003B Variable nr symbol description Geometry 1 ha Dike height 2 hg Berm height 3 B Berm width 4 h Toe height 5 tan Qub Angle outer slope top 6 tan Quo Angle outer slope bottom 7 tan Angle inner slope Overflow overtopping 8 Mc Model factor critical overflow discharge 9 k Roughness inner slope 10 fo Factor for determination Q 11 fn Factor for determination Qa 12 Mao Model factor for occuring overflow discharge 13 c Cohesion Clay layer inner slope 14 Q Friction angle Clay layer inner slope 15 p Soil density Clay layer inner slope 16 dk Layer thickness Clay layer inner slope Stability 20 Au Deviation water levels 21 c cohesion per layer 22 tan friction angle per layer 23 q Model uncertainty Bishop Heave piping 41 d Thickness covering layer 42 D Thickness sand layer 43 L Leakage length 44 0 rolling resistance angle 45 x d4o Factor Crear 46 dzo Grain size 47
22. Deliverable D7 1 Appendices 1to 5 mn 7 Contract No GOCE CT 2004 505420 ThLUOD Site V Appendix 5 User manual reliability tool APPENDIX V 1 INSTALLATION GUIDE Software Installation procedure These steps describe the procedure for installing the software Steps 0 if needed and 3 will require local administrator rights for your PC If you do not have such rights you must ask somebody who does have these rights to perform the step s 1 Create a folder on your hard disk and copy the files from the ReliabiltyCalculator zip file supplied See Table 3 Reliability Calculator files below for a list of the files The remainder of this document uses RelCalcFolder to refer to the path name of the folder which you have created here 2 If INET Framework has not been installed already on your PC install it now To check run Control Panel and choose Add or Remove Programs If you see Microsoft NET Framework 2 0 all is well You may have versions other than 2 0 but you must also have 2 0 If not you may download the package from http www microsoft com downloads details aspx FamilyID 0856eacb 4362 4b0d 8edd aabl5c5e04f5 amp DisplayLang en You must then install it This may require administrator rights If you see NET Framework 3 0 or 3 5 then you must also take an additional step see 5 3 Locate the file regasm exe on your PC Start with your Windows folder e g C Windows and look in the subfolder Microsof
23. GOCE CT 2004 505420 FLOODS ite which at first were schematized as dikes in PC Ring into a dune profiles in order to be able to calculate 4 dune sections Possibly a fifth dune section could have been calculated if a dune section was divided into two dune sections In consult with VNK it was decided not to do this Of these 3 sections are weak and 1 section is a strong section The strong section was chosen to see whether the outcomes in PC Ring provide the same picture of the safety as the present situation of the section Eventually the sections 7008 7010 7013 and 7027 were calculated as dune section See the following schematization also t needs noticing here that At first it was decided to calculate section 71 Breskens as well since this is a weak section which was indicated by the water board as well In consult with VNK however it was decided not to calculate this section because the section and the rest of the dune site of which section 71 is a part can t be schematized properly The site is in between two jetties that have a strong reducing effect on the waves Wave conditions are used that serve as input for the SWAN calculations for dunes For the Westerschelde these are the wave conditions for platform EUR This is a deep water location at a considerable distance from the coast In practice this means that the wind directions W to NNE are governing for the dunes This was assumed because other conservative loads wer
24. MDAA 1S Da4 1 7 loading Da4 2a model uncertainty factor mR MDA4 2AR Da4 2a x strength Da4 2a model uncertainty factor ms MDA4 2AS Da4 2a E loading Da4 2b model uncertainty factor mR MDA4 2BR Da4 2b i strength Da4 2b model uncertainty factor ms MDAA 2BS Da4 2b loading Da4 2c model uncertainty factor mR MDA4 2CR Da4 2c a strength Da4 2c model uncertainty factor ms MDAA 2CS f Da4 2c 7 loading Da4 3 model uncertainty factor mR MDA4 3R Da4 9 B strength Da4 3 model uncertainty factor ms MDAA 3S Da4 10 E loading TUZ OB 2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 muousife Contract No GOCE CT 2004 505 20 Distribution Distribution EN Distribution Example parameter 1 parameter 2 Standard Deviation Parameter Unique fortran name Description LSE mapping parameter 3 distribution Variation name coefficient Bal 5ai model uncertainty factor mR MBA 5AIR Bal 5ai lognormal 1 0 7 0 1 E strength Bal 5ai model uncertainty factor ms MBAI 5AIS Bal 5ai loading Thickness of the wate conductive D WaterCondLayer layer underneath the Bal 5ai normal 1 0 1 embankments Aa2 4 Bal 1 Bal 5bii Bb BermLevel Berm level in outside slope normal 1 0 2 Bal 5dii Ba2 5 switch for calculation damage level of rock armour from Sdrock SdRock RockErosionArea and Dn50 0 Bc2 1c Ab2 1b deterministic or for the indication of the damage
25. Make a random draw for each of the random h 1 8 m OD variables Calculate the value for each of the individual s 1 failure modes Z1 Z2 Z3 etc Check whether Z1 lt 0 Z2 lt 0 Z3 lt 0 etc Tp 2 10 s 7 If so nl 2 nl 1 starting from n1 0 n2 n2 1 starting from n2 0 SEE n3 n3 1 starting from n3 0 10000 times Check whether OR gates or AND gates are 0 When Z1 0 OR Z2 0 n12 n12 1 starting from n12 0 When Z1 0 AND Z3 lt 0 etc n13 n13 1 starting from n13 0 Etc Repeat number of Monte Carlo simulations Calculate probabilities of failure by dividing nl n2 n3 n12 n13 etc by m total number of simulations Calculate relative standard deviation of probability of failure by IP mPf Figure 10 Flow chart with steps to calculate fragility and the annual probability of failure TUZ 0B 2 Reliability Analysis D7 1 Appendix 10 April 2008 Probability of failure Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505420 T LUUPS 2 Indication extreme water level OD 5 0 m in 1953 Crest level OD 6 94m 1 0 0 9 08 Total 0 7 Overtopping Me i Uplifting 0 5 04 l Piping 0 3 EE MS Indication extreme 02 water level i l Crest level 0 1 i 0 0 T T 0 0 2 0 4 0 6 0 8 0 Water level m OD 10 0 Figure 11 Fragility for earth embankment section 4 The fai
26. Roughness inner slope k Factor for determining Q bf b Factor for determining Q_n f n Model factor occurring overflow discharge m qo Error position bottom Model factor Bretschneider for Hs Model factor Bretschneider for Ts Error in local water level Storm duration t s Level Lake IJssel Wind speed Schiphol Deelen null Discharge Lobith Discharge Dalfsen Discharge Olst Root depth grass d w Widht covering layer of clay L_K alfa 0 05100 0 00600 0 00000 0 00100 0 01000 0 00000 0 00300 0 05200 0 00700 0 02200 0 02700 0 05700 0 00000 0 00000 0 00000 0 00000 0 00000 0 32800 0 86200 0 37300 0 00000 0 00000 0 00000 0 00000 0 00000 alfa 2 0 00260 0 00004 0 00000 0 00000 0 00010 0 00000 0 00001 0 00270 0 00005 0 00048 0 00073 0 00325 0 00000 0 00000 0 00000 0 00000 0 00000 0 10758 0 74304 0 13913 0 00000 0 00000 0 00000 0 00000 0 00000 TO07 08 02 Reliability Analysis D7 1 Appendix 76 10 April 2008 ContectNeGOCET20S0S FLOOESite 26 Widht dike core on crest height L BK 0 00000 0 00000 27 Stone thickness D 0 00000 0 00000 28 Tangent alfa u 0 00000 0 00000 29 Tangent alfa i 0 00000 0 00000 30 Relative density stone 0 00000 0 00000 31 Coefficient stone pitching op klei c k 0 00000 0 00000 32 Coefficient grass c g 0 00000 0 00000 33 Coefficient erosion covering layer c rk 0 00000 0 00000 34 Coefficient erosion dike core c rb 0 00000 0 00000 35 Thickness granular filter
27. SWAN points are 100 meter from the coast 300m apart so the influence of the foreland will be partially included in these Foreland over 100 meter is of no use anyway Selection of profiles for sliding mechanism inner slope Because calculating the sliding mechanism is an elaborate process this calculation is not performed for all sections The district water board has made a selection of 7 cross section profiles out of a series of 40 that were used for the testing during the process of schematization From these only 1 matches with one of the 33 selected dike sections Therefore only one result will be calculated for the sliding mechanism of the inner slope Assessment of the water board In accordance with the Law on water retention 1996 the District Water Board Zeeuws Vlaanderen reported on the condition of the embankments in dike ring 32 to the County Council of the Zeeland Province at the end of 2000 This concerned the first report from a series of the 5 annual safety tests Dikes The assessment of the water board for dike ring 32 based on the results of the first test is summarized in table 2 2 for the selected sections In this table the Ht score represents the score for overflow and T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 20 Task 7 Deliverable D7 1 Appendices 1to 5 PSAN cft Contract No GOCE CT 2004 505420 rLOODSite wave run up STPI score represents the score for bursting and piping STBI
28. Start method adapted from 8 1 This results in a beta of 36 Covering seems very thick Possibly a number of samples need to be calculated Ov ov keeps functioning Initial value adapted from 8 gt 1 Result for covering but now ov ov does not function well Southern wind direction 180 degrees turned off Now result for both mechanisms TO07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 56 Task 7 Deliverable D7 1 Appendices 1to 5 FLkU0Rns f2 Contract No GOCE CT 2004 505420 T eiv 159 163 167 185 202 211 23 124 129 233 All Initial value adapted from 8 gt 1 Result for covering but now ov ov does not function well It is not allowed to turn off the northern wind direction and even with 6DS FI no good result follows This should be solved manually Thus ov ov with initial value 8 and covering with initial value 1 gt both with FORM DS Initial value adapted from 8 gt 1 This does lead to result for the covering Ov ov keeps going well Wind directions North 30 60 90 turned off not governing for ov ov gt now result for both mechanisms Initial value adapted from 8 gt 1 Result for covering but now ov ov does not function well It is not allowed to turn off wind direction 60 degrees and even with 6DS FI no good result follows This should be solved manually Thus ov ov with initial value 8 and covering with initial value 1 both with FORM DS Initial value adapted fr
29. This would mean that flooding is to be expected more than once each 11 years for dike ring area 32 Since the results have not been analysed thoroughly one can not speak of a so called reference sum of dike ring 32 in this case Mechanism COMBINI COMBIN2 Overflow overtopping 1 794 1 11312 Bursting and piping 1 30211 1 30211 Revetment damage and dike 1 22 1 574713 erosion Overflow and overtopping of 1 16920 1 16920 hydraulic structures Non closure of hydraulic 1 3984 1 3984 structures Structural failures of hydraulic 1 22 1 34364 structures Overall failure probability 1 11 1 1996 Table 4 6 Probability of flooding dike ring 32 according to DHV When the 6 weakest spots for the dikes 7167 097 dp290 for overtopping and wave overrun 7002 072 dp7 7009 020 dp16 7028 004 dp25 7258 074 dp99 and 7271 072 dp69 for covering damaging and erosion body of the dike and the weakest spot for the structures constructive failure of pumping station Othene are left out of consideration a probability of flooding of 1 2000 per year COMBIN 2 is calculated According to the water board this approaches the value it would expect T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 45 Task 7 Deliverable D7 1 Appendices 1to 5 mn ita Contract No GOCE CT 2004 505420 fkOOP sic In both cases the mechanism sliding is not taken into account in the calculated probability whilst it is clear that sta
30. arameter arameter Parameter Unique fortran name Description LSE mapping An R PED p parameter 3 distribution Variation Standard name coefficient Deviation unfavourable effects resulting from the way in which the load is applied Cofficient for determination of Cc2 1c Ca2 3 P ABeta2 horizontal wave load combined f Cc2 2a Cc2 2b with ABeta Area of reinforcement steel in As ReinStArea C c1 2d C c2 2b concrete Reduction factor of the RedFuFh RedFuFh horizontal wave pressures Cc2 2a Fh0 1 Area of reinforcement steel in As StorArea Bal 6 Da2 5 concrete A AdmWOver Storage area behind the structure Bal 6 Da2 5 deterministic Switch for admissable q 0 or h switch qadm ABroadShort 1 40 Bal 6 deterministic A for broad crest 0 or short switch A ResStreF Da4 1 crested weir 1 i Resulting overall strength against Fr PierWidth Da4 2a ship collision TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to5 mnm m ey Contract No GOCE CT 2004 505 20 Distribution Distribution Example uri o Distribution arameter arameter Parameter Unique fortran name Description LSE mapping Ac 5 DR d parameter 3 distribution Variation Standard name coefficient Deviation The width of the pier under 1ce Tm D HeadEcc M Da4 2b deterministic collision switch head y Switch for head on co
31. been added T07_08_02_Reliability Analysis D7 1 Appendix 10 April 2008 51 Task 7 Deliverable D7 1 Appendices 1to 5 mn T Contract No GOCE CT 2004 505420 FLOODS ite section loc code loc code 1 loc code 2 intpolation 230 7770028 7770028 7770029 85 851 7770028 7770028 7770029 40 1242 7770028 7770028 7770029 15 1401 7770028 7770028 7770029 5 With respect to the boundary conditions it applies that the dunes of dike ring 32 are to be coupled to the load model of the Westerschelde For this input files were added to the existing input files for the sandy coast With the adding of the files mentioned above two locations were added The locations concern 7770028 Bresken coordinates 27502 380752 test level 5 25 m NAP 7770029 t Zwin coordinates 15013 378273 test level 5 05 m NAP MHW check At the MHW check of DHV an error was found in the location codes in PC Ring As a result of this the right location codes were put in The MHW check was done once more by DHV see appendix B Calculations DHV The country setting of the computer is set to English to guarantee that the values that are put in the database are read correctly by PC Ring Decimal values have to be put in with a dot so 0 4 in stead of 0 4 The stochastic variables for the calculations with coverings were all switched on except for the deviation of wave direction All calculations for probabilities of flooding are per
32. board All profiles were schematized again because anomalies occurred between the measured and the schematized profiles DHV both did the initial calculation and a further analysis for dikes and dunes in principle With the calculations one ran into many difficulties concerning amongst others the schematization the complexity of the dike profiles the variation in loads and the programming due to which doing good calculations for this dike ring turned out to be difficult VNK checked and corrected all DHV s calculation for the dikes together with TNO This resulted in the fact that a probability of failure has been calculated for almost all mechanism for the selected sections The calculated probabilities of failure are discussed with the water board VNK processed the results of these discussions in this dike ring report The structures are assessed by DHV The results are tested and checked by VNK and the water board The MproStab calculations are performed by DHV and checked by GeoDelft Results of the calculations of the probability of flooding 1 1 5 Introduction In this section an insight is provided in the calculated probabilities of failure for dike ring 32 It concerns preliminary results since the results have not been analysed thoroughly These preliminary results have been discussed with the water board Because it concerns preliminary results a so called reference sum is not yet presented for dike ring 32 1 1
33. considered yet A first simple step to consider dependencies might be sensitivity calculations for different degrees of dependencies resulting in a range of possible failure probabilities However since the overall failure probability seems mostly dependent on section 8 overtopping dike inclusion of dependencies at this stage will probably not influence the result significantly T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 64 Task 7 Deliverable D7 1 Appendices 1to 5 PRAN cf Contract No GOCE CT 2004 505420 FLOODS ite T07_08_02_Reliability_Analysis_D7_1_Appendix 10 April 2008 65 Task 7 Deliverable D7 1 Appendices 1to 5 PERAN cera Contract No GOCE CT 2004 505420 FLOODS IV Appendix 4 Uncertainty database IV A FAILURE MECHANISMS The computer program PC RING is used in the Netherlands to failure probabilities of dike sections Steenbergen and Vrouwenvelder 2003 In order to calculate the failure probability a dike ring system is cut in several dike sections The reliability of the dike sections with respect to the failure mechanism is calculated after which the total failure probability of the dike ring is determined The following failure mechanism are examined in PC RING Steenbergen and Vrouwenvelder 2003B Overflow overtopping Slope instability Heave piping Erosion revetment and erosion dike body Piping structures Not closing structures Dune erosion Other mechanism have not been conside
34. d7o dho Uniformity 48 n White s constant 49 MarK x Apparent relative density of heaving soil 50 d Yu Relative soil density sand grainl 51 Mo Model factor heave 52 Mp Model factor piping 53 mn Model factor water level damping 54 k Specific permeability Revetment 61 dw Root depth grass T07_08_02_Reliability_Analysis_D7_1_Appendix 10 April 2008 73 oe eee eet HOGRSite Variable nr symbol description 62 Lk Width covering clay layer 63 Lek Width dike core at crest height 64 D Stone pitching thickness 65 tan ou Angle outer slope 66 tan ai Angle inner slope 67 A Relative density stone pitching 68 Ck Coefficient stone pitching on clay 69 Cg Coefficient grass 70 CRK Coefficient erosion covering layer 71 CRB Coefficient erosion dike core 72 d Thickness granular filter layer 73 Dis Grain size 15 percentile filer 74 S Crack width 75 Ct Coefficient stone pitching on filter 76 C Coefficient in determination leakage length 77 Cb Coefficient in determination leakage length 78 Ct Coefficient in determination leakage length 79 D Thickness asphaltic concrete 80 A Relative density asphaltic concrete 81 fucws Factor for normative water level 82 hews Level average discharge 83 B Angle in reduction factor r 84 Cot Coefficient strength stone pitching 85 Oz Acceleration factor
35. first row calculated by VNK based on the following failure mechanisms e Second row Overflow and wave overtopping e Third row Bursting piping e Fourth row Covering damage The reliability index of section 7249 for the mechanism sliding inner slope has been calculated as 2 1 Indices for dune erosion has been calculated for the following sections 7008 7010 and 7013 with beta values equal to 4 4 4 4 and 4 9 1 1 7 Sliding inner slope 7 Profiles have been selected for calculating the probabilities of failure for the failure mechanism sliding DHV calculated these 7 profiles with MproStab Only 1 profile is part of the 33 selected sections for the PC Ring calculations EMMA118 belongs to section 7249 076 dp124 A result for the mechanism sliding inner slope is incorporated in table 3 2 for only this section An overview of the calculated safety factors and reliability indices at different water levels for all 7 sections is provided in table 4 2 T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 37 Task 7 Deliverable D7 1 Appendices 1to 5 p Contract No GOCE CT 2004 505420 MUOESie DHV consequently considered with which of the profiles from table 4 2 each of the 33 sections matches best A profile is linked to each selected section and a probability of failure has been calculated for each section using PC Ring Since the used method is not correct the results are not displayed here The coupling is based on height of th
36. it could be that 1 or 2 wind directions do not converge These wind directions could than possibly be turned off For this one should first check whether these wind directions are not governing for overtopping wave overrun Action 1 all wind directions are turned back on for all selected sections For the land of Saeftinghe a foreland of 5 kilometer is put in the SWAN points however are located 100 meters in front of the coast and 300 meters apart Foreland of 5 kilometer is useless Foreland is turned off in the calculations of VNK If the foreland is located gt 4 meter there are no waves and thus no result At a MHW check water levels are related to the RVW book This is not correct There are different values with which should be checked For this a file has been delivered by TNO in the past No set up is ipc expected along the coast The values from table 4 are thus not correct Action 2 all wind set up has been removed in the database the dike section set up is set to zero everywhere as is the number of sections due to newer MHW check One does not save alterations because then the new assortment is not saved but the altered values are All MHW check performed according to the prescribed procedure Foreland is turned off at sections 211 220 233 249 258 271 See tab MHW check gt 2 per T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 55 Task 7 Deliverable D7 1 Appendices 1to 5 muousife Contract No GOC
37. layer d f 0 00000 0 00000 36 Grain size15 percentile filter 0 00000 0 00000 37 Crack width s 0 00000 0 00000 38 Coefficient stone pitching on filter c f 0 00000 0 00000 39 Coefficient in leakage length determination 0 00000 0 00000 c a 40 Coefficient in leakage length determination 0 00000 0 00000 c b 41 Coefficient in leakage length determination 0 00000 0 00000 ct 42 Thickness asphalt concrete D 0 00000 0 00000 43 Relative density asphalt concrete 0 00000 0 00000 44 Factorf MGWS 0 00000 0 00000 45 Heighth GWS 0 00000 0 00000 46 Angle in reduction factor r 0 00000 0 00000 47 Coefficient strength stone pitching c gf 0 00000 0 00000 48 Acceleration erosion alfa z 0 00000 0 00000 49 Damping factor alfa h 0 00000 0 00000 50 Coefficient c 0 00000 0 00000 51 Heighth fictive bottom 0 00000 0 00000 52 Parameter b 0 00000 0 00000 53 Nominal diameter 0 00000 0 00000 54 Upgrade factor 0 00000 0 00000 55 Stability parameter 0 00000 0 00000 T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 7 Task 7 Deliverable D7 1 Appendices 1to 5 FLOOD Contract No GOCE CT 2004 505420 Sum 1 46700 0 99972 T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 78 eee LOOBsite Table C 2 Sensitivity coefficients dike ring 32 Zeeuws Vlaanderen Variable Description alfa alfa 2 Dike height h d 0 0883 4 0 0078 Berm height h B 0 0563 4 0 0032 Berm width B 0 0000 0 0000 Toe height h t 0 0055 0 0000 Slope outer slope top tan al
38. level Sd 1 Angle of the lower part of the Aa2 4 Bal 1 Bal Sbii OutSlopeAngLow outside slope in case of a berm Bal 5dii Ba2 5 Angle in relation to gamma v to Aa2 4 Bal 1 Bal 5bii WallAng take the influence of a vertical crown wall into account Bal 5dii Ba2 5 NumRun Number of the wave run up Aa2 4 Bal 1 Ba2 5 TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 ELAARN T reuunsife Contract No GOCE CT 2004 505 20 Distribution Distribution Example ird a5 Distribution arameter arameter Parameter Unique fortran name Description LSE mapping eines E NS p parameter 3 distribution Variation Standard name coefficient Deviation model Bal 5bii Bal 5dii Factor taking the effect of wind Aa2 4 Bal 1 Bal 5bii Fwind on the overtopping discharge into account EurOtop Manual Bal 5dii Ba2 5 0 for 5 1 battered wall 1 for Aa2 4 Bal 1 Bal 5bii BatterWall510 10 1 battered wall Bal 5dii Ba2 5 Shape factor in overtopping Aa2 4 Bal 1 Bal 5bii gamma f models see Eurotop manual Bal 5dii Ba2 5 Ba2 1bii model uncertainty MBA2 IBIIR Ba2 1bii E factor strength Ba2 1bii model uncertainty 7 MBA2 IBIIS Ba2 1bii x factor loading Ba2 1biii model uncertainty a MBA2 IBIIIR Ba2 1bii E factor strength Ba2 1biii model uncertainty T MBA2 IBIIIS Ba2 1bii B factor loading MER Switch for ship wa
39. level directly loads the landward embankment h Inundation occurs in that case if the landward embankment fails T07 08 02 Reliability Analysis D7pb gpendixe failure mechanisms are relevant Explanation to bottom fault tree If AAPril 2008 3 Task 7 Deliverable D7 1 Appendices 1 to 5 mn P Contract No GOCE CT 2004 505420 FLOORSie 2 2 Reinforced concrete walls The primary function of reinforced concrete walls is flood defence in many cases the reinforced concrete wall is part of a larger earth embankment The reinforced concrete walls were built as part of flood defence improvements to the Thames Estuary in the 70s and 80s There are a number of different types of reinforced concrete walls along the Dartford Creek to Gravesend flood defence line The three types considered in the reliability analysis as well as a superficial picture are shown in figure 5 Sheet piles applied underneath the concrete structure prevent seepage piping or in some cases mobilise the soil between the piles for extra stability Table 2 contains an overview of the failure processes for reinforced concrete walls along the Dartford Creek to Gravesend flood defence line The table also indicates the failure mechanisms incorporated in the reliability analysis and reference to those in the Task 4 Floodsite report Figure 6 presents the fault tree applied to the reinforced concrete wall in the Dartford Creek to Gravesend reliability analysis Riverward Landwa
40. of seepage present or the height of the dike contribute to the calculated probability of failure This is an important factor for conducting sensitivity analyses The reliability index beta is often used for calculating with probabilities The probability of failure is a function of this reliability index PC Ring also calculates with betas The probabilities of failure of structures are calculated using different procedures without PC Ring The calculated probabilities of failure per structure do form input for PC Ring for calculating the probability of flooding of the entire dike ring based on the contributions of the distinguished dike sections and structures Statistic data of wind and water level are used for calculating the probability of flooding of dike sections Based on these data the load models are defined which are implemented in PC Ring The load models in question are adjusted to the valid hydraulic boundary conditions Please note that a clear difference has to be made between probability of exceedance probability of failure and probability of flooding The probability of exceedance is the probability that the water level at a dike section reaches higher than the test level This is used in the present safety approach The probability of failure is the probability that a dike section actually yields to one the failure mechanisms The probability of flooding is the probability that the dike ring floods as a result of failure
41. of a dike section on one or several places A comparison between these latter two probabilities T07_08_02_Reliability_Analysis_D7_1_Appendix 10 April 2008 32 Task 7 Deliverable D7 1 Appendices 1to 5 mn ita Contract No GOCE CT 2004 505420 fkOOP sic and the probability of exceedance is not possible The fact that in this report weak links are indicated when the probability of failure of that specific link is greater than 1 1250 does not relate to the fact that the probability of exceedance of this area is 1 1250 as well 1 1 2Failure mechanism dikes For calculating the probabilities of failure of dikes the hydraulic load of water levels and waves is confronted with the relevant characteristics of the embankment that are governing for the strength of the embankment Both the load and the characteristics of the embankment are described in terms of probability distributions Uncertainties in the input data are accounted for using these probability distributions Calculations of the probability of failure of a dike are based on the following failure mechanisms Overflow and wave overtopping Covering damage and erosion body of the dike Bursting piping Sliding inner slope Overflow and wave overtopping Whth this failure mechanism the dike fails because large amounts of water run or sweep over the dike In case of offshore wind of otherwise very small wave heights the yielding is described by the failure mechanism overflow In other c
42. of illustration the response database was built up using Sobek for a set of observed random boundary conditions In practice it is expected that the response database would be available Importance sampling is subsequently carried out for estimating the failure probability for each dike segment All the dike segments are assumed to be in series configuration and Cornell s bounds are computed for the system reliability These bounds are observed to be 2 56x107 and 8 75x10 The use of importance sampling in reliability analysis of the dike reveal that the sample size required is considerably less than full scale Monte Carlo simulations Overflowing probability 8 Lj r5 o ni a 4 PEP IES 49 69 9 69 49 SL PPS PEE PLS SP hainage Figure 3 9 Overflow probability of the 80km long dike Concluding Remarks The probability of overtopping of the 80 km long dike due to the occurrence of extreme sea levels and river discharge either concurrently or otherwise is estimated The reliability computations are carried out using importance sampling based Monte Carlo simulations A novel response surface based method based on already existing database is adopted while computing the performance functions The procedure shows promise in significantly reducing the computational effort T07_08_02_Reliability Analysis D7 1 Appendix 10 April 2008 31 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505420 Onsite ju
43. overtopping Figure A 1 Overflow overtopping Technical Advisory Committee on Water Defences 1998 The following variables above the geometry variables apply to the mechanism overflow overtopping see Table A 2 For more information about this mechanism is referred to Steenbergen and Vrouwenvelder 2003B Table A 2 Variables for overflow overtopping Steenbergen and Vrouwenvelder 2003B Variable nr symbol description 9 k Roughness inner slope 10 fo Factor for determination Q 11 fn Factor for determination Qa 8 Mc Model factor critical overflow discharge 12 Mo Model factor for occuring overflow discharge 13 c Cohesion Clay layer inner slope 14 Friction angle Clay layer inner slope 15 p Soil density Clay layer inner slope 16 dk Layer thickness Clay layer inner slope 1 1 1 4 Slope instability Slope instability occurs in case the dike becomes unstable and cannot supports its own weight anymore see Figure A 2 This mechanism usually occurs due to infiltration of water in the dike and or due to water pressure in sand layers below the dike Slope instability can occur both on the inner side and on the outer side However slope instability of the inner slope is usually assumed to be the dominant mechanism T07_08_02_Reliability_Analysis_D7_1_Appendix 10 April 2008 67 Task 7 Deliverable D7 1 Appendices 1to 5 m ita Contract No GOCE CT 2004 505420 FLOODS io
44. partly of stone Both types should be calculated The water board has indicated that transition structures often form a weak spot VNK does not calculate these Section Judgement water board VNK Calculations based on 7002 Stone revetment after inspection considered good Gras 7023 Stone revetment insufficient Stone 7024 Stone revetment insufficient Asphalt 7025 Stone revetment insufficient Asphalt 7042 Stone revetment excellent Stone 7071 Excellent grass Stone 7074 Excellent grass Stone TO07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 40 Task 7 Deliverable D7 1 Appendices 1to 5 x Contract No GOCE CT 2004 505420 Gasite 7075 Excellent grass Stone 7111 Stone revetment excellent Stone 7129 Excellent grass Stone 7136 Stone revetment excellent Stone 7139 Excellent grass Stone 7159 Asphalt insufficient Stone 7163 Asphalt insufficient Stone Table 4 4 Assessment of the water board based on preliminary results 2005 testing Sliding inner slope VNK assesses the sliding of the inner slope This mechanism is calculated correctly for 1 section 7249 076 dp124 for which a large probability of failure is calculated Other indicating calculations also indicate large probabilities of failure betas around 2 The water board has seen sliding of the outer slope but no real problems for the inner slope have ever arisen The cause of the bad results ca
45. station Paal 7 Sluice station Terneuzen Oostsluis 8 Sluice station Terneuzen Middensluis schutsluis 9 Sluice station Terneuzen Middensluis spuiriool T07_08_02_Reliability_Analysis_D7_1_Appendix 16 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 PRAN cf Contract No GOCE CT 2004 505420 FLOOD Site 10 Sluice station Terneuzen Westsluis 11 Sluice station Terneuzen Westsluis spuiriool 12 Discharge sluice station Braakman 13 Discharge sluice station Hertogin Hedwigepolder 14 Discharge sluice station Nol Zeven Table 2 1 Structures in dike ring 32 Division in 33 dike and 4 dune sections The dike ring area Zeeuws Vlaanderen was initially divided into 287 dike sections according to the VNK schematization These were mainly dikes but encompassed a number of dunes and structures as well Because calculating the probability of failure for this number of dike sections with PC Ring is very elaborate a selection has been made by DHV This selection is based on the presently existing sections in PC Ring Thus no routes with representative dike sections have been selected The chosen 33 dike and 4 dune sections are dike ring covering and are deemed to be representative for the total dike ring The dike ring area is divided into parts for the selection each with their own characteristic orientation One or more dike sections are selected within these parts where thought is given to
46. the result of VNK will have to be analysed further Result from VNK mainly agrees with the assessment of the water board and the testing TO07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 39 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505420 muunsize Covering damaging and erosion body of the dike The sections 7002 024 dp7 for grass 7009 020 dp16 for stone 7028 004 dp25 for stone 7258 074 dp99 for grass and 7271 072 dp69 for grass score relatively bad for the mechanism covering damaging and erosion body of the dike For the section 7028 004 dp25 as for 7038 139a dp17 insufficient data for the stone covering were initially put into the overall spreadsheet to calculate a result with PC Ring For dike section 7028 the data were copied from dike section 7042 after consult with the water board concerning the type of stone covering This results in a large probability of failure By principle it should be verified whether the copied data match the reality The water board indicates that this section is nominated for improvement concerning the stone coverings Thus the bad result is identifiable gt The water board thinks that the present result should not be taken into the calculation for the probability of flooding of the dike ring because the section is part of a running improvement project No other data have been put in for dike section 7038 and thus no result has bee
47. wall The probability of failure does not always cover all the relevant failure mechanisms or the probability of breach The probability of failure of earth embankments does not take slope instability into account The probability of failure of reinforced concrete walls does not take failure of the embankment underneath the concrete wall into account and therefore does not T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1 to 5 FLOOR Contract No GOCE CT 2004 505420 represent the probability of breach The probability of failure of anchored sheet pile walls represents the probability of ground instability and damage to the assets behind the anchored sheet pile wall T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 Tesk7 Deliverable D7 1 Appendices 1t0 5 Littlebrook power station Figure 9 Flood defence sections 1 to 75 in the Dartford Creek to Swanscombe Marshes flood defence system sections 1 to 67 are included in the time dependent system reliability analysis TUZ 08 C2 Reliability Anelysis D7 1 Appendix 10 April 2008 5 Annual probability of failure Central calculation method Calculation loops for wind NE SE SW NW Read joint sea water level and wind speed at mouth Thames Estuary from JoinSea Extra calculation loops for h Hs Tp A Calculate local h by interpolation Fetch depth Calculate local Hs and Tp with Bretschneider
48. 1 Appendices 1to 5 muousife Contract No GOCE CT 2004 505 20 Distribution Distribution enn ul io Distribution arameter arameter Parameter Unique fortran name Description LSE mapping AM E NS p parameter 3 distribution Variation Standard name coefficient Deviation pa RhoA Density of the revetment kg m Bcl 4 Bc2 1j Bc2 3a normal 1 pg RhoG Density of the subsoil kg m Bcl 4 normal 1 0 05 Cc1 2aii Cc1 2b pr rc ConcreteDensity Mass density of rock concrete kg m normal 1 Ca2 2a Ca2 2b Cc2 2a 0 05 pt RhoT Density of the top layer kg m Bc2 1k normal 1 0 05 Bal 5aiBc2 3a Bc2 3b Bc2 1k Bc2 3d Ca2 2a pw RhoW Mass density of sea water kg m Ca2 2b Cc2 2a Cc2 2b normal 1 Ca2 3 Cc1 2c 0 05 v Upsi Kinematic viscosity m s deterministic Bal 5ai lognormal 0 15 normal 1 2 ON IntFriction Angle of internal friction normal 2c 0 1 Bcl 5 normal 2a 0 2 tan q tan IntFriction Bcl 5 lognormal 5 0 15 ON SoilAngleFriction Effective soil angle of friction id Bal 4 Bb1 4 Bcl 4 lognormal TU7 OB 2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505 20 Parameter Unique fortran name Description LSE mapping Bc2 3a Ba2 4b Example distribution Distribution Distribution SET Distribution parameter 1 parameter 2 Standard Deviation PRA parameter 3
49. 1a model uncertainty factor mR MBC2_1AR Bc2 la z strength Bc2 1a model uncertainty factor ms MBC2 1AS Bc2 1a loading TU7 OB 2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 mn m e oy fLOOP Sic Contract No GOCE CT 2004 505 20 Distribution Distribution SET Distribution Example arameter 1 parameter 2 Parameter Unique fortran name Description LSE mapping i E d parameter 3 distribution Variation Standard S name coefficient Deviation Bc2 1b model uncertainty factor mR MBC2_1BR Bc2 1b strength Bc2 1b model uncertainty factor ms MBC2 IBS Bc2 1b B loading Bc2 1c model uncertainty factor mR MBC2_1CR Bc2 1c E strength Bc2 1c model uncertainty factor ms MBC2 IDR Bc2 1d loading Bc2 1d model uncertainty factor mR MBC2 1CS Bc2 1c i strength Bc2 1d model uncertainty factor ms MBC2 IDS f Bc2 1d e loading Bc2 1g model uncertainty factor mR MBC2_1GR Bc2 1g B strength Bc2 1g model uncertainty factor ms MBC2_1GS Bc2 1g loading Bc2 1h model uncertainty factor mR MBC2 IHR Bc2 1h strength Bc2 1h model uncertainty factor ms MBC2 1HS Bc2 1h loading T07_08 2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to5 mn m e oy fLOUPSHc Contract No GOCE CT 2004 505 20 Distribution Distribution SET Distribution Example aramete
50. 2 Standard Deviation Parameter Description LSE mapping Unique fortran name arameter 3 distribution Variation P name coefficient The level of the longest sheet L1 SPileLong m Cc1 2aii Cc1 2b normal 1 0 1 pile cut off L1 SPileToeLev The toe level of the sheet pile m Cb1 2a Cb1 2c Cb1 2d normal 1 0 1 The level of the shortest sheet L3 SPileShort m Cc1 2aii Cc1 2b normal 1 0 1 pile cut off LKh HorSeepageLength horizontal seepage length m Bal 5ai Bal 5aii Cc1 5 normal 1 0 1 LKv Lkv Vertical seepage length m Bal 5aii Ccl 5 normal 1 0 1 Ls ShipL Length of the ship m Bc2 1b Bc2 1d it T Partial length of the dike at the Aa2 4 Ba2 4d Ba2 3 inner toe Bal 5bii Bal Sdii Aa2 1b Bc2 1b Bc2 1j normal 0 15 m MOutS Mean outer slope Bc2 3 Ba2 4d Bc2 3a m normal 1 0 05 Ba2 1bii Ba2 1biii Ratio of tangential force to m TangF Da4 2b normal force in the contact area M MIceF Mass of the ice feature kg Da4 2b Aa2 4 Ba2 4d m0 m0Flow m0 coefficient E bs Bal 5bii Bal Sdii mf Mf Mass of the fluid displaced by kg Da4 3 TO7_08 CX Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505 20 Distribution Distribution enn ul a5 Distribution arameter arameter Parameter Unique fortran name Description LSE mapping eines
51. 20007277 13200707 TA ppp ps AAT 136 Cz Dp1h DOCz Dp1D Tege Dp11 ee Lp ls ine i 25 13S9a Dpi7 1 JH Dp2 1 ib Jp 13 fa D 24 153a Dpr7 322a Dp74 TS TD py 1276 Ape 123 DpJ amp 122 Dp165 121a Dp3 15 3 DpB8 ATa D nz 1825 i3p 1n 103 271 Ja Dp 330 Da35sz Dp 3T5 Oagh Lindi OG7 Omen OSM psa e85s Dp20 C833a Dp185 OSia Lind gt Gz8 DnTasR 76 Da 24 Dz4 Dann wf 2 LipEd Seepage path lengths for all sections first column given by Water board 3 column and based on the dike geometry 4 column T07_08_02_Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505420 5 Onsite APPENDIX II B ADAPTATIONS BY VNK VNK has performed all calculations again based on the database of DHV In the scheme below it is indicated per section what has been altered relative to the database that was supplied by DHV Section Actions notes related to database after delivery by DHV All In the calculations of DHV many wind directions have been turned off because these would not be relevant for the mechanism overtopping wave overrun this would be valid for offshore wind Offshore wind can be neglected in the river area Bretschneider 1s used in that case Along the coast the boundary conditions are determined using SWAN in which also heave and diffraction etc are present Herewith
52. 2004 505420 FLOOESive CONTENTS Document Information Document History Disclaimer Contents i I APPENDIX 1 DETAILS OF THE PRA THAMES 1 II APPENDIX 2 DETAILS OF THE PRA SCHELDT 12 III APPENDIX 3 DETAILS OF THE PRA GERMAN BIGHT 60 IV APPENDIX 4 UNCERTAINTY DATABASE 66 V APPENDIX 5 USER MANUAL RELIABILITY TOOL 84 T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1 to 5 Een cis Contract No GOCE CT 2004 505420 FLOOPSHe T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1 to 5 FLOOR Contract No GOCE CT 2004 505420 l Appendix 1 Details of the PRA Thames This appendix describes the reliability analysis applied to the Dartford Creek to Gravesend flood defence system in more detail Section 1 provides a site description which includes a definition of the floodplain boundaries and main structure types Section 2 discusses the failure mechanisms and fault trees of the structure types in more detail To enable probabilistic calculations the flood defence line is discretised into sections which are each over the whole length characterized by one cross section The discretisation into flood defence sections and the probabilistic calculations 1s described in section 3 The results of the probabilistic calculations are discussed in section 4 1 Site description The Dartford Creek to Gravesend flood defence line protects one f
53. 7 1 Appendix 10 April 2008 61 Task 7 Deliverable D7 1 Appendices 1to 5 mn i Contract No GOCE CT 2004 505420 muGnsife 12 0 4 Dunes up to 18m 10 0 o t Overf low dike 6 0 Height mNN 4 0 i i Design flood level iO Lp pee samine 1 i Dike height 0 0 9e e e e e e 9 127 0 129 0 131 0 133 0 135 0 137 0 139 0 141 0 143 0 Kilometrierung km Figure Error No text of specified style in document 13 Height of costal defence structures at pilot site German Bight Risk sources at the German Bight are resulting from storm surges in the North Sea associated with high water levels and storm waves at the flood defences Typically storm surges last not longer than 12 to 24 hours but may increase the water level considerably up to 3 5 m in the North Sea The interaction of normal tides water level differences in the range of 1 2 m are normal in the North Sea region storm surges and waves is crucial for the determination of the water level at the coast In addition the foreshore topography plays a major role when determining the waves at the flood defence structure In case of the German Bight the limited water depths over a high foreland will cause the waves to break and will therefore limit the maximum wave heights which reach the flood defence structures However the PRA has only considered single probability distributions for each of the governing variables such as water leve
54. 7009 020 dp16 7111 122 dp16 7116 121a dp9 7124 113 dp87 7167 097 dp290 7211 083a dp186 and 7233 078 dp148 the water board separately indicated that these score well for height in the preliminary results of the 2005 testing A number of these sections scored unsatisfactory in the 2000 testing see table 2 2 The section 7152 100a dp330 scored unsatisfactory in the 2000 testing but is strong according to the VNK calculations If this section still appears to be unsatisfactory in the new testing the result of VNK will have to be examined further Bursting and piping The results of VNK do not indicate weak spots for the mechanism bursting piping A number of sections scored unsatisfactory with the first testing No improvement works related to the phenomenon bursting piping have been executed since Works have been executed to drainage and better soil research has been done For now a number of sections do not yet score satisfactory for this mechanism with the second testing For the sections 7109 123 dp26 7111 122 dp16 7220 081a dp175 the water board has separately indicated that they score well for the mechanism bursting and piping in the preliminary results of the 2005 testing The section 7223 078 dp148 scored unsatisfactory in the 2000 testing Both sections are strong according to the VNK calculations If it appears from the final results of the new testing that these sections still score unsatisfactory
55. B Table A 5 Variables for erosion revetment Steenbergen and Vrouwenvelder 2003B Variable nr symbol description 62 Lk Width covering clay layer 63 Lak Width dike core at crest height 65 tan ou Angle outer slope 66 tan ai Angle inner slope 70 CRK Coefficient erosion resistance covering layer 71 CRB Coefficient erosion resistance dike core 85 Oz Acceleration factor erosion rate 86 Oh Declination erosion speed 83 B Angle in reduction factor r Grass 61 dw Root depth grass 69 Cg Coefficient erosion resistance grass Stone pitching directly on clay 64 D Stone pitching thickness 67 A Relative density stone pitching 68 Ck Coefficient stone pitching on clay Stone pitching with granular filter 64 D Stone pitching thickness 67 A Relative density stone pitching 72 d Thickness granular filter layer 73 Dns Grain size 15 percentile filer 74 S Crack width 75 Ct Coefficient stone pitching on filter 76 Ca Coefficient in determination leakage length 77 Cb Coefficient in determination leakage length 78 Ct Coefficient in determination leakage length 84 Cot Coefficient strength stone pitching 87 C Coefficient Asphalt revetment 79 D Thickness asphaltic concrete 80 A Relative density asphaltic concrete 81 fucws Factor for normative water level 82 hews Level average discharge 88 hio Height fictive bottom 89 b Parameter 90 Dnso Nominal a
56. Bal 5d Bc2 1d Bc2 1k d LayerThick m lognormal 1 element Bc2 3a Bc2 3b 0 3 D BaskThick Basket or mattress thickness m Bc2 1m Bal 5c Aa2 1b D50 D50 Sieve diameter diameter of stone e loses which exceeds the 5096 value of Ab2 1a Ba2 lbii TO7_08 CX Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505 20 Distribution Distribution enn ul a5 Distribution arameter arameter Parameter Unique fortran name Description LSE mapping eines E NS p parameter 3 distribution Variation Standard name coefficient Deviation sieve curve Ba2 1biii d70 D70 70 pass grain diameter Bal 5ai lognormal 1 0 15 D85 D85 85 value of sieve curve Bal 5c 15 non exceedance diameter of filter layer from grading curve Df15 Df15 jE a Bal 5c lognormal 1 0 02 indicating permeability of the filter 0 1 Thickness of remaining clay Bcl 1 Ba2 1a dk Dk layer Ba2 1b Bc2 1h General erosion long term dgen Dgen degradation of the bed level Nominal mean diameter Dn50 Dnl5 Dn15 Bcl 5 normal 3 M15 rr1 3 0 25 Bc1 5 Aa2 4 Ab2 1b Bal 1 Bal 5bii Dn50 Dn50 Nominal mean diameter Dn50 Bal 5dii Ba2 4c omit M50 r1 3 Ba2 4iii Ba2 5 Bc2 1a Bc2 1c Bc2 1g Bc2 1m Bc3 1b Bc3 1c TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 App
57. Bc2 1h Bc2 1m Bc2 3b Ca2 1a Ca2 1b Ca2 2aCa2 2b Ca2 3 Cc2 2a Cc2 2b Da2 5 hb Hb5Hd Water depth at 5Hd dinstance of Ca2 2a T07_08 2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505 20 Parameter Unique fortran name Description the wall LSE mapping Example distribution Distribution Distribution Mop Distribution parameter 1 parameter 2 Standard Deviation PRA parameter 3 Variation name coefficient Width of the foundation of the Ca2 2a Ca2 2b Cc2 2a Bstruc WidthFoun normal 2d 0 004 structure Cel 2aii Ccl 2b lg Lr Effective width Ba2 3 V VolStruc Volume of the structure Ca2 2a Ca2 2b Cc2 2a Aa2 4 Ab2 1b Bal 1 Bal 5bii Bal 5dii Ba2 1a Ba2 1b Ba2 3 Ba2 1bii Ba2 1biii Level of the foundation of the Ba2 3 Ba2 4iii Ba2 5 TI normal 1 0 1 WallToe structure Bc2 1a Bc2 1b Bc2 1c Bc2 1g Bc2 1h Bc2 1j Bc2 1m Bc2 3a Bc2 3b Ca2 1a Ca2 2a Ca2 2b Ca2 3 Cc2 2a Cc2 2b Da2 5 coefficient which takes account a LongTAlp of the long term effects on the Ca2 3 compressive strength and of the TO7_08 C2 Reliability Analysis D7 1 Appendix JO April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 muousife Contract No GOCE CT 2004 505 20 Distribution Distribution Pune ul 5 Distribution NS
58. Bw StrucWidthToeLevel Structure width often at toe level m Bal 6 Da2 5 C AMassCoeff Added mass coefficient Da4 3 C Chezy Chezy coefficient m s Ba2 4b Bc3 1c A22 4 Bal 1 Bal 5bii C WaveProp Propagation celerity of waves m s Bal 5dii Ba2 5 normal 2b 0 2 0 5 C c DrainedSoilCohesion Drained cohesion of soil N m lognormal 1 Bal 4 lognormal 2c 0 3 TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505 20 Distribution Distribution EN Distribution Example parameter 1 parameter 2 Standard Deviation Parameter Description LSE mapping Unique fortran name OSA CEU parameter 3 distribution Variation i name coefficient Grass quality after Verheij et al cE GrassQual m s Ba2 1a Ba2 1biii 1998 Cf SBedFri Friction of sand bed Aa2 4 Ba2 4d Bal 5aii Ccl 5 Ck CreepCoeff Creep coefficient 5 normal 1 0 1 Bal 5bii Bal 5dii Factor representing the erosion cRK CRK ME Ba2 1b Bc2 1h lognormal 1 sensitivity of the clay cover 0 3 cT TurbCoeff turbulence coefficient Bc3 1b Bc3 1d cV DissCoeff Dissipation coefficient Aal 1 Bal 1 Ba2 5 cw CW Cohesion due to root penetration kPa Ba2 3 d FriCoeff Friction coefficient Cc1 2aii d PileDia Diameter of the pile m Da4 2c d Depth Depth m Ba2 4b normal 0 1 Thickness of certain layer
59. Distribution parameter 1 parameter 2 Standard Deviation R parameter 3 Variation d name coefficient Factor for mean force due to fA Fa Ba2 3 wave impact Cubic tensile strength of the fb ConcTensStrength kN m Ccl 2d lognormal 2d concrete 0 2 fc CurFriFact Friction factor for current Ba2 41 Quality factor for grass F fg GrassRevQ Aal 1 Bal 1 Ba2 41 normal 0 2 revetment i Factor for force due to mass of fG SoilMassF Ba2 3 soil Peak frequency of wave fp Fp s 1 Ba2 3 spectrum fpmax Fpmax Factor for pmax Ba2 3 Yield stress of the steel net of Cb1 2a Cb1 2 fs YieldStress kN m lognormal 2d any factoring Cc1 2c Cc2 2b Ca2 3 0 1 Aal 1 Bal 1 Bel 1 Gravitational acceleration 9 81 MM g Grav m s Cel 2aii Cc1 2b deterministic m s2 Cc1 2c Aa2 1b Ab2 1a TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 FLkU0Rns f2 Contract No GOCE CT 2004 505420 Distribution Distribution K Distribution Example arameter 1 parameter 2 Parameter Unique fortran name Description LSE mapping E p parameter 3 distribution Variation Standard Us name coefficient Deviation Ab2 1b Bc2 1c Bc2 1d Bc2 1h Bc2 1m Ca2 2a Cc2 2a Cc2 2b Ba2 3 Bc2 3b Ca2 3 Aa2 4 Ba2 4i Ba2 4111 Ba2 4b Ba2 4c Ba2 4d Ba2 5 Da2 5 Bc3 1b Bc3 1c Bc3 1d Ba2 1a Ba2 1b Bc2 1a
60. E CT 2004 505420 r eiv 249 109 38 24 25 42 74 116 139 152 New PCR file produced gt results VNK gt let everything be calculated by FORM DS with 1 foreland point If no good result was obtained one looked whether this could be solved by adapting various things Result seems to be caused by a strong covering With the DS calculation the number of 5000 samples is too little with the higher beta values These calculations thus have to be performed again with 100000 samples Sections will not suddenly turn out weak No result for overtopping wave overrun initial value altered from 8 to 1 and calculated with 6DS FI instead of SFORM DS No data input on stone covering thus no result Level of GWL was set as 5 35 This should be 0 is sea Has been adapted Factor fmGWS was set as 0 15 River has been adapted into 0 25 Sea Covering at wind direction 330 the residual strength crashed 330 is governing wind direction for wave overrun can not be turned off Start method has been adapted from 8 gt 1 Ov ov keeps functioning Covering crashes on the residual strength with the Combin calculation Start method adapted from 8 1 Ov ov keeps functioning Covering crashes at residual strength Start method has been adapted from 8 1 Result for covering but now ov ov does not function well Southern wind direction 180 degrees turned off Now result for both mechanisms Gets stuck on residual strength
61. E NS p parameter 3 distribution Variation Standard name coefficient Deviation the object ml MI Mass of the storm debris kg Da4 3 Aa2 1b Ba2 4i n MInSlope Mean inner slope 0 i s normal 0 05 Ba2 1bii Ba2 1biii nf D15f D50b where nf N N f Bc1 5 porosity of filter material Number of waves over the Ab2 1b Bc2 1c N NbWaveStorm duration Tr of a storm record or Ba2 4c test N Tr Tm Number of displaced units per Bc2 la Nod Nod width Dn across armour face Ca2 1b p ICP Effective ice crushing pressure kN n Da42a Net uniformly distributed pressure acting on the member in the case of the front wall p 1s the p P Mpa Ca23 arithmetic sum of the applied wave loading and the internal cell pressure Particle size or typical Aa2 4 Bal 1 Ba2 5 dimension Bal 5bii Bal 5dii TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505 20 Distribution Distribution Distribution x Example parameter 1 parameter 2 Parameter Unique fortran name Description LSE mapping UPS PED parameter 3 distribution Variation Standard name coefficient Deviation PartSize ic i Aa2 4 Ba2 4d ENN istic p ba cone Bal 5bii Bal Sdii ee Permeability parameter P Pparam Ab2 1b Bc2 1c 0 1 P 0 6 q MORate Mean overtopping rate l m Ba2 4d Mean overtoppin
62. HV results These results provide an indication of the probabilities of failure to be expected Before the results are incorporated in the calculation of the probability of failure for dike ring 32 it should be checked whether coupling of the dike sections from PC Ring to representative profiles with another MHW is possible 1 1 8 Feedback results per section to water board The results of the calculations per dike section are discussed with the water board An overview of its findings per mechanism is given below The results are compared with the results of the testing in 2000 table 2 2 and the preliminary results of the 2005 testing as far as these are available As a result of this it is concluded that a number of results should left out of consideration for the time being these results are shaded grey in table 4 1 Overtopping and wave overrun Dike section 7167 097 dp290 Molenpolder has a relative bad score for the mechanism overtopping wave overrun beta is 3 03 This result is not recognisable for the water board Possibly the sandbank ahead is not schematised correctly this is no foreland due to which too little wave reduction is accounted for Other cause could be the calculated profile A further analysis of required here The water board thinks the present result should not be considered in the calculations of the probability of flooding of the dike ring because it doesn t recognise the results For sections
63. Integrated Flood Risk Analysis and Management Methodologies Reliability Analysis of Flood Sea Defence Structures and Systems APPENDICES 1 TO 5 Date April 2008 Report Number T07 08 02 Revision Number 1_2 P0 Deliverable Number D7 1 Due date for deliverable February 2008 Actual submission date April 2008 Task Leader 12 FLOODsite is co funded by the European Community Sixth Framework Programme for European Research and Technological Development 2002 2006 FLOODsite is an Integrated Project in the Global Change and Eco systems Sub Priority Start date March 2004 duration 5 Years Document Dissemination Level PU Public PP Restricted to other programme participants including the Commission Services RE Restricted to a group specified by the consortium including the Commission Services co Confidential only for members of the consortium including the Commission Services Co ordinator HR Wallingford UK EF Project Contract No GOCE CT 2004 505420 SIXTH FRAMEWORK Project website www floodsite net HevsELU UE DOCUMENT INFORMATION Reliability Analysis of Flood Sea Defence Structures and Systems Title Appendices 1 to 5 Lead Author Pieter van Gelder TUD Foekje Buijs Cong Mai Van Wouter ter Horst Wim Kanning Mohammad Nejad Sayan Gupta Reza Shams Noel van Erp Contributors HRW Ben Gouldby Greer Kingston Paul Sayers Martin Wills LWI Andreas Kortenhaus Hans J rg Lambrecht Distribution
64. MBA2 AIIIR Ba2 Aiii strength Ba2 4iii model uncertainty factor ms MBA2 AIIIS Ba2 Aiii loading Ba2 5 model uncertainty factor mR MBA2 5R Ba2 12 i strength Ba2 5 model uncertainty factor ms MBA2 5S Ba2 13 B loading Ba3 1 model uncertainty factor mR MBA3 IR Ba3 1 i strength Ba3 1 model uncertainty factor ms MBA3 IS i Ba3 1 D loading Bb1 2 model uncertainty factor mR MBBI 2R Bb1 2 B strength Bb1 2 model uncertainty factor ms MBBI 2S Bb1 2 B loading TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 muousife Contract No GOCE CT 2004 505 20 Distribution Distribution SET Distribution Example parameter 1 parameter 2 Parameter Unique fortran name Description LSE mapping ERU A PRA parameter 3 distribution Variation Standard name coefficient Deviation Bb1 4 model uncertainty factor mR MBb1 4R Bb1 4 E strength Bb1 4 model uncertainty factor ms MBbI 4S Bb1 4 E loading Bc1 1 model uncertainty factor mR MBCI IR Bcl 1 B strength Bcl 1 model uncertainty factor ms MBCI 1S Bcl 1 loading Bcl 4 model uncertainty factor mR MBCI 4R Bcl 4 us strength Bc1 4 model uncertainty factor ms MBCI 4S f Bcl 4 5 loading Bc1 5 model uncertainty factor mR MBCI 5R Bcl 5 B strength Bc1 5 model uncertainty factor ms MBCI 5S Bcl 5 loading Bc2
65. Public Document Reference T 0 7 08 01 Appendix DOCUMENT HISTORY Date Revision Prepared by Organisation Approved by Notes 01 01 08 1 1 p12 P van Gelder TUD 01 04 08 1 1 p12 C Mai Van TUD 09 04 08 1 1p12 P van Gelder TUD Corrupted Word version replaced 10 04 08 1 2P01 Paul Samuels HR Wallingford Formatted as a deliverable ACKNOWLEDGEMENT The work described in this publication was supported by the European Community s Sixth Framework Programme through the grant to the budget of the Integrated Project FLOODsite Contract GOCE CT 2004 505420 DISCLAIMER This document reflects only the authors views and not those of the European Community This work may rely on data from sources external to the FLOODsite project Consortium Members of the Consortium do not accept liability for loss or damage suffered by any third party as a result of errors or inaccuracies in such data The information in this document is provided as is and no guarantee or warranty is given that the information is fit for any particular purpose The user thereof uses the information at its sole risk and neither the European Community nor any member of the FLOODsite Consortium is liable for any use that may be made of the information FLOODsite Consortium T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1 to 5 E ec cis Contract No GOCE CT
66. Select Medium or Low but this is not recommended Now choose Tools Macro Visual Basic Editor In the Project Explorer window click on the line VBAProject ReliabilityCalc xls or any of the Microsoft Excel Objects below From the Tools menu on the Visual Basic Editor window choose References Click Browse find your RelCalcFolder choose RelCalc tlb and click Open You should now see RelCalc with a tick under Available References From the File menu choose Close and return to Microsoft Excel Save the spreadsheet and your Reliability Calculator is available for use Name Description UserGuide doc This document Reliability Calc xls EXCEL spreadsheet which forms the user interface RelCalc dll Reliability Calculator engine used by the Reliability Calculator spreadsheet RelCalc dll Type library which defines the COM interface provided by LSESupport dll DLL which includes the IndexOf function used by the LSE functions LSESupport lib Library file used when building the LSE function DLL Task7 LSEs dll DLL built using latest release of LSE functions from TU Delft Also includes interim dummy LSE functions for Bb1 3a and Bb1 3b for use with the current Sheet Pile Wall fault tree These interim functions always return zero i e no fail StatFunc dll A DLL containing functions for generating random numbers according to specified distributions FailureMode csv A CSV file defining the Failure Modes and their
67. a as well as the river discharges from the Scheldt There are four surrounding dike ring areas along the Western Scheldt no 29 to 32 T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 13 Task 7 Deliverable D7 1 Appendices 1to 5 7 Contract No GOCE CT 2004 505420 fkOEsc Figure 1 Dike ring areas in the southern part of the province Zeeland along the estuary Western Scheldt no 29 Walcheren no 30 Zuid Beveland West no 31 Zuid Beveland Oost no 32 Zeeuwsch Vlaanderen The water board Zeeuwse Eilanden http www wze nl has provided the problem identification and data with respect to problematic dike sections along the western Scheldt The study of VNK Ministry of Water Management will serve as a basis for further investigations of this test pilot site T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 14 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505420 5 Onsite II 2 PILOT SITE SCHELDT This section provides a description of dike ring area 32 Zeeuws Vlaanderen and the schematizations of the various dike sections The assessment of the water board is given in this section as well Section 2 1 provides general information concerning the location and the characteristics of the dike ring followed by an overview of the dikes and structures in section 2 2 Sections 2 3 to 2 8 take a closer look at the schematization of the dikes and dunes Section 2 9 fi
68. a2 1a Ba2 1bii Ba2 3 Ba2 lbiii Ba2 4b Hs WaveHeight Significant wave height Ba2 4c Ba2 4d Ba2 4i Ba2 4i11 Ba2 5 Bc2 1a Bc2 1b Bc2 1c Bc2 1g Bc2 1h Bc2 1m Bc2 3b Bc3 1a Ca2 1b Ca2 2a Ca2 2b Ca2 3 Cc1 2c Cc2 2a Cc2 2b Da2 5 TO7_08 C2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505 20 Distribution Distribution eee Distribution Example parameter 1 parameter 2 Standard Deviation Parameter Unique fortran name Description LSE mapping parameter 3 distribution Variation name coefficient ht DStrToe Water depth at structure toe m Bal 5b hwlr Hwilr Allowable water level rise m Bal 6 Da2 5 1 HydG hydraulic gradient Bal 4 Bb1 4 Bc3 1c E Effective contact stiffness of the k CollisionContact NS kg s2 Da4 3 collision normal lognor Permeability coefficient mg clay 1 0E 08 0 2 k Darcy m s according to Darcy lognormal 2b sand 1 0E 05 1 6 Bal 5ai lognormal 2b 0 5 o normal lognor k CorePerm Permeability of core material m s Bal 5b mal 0 2 normal 0 1 k RouF Roughness factor by Strickler m Aal 1 Bal 1 0 015 lognormal 0 25 Empirical factor e g k 1 0 Aa2 1b Ba2 1bii k EmpF 2x see Sch ttrumpf 2001 Ba2 1biii Cb1 2a Cb1 2c Cb1 2d Coefficient for active horizontal Ka ActGrainFCoeff lognormal 4 0 1 grain force Cel 2aii Ccl 2b
69. ailure mechanisms refer to the Task 4 Floodsite report on flood defence failure mechanisms gamma sl gamma s2 gamma s3 gamma s4 gamma s5 gamma s Figure 7 Example of a sheet pile wall along the Dartford Creek to Gravesend defence line Table 3 Site specific failure processes and failure mechanisms implemented in the reliability analysis of anchored sheet pile walls T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1 to 5 FLOOR Contract No GOCE CT 2004 505420 Structural failure Breaking sheet Instability due to pile wall anchor failure bending moments L l kaa cana sheet pile toe Failure of anchor Slip failure of anchor Anchor breaking Figure 8 Simplified fault tree for anchored sheet pile wall as applied in reliability analysis Site specific failure processes Failure mechanisms implemented in the reliability analysis Accelerated Low Water Corrosion in e Breaking of the ground anchor Cb1 2a the splash zone Sliding of the ground anchor due to insufficient shear strength of the soil not Corrosion of the ground anchors included in Task 4 report Breaking of the sheet pile cross section Cbl 2c Rotational failure of the sheet pile after failure of the ground anchor Cb1 2d 3 Discretisation and probab
70. and wave run up the third row shows the STPI score which represents the score for bursting and piping and the fourth row the STBI score which represents the score for stability of the inner slope Suf stands for sufficient and Insuf for insufficient Dunes Recent research established that one has to reckon with heavier wave action than was assumed so far along the Dutch coast This could imply that embankments of Zeeuws Vlaanderen no longer comply with the legal requirements The calculated weak spots based on the given boundary conditions T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 21 Task 7 Deliverable D7 1 Appendices 1to 5 mn Contract No GOCE CT 2004 505420 FLOOESH provide a true representation of the locations with the greatest strength deficiencies These are determined by the water board and the assessment of the water board based on unambiguity in boundary condition sections and the shape of the coastal sections leads to the following strength deficiencies see figure 2 3 The dune area of Cadzand west of the outlet with the adjoining sea dike of the Kievitspolder East coastal length 940m test crown height deficiency 2 00m Figure 2 4 top left The sea dike of the Jong Breskenpolder between Nieuwe Sluis and the lighthouse coastal length 1060m test crown height deficiency 0 50 to 1 00m Figure 2 4 top right The addition to the artificial dune in Breskens at the Veerhaven coastal le
71. ase the extreme levels of the sea and river discharge and the time of their occurrence implies the necessity for using probabilistic methods for the analysis Wo 7 s discahrge DIS Pak a Sealevel Tidal Reach Dike under influence of both River Discharge and sea Figure 3 11 Dike on tidal reach of a river subjected to both discharge and sea level variations Use of Monte Carlo simulations for reliability analysis lead to accurate estimates of the failure probabilities Here the basic steps involved are 1 digital generation of an ensemble of loading conditions that obey specified probabilistic laws 11 treatment of each realisation of the problem using deterministic procedures and iii statistical processing of the ensemble of sample solutions for the problem leading to estimates of the failure probability Thus in principle the method is applicable to any problem where it is possible to digitally generate an ensemble of loading T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 24 Task 7 Deliverable D7 1 Appendices 1to 5 PSAN cft Contract No GOCE CT 2004 505420 FLOODS ite conditions and deterministic solution methods for a sample problem are available The method however can be computationally intensive For the river dike problem considered in this study the water levels along the dike segment are computed using a hydrodynamic model This requires nontrivial computational effort In M
72. ases the yielding is described by the failure mechanism wave overtopping Covering damage and erosion body of the dike With this failure mechanism the dike fails because the covering is damaged by wave action first after which the cross section of the dike core is diminished by erosion Bursting piping With this failure mechanism the dike fails because the sand is washed away from underneath the dike The sealing layer if present will first burst due to the pressure of the water Consequently so called pipes can occur causing the sand to be washed away and the dike to collapse Sliding inner slope Whth this failure mechanism the dike fails because a part of the dike becomes unstable as a result of high water levels for a long period of time and consequently slides T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 33 Task 7 Deliverable D7 1 Appendices 1to 5 Pann eere Contract No GOCE CT 2004 505420 fkOOP sic The possible failure mechanisms liquid settlement buoyancy sliding of the foreland sliding of the outer slope micro instability and weakening are not taken into account because these failure mechanisms do not directly result in flooding An assessment model is used per failure mechanism in order to be able to compare loads and strengths or otherwise to be able to calculate the probability of failure for the failure mechanism in question 1 1 3Failure mechanisms structures For determining the pr
73. ated to prediction models and limit state equations by means of a detailed top down analysis iii the uncertainties which are worth reducing by the generation of new knowledge iv the priorities with respect to the allocation of research efforts for the various topics to be addressed in the other sub projects v the areas of high low and medium uncertainty There is potential for significant differences in the PRA approach between the 3 pilot studies TUD HR LWI need to review before any work starts to ensure that at minimum there is a common understanding of each PRA approach and at best that a common approach is adopted for all three The preliminary analysis in this report will assess the probabilities of flooding and related uncertainties in the south western province of the Netherlands Dike ring area 32 will be examined to see how reliable the flood defences are and to identify any weak points In particular attention will be paid to the special elements in the dike rings hydraulic structures such as locks weirs and pumping stations To date little is known about the safety of these elements Existing techniques among others the PC Ring approach will be applied in first instance Refined techniques will be proposed in case the resulting failure probability from PC Ring is too inaccurate The Western Scheldt forms the entrance to the harbour of Antwerp Belgium Water levels are influenced by the wind surges on the North Se
74. ater kN m3 normal 1 0 1 Cc1 2b Cc1 2d Ba2 4b Bal 5aiii Ba2 3 lognormal 1 0 3 Dimp Dimp thickness impermeable layers Bal 5aiii normal 2b 0 2 h Height Height of a element segment m Cc1 2e Cc2 2a Cc2 2b TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 muousife Contract No GOCE CT 2004 505 20 Distribution Distribution K Toaniple ird as Distribution arameter arameter Parameter Unique fortran name Description LSE mapping xu E PRA p parameter 3 distribution Variation Standard name coefficient Deviation h IceThick Ice thickness m Da4 2a Da4 2c Aal 1 Aa2 1a Aa2 4 Ab2 1a Ab2 1b Bal 1 Bal 5ai Bal 5aii Bal 5aiii Bal 5b Bal 5bii Bal 5dii Bal 6 Ba2 1a Ba2 1b Ba2 1bii Ba2 1biii h WaterL River water level m Pea A deterministic Ba2 4i Ba2 Aiii Ba2 5 Bc2 1a Bc2 1b Bc2 1c Bc2 1g Bc2 1h Bc2 1j Bc2 1m Bc2 3a Bc2 3b Ca2 1a Ca2 2a Ca2 2b Ca2 3 Cb1 2a Cb1 2c Cb1 2d Cc1 2ai1 Da2 5 Cc1 2b Cc1 2c Cc1 2d T07_08 2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 muousife Contract No GOCE CT 2004 505 20 Distribution Distribution K Distribution Example arameter 1 parameter 2 Parameter Unique fortran name Description LSE mapping E p parameter 3 distribution Variation Standard Us name coefficient Deviation
75. bility problems are a real threat in this case because the dikes are high and steep and stand on weak layers in the sub soil Because of the reasons a probability of flooding of 1 100 for dike ring area 32 is presented in the main report and the management summary of the project VNK Herewith it is indicated that the probability of flooding is mainly determined by stability problems at the pumping station or at the dikes In relation to the pumping station it is consequently also indicated that this can be approved based on recent information with the second testing Possibilities of sensitivity analyses For dike ring area 32 no sensitivity analyses have yet been performed In the section discusses in which way it can be determined which sensitivity analyses can be of interest The calculated probability of flooding of the dike ting is determined by a large number of dike sections dune sections and structures various failure mechanisms and a large number of stochastic variables per failure mechanism The possible number of sensitivity analyses is in that way endless It is therefore important to focus the sensitivity analyses on those factors that determine the level of probability of flooding most For the dike sections it concerns the relatively weak dike sections For those dike sections the attention is consequently given to the failure mechanisms that contribute to the probability of flooding most And for those failure mechanisms the
76. cies of limit state equations have been considered Figure 3 8 shows a simplified version of the fault tree used for one of the sections at German Bight Coast for a typical sea dike Most of the required input parameters for the failure modes are of stochastic nature which means that not only mean or design parameters but also a statistical distribution of this parameter describing the uncertainty is provided The result of this analysis is an annual probability of flooding of the hinterland for each dike section which has been selected These flooding probabilities were typically found to range from a probability of 10 to 10 which means a return period of flooding in the range of 10 000 or 1 000 000 years The overall flooding probability using a fault tree approach for all sections results in P 4 10 Flooding of hinterland 5 4105 A Non structural failure 5 1 105 Overflow Wave Overtopping 1 1 105 4 0 105 Failure dike top 1 0 1010 Sliding 3 0 10 6 Figure Error No text of specified style in document 14 Typical fault tree for a dike section at German Bight Coast The following lessons have been learned from performing this study for the German Bight Coast pilot site e The given results should only be used carefully since results depend on variations of parameter settings which still have to be performed e A limit state equation for dunes is still missing and needs to be implemented e The wi
77. de foreland in the German Bight Coast will induce heavy wave breaking under design conditions and also for lower water levels of course Results might therefore be dependent on morphodynamic processes and changes of these forelands Breaker criteria should always be used when waves approaching the structure e Updated and harmonised limit state equations are needed to compare reliability calculations of pilot sites to each other e A wide range of input parameters are not directly available and had to be estimated Therefore sensitivity analyses of the influences of parameters have to be performed T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 63 Task 7 Deliverable D7 1 Appendices 1to 5 mn Contract No GOCE CT 2004 505420 FLOODS do e Criteria for splitting the defence line into various sections need to be automatically derived in the model Up to now this is done semi automatic with some manual checks of the section at the end Any change in key parameters of a dike section is therefore not directly leading to a re calculation of the distinction of all the sections e Distinction between different sections was based on the assumption that the sections can be treated independently when calculating the overall failure probability of the system This still needs verification or improved methods considering the length effect between sections e Dependencies between failure mechanisms or scenarios have not been
78. e buoyant density of Aa2 4 Ab2 1b Bal 1 A BuDen material i e for rock A pr pw Bal 5bii Bal 5dii lognormal 1 Ba2 Aiii Ba2 4c 0 02 TUZ OB Q2 Reliability Analysis DZ 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505 20 Distribution Distribution Distribution MN t Example parameter 1 parameter 2 Parameter Unique fortran name Description LSE mapping Nut tae MA parameter 3 distribution Variation Standard name coefficient Deviation Bc2 1c Ba2 4d Bc2 1g Bc1 1 Bc2 1a Bc2 1d A CoD Relative density of cover layer A i T oDen ognorma pr pw pw Bc2 1m Bc3 1b 8 Bc3 1c Bc3 1d 0 02 Coefficient of proportionality Be2 1b is ShipGeo et representing the ship s geometry Bc2 1d Drag force factor Constante of n WhiteConst Bal 5ai lognormal 1 0 15 White Rolling resistance angle of sand 0 Rolling i 7 Bal 5ai lognormal 1 3 grains RootAng Root angle of shear rotation s Ba2 4b A LeaLen Leakage length m Bcl 5 Modification factors depending Al A2 A3 Lambda on the geometry and the nature Ca2 2a of the wall u SlidF Sliding factor Ca2 2a Ca2 2b Cc2 2a p SandDensity Density of the sand kg m Bal 5ai normal 1 0 05 p Rho Volumetric weight of the soil kN m Bal 4 Bb1 4 normal 1 0 05 TO7_08 C2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7
79. e crown gradient of the inner slope MHW and thickness of the covering layer but doesn t account for the structure of the soil The coupling of the sections and the profiles does thus not match the routes for which the profiles are deemed to be representative according to the water board Num a p vcobpnmsB s 1 31 2 53 7 33 2 68 1 36 2 80 1 24 2 13 1 28 2 29 1 08 1 27 UE 155 131 2 433 1 347 2 53 0 93 0 21 0 96 0 42 0 94 0 29 1 17 1 874 1 197 1 95 1 237 2 18 Table 4 2 Comparison safety factors according to Bishop from MStab and MproStab results by VNK When considering this latter next to section 7249 076 dp124 DHV made the right coupling for sections 7109 123 dp26 7111 122 dp16 7116 121a dp9 7233 078 dp148 7258 074 dp99 and 7271 072 dp69 For the latter three sections the MHW almost matches with the MHW of the representative profile This is not the case for the first three The probabilities of failure that DHV calculated for these sections are provided in table 4 3 afschuiven 7109 123 Dp26 7111 122 Dp16 1 906 7258 074 Dp99 7271 072 Dp69 B ta afschuiven Faalkans T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 38 Task 7 Deliverable D7 1 Appendices 1to 5 mn Contract No GOCE CT 2004 505420 FLOODS ite Table 4 3 Reliability index Beta and the failure probability for the mechanism sliding D
80. e not yet available The deep water waves will in reality not reach the foot of the dune as a result of protection by the dams and possibly also because the coast is located in the shade of Walcheren orientation of the dune is northerly zero degrees Expectation is thus that the wave loads will be less than follows from the calculations Next to that stone covering is present at the foot of the dune All considered section 71 can not be taken into account properly in the calculations at this moment despite it being a weak section Section 1354 was chosen as well After consult with the water board this appeared to be a dike Dune section 1401 is put in on the original dike section 7008 Originally the x y coordinates of dike pole 020 dp15 were used for 7008 in the schematization of dike ring 32 in PC Ring This dike pole is located roughly 500 meters east of section 1401 Because of this one calculated with wave conditions specified for a location 500 meters away for section 1401 In connection with the schematization several adjustments to PC Ring have been carried out Two dike sections are converted to a dune by Putting in type 2 for DS Profile type is 7 River normal 999 All fetch sections switched on landside also Location codes are adjusted e Profile overwritten by the Jarkus profiles For the four dune sections the right location codes have been put in the table dike section with this also load model 10 has
81. e schematization is conservative At firs it was agreed upon to calculate 5 dune sections With this it was agreed upon to calculate the dune sections of 2004 These provide a conservative picture because the next 5 annual suppletion is carried out in 2005 The choice of the dune sections to be calculated was made based on the Base Coastline Report of the RIKZ The choice is based on a comparison of the base coastline BCL the coastline to be tested TCL and the trend the BCL has When a probability of failure is calculated that contributes a lot to the total probability of flooding of the dike ring possible nuances can be made based on information from the report on the base coastline The calculated profile namely provides a lower limit of the probability of failure of the dune in question dunes are calculated based on section measurements of a weak year For the location of the dune sections one considered the maps of the Base Coastline Report and maps of the water board with the location of the dikes and dunes on them The dike ring schematization in PC Ring was also considered where is a dune and where is a dike schematized Eventually considering the schematization present in PC Ring this lead to the choice of 4 dune sections to be calculated In consult with VNK one has chosen to convert 2 profiles TO07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 50 Task 7 Deliverable D7 1 Appendices 1to 5 mn Contract No
82. egistration failure alarm etc Failure of mobilization operating personnel is not present at the retaining structure in time Failure due to operating errors faulty or omitted acts Technical failure of the closing elements motion device fails etc Constructive failure T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 34 Task 7 Deliverable D7 1 Appendices 1to 5 mn ita Contract No GOCE CT 2004 505420 fkOOP sic With the failure mechanism constructive failure the structure fails as a result of loss of strength or stability of parts of the structure The assessment of the structure is based on a consideration of constructive strength and stability of the structure in relation to the loads when retaining high water For this assessment the following mechanisms are applicable Constructive failure of the retaining devices resulting from drop load Constructive failure of the concrete construction Constructive failure of the foundation Chance of loss of stability due to instability of the bottom protection Failure due to loss of stability as a result of a collision Failure due to general loss of stability Failure due to under or rear seepage piping Method of assessment Within the project VNK a method has been developed for several types of structures to calculate the probability of flooding for different failure mechanisms It concerns the following types of structures navigation locks disc
83. endices 1to 5 Contract No GOCE CT 2004 505 420 Distribution Distribution enii uri ps Distribution hee arameter arameter Parameter Unique fortran name Description LSE mapping AME E PR p parameter 3 distribution Variation Standard name coefficient Deviation Bc3 1d Ca2 1b The water level in front of or at do DO m Bcl 1 upstream of the dyke dr GapDepth Depth of gap m Ba2 3 dw GrassRootsDepth Depth of the grass roots m Ba2 3 lognormal 1 0 2 dw DuneWidth Width of the dune m Aa2 la Dead weight tonnage of the DWT DWT t Da4 1 vessel d zs Dzs Depth of slope affected by flow Bal 4 Bb1 4 e bu ConcStrain Ultimate strain of the concrete Cc1 2c Cc2 2b e pl ConcPlast Plasticity strain of the concrete Cc1 2c Cc2 2b es Es Fraction of air pore Bal 5dii Stability coefficient general f StabCoeff mainly dependent on structure Bc2 1d Bc2 1m type tana and friction Cubic pressure strength of the Cc1 2e Cc1 2d fb ConcreteStrength kN m lognormal 2d concrete Cc2 2b Ca2 3 0 15 f2 DecCoeff Coefficient for deceleration of E Aa2 4 Ba2 4d TU7 OB 2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505 420 Parameter Unique fortran name Description erosion process LSE mapping Bal Sbii Bal Sdii Example distribution Distribution Distribution K
84. ent rN A N Insufficient Insufficient Overall Overturning of infi i rotational Sliding of the Temforeement capacity to take the concrete Bue AT strength bending on shear force slip wall moments Not taken into account in reliability analysis Uplifting impermeable layers Figure 6 Simplified fault tree for reinforced concrete wall as applied in reliability analysis top 2 3 Anchored sheet pile walls The primary function of anchored sheet pile walls is a ground retaining frontage which was previously used as docks Sheet pile walls were refurbished as part of the Thames Estuary flood defence improvements in the 70s and 80s Figure 7 shows an example of an anchored sheet pile wall applied along the Dartford Creek to Swanscombe Marshes defence line In some cases old frontages in the form of for instance masonry walls are still present in the ground behind the current sheet pile walls the space in between the walls backfilled with concrete In other cases the old frontage was used to anchor the sheet pile walls or the rubble of the old frontage was used as backfill material The failure mechanisms are organized in a fault tree according to figure 8 Table 3 presents the site specific failure processes and the failure mechanisms taken into account in the Dartford Creek to Gravesend reliability analysis The f
85. er 2 Parameter Unique fortran name Description LSE mapping parameter 3 distribution Variation Standard S name coefficient Deviation Ba2 1b model uncertainty factor mR MBA2 IBR Ba2 1b strength Ba2 1b model uncertainty factor ms MBA2 IBS Ba2 1b E loading Ba2 3 model uncertainty factor mR MBA2 3R Ba2 3 E strength Ba2 3 model uncertainty factor ms MBA2 3S Ba2 3 loading Ba2 4b model uncertainty factor mR MBA2 4BR Ba2 4b ES strength Ba2 4b model uncertainty factor ms MBA2 4BS Ba2 4b E loading Ba2 4c model uncertainty factor mR MBA2 4CR Ba2 4c i strength Ba2 4c model uncertainty factor ms MBA2 4DR i Ba2 4d loading Ba2 4d model uncertainty factor mR MBA2 4CS Ba2 4c E strength Ba2 4d model uncertainty factor mS MBA2 4DS Ba2 4d E loading TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 mnm m ey Contract No GOCE CT 2004 505 20 Distribution Distribution Distribution d NS t Example parameter 1 parameter 2 Parameter Unique fortran name Description LSE mapping ANE toi e ND parameter 3 distribution Variation Standard name coefficient Deviation Ba2 41 model uncertainty factor mR MBA2 4IR Ba2 4i i strength Ba2 41 model uncertainty factor mS MBA2 4IS Ba2 4i loading Ba2 4iii model uncertainty factor VA mR
86. erosion rate 86 Oh Declination erosion speed 87 C Coefficient 88 hio Height fictive bottom 89 b Parameter 90 Dhns0 Nominal diameter 91 Va Revaluation factor 92 Psw Stability parameter No closure structure 101 A Cross section discharge 102 B Width structure 103 hok Water level in open condition 104 Akom surface retention area 105 hwv Level raise 106 m Discharge coefficient 107 Mkom Model factorVkom 108 Min Model factorVin 109 C Coefficient 110 Pas Probability of no closure Piping structures 111 Ly Vertical leakage length 112 Ln Horizontal leakage length 113 CL Lane s constant 114 m Model factor 115 Me Model factor Dunes 121 ha Dune height 122 Mp Model factor 123 dm Median grain size General T07_08_02_Reliability_Analysis_D7_1_Appendix 10 April 2008 74 oe eee eet HOGRSite Variable nr symbol description 131 Ad Error in determination ground level 132 MgH Model factor Bretschneider for Hs 133 Mgt Model factor Bretschneider for Ts 134 Aliioc Error in local water level 135 J Deviation wave direction 136 ts Storm duration 137 hy Inner water level Loads 140 UA Parameter magnitude discharge Lobith 141 Up Parameter slope discharge Lobith 142 u Parameter h North Sea 143 o Parameter h North Sea 144 y Parameter h North Sea 145 A Parameter wind 146 B Parameter wind 147 hum Water level Maasmond 148 v Wind speed
87. etas The influence coefficients for both the mechanism constructive failure and not closing of the closing elements were weighed based on probability contributions of the related mechanism and the residual strength The values used for the mechanism overtopping and wave overrun and the other mechanisms are based on the ISO norm The sto files serve as input for PC Ring which calculates the values of the alphas betas and the probabilities of failure These values are consequently combined with the other probabilities to determine the total probability of flooding for the dike ring After studying the values in the overall spreadsheet of the Water Board Zeeuws Vlaanderen it appeared that the length of the seepage path was not represented correctly The significantly greater length of the seepage path was determined based on the geometry of the dike for several dike sections one assumed that the seepage path is minimal from toe to toe T07_08_02_Reliability Analysis_D7_1_ Appendix 10 April 2008 53 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505420 FLOGDSite 1320007009 320007023 1320027024 1370007025 1320007028 1326007035 1320007042 132 0007047 13272007053 122000707 1 1320007074 1320202294 1 FAO 102 1320007111 132007116 1320007124 3200071 Jw 132009 139 13200077 5z 132600771 NE 179527007 63 1 FAAATA 1220007185 1320007203 ERER Pl 1320007770 1320007233 132012229 132000724 13
88. fa b 0 0185 0 0003 Slope outer slope bottom tan alfa o 0 0150 0 0002 Slope outer slope tan alfa i 0 0061 0 0000 Roughness inner slope k 0 0802 0 0064 Factor f b for determination Q b breaking waving 0 0106 0 0001 Factor f n for determination Q n non breaking waving 0 0787 0 0062 Model factor critical overflow discharge m qc 0 0075 0 0001 Model factor occurring overflow discharge m qo 0 0827 0 0068 Cohesion clay inner slope c 0 0000 Friction angle clay inner slope phi 0 0000 Soil weight clay inner slope rho 0 0000 Layer thickness clay inner slope d_k 0 0000 Error in position bottom Delta_d 0 0000 0 0000 Model factor Bretschneider for wave height m_gH 0 0000 0 0000 Model factor Bretschneider for wave period m gT 0 0000 0 0000 Error in local water level Delta hlok 0 0000 0 0000 Error in wave direction beta 0 0000 Storm duration t s 0 0299 0 0009 Thickness covering layer d 0 3170 0 1005 Apparent weight soil with respect to uplift 0 0005 0 0000 Model factor uplift m o 0 0009 0 0000 Model factor damping m h 0 0009 0 0000 Root depth grass d w 0 0000 0 0000 Width covering clay layer outer slope L K 0 0210 0 0004 T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 79 AE RUNE FOODS Width dike core at crest level L BK 0 0216 0 0005 Stone pitching thickness D 0 0000 Slope outer slope dike core tan alfa_u 0 1001 0 0100 Slope inner slope dike core tan alfa 1 0 0000 Relative density stone Delta 0 0000 Coefficien
89. factor mS MCA2 3S Ca2 3 E loading Cb1 2a model uncertainty factor mR MCBI 2AR Cb1 2a strength Cb1 2a model uncertainty factor mS MCBI 2AS Cb1 2a loading TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 mn m m ey crita Contract No GOCE CT 2004 505420 FLOGS Distribution Distribution SET Distribution A hoe Example ter 1 D Parameter Unique fortran name Description LSE mapping i ie agaist rd UE parameter 3 distribution Variation Standard S name coefficient Deviation Cb1 2c model uncertainty factor mR MCBI 2CR Cb1 2c E strength du MCB1 2CS bh ee model uncertainty factor Cbl2c E loading Cb1 2d model uncertainty factor mR MCBI 2DR Cb1 2d strength Cb1 2d model uncertainty factor mS MCBI 2DS Cb1 2d loading Cc1 2aii model tainty fact mR MCC1 2AIIR VE A QE Cel 2aii strength mS MCC1 2AIIS cele model uncertainty factor Col 2aii loading Ccl_2b model uncertainty factor mR MCCI 2BR Cc1 2b B strength m MCCI 2BS COT D model uncertainty factor Ccl2b loading oe MCCI 2CR Ccl 2c model uncertainty factor Nos B strength m MCCI 2CS CELA model uncertainty factor Cel2c 5 loading TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Pann era Contract No GOCE CT 2004 505 20 FLOGS Distribution Distribution SET Distribution
90. flood related processes Any new knowledge developed in FLOODsite will be developed and tested at selected pilot sites in Europe which will help to identify missing elements in research These pilot sites are River Elbe Basin River Tisza Basin Flash Flood Basins o the C vennes Vivarais Region France o the Adige River Italy o the Besos River and the Barcelona Area Spain o the Ardennes Area Trans national River Thames Estuary River Scheldt Estuary River Ebro Delta Coast German Bight Coast It can be seen that pilot sites are well distributed over the types of waters like rivers estuaries and coasts as well as types of floods like plain and flash floods For each of those sites at least two pilot areas with different properties have been selected to test as many newly developed tools as possible The Scheldt has been selected as a typical North Sea area which is protected against coastal flooding by means of different flood defence structures such as forelands sea dikes dunes and other constructions The methodologies developed under FLOODsite are partly based on a probability based risk analysis This analysis will require a set of failure modes and related limit state equations for each of the flood defence structures under question The aim of this report 1s to provide a first calculation of the overall failure probability of flood defence structures in the Scheldt area The limit state equations which will be used within t
91. formed with the FORM DS calculations method VNK has used other techniques later on as well For the covering calculations certain model settings were used Possible adjustments are indicated by VNK in appendix B According to PC Ring VNK asphalt has to be chosen at all time for residual strength calculations however this doesn t lead to any results DHV has switched off a large number of wind directions by putting the number of fetch sections to zero in order to obtain results for the stone coverings This lead to the fact that TO07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 52 Task 7 Deliverable D7 1 Appendices 1to 5 mn T Contract No GOCE CT 2004 505420 FLOODS ite sometimes only three to seven wind directions were taken into account for calculating the probability of failure this is in principal not correct VNK has switched all these wind direction back on again at the start of its calculations For the results for stone coverings this doesn t this didn t matter too much For the overtopping wave overrun it did have somewhat more influence With the help of the program MProStab calculations were done from which the probabilities of failure followed for the mechanism sliding which can be combined with the other mechanisms of failure The probabilities for the structures are determined using a method by hand The sto files obtained this way serve as input for PC Ring which calculates the alphas en b
92. g discharge per Q MeanOvertopDis m3 s m Bal 6 Da2 5 metre run of crest qG Qg Grass quality between 0 and 1 Bc2 1b normal 0 2 qM Qm Material quality 1 0 for Sand Bc2 1b Reduction factor for oblique e r R Bc2 1h deterministic wave attack Distance to the center of gravity R GraDis i m Da4 2b from the point of impact R HyRad Hydraulic radius m Bc3 1c Radius of gyration of the ice Rg Rg feaure about the vertical axis m Da4 2b through its center of gravity IT SteelArea Area ratio of steel reinforcement Ca2 3 normal 0 01 TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505 20 Parameter Unique fortran name Description with respect to the concrete cross sectional area D LSE mapping Example distribution Distribution Distribution MET Distribution parameter 1 parameter 2 Standard Deviation WA parameter 3 Variation 1 name coefficient A reduction factor depending on Rw Rw Bc2 1j Bc2 3a on the slope angle ud Distance from the ship s sailing S DiShip m Bc2 1b Bc2 1d line sB CtrlVar Control variable inner slope Ba2 4i Non dimensional damage Sd Ae Dn50 calculated from Ab2 1b Sd Sd normal 6 0 02 mean profiles or separately for Be2 1c each profile line then averaged Undrained shear strength of the Su Su kN m Bb1 2 n
93. h For maximum efficiency use Yes for Optimise and No for both Log each Sample and Log each LSE call Events per Used as a multiplier of failure rate to produce Annual Probability of year Failure If the assumption is that the load occurrence rate is 1 per year then this is set to 1 and the failure rate annual probability of failure Optimise Should normally be set to Yes This ensures that the Calculator will only call an LSE function if it s essential For example if it evaluates an OR gate and the LSE function for the first Failure Mode within the gate indicates failure then there is no need to call the LSE functions for the other Failure Modes within the same gate You may find it useful to set Optimise to No if you are logging each LSE call see section 0 and you want to monitor the results of the LSE function even where they are irrelevant Log each Controls how much information is logged by the Calculator see section Sample 0 Log each ditto LSE call 7 Press Calculate button The values under RESULTS will clear while the calculation proceeds When the calculation completes the results will be shown as follows Annual Reliability 1 0 minus the Annual Probability of Failure T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 146 Task 7 Deliverable D7 1 Appendices 1 to 5 FLUOR ita Contract No GOCE CT 2004 505420 Annual Probability Calculated Fail rate times Events per year of Failure
94. harge sluices cuttings tunnels and pumping stations The failure of a structure by overflow and wave overtopping or not closing of the closing elements does not inevitably result in the arising of a breach in the embankment and with that the flooding of a dike ring area The water flowing in can often be stored in the adjacent water system behind the structures that are linked to the inland water without resulting in flooding Also the structures can often handle large flows without loss of stability Therefore the initially calculated probabilities of failure as a result of overflow and wave overtopping and not closing of the closing elements respectively are tightened in the assessment system to probabilities where the start of a breach occurs These are smaller probabilities by definition This tightening requires extra effort and is thus only executed when the first approach results in relative large probabilities compared to the existing standard frequency for design water levels Whth the mechanism constructive failure it is assumed that the stability is directly lost when breaching occurs The corresponding probability of failure is therefore considered the probability of breaching 1 1 4 Probability of flooding of the dike ring area The probability of flooding of a dike ring area is made up of the calculated probabilities of failure of the dikes dunes and structures in question First the probability of failure is determined per dike sec
95. hem If possible this should be the full pathname of the file including drive and directories If not the file must be in the same directory as the Reliability Tool T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 150
96. his report is based on available LSEs outside FLOODsite These equations will be updated when more information is available from Task 4 of FLOODsite At the beginning of a reliability analysis of a flood defence system a very limited physical knowledge will be available on failure modes their interactions and the associated prediction models including the uncertainties of the input data and models Therefore a detailed flood risk assessment based on a T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 12 Task 7 Deliverable D7 1 Appendices 1to 5 Ep ita Contract No GOCE CT 2004 505420 FLOODS ifo sound physical understanding of the failures and the possible flooding of the protected area will not be feasible at this stage Therefore initially the reliability analysis focuses on providing support to feasibility level decisions In order to identify the relative importance of the gaps in the existing knowledge and to help to optimise research objectives it is necessary to perform a very preliminary flood risk analysis using a holistic approach feasibility level For this purpose three selected pilot sites in different countries and from different areas coast estuary river will be used HRW TUD and LWI The main outputs and benefits from this preliminary study will identify more precisely 1 the relative importance of the uncertainties and their possible contributions to the probability of flooding ii the gaps rel
97. ient Sm req Required scour width Ca2 2a Water depth above the top layer of the sill Goda Ca2 2a Ca2 2b Bc3 1a Kdiff Diffraction coefficient Aa2 1a Aa2 1b Aa2 4 Ab 2 1a Ab2 1b Bal 1 Bal 5bii Bal Sdii Ba2 1a Ba2 1b Ba2 3 Ba2 4biii Bc2 1a Ba2 5 Bc2 1b Bc2 1c Bc2 1g Bc2 1h Bc2 1m Bc2 3b Ca2 1a Ca2 1b Ca2 2a Ca2 2b Ca2 3 Cc2 2a Cc2 2b 00 Theta0 angle between wave crests and depth lines on deep water Aal 1 Aa2 1b Aa2 4 Bal 1 Aa2 1a Aa2 1b Bal 5bii Bal 5dii Aa2 4 Ab2 1a Ab2 1b Bal 1 Ba2 1a Ba2 1b Bal 5bii Bal 5dii TO7_08 C2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 mn mm ey reuunsife Contract No GOCE CT 2004 505 20 Distribution Distribution EN Example irt Tu Distribution arameter arameter Parameter Unique fortran name Description LSE mapping Ac 5 PRA p parameter 3 distribution Variation Standard name coefficient Deviation Ba2 lbii Ba2 Ibiii Ba2 3 Ba2 4iii Ba2 5 Bc2 1a Bc2 1b Bc2 1c Bc2 1g Bc2 1h Bc2 1m Bc2 3b Ca2 1a Ca2 1b Ca2 2aCa2 2b Ca2 3 Cc2 2a Cc2 2b Da2 5 Aal 1 Aa2 1b Aa2 4 Bal 1 Aa2 1a Aa2 1b Bal 5bii Bal 5dii Ab2 1a Ab2 1b Bal 1 Bal 5bii Bal 5dii di angle Dermeen WANE get and Ba2 1a Ba2 1b depth lines on location of interest Ba2 1bii Ba2 1biii Ba2 3 Ba2 Aiii Ba2 5 Bc2 1a Bc2 1b Bc2 1c Bc2 1g
98. ient Bal 6 Cm CoeffM Coefficient Bal 6 Cn CoeffN Coefficient Bal 6 kd Kd KdKf Coefficient for the consideration Bal 6 TU7 OB X2 Reliability Anelysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505 20 Distribution Distribution enn ul a5 Distribution arameter arameter Parameter Unique fortran name Description LSE mapping eines E NS p parameter 3 distribution Variation Standard name coefficient Deviation of the crest width Bk Coefficient for the kf Kf KdKf Bal 6 Ys sharpcrestedness of the weir Rk KD coefficient in Hudson s normal 6 4 8 KD KdHudson Ab2 1b Bc2 1c breaking waves formula normal 6 3 5 8 d zs Depthl Depth of slope affected by flow Bcl 1 The water level at downstream of MT di FreeSubmerged Bel 1 deterministic the dike Free Switch to indicate a free or Submerged Hvert Bcl 4 i submerged weir weir Water level difference between H Lhor outside water level and level in Bc1 4 embankment Horizontal distance between intersections between 1 inside 1 PresSwitch water level and revetment Bc2 1j Bc2 3a 2 outside water level and revetment TO7_08 C2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 muousife Contract No GOCE CT 2004 505 20 Distribution Distribution K Distribution Example parameter 1 parameter 2 Standard
99. ignificant influence on the result For a parameter with a small variation coefficient however the value of this parameter is relatively certain This means that it can not be expected that the average value will change a lot as a result of new insights Varying the average values of those kinds of T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 46 Task 7 Deliverable D7 1 Appendices 1to 5 PR Amm Contract No GOCE CT 2004 505420 FLOODS A d parameters is possibly interesting for the calculating of measures The alpha values influence coefficient are not beatific Sensitivity analyses and influence coefficients are to be considered together The alphas thus represent the contribution of the stochastic variable to the probability of failure for a sub mechanism These can take effect both on the side of the load negative alphas and on the positive side positive alphas T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 47 Task 7 Deliverable D7 1 Appendices 1to 5 muousife Contract No GOCE CT 2004 505420 Li eiv REFERENCES 1 Bucher C G 1988 Adaptive sampling an iterative fast Monte Carlo procedure Structural Safety 5 119 126 2 Cornell C A 1967 Bounds on the reliability of structural systems Journal of Structural Division ASCE 93 ST1 171 200 3 Kahn H 1956 Use of different Monte Carlo sampling techniques Symposium on Monte Carlo methods
100. ike overtopping Response database Despite adopting an importance sampling strategy computation of the water level at the dike section requires significant computational effort In this study we explore the possibility of further reduction in computational time using a response database This is possible if there exists a database of observations of water levels corresponding to different boundary conditions During Monte Carlo simulations first the program searches into the database for the set of boundary conditions which have the closest correspondence to the particular realization The local water level is then calculated by interpolation This strategy for computing the river water level ensures a that the costly computations through the hydrodynamic model can be avoided and b the database of observations already existing is of use Figure 4 illustrates a schematic framework for the use of response database instead of probabilistic loop in this study Variable distributions Bootstrap simulation of Random generation of simulation variables Data Query Routine Limit State Function Evaluation LSFE Additional procedures for sampling techniques Monte Carlo Simulation Figure 3 3 Block Diagram of conceptual framework for response database used in Monte Carlo simulation The method of estimating the river water levels along the dike sections through interpolations from the response database is s
101. iles in PC Ring and to put them in based on recent measurements by the water board Adjustments of profiles resulted in the fact that the profiles used for calculations in this report differ from the profiles used for the first calculation For the new schematization the following assumptions were made For the toe of the dike one assumed the sand line If no foreland is present the second point is the toe An extra point appears than which is located 2 meter in front of the toe on the same level as the toe The choice between a bend or not on the crown is made based on a visual estimation If a berm is indicated in the file of the water board but is it steeper than 1 15 it has to be adjusted for this schematization In PC Ring a slope than has to be steeper than 1 10 and a berm than less than 1 15 Everything steeper than 1 10 is considered a slope With this berm T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 49 Task 7 Deliverable D7 1 Appendices 1to 5 FLOORS Contract No GOCE CT 2004 505420 eiv Dunes disappears and one obtains a slope with a bend Everything below 1 10 becomes a berm One considered up to one digit behind the comma for this With the downgrading to a berm an adjustment in height is made for the lowest point on the berm This way the gradient of the attacked upper slope stays the same The shift is not done in the line of the slope The y point is vertically lifted of lowered i
102. ilistic calculations After the site description the reliability analysis proceeds with the process model definition of the failure mechanisms for each structure type In order to carry out the probabilistic calculations the relevant flood defence information needs to be extracted To this end the flood defence line is discretised into flood defence sections with similar characteristics Each flood defence section is represented by one cross section in terms of its geometry revetment soil properties hydraulic boundary conditions etc The information requirements are determined by the failure mechanisms that are taken into account for the structure type of the flood defence section Figure 8 presents the flood defence sections in which the flood defence line is discretised Figure 9 shows a flow chart for the calculations of the annual probability of failure and the fragility of the earth embankments reinforced concrete walls and anchored sheet pile walls 4 Discussion of the results of the reliability analysis Figure 10 and 11 present fragility curves for earth embankments and reinforced concrete walls Anchored sheet pile walls are more likely to fail for lower water levels During a storm with increasing water levels the probability of failure therefore remains equal to the initial failure probability The probability of failure of the anchored sheet pile wall equals 0 15 due to jointly anchor breaking and rotational failure of the sheet pile
103. k 7 Deliverable D7 1 Appendices 1to 5 muousife Contract No GOCE CT 2004 505 20 Distribution Distribution Exunil uci ur Distribution arameter arameter Parameter Unique fortran name Description LSE mapping RA i k AA d parameter 3 distribution Variation Standard name coefficient Deviation Cc1 2d dimension concrete wall sheet z d3 d3 Cc1 2aii Cc1 2b normal 2d 0 004 pile wall dimension concrete wall sheet Cc1 2aii Ccl 2b d4 d4 t normal 2d 0 004 pile wall Ccl 2d dimension concrete wall sheet d5 d5 Cc1 2aii Cc1 2b normal 2d 0 004 pile wall dimension concrete wall sheet ee d6 d6 Cc1 2aii Cc1 2b normal 2d 0 004 pile wall dimension concrete wall sheet d7 d7 Cc1 2aii Cc1 2b normal 2d 0 004 pile wall dimension concrete wall sheet d8 d8 Cc1 2ai1 Cc1 2b normal 2d 0 004 pile wall dimension concrete wall sheet p d9 d9 i Cc1 2ai1 Cc1 2b normal 2d 0 004 pile wall distance from outer concrete Cc1 2c Cc2 2b l 112 0 01 ds ds fibre to heart of the ene e reinforcement Ca2 3 ee Level of elevation in front of hl hl Cb1 2a Cb1 2c Cb1 2d normal 1 0 1 riverside of concrete wall T07_08 2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 muousife Contract No GOCE CT 2004 505 20 Distribution Distribution Exan
104. l wave height and wave period No joint or conditional probability density functions were considered As for risk pathways in the German Bight Coast pilot site flood defences comprise more than 12 km of dikes grass and asphalt dike and a dune area of about 2 5 km length The PRA has however focussed on the dikes as the key flood defence structure since the dune belt is extraordinary high and wide and is regarded as significantly safer than the dike protection Before starting the probabilistic analysis the dike geometry and laser scan data have been used to define different sections of the flood defences Criteria for distinction of different sections were the type of flood defence its height its orientation the key sea state parameters like water level and waves and geotechnical parameters Thirteen sections have been identified using these criteria see Kortenhaus amp Lambrecht 2006 Each of these sections is assumed to be identical over its entire length and hence will result in the same probability of failure T07_08_02_Reliability Analysis D7 1 Appendix 10 April 2008 62 Task 7 Deliverable D7 1 Appendices 1to 5 mn i Contract No GOCE CT 2004 505420 TLUOPS Uc The PRA has used a full probabilistic approach starting from the input parameters at the toe of the dike and applying early versions of the failure modes and fault trees which have been developed under FLOODsite for the specific type of flood defences Time dependen
105. l 2008 29 Task 7 Deliverable D7 1 Appendices 1to 5 mn it Contract No GOCE CT 2004 505420 FLOODS ile Figure 3 7 Change in location and scale parameter with different POT values The dike length is discretized into segments such that each segment could be considered independent of each other The length of each segment was taken equal to the correlation length of the random process modelling the spatial randomness of the dike height The autocorrelation function considered is as follows BE Px x L L e gt 6 where D is the fluctuation scale given by D p DAL 7 0 Figure 3 8 illustrates the auto correlation function for the dike height The fluctuation scale is found to be 3532 m and the dike segments were taken to be of length 3500m AutoCorrelation Plot i amp o S E E E o o Autocorrelation Length m Figure 3 8 Autocorrelation for dike height A new sea level is assumed to take place every 2 days 48 hours The typical travel time of a flood wave along the length of the dike is approximately one hour Thus the river water levels along the dike are measured every hour Calculations through the hydrodynamic model are carried out using SOBEK A node is selected in each dike segment in SOBEK 1D schematisation T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 30 Task 7 Deliverable D7 1 Appendices 1to 5 mn i Contract No GOCE CT 2004 505420 muGnsize For the purpose
106. lated in the calculations for dike ring 32 Dike sections 7002 024 Dp7 7258 074 Dp99 and 7271 072 Dp69 for grass covering Dike sections 7024 006a Dp11 and 7025 006a Dp15 for asphalt covering The other sections for stone covering The types of covering for which the various sections have been calculated are familiar to the water board There are 2 options for schematization in case more than one type of stone covering is present in 1 section Take the average along the total section T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 19 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505420 Gasite E Take the worst part for a shorter length of the section In order to be able to compare the results it should be possible to insert both values in the overall spreadsheet Schematization of dunes It was agreed upon with engineering bureau VNK to perform calculations on the measured dune sections of 2004 5 pieces because these provide a conservative image a 5 annual supplement is not planned until 2005 The choice of dune sections to be calculated is done based on the 2004 report of RIKZ The choice is commented on in appendix A Schematization foreland of Saeftinghe Shallow foreland is present in the land of Saeftinghe 6 most easterly located sections 7211 to 7271 This foreland is not accounted for in the calculations in this dike ring report The boundary condition points
107. latter section are not part of the 33 dike sections that are selected for calculation For section 7025 006a dp15 the water board also indicated that it needs advanced testing for the mechanism sliding of the inner slope The profile of this section is not assessed on this mechanism within VNK At the 200 testing none of the selected section scored unsatisfactory for the mechanism sliding of the inner slope It is recommended to couple the other selected profiles which do not match the selected sections to the right section 1n PC Ring section that is thus not in the selection It concerns the sections 7012 019a dp20 7014 013 dp8 7052 137 dp23 7079 130 dp16 7204 084 dp199 7226 080 dp169 Next to that it is recommended to use the results of the additional soil research for these calculations T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 41 Task 7 Deliverable D7 1 Appendices 1to 5 PERAN cra Contract No GOCE CT 2004 505420 fLOUOP Sic The water board thinks that the present results should not be taken into the calculations of the probability of flooding since research is now being done to improve the input data Dunes No single dune sections scores unsatisfactory in the 2005 testing with the new graver boundary conditions for waves also see section 2 9 The results of the 4 sections that VNK calculated with the old lees grave boundary conditions seem to be correct beta 4 37 to 5 26 Asu
108. llission 0 ResStrePier MN Da4 2a eccentric or eccentric impact 1 Overall strength pier in ice Fr ResStreColl i Da4 2b accumulation circumstances Overall strength structure in Fr ResStreAttach Nd Da4 2c colliding ice circumstances Overall strength structure in ice Fr ResStreDebris Da4 3 attachment circumstances Overall strength structure in Fr Ktheta A Cc1 2d debris circumstances k0 SectionMod coefficient Cb1 2c lognormal 2d 0 01 Section modulus of sheet pile f Z OverPercen Aal 1 Bal 1 Ba2 41 wall Proportion of time overtopping od P BreakSlope i Bc3 la deterministic during the storm duration D20f D20f 20 value of sieve curve filter Bal 5c 20 value of sieve curve base D20b D20b Bal 5c layer TO7_08 C2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505 20 Parameter Unique fortran name Description LSE mapping Example distribution Distribution Distribution EN Distribution parameter 1 parameter 2 Standard Deviation TA parameter 3 Variation d name coefficient cl FilterC1 Filter coefficient c1 Bal 5c c2 FilterC2 Filter coefficient c2 Bal 5c Switch whether the material is g Grading i Bal 5c uniformly 0 or wide graded 1 Aal 1 model uncertainty factor mR MAAI IR Aal 1 strength Aal 1 model uncertainty factor ms MAAI 1S Aal 1 l
109. loodplain and consists of a wide variety of flood defence structures figure 1 i EET P 3 a BASILDON gt lt ROMFORD N pe j ILFORD SOUTHEND ON SEA DAGENHAM 1 p Shoebury Ness CITY OF LONDON Greerwicn Ec Purfleet r GRAYS Aina iow Grain TILBURY eet Saanepombe Na DARTFORD GRAVESEND SHEERNESS M Dartford Creek 25 S S os a 4 25km EP dE a CHATHAM GILLINGHAM Ss f UfinGTON CROYDON Figure 1 The location of Dartford Creek Gravesend and the Thames barrier at Greenwich in the Thames Estuary The flood defence line in the reliability analysis is 10 6 km long whereby the structure types represent the following proportions e Earth embankments 6 7 km e Reinforced concrete walls 1 9 km e Anchored sheet pile walls 2 1 km The elevation of the crest levels is shown in figure 2 The structure types and failure mechanisms are described in more detail in the following section The hydraulic boundary conditions along the Dartford Creek to Gravesend flood defence line are governed by the tidal conditions rather than the fluvial discharges A Monte Carlo simulation of joint wind speed and tidal water levels at the mouth of the Thames Estuary is combined with iSIS predictions to derive inner estuarial local water levels A simple predictive model is applied to derive local wave conditions The soil conditions are generally represented by a clayey peaty layer o
110. lure mechanism driven by a combination of uplifting and piping Section 16 0 8 0 6 0 4 0 2 Probability of failure 0 0 T T 0 0 2 0 4 0 6 0 8 0 10 0 Water level m OD Total Uplifting Piping Sliding Overturning Reinforcement failure Shear failure Piping toe Aeman Crest level Indication extreme water level Section 26 Probability of failure oooooooooon2 OLionauuooo 0 0 2 0 4 0 6 0 8 0 10 0 1 0 g 0 9 508 07 0 6 G 205 0 4 0 3 9 0 2 0 1 0 0 4 0 6 0 Water level imi OD Tuta Uplifting Pinirq Reinforcement fail re Shwear alum Indicuburi ol cxarcrmc water level falu hw bia iiie Ouwe y Pip nig n Crest level hrabu exlre nie water beoe Figure 12 Fragility curves for three different types of reinforced concrete walls T07 08 02 Reliability Analysis D7 1 Appendix 11 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505420 Onsite ju F Il Appendix 2 Details of the PRA Scheldt II 1 INTRODUCTION Background FLOODsite is aiming for Integrated Flood Risk Analysis and Management Methodologies New research efforts in this field will be undertaken to fill gaps in knowledge and to achieve a better understanding of the underlying physics of
111. mpling density function could be Gaussian or non Gaussian and is centred over an appropriately defined multi dimensional region covering the region of likelihood around the design point Shinozuka 1983 Considering non Gaussian importance sampling functions however lead to difficulties when the random variables are mutually correlated These problems can be circumvented by transforming the problem to the standard normal space and constructing Gaussian importance sampling functions Schueller and Stix 1987 This is especially true when the location of the design point is not known apriori Bucher 1988 Model setup The overflowing of the dike triggers erosion in inner slope breach starts to grow which leads to the ultimate failure of the dike Thus in the study reported in this paper failure is defined as the overtopping of the dike and the performance function is taken to be of the form 2 h h 0 h h h Q 5 where A is crest height of dike and A is the local water level obtained as a function of and Q representing respectively the extreme sea level and extreme river water discharge The relationship between the local water level and the boundary parameters and Q is through a nonlinear hydrodynamic model The parameters j As and h are modeled as mutually independent random variables The extreme values of the sea water levels and the river discharges are assumed to be non Gaussian random variables The dike cres
112. n True or False An additional message is produced for each LSE function call if you have chosen Yes for Log each LSE call This gives the Failure Mode name the LSE function name and the value returned by the function If your Optimise value is Yes you may find that some Failure Modes are not evaluated since they were found to be unnecessary for the final result T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 147 Task 7 Deliverable D7 1 Appendices 1 to 5 T ta Contract No GOCE CT 2004 505420 fLOGESe APPENDIX V 6 SPECIFICATION CSV FILES Parameter csv The file Parameter csv contains one line per parameter value which specifies the unique name of the parameter The order of the parameters determines the order in which they are stored in the VALUES array passed to an LSE function However an LSE Function should not rely on this It should always call I ndexOf Failure Mode csv The fileFai ureMode csv contains one line per Failure Mode Each line has 3 columns as follows Column Description FailureModeName A unique name for the Failure Mode This name must match that used in the FTA Event Database LSEFuncDLL The Filename of the DLL which includes the LSE function for this Failure Mode If possible this should be the full pathname of the DLL including drive and directories If not the DLL must be in the same directory as the Reliability Tool or in one of the directories listed in
113. n calculated The testing is being performed now On this moment additional data are gathered for an advanced testing amongst others on the grass quality The water board has already indicated the state of affairs of the preliminary results of the 2005 testing for a number of sections In many cases the type of covering for which these sections were tested differs from the type that VNK has calculated and which was identifiable for the water board see section 2 5 This assessment of the water board with the mechanism for which the section is calculated at VNK next to it is given in table 3 5 Comments can thus be given on the results for the mechanism covering damaging and erosion of body of the dike Further research on the various types of covering a dike is always constructed from a combination of multiple types of covering dry stone stone asphalt and grass that are present on a dike section seems necessary All types will need to be calculated separately and consequently it has to be determined which one is governing also in relation to the associated design criteria Even better would be if multiple types of covering on 1 dike section could be calculated with PC Ring For section 7159 099a dp319 it is indicated that it is nominated to be improved With testing this section doesn t make it based on its age The water board thus doubts the calculated result which is relatively good beta 5 2 The section partly consists of asphalt and
114. n be found in the conservative data that are used for the 1 testing due to a lack of data These data were also used for VNK This results in a pessimistic picture On this moment one is busy doing additional soil research for the 2 testing gathering of test samples borings measurements of water pressures foundation The sub soil is mapped out better with these methods It 1s expected that this will lead to better results for sliding The model of the sub soils used for the Mstab calculations also seems conservative For the long term the water board expects to be able to take this into account better and consequently calculate better results Apart from that it needs noticing that the dikes around dike ring 32 are high and steep and that additionally the sub soil is not very good weak layers are present Based on that fact it is not unlikely that sliding will appear as a relatively weak mechanism For less conservative data as well it is expected that this mechanism will score relatively bad beta around 2 5 3 For sections 7012 019a dp20 7052 137a dp23 7079 130 dp16 the water board has separately indicated that these score well for the mechanism sliding in the preliminary results of the 2005 testing For the sections 7226 080 dp169 2749 076 dp124 applies that they need advanced testing for the mechanism sliding of the inner slope These are sections that are part of the selected cross sections and thus except for the
115. n size diameter d m 0 0000 Sum 0 8454 0 3413 TO7 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 81 ContactNo GOCE CT 2004505420 FOOT Site Table C 3 Sensitivity coefficients dike ring 36 Land van Heusden De Maaskant Variable Description alfa alfa 2 1 Dike height h d 0 00200 0 00000 2 Bermheighth B 0 00000 0 00000 3 Berm width B 0 00000 0 00000 4 Toe height h_t 0 00000 0 00000 5 Slope outer slope top 0 00000 0 00000 6 Slope outer slope bottom 0 00000 0 00000 7 Slope outer slope 0 00000 0 00000 8 Model factor critical overflow discharge m_qc 0 00000 0 00000 9 Roughness inner slope k 0 00000 0 00000 10 Factor for determination Q bf b 0 00000 0 00000 11 Factor for determination Q_n f_n 0 00000 0 00000 12 Model factor occuring overflow discharge m qo 0 00100 0 00000 13 Error in position bottom 0 00400 0 00002 14 Model factor Bretschneider for Hs 0 02700 0 00073 15 Model factor Bretschneider for Ts 0 00000 0 00000 16 Error in local water level 0 06100 0 00372 17 Storm duration t_s 0 01100 0 00012 18 Water level Maasmond 0 01000 0 00010 19 Discharge Lobith 0 90600 0 82084 20 Discharge Lith 0 25100 0 06300 21 Wind speed Schiphol Deelen 0 02100 0 00044 22 null 0 16100 0 02592 23 Prediction error water level MK 0 00700 0 00005 24 Thickness covering layer d 0 02800 0 00078 25 Thickness sand layer D 0 01100 0 00012 26 Length leakage length L 0 06600 0 00436 27 Rolling friction angle theta 0 05300 0 00281 28 Fact
116. nally gives an overview of the assessment of the water board Calculations have been made by DHV with checks by VNK and assessments by WZE Location and characteristics Dike ring area 32 encompasses all of Zeeuw Vlaanderen with primary embankments of category a these are embankments that enclose the dike ring areas either with or without high grounds and directly retain outside water along the North Sea and Westerschelde The length of primary embankments in Zeeuws Vlaanderen amounts to 85 kilometers of which 8 kilometers of dune coast The exceedance frequency for this area equals to 1 4000 years The dike ring is border crossing with Belgium The embankments in Belgium are of category d Its length is unknown A system of regional secondary embankments is situated at a variable distance from the primary embankments along the whole North Sea coast and Westerschelde An overview of the dike ring area is given in figure 2 1 The dike ring is enclosed by the following embankments The dike along the Westerschelde The dike along the Schelde The high grounds in Belgium and Northern France The sea retaining dunes or dikes of Belgium Northern France and the Netherlands Dikes dunes and structures An overview of the embankments in dike ring 32 is given on the overview map primary and regional embankment of dike ring area 32 The following important water retaining structures can be distinguished Dike with stone covering
117. nd minima of Sea Water level at Vlissingen Western Scheldt Comparison of Input Distribution and Pareto 14 52 2 51e 2 0 06 Input 0 03 B Pareto 05 28 31 33 36 39 Valuesin 10 2 Figure 3 5 Pareto distribution representing sea level flactuation T07_08_02_Reliability_Analysis_D7_1_Appendix 10 April 2008 28 Task 7 Deliverable D7 1 Appendices 1to 5 muousite Contract No GOCE CT 2004 505420 FLOODS A family of Pareto distributions were obtained depending on the threshold level selected while constructing the Pareto distributing using peak over threshold POT analysis see figure 3 6 fx POT300 fx POT200 s fx POT 250 2d a extended 200 Figure 3 6 Effect of Choice of POT value on distribution Parameters of exponential distribution calculated by Bestfit are based on zero position of the location parameter For corresponding 2 days maxima POT analysis is carried out by changing location and scale parameters successively Figure 3 7 illustrates the effect of changing the threshold during POT analysis on the location and scale parameters Change of parameters on selection of POT for discharge series _ scale b Location a Ployb 200 300 Threshold T07_08_02_Reliability Analysis D7 1 Appendix 10 Apri
118. nged since the sheet was last shown Choose a structure from the drop down box If you have changed the structure click on Load Parameter Names This will regenerate the names under PARAMETERS to reflect the parameters used by the LSE functions for the failure Modes used in the fault Tree for the structure Currently the values from the previous structure are not cleared automatically You need to do this yourself For each parameter enter a value and optionally a Statistical distribution and its parameters Currently the permitted values in the Distribution column are blank fixed value N Normal and LN Log Normal Other options will be added at a later release The value in the Value column is either the fixed value Distribution is blank or the Mean value other Distributions The Distribution Parameters columns allow up to 4 further parameters for a statistical distribution The current distributions Normal or Log Normal only require the first column to be filled with the standard deviation for the distribution Review the Convergence Control parameters and change them if you wish These are used to control how the Reliability Calculator determines whether it has a converged value for the annual probability of failure For any sample the Calculator generates values for each parameter either the fixed value of a random value according to the prescribed distribution It then evaluates the Fault Tree for the str
119. ngth Bal 5aii model uncertainty factor ms MBAI 5AIIS Bal 5aii 7 loading TU7 OB 2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505 20 Parameter Unique fortran name Description Bal 5aiii model uncertainty LSE mapping Example distribution Distribution Distribution Distribution parameter 1 parameter 2 Variation coefficient Standard parameter 3 ES name Deviation mR MBAI SAIIIR Bal 5aiii lognormal 1 1 2 0 1 factor strength Bal 5aiii model uncertainty Pr mS MBA1 SAIS Bal 5aiu factor loading Bal 5b model uncertainty factor mR MBAI S5BR Bal 5b strength Bal 5b model uncertainty factor ms MBAI SBS Bal 5b loading Bal 5d model uncertainty factor mR MBAI 5DR Bal 5d strength Bal 5d model uncertainty factor ms MBAI 5DS Bal 5d gt loading Bal 6 model uncertainty factor mR MBAI 6R Bal 6 i strength Bal 6 model uncertainty factor ms MBAI 6S Bal 6 loading Ba2 1a model uncertainty factor mR MBA2_1AR Ba2 la x strength Ba2 1a model uncertainty factor ms MBA2 1AS Ba2 la loading TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 mn m e oy fLOOP Sic Contract No GOCE CT 2004 505 20 Distribution Distribution SET Distribution 9 eee Example parameter 1 paramet
120. ngth 470m Figure 2 4 bottom left 4 junctions of constructions of sea dikes and or dune toe defense on the adjacent dune area coastal length 600m at Schoneveld the Kruishoofd and Nieuwe Sluis The slopes of stone on sea dikes and connection constructions coastal length 8100m tested under Project Zeeweringen 1 The dune area of Cadzand west of the outlet with the adjoining sea dike of the Kievitspolder East 2 The sea dike of the Jong Breskenpolder between Nieuwe Sluis and the lighthouse 3 The addition of the artificial dune in Breskens at the Veerhaven Figure 2 3 Weak spots according to the assessment of the water board T07_08_02_Reliability Analysis_D7_1_ Appendix 10 April 2008 22 Task 7 Deliverable D7 1 Appendices 1to 5 ita Contract No GOCE CT 2004 505420 TOCESiie i r C dzald Bad T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 23 Task 7 Deliverable D7 1 Appendices 1to 5 mn Contract No GOCE CT 2004 505420 FLOORS ita view II 3 LEVEL III PROBABILITY OF OVERTOPPING CALCULATION DIKE RING AREA 32 The probability of a dike failure due to overtopping is considered of dike ring 32 Overtopping is assumed to take place due to extreme sea levels extreme river discharge or a coincidence of both The levels of the river and sea are modelled as random variables and the water level along a dike section is obtained as a nonlinear function of these random variables The height of
121. o 5 m ita Contract No GOCE CT 2004 505420 rhUGDsize The following variables above the geometry variables apply to the mechanism structure not closed see Table A 7 For more information about this mechanism is referred to Steenbergen and Vrouwenvelder 2003B Table A 7 Variables for structure not closed Steenbergen and Vrouwenvelder 2003B Variable nr symbol description 110 Bns Reliability closure 107 Mom Model factorVkom 108 Min Model factorVin 109 C Coefficient 104 Akom surface retention area 105 hov Level raise 102 B Width structure 103 hok Water level in open condition 101 A Cross section discharge 106 u Discharge coefficient 1 1 1 9 Dune erosion The flood defence fails due to dune erosion in case the cross section is eroded below a threshold due to wave attack see Figure A 8 Figure A 8 Dune erosion Technical Advisory Committee on Water Defences 1998 The following variables above the geometry variables apply to the mechanism dune erosion see Table A 8 For more information about this mechanism is referred to Steenbergen and Vrouwenvelder 2003B Table A 8 Variables for dune erosion Steenbergen and Vrouwenvelder 2003B Variable nr symbol description 122 Mp Model factor 123 dso Median grain size T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 72 Task 7 Deliverable D7 1 Appendices 1to 5 FLOOD
122. oading Aa2 la model uncertainty factor mR MAA2_ 1AR Aa2 la B strength Aa2 1a model uncertainty factor ms MAA2 1AS Aa2 la z loading Aa2 1b model uncertainty factor mR MAA2 1BR Aa2 1b strength Aa2 1b model uncertainty factor ms MAA2 IBS Aa2 1b loading Aa2 4 model uncertainty factor mR MAA2 4R Aa2 4 strength Aa2 4 model uncertainty factor ms MAA2 4S Aa2 4 loading TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 mn m Ae 7 fLOOP Sic Contract No GOCE CT 2004 505 20 Distribution Distribution SET Distribution Example arameter 1 parameter 2 Parameter Unique fortran name Description LSE mapping i E d parameter 3 distribution Variation Standard US name coefficient Deviation Ab2 1a model uncertainty factor mR MAB2 IAR Ab2 1a E strength Ab2 1a model uncertainty factor ms MAB2 1AS Ab2 1a B loading Ab2 1b model uncertainty factor mR MAB2 IBR Ab2 1b strength Ab2 1b model uncertainty factor ms MAB2 IBS Ab2 1b loading Bal 1 model uncertainty factor mR MBAI IR Bal 1 E strength Bal 1 model uncertainty factor ms MBAI IS Bal 1 loading Bal 4 model uncertainty factor mR MBAI 4R Bal 4 i strength Bal 4 model uncertainty factor ms MBAI 4S j Bal 4 B loading Bal 5aii model uncertainty factor mR MBAI SAIIR Bal 5aii 5 stre
123. obabilities of failure for structures the exceedance frequency line of water levels is confronted with the strength of the embankment For the structures the uncertainties in the input data are also accounted for explicitly For determining the probability of failure of a structure the following failure mechanisms are accounted for e Overflow and wave overtopping e Not closing of the closing elements e Constructive failure The failure mechanisms are briefly described below Overflow and wave overtopping With the failure mechanism overflow and wave overtopping the structure fails because water runs over the structure The assessment of the structure is based on a comparison of the retaining height in relation to the exceedance frequency line of the outside water level Not closing of the closing elements With the failure mechanism not closing of closing elements the structure fails as a result of the closing elements not being closed off in good time The assessment of the structure is based on a comparison between the exceedance frequency line of the outside water level and the open retaining level OKP taking into account the probability of the not closing of the closing elements For determining the probability of not closing of the closing elements the VNK method follows the Guideline Structures 2003 This guideline distinguishes four main causes of failure Failure of the high water warning system failure water level r
124. om 8 gt 1 Ov ov keeps functioning Now result for the covering Initial value adapted from 8 1 No result yet for covering stops at 150 Ov ov keeps functioningSouthern wind direction turned off now result for covering Initial value adapted from 8 gt 1 Result follows for both covering and overtopping wave overrun Covering crashes Start method adapted from 8 1 Ov ov keeps functioning Covering crashes Start method adapted from 8 1 Ov ov keeps functioning Covering crashes Start method adapted from 8 1 Ov ov keeps functioning Covering crashes Start method adapted from 8 1 Ov ov keeps functioning Number of samples adapted from 5000 gt 10 000 for all sections T07_08_02_Reliability Analysis D7 1 Appendix 10 April 2008 57 Task 7 Deliverable D7 1 Appendices 1to 5 muousife Contract No GOCE CT 2004 505420 Li eiv Selection For selected sections checked whether level GWL and Factor fmGWS are filled in correctly Appeared this was often not the case Adapted if necessary Level GWL was set at 5 35 This should be 0 is sea Has been adapted Factor fmGWS was set at 0 15 River has been adapted to 0 25 Sea 28 38 For these sections no covering data were put in 28 For 28 data input based on section 42 Width stone 0 gt 0 2 Length stone 0 gt 0 2 Porosity filter 0 gt 0 35 This results in a beta of 0 2 Section 28 has a very high toe All Calculate all sections for all mechanism Initially all
125. omewhat in principle similar to the response surface method It must be noted that the response surface based methods are used to develop approximating functions that surrogate for long running computer codes Khuri and Cornell 1987 In this study the interpolation functions used to estimate the water levels along the dike sections can be viewed as response surface functions for the particular realization T07_08_02_Reliability_Analysis_D7_1_Appendix 10 April 2008 27 Task 7 Deliverable D7 1 Appendices 1to 5 Gok it Contract No GOCE CT 2004 505420 ri SHO Simulation details and results The overflowing failure mechanism of dike ring No 30 31 from Western Scheldt Province of Zeeland is studied The water levels of North Sea recorded at station Vlissingen were used to construct probability distribution functions of downstream levels The data analysed are daily records from 1863 to 2004 see figure 3 4 Bestfit package was used to rank the distribution and find the parameters based on method of moments A Pareto distribution was observed to lead to a realistic description for the observed data see figure 3 5 Annual Maxima of Sea WL 500 400 0 HHH kL tte eed ae HT mane i SCCM m SECC see COO DEETAN t ETT ane fae a alc allt tla lint Lt blade 200 WLicm m CELLET ELLLELELE LL LELELEIL L 400 1850 1880 1900 1920 1940 1950 1980 2000 Year Figure 3 4 Annual maxima a
126. onte Carlo simulations repeated analysis of the hydrodynamic model for each realization of the random boundaries makes Monte Carlo simulations very expensive This implies that there is a need to explore the use of alternative less computationally intensive techniques for reliability analysis One such method the importance sampling technique is used in the study carried out in this paper The method is applied to estimate the two days overflowing probability of a dike of length 80 km along the Western Scheldt Province of Zeeland The Netherlands Three variables namely the dike height sea level and Scheldt river discharge are considered as randomly distributed variables The limit state is idealized as a function of these three mutually independent random variables Probability distributions for these three random variables are constructed from analysis of data based on observations from the site Pandey et al 2003 Calculations through the hydrodynamic model are carried out with a commercially available software SOBEK Additionally the use of a response database in lieu of the hydrodynamic model for calculating the water level along the dike is explored Dahal 2005 Importance sampling First a brief review of the method of importance sampling is presented Assume that the uncertainties associated with the problem are represented through a vector of random variables X The performance function is given by g X such that g X 0 indicate
127. or C Bear 0 00000 0 00000 29 Grain sized 70 0 11000 0 01210 30 Uniformity d 70 d 10 0 00000 0 00000 T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 a Task 7 Deliverable D7 1 Appendices 1to 5 FLOODBz f2 Contract No GOCE CT 2004 505420 31 Constant van White 0 19200 0 03686 32 Apparent Relative volumetric mass soil 0 00200 0 00000 33 Relative volumetric weight sand 0 02100 0 00044 34 Model factor uplift 0 00500 0 00003 35 Model factor piping 0 11400 0 01300 36 Model factor damping 0 00500 0 00003 37 Specific permeability 0 11000 0 01210 38 Inner water level h_b 0 03800 0 00144 39 Root depth grass d_w 0 01000 0 00010 40 Width covering layer of clay L_K 0 00000 0 00000 41 Width dike core on crest height L BK 0 00000 0 00000 42 Tangent alfa_u 0 00000 0 00000 43 Coefficient grass c_g 0 00500 0 00003 44 Coefficient erosion covering layer c_rk 0 00000 0 00000 45 Angle in reduction factor r 0 00000 0 00000 46 Acceleration erosion process alfa_z 0 00000 0 00000 47 Damping factor alfa_h 0 00000 0 00000 48 Unavailable wave direction 0 00100 0 00000 Sum 0 93100 0 99914 Dike ring 36 is not threatened by the river Rhine which is measured in Lobith but due to the structure of the load models in PC Ring the discharge of the Rhine in Lobith plays a fictive role In fact the squared alfa value for the river Meuse 2 EN p should be O oue Arin Lobith T07_08_02_Reliability_Analysis_D7_1_Appendix 10 April 2008 83 Task 7
128. or closing almost daily on one hand and the presence of 2 flood defences on the other hand The failure situation concerns the blocking of the mitre gates due to sedimentation or obstacles after which the emergency gate can t be closed in time Improving the situation 1s possible by installing an additional set of mitre gates Next to that one can think of further investigating the probabilities of failure for the not closing advanced method possibly in combination with optimizing the controls 2 Further inspection showed that this inflow is not possible due to which a lower probability of failure than is now calculated can be expected This result can thus be left out of consideration For the pumping station Cadzand the result of VNK seems too good for the mechanism overtopping and wave overrun beta is 4 39 probability of failure lt 1 100 000 From the structures report the following follows VNK calculates a large probability of failure for this structure but this probability of failure is adjusted to a much lower probability of flooding With failure water waves runs over the valve chamber This overrun flow does not directly result in a loss of stability of the structure and thus to flooding The overrun flow ends up on a hardened surface the behind the valve chamber and on both sides runs into the outlet channel lying behind The stability of the structure is not lost until a flow runs over that is associated with a much highe
129. ormal 0 2 fine grained soil t T Period of constant loading S Bal 6 Da2 5 Start time of erosion if inner Aa2 4 Ba2 4d t0 TO h n bs slope Bal 5bii Bal Sdii Thickness of armour and ta tu tf TIT underlayer or filter layer in m Bc2 1k direction normal face tano Tano Slope of the initial dune profile s Aa2 la TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 muousife Contract No GOCE CT 2004 505 20 Distribution Distribution K Distribution Example parameter 1 parameter 2 Parameter Unique fortran name Description LSE mapping BAG PUE parameter 3 distribution Variation Standard PA name coefficient Deviation simplified Afi Aa2 1a Aa2 4 Bal 1 Slope of the initial dune profile tans Tans i Bal 5bii Bal 5dii simplified Ba2 5 t TI Toe level of initial dune profile m Aa2 1a Ab2 1a Ba3 1 normal 0 2 A22 4 Bal 1 Tm Tm Mean wave period S Bal 5bii Bal 5dii Ba2 41 Ba2 5 Bc2 3b Aa2 1b Aa2 4 Bal 1 Bal 5bii Bal Sdii Tm 1 0 EneWPer Spectral wave period Be called Bal 2bii Bal 2biii the energetic wave period Ba2 4i Ba2 4iii Ba2 4b Ba2 4d Ba2 5 A22 1a Aa2 1b Ab2 1a Ab2 1b Ba2 1a Ba2 1b Du Ba2 1bii Ba2 1biii Spectral peak period inverse of Tp WavePeriod s deterministic peak frequency Ba2 3 Ba2 4d Bc2 la Bc2 1b Bc2 1c Bc2 1g Bc2 1h Bc2 1m Ca2 1a
130. osion and slope instability Wave overtopping and erosion Aal 1 Ba2 4i Uplifting and piping Combination of uplifting and piping Bal 5aii and Fissuring cracking Bal 5aiii e Long term crest level settlements compressible layers and estuarial settlements Short term crest level settlements off road cycling Bathymetrical changes of Thames e Third party activities loading embankment slopes X OR gate h gt h Breach N J AND gate Failure landward Piping underneath ig 63 INHIBIT gate embankment embankment Er f a Pipi Slope Failure due to E n PE instability overtopping Uplif ng s A impermeable layers hsh gt Failure riverward Breach embankment T Failure both Piping underneath es t jac a embankments embankment y ppg F Failure landward F ir ca ss embankment _ ab Piping T Failure riverward Uplifting t Failure landward embankment impermeable layers embankment Slope instability Not taken into account in reliability analysis Figure 4 Fault trees for double crested earth embankments underpinning the reliability analysis Explanation to top fault tree if the water level is higher than the riverward crest level h then the water
131. ow the Failure Mode to be used in any fault Trees that you develop T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 89 Task 7 Deliverable D7 1 Appendices 1to 5 ran crta Contract No GOCE CT 2004 505420 fkOOP sic If you want to use the Failure Mode as a conditioning event create a second Failure Mode with the same name but followed by a question mark Again this Failure Mode should be added to both FailureMode csv and FailureMode ped Failure Mode Parameters Use the sheet labelled Fm Parameters You will need to update this sheet and save the file if there is any change to the parameters used by an LSE function Otherwise you may not be able to supply values for the parameter T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 90 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505420 muunsize APPENDIX V 4 PARAMETER DESCRIPTION AND LSE MAPPING T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 91 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505 20 Parameter Unique fortran name Description Slope of the pier from the LSE mapping Distribution parameter 1 Variation coefficient Distribution NETS Distribution parameter 2 Standard parameter 3 uS name Deviation a PierSlope i i Da4 2a downstream horizontal lt 75 A Area Area m Cb1 2a Coefficient used in various
132. plain m Mo model uncertainty factor my model uncertainty factor for damping Loading equations Resistance strength equations Ah h h h Ywa Vw Yw Parameter definitions Ywet saturated volumetric weight of the impermeable soil layers Yw volumetric weight of the water d thickness of the impermeable layers h water level on the river m hy water level in the floodplain m Sources of failure mechanism equations methods T07_08_02_Reliability Analysis_D7_1_ Appendix 10 April 2008 87 Task 7 Deliverable D7 1 Appendices 1to 5 FLOOR Contract No GOCE CT 2004 505420 ok Vrouwenvelder et al 2001 Sources of uncertainties in failure equations input parameters Vrouwenvelder et al 2001 Remarks T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 88 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505420 5 Onsite APPENDIX V 3 EXTENDING THE CALCULATOR The key point about the Extensibility of the Calculator is that it is driven by the content of the four files defined in the Framework document Parameter csv FailureMode csv FailureModeParam csv Structure csv These files must be in the same folder as the file ReliabilityCalc xls These are simple CSV files and code has been supplied to TU Delft to enable them to generate the files from their internal spreadsheets They can also be edited using simple text editor
133. ppletion policy is pursued along the whole North Sea coast for both the dunes and the dikes to maintain the basic coastline VNK can t directly calculate such dikes One should assume a coupled failure mechanism the dike 1s addressed only after the dune is swept away Sliding outer slope Stability outside the dike dike and shore drops is not considered by VNK The water board expects that especially this mechanism is a threat to the safety of dike ring area 32 and consequently has a large influence on the probability of flooding Sliding of the outer slope occurs at low tide Depending on the degree of sliding this leads to a threat to safety or not The water board indicates that dike ring area 32 has a closed system of regional flood defences with closable constructions to counteract this phenomenon This system is controlled and maintained by the water board Results per structure The results per structure are given in table 4 5 The results that can be left out of consideration in connection with consult with the water board are shaded grey here as well No Structure Overflow Non Structural and closure failure overtopping 1 Pumping station Cadzand 4 4 6 0 4 5 2 Pumping station Campen 5 5 4 4 3 Pumping station Nieuwe Sluis 5 9 6 1 4 Pumping station Nummer Een 6 0 4 9 5 Pumping station Othene 4 3 5 1 7 6 Pumping station Paal 5 0 5 8 4 7 7 Sluice station Terneuzen Oos
134. r 1 parameter 2 Parameter Unique fortran name Description LSE mapping i E d parameter 3 distribution Variation Standard S name coefficient Deviation Bc2 1j model uncertainty factor mR MBC2 IJR Bc2 1j E strength Bc2 1j model uncertainty factor ms MBC2 1JS Bc2 1j P loading Bc2 1k model uncertainty factor mR MBC2 IKR Bc2 1k strength Bc2 1k model uncertainty factor ms MBC2 IKS Bc2 1k loading Bc2 1m model uncertainty factor mR MBC2_1MR Bc2 1m B strength Bc2 1m model uncertainty factor ms MBC2 1MS Bc2 1m F loading Bc2 3a model uncertainty factor mR MBC2_3AR Bc2 3a strength Bc2 3a model uncertainty factor ms MBC2 3AS Bc2 3a B loading Bc2 3b model uncertainty factor mR MBC2 3BR H Bc2 3b strength Bc2_3b model uncertainty factor mS MBC2_3BS m Bc2 3b i loading TU7 OB 2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 muousife Contract No GOCE CT 2004 505 20 Distribution Distribution SET Distribution Example arameter 1 parameter 2 Parameter Unique fortran name Description LSE mapping i E d parameter 3 distribution Variation Standard S name coefficient Deviation Bc3 la model uncertainty factor mR MBC3_1AR Bc3 la E strength Bc3 1a model uncertainty factor ms MBC3 1AS F Bc3 la
135. r Ording is a large community at the Schleswig Holstein North Sea coast with the character of a tourist seaside resort The community is located on the west exposed coast of Eiderstedt peninsula Figure 3 6 The size of the study area is approximately 6000 ha from these about 4000 ha are considered to be flood prone with the respective height distribution NN Ordinance Datum regional Mean Water Level T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 60 Task 7 Deliverable D7 1 Appendices 1to 5 j Contract No GOCE CT 2004 505420 PLOOESZe I NACE TT CRANE Forschungs und Technologlezentnuem Viestk ste AQ Kumengeograpzis Dip Geegr Stetan Rene eae 05 6 0 amp 8 4 15 2 DOK S und Guantebarcerungen Q Figure Error No text of specified style in document 12 Map of pilot site German Bight red line illustrates the coastal dike The territory of the community amounts to 2800 ha with about 6300 inhabitants In this area the irregular topography with intermittent small hills and dunes makes it difficult to draw flood distance boundaries Presently flood protection is provided by a major dike 12 5 km long about 8 0 m high as well as dune structures 800 m about 10 m and up to 18 0 m high surrounding the community on three sides over a length of more than 15 km The height of the dike line is not constant as shown in Figure 3 7 T07 08 02 Reliability Analysis D
136. r that much according to the water board both weak and 7008 has later been converted into a dune Consequently the total number of dike sections amounted to 33 Adjusting profiles The profiles used in the first calculation in PC Ring were based on old measurements by the water board Next to that several adjustments were done in the profile in the first calculation to be able to calculate them in PC Ring without data of the water board at hand to check the adjustments in the profiles Because of that the input profiles were still compared to the recent measurements provided by the water board The recent measurements of the water board are based on a hectometering of the dike after a recent merging the water board switched from dike pole numbering to hectometering The dike pole numbering has been re numbered to a hectometering based on a conversion table provided by the water board With this a difference occurs in the exact position of the dike profiles of less than 50 meters On a location a difference of 80 meters occurs From the comparison it appeared that there were differences between the schematization and the recent profile measurements at several points For more than one profile the crown height differed 20 to 70 cm For more than one profile there were differences in sloping On several points the profile type in PC Ring didn t quite match reality In consult with VNK it was decided to adjust all 33 prof
137. r water level and with that a much smaller probability of failure than the water level at which failure overrunning of the structure occurs T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 43 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505420 Onsite ju F Overall probability of flooding dike ring 32 Let us assume that a dike stretch of length L is schematised into n sections by Ax L n If the following autocorrelation function for the dike strength R at section x is assumed ie p R x R x Ax e And the reliability index for the i th section is beta for i 1 n P F 2 6 B Then we can write the overall failure probability as P F 6B n 1 4 6B 20 v p zt 8j AX Since max P F en F P F en F and p e a x rit as well as j i 2 p 1 whereas u Lg forsmallu 2 adm Therefore P F c t d and Led ae pi nd Jn dx T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 L since Ax 2 n gt n 44 Task 7 Deliverable D7 1 Appendices 1to 5 Eran erg Contract No GOCE CT 2004 505420 fkOOP sic Therefore pepe E PF 9 bit Which is independent of the number of sections n If all results from table 4 1 and table 4 5 are taken into consideration a preliminary probability of flooding of gt 1 11 per year COMBIN 1 is calculated for dike ring area 32 Zeeuws V laanderen
138. rd Thames c Landward Figure 5 The three reinforced concrete wall types implemented in the reliability analysis left a picture of reinforced concrete walls along the flood defence line right Table 2 Overview of site specific failure processes and failure mechanisms implemented in the reliability analysis Site specific failure processes Failure mechanisms implemented in reliability analysis Damage by residential developments concrete cracking Uplifting and piping underneath overall earth embankment only for types 1 and 2 Bal 5aii and Bal 5aiii Sliding of the concrete wall Cc1 2aii Overturning of the concrete wall Cc1 2b Reinforcement failure in the vertical concrete slab Cc1 2c Shear failure in the vertical concrete slab Cc1 2d Piping directly underneath seepage screen Cc1 5 joint failure and settlements T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1 to 5 Een cis Contract No GOCE CT 2004 505420 FLOOPSHe T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1 to 5 Contract No GOCE CT 2004 505420 fkOUP Sie Breach Structural failure of the concrete f E I Y 1 Instability of the concrete wall T Piping directly underneath concrete sheet pile toe Piping underneath embankm
139. red important enough to incorporate The failure mechanism are elaborated in the following sections The remaining part of this appendix 1s based on Steenbergen and Vrouwenvelder 2003A and Steenbergen and Vrouwenvelder 2003B A list with all the random variables in PC RING is provided in Appendix B The variable numbers in Appendix B correspond to the variable numbers below 11 1 General The geometry parameters apply to more than one failure mechanism The geometric variables are listed in Table A 1 Table A 1 General parameters Steenbergen and Vrouwenvelder 2003B Variable nr symbol description 1 ha Dike height 4 hi Toe height 5 tan Qu b Angle outer slope top 6 tan Quo Angle outer slope bottom 7 tan i Angle inner slope 2 hg Berm height 3 B Berm width 131 Ad Error in determination ground level 1 1 1 3 Overflow overtopping The mechanism overflow overtopping occurs in case to much water is flowing or topping over the dike see Figure A 1 Failure due to overflow overtopping occurs either if the revetment of the inner T07_08_02_Reliability_Analysis_D7_1_Appendix 10 April 2008 66 Task 7 Deliverable D7 1 Appendices 1to 5 AAN cit Contract No GOCE CT 2004 505420 FLOORS ite slopes fails or due to saturation of the inner slope Saturation occurs when the overflow overtopping discharge is larger than the critical discharge and when the inner slope slides overflow
140. rs E Floodplain Zone 3 h hc2 he2 owe Zone 2 hcl lt h lt hc2 Groundwater level in floodplain Zone 1 h hcl 7 Water conductive CF AEA SAPP PP RAAP EES sand layer in contact with the river Thames Figure 3 Representation of double crested earth embankments Characteristics of process models or fault trees change according to the three different water level zones T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1 to 5 AAN crf Contract No GOCE CT 2004 505420 FLOORS The embankments are generally founded on impermeable layers overlaying a water conductive sand or gravel layer At some locations the water overpressures in the sand gravel layer are drained by a pipe see figure 3 The failure processes associated with the embankments along the Dartford Creek to Gravesend flood defence system are listed in table 1 along with the failure mechanisms that are implemented in the reliability analysis Table 1 refers to the failure mechanisms in the Task 4 Floodsite report The process models for grass erosion are slightly different from those applied in the reliability analysis Table 1 An overview of the site specific failure processes and the failure mechanisms included in the Dartford Creek to Gravesend reliability analysis Site specific failure processes Failure mechanisms in reliability analysis Overtopping overflow causing er
141. s failure g X gt 0 indicates safe region and g X 0 denotes the limit state Using Monte Carlo simulations an estimate of the failure probability P is obtained as P IteQO SO GOdx 3 Ie X 09 Here is an indicator function which takes values of unity when g X 30 and zero otherwise The minimum number of samples required for target coefficient of variation V P is given by N 3 ee 2 POPE es Thus it follows that to reduce the estimate of variance to acceptable levels for low failure probability levels sample size N needs to be large This has led to the development of a number of variance reduction techniques Kahn 1956 In implementing the importance sampling technique Eq 1 is rewritten as _ eX S 0 py x P zi he x hy x dx 3 and an estimate of the failure probability is obtained as 1 Sifg X lt 0 rA p pO 4 i l T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 25 Task 7 Deliverable D7 1 Appendices 1to 5 muousite Contract No GOCE CT 2004 505420 FLOOD Site Procedures that estimate P with specifically chosen hy x as sampling density functions are called important sampling procedures and Ay x is called the importance sampling function Here the sampling is done in the y x region rather than px x A major step in implementing the procedure lies in choosing an appropriate importance sampling probability density function hy x The importance sa
142. s such as Notepad However a worksheet is also included for each file to enable you to edit the files within the Reliability Calculator Each sheet shows the values from the file and has a Load button and a Save button You need to be aware that the values in the sheets do not necessarily reflect those in the file the file may have been edited outside the spreadsheet or the user didn t click Save having made some changes So before you do any work on one of these sheets click Load This will load the values from the file into the sheet When you have made some changes click Save This will save the changes to the csv file When you next move to the Calculate sheet click Reset to tell the Calculator to use the changed files Any additional advice for specific files is given in the following sub sections Structure Initially only one example structure SheetPileWall will be available from the structure drop down list in the Calculate sheet To add additional specific structure fault trees to the calculator the name of the structure and the name and location of the fault tree text file fta file must be entered as a list in the Structure File sheet of the Reliability Calculator When the Save button is clicked this information is saved to the Structure csv file Failure Mode If you add a Failure Mode to FailureMode csv you must also add one of the same name to the OpenFTA Event Database FailureMode ped This will all
143. score represents the score for stability of the inner slope In case of an even score one can assume that the overflow and wave run up mechanism is governing For the covering damage and erosion body of a dike mechanism the result of the old testing is not provided The calculated probabilities of failure for this mechanism are discussed during consults with the water board and related to the temporary results of the new testing see section 4 7002 7009 7023 7024 7025 7028 7038 7042 7047 7053 7071 7074 7075 7094 7109 suf insuf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf insuf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf 7111 7116 7124 7129 7136 7139 7152 7159 7163 7167 7185 7202 7211 insuf insuf insuf suf suf suf insuf suf suf suf suf suf insuf insuf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf suf 7220 7233 7249 7258 7271 suf suf suf suf suf insuf insuf suf suf insuf suf suf suf suf suf Table 2 2 Assessment of the water board for dikes in dike ring 32 the first row shows the section number the second row the Ht_score which represents the score for overflow
144. stochastic variables are looked at that have the largest contribution to the probability of flooding On top of that it 1s important that these stochastic variables can be decreased by means of further research in reasonable time and with reasonable effort The latter is an important restriction for dike ring 32 the stochastic variable Water level Vlissingen contributes most by far to the probability of failure for the mechanism overtopping and wave overrun It is however a stochastic variable for which further research will generate little new insights Even 10 years of additional observations will only be of limited influence on the stochastic variable insecurity with which this stochastic variable is afflicted Decreasing the probability of flooding by reducing insecurities by means of additional research will thus have to focus on other stochastic variables Information on the most influential stochastic variables can be derived from PC Ring PC Ring calculates an influence coefficient alpha per stochastic variable also called sensitivity coefficient The magnitude of the alpha value is determined by a combination of the influence of the average value and the magnitude of the standard deviation or variation coefficient A low alpha value for a parameter does not inherently mean that this parameter has little influence on the result For a small variation coefficient or standard deviation the variation of the average value can still have a s
145. strength model gt 6 as in 42 T07_08_02_Reliability Analysis_D7_1_ Appendix 10 April 2008 59 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505420 Gasite ju F Ill Appendix 3 Details of the PRA German Bight Within Action 3 of Activity 1 a preliminary reliability analysis PRA of the pilot site German Bight Coast was performed the results of which are summarised in Kortenhaus amp Lambrecht 2006 The reliability analysis was performed using the German ProDeich model for coastal dikes as described in Kortenhaus 2003 and laser scan data of the flood defences made available by the coastal authorities of Schleswig Holstein This section describes the approach to derive the overall probability of failure for all flood defences in the area This comprises e a description of the flood prone area and the flood defence structures e the methodology to obtain geometrical parameters from laser scan measurements of the defence line e the development of an algorithm how the defence line can be split into different sections which can be treated independently e the calculation of the failure probability for each section of the flood defence line The methodology applied here is following the source pathway receptor model used in FLOODsite The result of assessing the risk sources and the risk pathways is the probability of the flood defence failure as highlighted in this figure St Pete
146. t NET Framework v2 0 50727 4 Start a Command Prompt window and run the following command C windows Microsoft NET Framework v2 0 50727 regasm codebase RelCalcFolder RelCalc dll Replace C windows Microsoft NET Framework v2 0 50727 with the path where you found regasm exe 5 Skip this step unless you detected NET Framework 3 0 or 3 5 at step 0 this step needs administrator permissions Locate the folder containing EXCEL EXE maybe C Program Files Microsoft Office OFFICE11 Copy the file Excel exe config from RelCalcFolder to this folder This file ensures that NET objects used by Excel Visual Basic use NET Framework 2 0 If there is an existing Excel exe config file please seek advice before replacing it 6 Edit the file RelCalcFolder Structure csv with a Text editor such as Notepad or TextPad Replace the string d work Reliability with the path for your RelCalcFolder and save the file 7 Start Microsoft Excel and open the file ReliabilityCalc xls in RelCalcFolder You can expect to see some error messages until you carry out the following T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 84 Task 7 Deliverable D7 1 Appendices 1to 5 mn it Contract No GOCE CT 2004 505420 FLOODS ile The following instructions are valid for Excel 2003 There should be corresponding features in other Excel versions To ensure that you can run Visual Basic Macros choose Tools Options Security Macro Security
147. t for strength stone pitching on clay c k 0 0000 Coefficient for erosion resistance grass c g 0 0000 0 0000 Coefficient for erosion resistance of covering layer c RK 0 3445 0 1187 Coefficient for erosion resistance of the dike core c RB 0 0000 Thickness granular filter layer d f 0 0000 Grain size 15 percentile weight filter material D f15 0 0000 Crack width s 0 0000 Coefficient for strength stone pitching on filter c f 0 0000 Coefficient in determination leakage length c a 0 0000 Coefficient in determination leakage length c b 0 0000 Coefficient in determination leakage length c t 0 0000 Thickness asphalt layer D 0 0000 Relative density asphalt layer 0 0000 Factor f MGWS 0 0000 Heighth GWS 0 0000 Angle of wave attack Beta r 0 0005 0 0000 Coefficient for strength stone pitching c gf 0 0000 Measure of erosion acceleration in dike core alfa z 0 0000 0 0000 Measure of erosion decrease with height alfa h 0 0000 0 0000 Coefficient c 0 0000 Height of fictive bottom h fo 0 0000 Parameter b 0 0000 Nominal average diameter of pitching D 050 0 0000 Upgrade factor Psi u 0 0000 Stability parameter Phi sw 0 0000 T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 80 n Me Us ROGBsite Error in bottom determination Delta d 0 0000 0 0000 Error in local water level Delta hlok 0 0000 0 0000 Error in wave direction beta 0 0084 0 0001 Storm duration t s 0 2811 0 0790 Dune height h d 0 0000 Model factor m D 0 0000 Median grai
148. t height along the entire stretch of the dike is modeled as a Gaussian random process with a specified auto correlation function The length of the dike is discretized into smaller segments The dike crest height is assumed to be constant throughout each segment and is modeled as a Gaussian random variable The probability of overtopping is calculated for each segment using the performance function in Eq 5 The dike segments are assumed to be in series and the bounds on the failure probability estimates for the series system are obtained Cornell 1967 Extreme discharge Distribution Other variables distribution Extreme Sea level Distribution Monte Carlo based simulation Code to calculate pf Figure 3 2 Probabilistic loops through hydrodynamic model for stochastic simulation T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 26 Task 7 Deliverable D7 1 Appendices 1to 5 mn 7 Contract No GOCE CT 2004 505420 ThLOGD site During Monte Carlo simulations first an ensemble for the random variables are generated and deterministic calculations are carried out using the hydrodynamic model is necessary for each realization Figure 3 2 illustrates a schematic diagram of the simulation procedure and loop through hydrodynamic model The computation time for one sample realization through the hydrodynamic model is non trivial An importance sampling based Monte Carlo approach is adopted for estimating the probability of d
149. t is one or the other because for extending the line of the lower slope the berm width changes it is better to adjust the gradient of the lower slope Applying a bend in the outer slope is done in such a way that the gradient of the upper slope doesn t change The upper slope point is used as point of inflection This is done for Ds numbers 7094 and 7159 For the point on the inner slope one assumed the first point on the inner slope that is given by the water board Of Ds number 7028 the adjustment is done differently in order to be able to fit the profile in one of the schematizations The berm has been lengthened increasing the gradient of the upper slope and can be considered a slope Other options for adjusting would mean adjustments for several points due to which the profile would differ even more from reality Ds number 0747 can t become category 8b because the gradients of the crown are too high in case of an extra point on the crown Thus one did eventually decide for a category 7a Of the sections 7047 7094 7139 one would say that there s a bend in the crown Copying 1 on 1 however means that the gradients of the crown planes become too steep steeper than 1 15 Getting the gradients below 1 15 however means that one has to adjust the points in such a way that it either won t work or the bend becomes next to nil In these cases one has chosen for a flat crown and the levels are adjusted in such a way that th
150. the PATH environment variable LSEFuncName The name of the LSE Function Failuremodeparam csv The file Fai ureModeParam csv contains one or more lines per Failure Mode and identifies which Parameters are used by the LSE Function for the Failure Mode Each line has 2 columns as follows Column Description FailureModeName The unique name for the Failure Mode This name must be one of those included in the Failure Mode file see 0 ParamName The name of a Parameter used by the LSE Function for this Failure Mode T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 148 Task 7 Deliverable D7 1 Appendices 1 to 5 FLOOR kn Contract No GOCE CT 2004 505420 The name must match one of those in the Parameter file see 0 This file may not be used in early implementations of the Reliability Tool software Instead a sheet will be provided for the user showing a grid of Parameter Names versus Failure Modes T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 149 Task 7 Deliverable D7 1 Appendices 1 to 5 ita Contract No GOCE CT 2004 505420 FLOODS ite Structure csv The file Structure csv contains one line per structure type Each line has 2 columns as follows Column Description StructureName The unique name for the Structure Type This name will be used by the Reliability Tool user FTAFile The filename of the OpenFTA file which defines the relevant Failure Modes and the logic which links t
151. the dike is assumed to have spatial uncertainty variation A Monte Carlo simulation based approach is considered for the reliability analysis of the dike The computation of the local water level involves calculation through a computationally intensive hydrodynamic model and is carried out using commercially available software Efforts to reduce computational time in the reliability analysis are explored through the use of importance sampling technique Further reduction in computational efforts is achieved by adopting a novel response surface based method This strategy involves using available response database for the local water levels corresponding to observed boundary conditions In the importance sampling based Monte Carlo simulations carried out in this study the local water levels are computed by interpolating from the available response database rather than using the hydrodynamic model The proposed method is observed to bring about significant reduction in computational efforts Introduction The reliability analysis of a dike at a lower reach of the tidal Scheldt river is considered In this study it is assumed that dike failure occurs due to overtopping only Overtopping of the dike is assumed to take place due to a extreme sea levels b extreme river discharge and c coincidence of both of the above extremal events This has been illustrated by the schematic diagram in Figure 3 1 The stochastic nature of the input variables in this c
152. the following aspects Length of the dike section Height of the crown Height of the toe Orientation of the dike section Presence of shoulder and or bend in other words type of dike section Dike covering The results of the already calculated overflow wave run up and bursting piping of PC Ring are considered for the choice of dike sections The dike sections with a significant higher probability of failure have been selected It was decided to add two more weak links in consultation with the District Water Board Zeeuws Vlaanderen These are dike sections 7009 and 7023 This brings the total number of sections that are taken into account in PC Ring to 37 of which 33 dike and 4 dune sections This T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 17 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505420 FLOORS number is without the water retaining structures 14 structures The location of the selected dike sections is shown in figure 2 1 in which dike section 2 represents dike section number 7002 etc The selected dune sections are given in figure 2 2 dune section 8 represents dune section number 7008 etc Figure 2 1 Selected dike sections Figure 2 2 Selected dune sections The 33 dike sections are numbered according to the following distances in kilometer
153. tion of structure based on the contributions of the various failure mechanisms Consequently the probability contributions of the various dike sections and structures are combined into the probability of flooding of the dike ring With combining the various contributions possible dependencies in probabilities of failure of nearby dike sections are accounted for T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 35 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505420 5 Onsite Process description The collecting of data on dike ring 32 is done by the water board in cooperation with VNK The quality of the data is checked by both VNK roughly and the Bouwdienst during the conversion of the data from the overall spreadsheet to the database The result of this is recorded in various checklists and reports overall spreadsheet dike ring 32 With executing the first calculations for dikes and dunes several adjustments to the PC Ring database were performed The greatest adjustments concerned the selection of dike sections see section 2 3 and the schematization of the dike profiles With the selection of dike sections 33 dike sections and 4 dune sections were chosen out of 287 sections that were schematized by the water board With the schematization of the profiles the schematized profiles done by the water board in the PC Ring database were compared with recently measured cross sections of the water
154. tsluis 4 6 6 5 3 8 Sluice station Terneuzen Middensluis 3 9 4 8 4 3 schutsluis T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 42 Task 7 Deliverable D7 1 Appendices 1to 5 ELAAN eita Contract No GOCE CT 2004 505420 FLOOR Size 9 Sluice station Terneuzen Middensluis 5 1 4 3 spuiriool 10 Sluice station Terneuzen Westsluis 3 9 5 3 5 2 11 Sluice station Terneuzen Westsluis 5 2 5 2 spuiriool 12 Discharge sluice station Braakman 4 7 4 3 4 5 13 Discharge sluice station Hertogin 4 0 4 7 5 2 Hedwigepolder 14 Discharge sluice station Nol Zeven 4 6 4 5 4 5 Table 4 5 DHV Results of the assessed structures in dike ring 32 Structures The pumping station Othene scores very bad for the mechanism constructive failure beta of 1 96 probability of failure 1 22 This has to do with the mechanism bursting and piping This appears to be a problem if one assumes that the ground sills and aprons are not fully watertight In case one can prove this is the case or if physical measures are taken to achieve this the norm can be complied with 2 In consultation with GeoDelft this structure has been tested correctly in the meanwhile The result can thus be left out of consideration For the mechanism not closing the pumping station scores relatively bad beta of 3 5 probability of failure of 1 4300 Not closing results in a high probability of failure due to the large number of requests f
155. ucture using these values for the parameters passed to the LSE functions and determines whether or not the structure will fail After a number of samples the non annualised failure rate is simply the number of failures divided by the number of samples After Min samples and every Interval samples the failure rate is calculated and compared with the previously calculated value If the 2 failure rates satisfy the following relationship then convergence has been detected at this interval T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 145 Task 7 Deliverable D7 1 Appendices 1 to 5 FLOOR Contract No GOCE CT 2004 505420 Ratel Rate ani lt Factor Factor is the value of the Convergence Factor If Rate2 is very small less than Factor then the relationship is Abs Ratel Rate2 lt Factor In normal use once convergence has been detected then the calculator has a value However you can ensure tighter convergence control by supplying an integer value greater than one for Successive Intervals In this case convergence must be detected for that number of successive intervals before a final value is determined This can be used if you suspect that convergence is being detected too early by chance The Calculator will also terminate if it reaches MaxSamples The rate of failure value at the last interval will be returned 6 Review the other Control parameters and change them if you wis
156. ul url as Distribution arameter arameter Parameter Unique fortran name Description LSE mapping A E NS d parameter 3 distribution Variation Standard name coefficient Deviation Cc1 2aii Ccl 2b Cc1 2d Level of elevation of ground Cb1 2a Cb1 2c Cb1 2d h3 h3 behind concrete wall on Ccl 2aii Ccl 2b normal 1 0 1 landward side Ccl 2d 1 Vrouwenvelder et al 2a CUR 140 2b CUR 141 2c CUR 162 2d CUR 190 3 IGBE 4 Christian amp Baecher 5 Leidraad 6 lecture notes CT5310 probabilistic design in hydraulic engineering TU7 0B Q2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 muousife Contract No GOCE CT 2004 505 20 T07_08 2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1 to 5 FLOOR ita Contract No GOCE CT 2004 505420 WM APPENDIX V 5 DETAILED INSTRUCTIONS FOR INTERFACE OPERATION The Calculator is started by opening the EXCEL spreadsheet ReliabilityCalc xls There are five tabbed worksheets of which the prime one is that labelled Calculate as shown in Figure 3 The remaining four are used to update the files defining Failure Modes Parameters etc These may be used for extending the Calculator see Appendix 3 In outline the process is 1 Q 3 4 5 Click Reset button if any of the underlying files Parameter csv FailureMode csv FailureModeParam csv Structure csv have cha
157. ve Ba2 4b overtopping u g delta Dimensionless flow velocity DimFlowU Ba2 4iii Dn50 rc RcBend radius of curvature of bend Ba3 1 Area of the saturated part of the A sat ASat Bb1 2 i embankment Area of the unsaturated part of A_unsat AUnsat Bb1 2 the embankment Area of the fine grained soil A_fg AFg Bb1 2 T underneath the embankment indi choice for 0 drained or indicato DrainUndrain i n Bb1 2 deterministic drained undr 1 undrained condition TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 mn AR Contract No GOCE CT 2004 505 20 FLOGS Distribution Distribution Fxample ied ues Distribution 3 Ao arameter arameter Parameter Unique fortran name Description LSE mapping M z 5 PRA p parameter 3 distribution Variation Standard name coefficient Deviation ained k Admissable wave overtoppin l s m Qadm Qadm EE Ba2 5 rate P AngleShearGap Angle of shear gap rad or Ba2 3 fF Ff Coefficient Ba2 3 Admissable cohesion in local Cu adm CuAdm Ba2 3 clay failure due to wave impact angle of wave obliquity for TU Ba 1b Ba2 Ibii beta r betaR which reduction is taken into P ES Ba2 1biii account flow velocity at the riverside v0 Vel0 Bal 6 crest of the embankment Cw CoeffW Coefficient Bal 6 CL CoeffL Coefficient Bal 6 CR CoeffR Coeffic
158. velder 2003B Variable nr symbol description 41 d Thickness covering layer 137 hy Inner water level 49 at wy Apparent relative density of heaving soil 50 Hl Yu Relative soil density sand grain 43 L Leakage length 42 D Thickness sand layer 45 x dio Factor Chear 47 d o d4o Uniformity 44 0 rolling resistance angle 46 dzo Grain size 48 n White s constant 54 k Specific permeability 51 Mo Model factor heave 52 Mp Model factor piping 53 mn Model factor water level damping P Clay J Figure A 4 Part of the variables in heave piping Steenbergen and Vrouwenvelder 2003B 1 1 1 6 Erosion revetment and erosion dike body The mechanism erosion revetment dike body occurs when first the revetment of a dike is eroded and secondly the body of the dike is eroded away see Several types of revetment have been considered grass stone pitching without filter stone pitching with granular filter and asphalt Erosion outer slope Figure A 5 Erosion revetment and erosion dike body T07_08 02 Reliability Analysis D7 1 Appendix 10 April 2008 69 Task 7 Deliverable D7 1 Appendices 1to 5 FLOOD Contract No GOCE CT 2004 505420 The following variables above the geometry variables apply to the mechanism erosion revetment and dike body see Table A 5 For more information about this mechanism is referred to Steenbergen and Vrouwenvelder 2003
159. verage diameter revetment 91 Va Revaluation factor 92 Psw Stability parameter 1 1 1 7 Piping structures Piping of structures occurs in case sand below a hydraulic structure for instance a sluice is flushed away due to a hydraulic gradient see Figure A 6 T07_08_02_Reliability_Analysis_D7_1_Appendix 10 April 2008 70 Task 7 Deliverable D7 1 Appendices 1to 5 ta Contract No GOCE CT 2004 505420 MUOESie Piping structure Figure A 6 Piping under a structure FLORIS 2006 The following variables above the geometry variables apply to the mechanism piping structures see Table A 6 For more information about this mechanism is referred to Steenbergen and Vrouwenvelder 2003B Table A 6 Variables for piping structures Steenbergen and Vrouwenvelder 2003B Variable nr symbol description 114 m Model factor 115 Me Model factor 111 Ly Vertical leakage length 112 Lh Horizontal leakage length 113 CL Lane s constant 137 hp Inner water level 1 1 1 8 Structure not closed The failure mechanism structure not closed occurs when the structure is not closed and when there is too much water flowing through the structure for the surface of the retention area behind the structure see Figure A 7 No closure h Figure A 7 Structure not closed FLORIS 2006 T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 71 Task 7 Deliverable D7 1 Appendices 1t
160. verlying a water conductive gravel or sand layer T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1 to 5 mn 4 Contract No GOCE CT 2004 505420 POOR Sie Upstream Thames Downstream Thames 8 towards London towards Southend 7 5 SARE Overall just improved Overall survey 92 earth 92 Concrete 92 Sheet 92 45 4 0 000 2 000 4 000 6 000 8 000 10 000 Di ink Dartford Creek eee Figure 2 Elevation of the defence line between Dartford Creek to Gravesend after 70s 80s improvements in black versus the recently surveyed defence line dashed purple The latter indicates the stretches of the different flood defence types 2 Structure types and failure mechanisms 2 1 Earth embankments The primary function of earth embankments is flood defence Two types of earth embankments occur along the Dartford Creek to Gravesend defence line a combination of a riverward and landward earth embankment referred to as double crested and the regular earth embankment referred to as single crested Figure 3 shows a drawing of the double crested embankments The basic failure mechanisms and equations of the single and double crested earth embankment are similar Differences occur between fault trees and some of the details in the failure mechanisms River Thames Impermeable compressible laye
161. ves 0 or ShipWindWaves f Bc2 1b wind waves 1 MBAI 5BIIR Bal 5bii model uncertainty Bal 5bii TU7 OB 2 Reliability Analysis D7 1 Appendix 10 April 2008 Task 7 Deliverable D7 1 Appendices 1to 5 Contract No GOCE CT 2004 505 20 Parameter Unique fortran name Description factor strength LSE mapping Example distribution Distribution Distribution K Distribution parameter 1 parameter 2 Standard Deviation A parameter 3 Variation x name coefficient Bal 5bii model uncertainty MBAI SBIIS Bal 5bii B factor loading Bal 5dii model uncertainty F MBAI 5DIR Bal 5dii n factor strength Bal 5dii model uncertainty MBAI SDIIS 5 f Bal 5dii E factor loading Df50 Df50 50 percentile in the filter Bal 5c FineGrainLayer Layer thickness of fine grained Bb1 2 Bottom level at deep water to BottomLevelDeep Ca2 2a derive water depth at deep water the type of embankment OR A22 4 Bal 1 Ba2 5 num Num determining the type of T i f Bal 5bii Bal 5dii overtopping model the number of knots in the grid in no k hor KnotsNumber d Ab2 1b Bc2 1c n horizontal direction dimension concrete wall sheet s dl dl i Cc1 2aii Cc1 2b normal 2d 0 004 pile wall dimension concrete wall sheet d2 d2 Cc1 2aii Ccl 2b normal 2d 0 004 pile wall TO7_08 O2 Reliability Analysis D7 1 Appendix 10 April 2008 Tas
162. with FORM DS initial value 8 or 1 resulting from action for 23 42 74 116 124 139 152 163 167 185 202 211 and 233 For the following sections no or odd results have been calculated 28 covering beta is 0 2057 109 covering 136 overtopping wave overrun covering 159 overtopping wave overrun 167 overtopping wave overrun 159 167 Adapt initial value from 1 8 then a result for overtopping wave overrun adapting manually per mechanism Thus covering with initial value 1 and overtopping wave overrun with initial value 8 109 136 Calculate with DS FI with initial value 8 gives a result for all mechanisms T07 08 02 Reliability Analysis D7 1 Appendix 10 April 2008 58 Task 7 Deliverable D7 1 Appendices 1to 5 mn Contract No GOCE CT 2004 505420 FLOODS ite 28 If 3 residual strength not relevant is chosen in stead of 6 for the type number of the residual strength model the same beta is calculated Thus there are no data concerning residual strength in calculation Based on 42 more data adapted Measure for acceleration of erosion in core of the dike 1 equal to the cover layer gt 3 sand core Crack width 0 001 gt 0 015 is standard in overall spreadsheet was not put in for 28 Relative density stone average value 1 gt 1 8 Thickness granular filter layer 0 04 gt 0 1 Grain size 15 fraction of the filter material 0 001 gt 0 02 is standard in the overall spreadsheet was not put in for 28 Residual

T07-08-02_Reliability_Analysis_D7_1_Appendix

Contents

Download Pdf Manuals

Related Search

Related Contents