84661385 Deter Minis Tic Fatigue Analysis How To

Deterministic Fatigue Analysis How To - Using SACS 5.1

25-Jun-02

Projects Oil & Gas Level 6, QV1 250 St Georges Terrace Perth, Western Australia Australia 6000 Tel: +61 8 9278 8111 Fax: +61 8 9278 8110 Web: http://www.worley.com.au Worley Pty Ltd ABN 61 001 279 812 © Copyright 2002 Worley Pty Ltd

HOW TO DO A DETERMINISTIC FATIGUE ANALYS IS HOW TO - USING SACS 5.1

SYNOPSIS This document describes the steps taken in performing a deterministic fatigue analysis using the SACS 5.1 package. It should be noted that at the time of writing, the most recent edition of API RP2Ast WSD (21 Edn, 2000) recommends that a spectral (rather than deterministic) fatigue analysis be undertaken. This is particularly important for small platforms with less redundancy (eg tripods or monopods) and/or large natural periods. This How-To should be used in conjunction with the SACS IV, Fatigue and Seastate manuals.

Disclaimer

This report has been prepared on behalf of and for the exclusive use of Worley. Worley Pty Ltd accepts no liability or responsibility whatsoever for it in respect of any use of or reliance upon this report by any third party. Copying this report without the permission of Worley Pty Ltd is not permitted.

PROJECT 000/00000 - HOW TO DO A DETERMINISTIC FATIGUE ANALYSIS REV

A

DESCRIPTION

ORIG

Issued for internal review

REVIEW

WORLEY APPROVAL

N/A M Stock

DATE

25-Jun-01

CLIENT APPROVAL

N/A

DATE


CONTENTS 1.

Introduction ................................................................................................ 1

2.

Setting Up .................................................................................................. 2 2.1

SACS Files Required ..................................................................................................... 2

2.2

Create Fatigue Model File ............................................................................................. 2

2.3

Create Fatigue W ave Definition Files ............................................................................2

2.4

Calculate DAFs for Fatigue Waves ............................................................................... 4

2.5

3.

2.4.1

Calculate Centre of Damage Wave Height .............................................................. 4

2.4.2

Create Centre of Damage Superelement .................................................................5

2.4.3

Run Modal Analysis with Centre of Damage Superelement ....................................5

2.4.4

Calculate DAFs .........................................................................................................5 Create Fatigue Run Input File........................................................................................5

Running the Fatigue Analyses ................................................................... 9 3.1

Create CSF Files ........................................................................................................... 9

3.2

Combine SCF Files for Fatigue Life Determination .......................................................9

Appendices


1.

INTRODUCTION

Deterministic fatigue analysis involves specifying many fatigue waves (several heights from each direction), generating corresponding stress cycles in each joint, and using height- and directionspecific wave frequency data and Miner’s rule to determine the amount of cumulative damage done to each joint through application of an S-N curve.


2.

SETTING UP

2.1

SACS Files Required

Before starting to set up a deterministic fatigue analysis, the following files will be required:

1.

Inplace analysis model file (e.g. sacinp.dat)

2.

Soil properties file (e.g. psiinp.dat)

2.2

Create Fatigue Model File

A fatigue model file with no Seastate loading is required, in order to apply the fatigue waves and determine the resulting stresses in each joint from each wave. This file is created by following these steps:

1.

Save a copy of the Inplace file as ftginp.dat

2.

Open this file and remove the first LOAD line and all lines after this line. Also, remove any Seastate and load case lines at the start of the file (such as LCSEL, HYDRO, UCPART, AMOD AND LDOPT) such that the first two data lines of the fatigue model file (ftginp.dat) are OPTIONS and SECT. Leave member & group overrides, marine growth definitions and so on in this file (see below).

3.

The SACS Fatigue Module manual recommends including DL in columns 19-20 of the OPTIONS line in order to condense distributed loads and reduce the size of the common solution file.

4.

Set the drag and mass coefficients in accordance with the fatigue-specific information in Chevron Offshore Thailand Ltd Specification 22.01 (Environmental Criteria for the Gulf of Thailand), as shown below:

2.3

CDM

1.00 0.500

2.000

0.800

2.000

CDM

500.00 0.500

2.000

0.800

2.000

Create Fatigue Wave Definition Files

Generally, several (usually four) separate wave definition files are required, in order to keep each resulting Seastate Common Solution File to a manageable size. For example, with a four-legged


platform, each fatigue wave definition file may contain wave data for two of the eight wave directions and several wave heights for each of those two directions. For each wave height and direction, four load cases (wave positions) should be defined – maximum base shear, minimum base shear, maximum uplift and maximum down force. From these four wave positions, the Fatigue module will determine the maximum possible stress range at each joint (see Section 2.5 for details on this). See the partial example below:

LDOPT

IN

NF+Z

1.031

7.850

-71.50

74.09GLOBMN

MPTNP

FILE ****** FATIGUE WAVES **** **** NORTH LOAD LOADCN

1

LOADLB

1 WAVE LOAD NORTH , H 0.3M

MAX SHEAR

WAVE WAVE

STRN

0.3

4.5

LOADCN

2

LOADLB

2 WAVE LOAD NORTH , H O.3 M

270.

D

0.0

10.

18MS

1

D

0.0

10.

18NS

1

D

0.0

10.

18MU

1

D

0.0

10.

18MD

1

D

0.0

10.

18MS

1

D

0.0

10.

18NS

1

D

0.0

10.

18MU

1

0.0

10.

18MD

1

MIN SHEAR

WAVE WAVE

STRN

0.3

4.5

LOADCN

3

LOADLB


270.

MAX UPLIFT

WAVE WAVE

STRN

0.3

4.5

LOADCN

4

LOADLB


270.

MAX DOWNFORCE

WAVE WAVE

STRN

0.3

4.5

270.

**** LOADCN

5

LOADLB


MAX SHEAR

WAVE WAVE

STRN

0.91

5.3

LOADCN

6

LOADLB


270.

MIN SHEAR

WAVE WAVE

STRN

0.91

5.3

LOADCN

7

LOADLB


270.

MAX UPLIFT

WAVE WAVE

STRN

0.91

5.3

LOADCN

8

LOADLB


270.

MAX DOWNFORCE

WAVE WAVE

STRN

0.91

5.3

270.

D

**** * AND SO ON WITH INCREASING WAVE HEIGHTS, FOLLOWED BY DEFINITIONS FOR NORTH-WEST WAVES..


The example above shows just two wave heights (0.3m and 0.91m) for one direction (North). The full file would contain several wave heights f or Northerly waves, followed by several wave he ights for North-West waves. The second wave definition file would contain West and South-West waves, and so on. Note that each of these (typically four) files should contain identical headers like that shown above (i.e. with and LDOPT line, and the FILE line to signify that the model geometry is stored in a separate file – that described in Section 2.2). Furthermore, the load cases in every file should be labelled from 1, 2, 3 onwards (rather than continuing the numbering from one file to the next). Each wave definition file need not contain exactly the same number of load cases (for example, large waves may occur only from certain directions, and therefore would not be defined in each of the wave definition files).

2.4

Calculate DAFs for Fatigue Waves

Before combining the effects of each of the fatigue waves, it is necessary to calculate a separate DAF for each fatigue wave height. This can be achieved by using the single degree of freedom DAF equation given in Section 2.4.4 below. However, due to the non-linear nature of soils, the natural period of the structure is dependant on the wave height used to develop the linearised pile-soil superelement used in the modal analysis. For a fatigue analysis, the superelement should be generated from a wave height equal to that of the “centre of damage” fatigue wave.

2.4.1 Calculate Centre of Damage Wave Height The centre of damage wave is an equivalent wave that represents the spectrum of fatigue waves to which a structure is subjected. For n wave heights, the centre of damage height is calculated as: n

 damagei  height i COD Height 

i 1 n

 damagei i 1

where damagei  number of wavesi  height i

bm

The constants b and m are defined by the joint stress-wave height relationship and S-N curve respectively. If it is assumed that stress is related to wave height as follows:

  H b Then, generally, 1.0 ≤ b ≤ 2.0. The value of b is platform-specific, and can be calculated by plotting the axial plus bending stresses of several primary members under the action of waves of various heights (with no current, dead loads, etc). The other constant, m , is equal to the negative inverse slope of the log-log S-N curve. For API RP2AWSD X and X’ curves, the values of m are 4.38 and 3.74 respectively. The fatigue analysis will be most conservative when the product b · m is large, since this gives a larger centre of damage wave height, lower pile stiffness, longer natural period, and hence larger DAFs.


2.4.2 Create Centre of Damage Superelement A superelement creation run should then be executed, using the centre of damage wave height (plus long term dead load, operating current and so on) load cases. For this run, the increased C d and Cm values given in Section 2.3 should be used, and a line similar to the following should be inserted into the soil properties file: PILSUP AVG

60

60

62

62

64

64

66

66

In the above example, the superelement is created from the average of several attack directions; load case 60 at 0°, load case 62 at 90°, load case 64 at 180°, and load case 66 at 270°.

2.4.3 Run Modal Analysis with Centre of Damage Superelement More information on running a modal analysis can be found in the appropriate How-To document. The principal reason for running the modal analysis is to determine the natural period of the structure under the action of the centre of damage wave, in order to calculate the DAFs for each individual fatigue wave height.

2.4.4 Calculate DAFs The DAF for each fatigue wave height (i.e. each period) can be calculated from the single degree of freedom equation given below:

DAF 

1 2

  T n  2    T n   2 1       2      T     T     where T n  natural period of structure T  period of wave   0.02 Note that similar wave heights may have slightly different periods for different approach directions. The DAFs calculated here are used in Section 2.5 when combining the various wave cases.

2.5

Create Fatigue Run Input File

The fatigue run input file defines the number of waves of each height impinging on the structure from each direction, the design life of the structure, S-N curve options, SCF options, member selections, joint overrides and so on. A partial example of such a file is shown below:


FTOPT 3

40.

FTOPT2

1.

1.0IAINAXP

SK

SKNE

5.08

TLPEFT

PTPT

* *

JOINT OVERRIDES

JNTOVR 1804

25.000

JNTOVR 6885 AXX

15.000

* GRPSEL RM

CD1 CD2 CD3 MD1 MD2 MD3 COL BRC

* * SCF OVERRIDES JNTSCF

2.5

6011 6081 6811 6881

GRPSCF

6.0

S12 S16 S18

* RELIEF *** *** WAVE LOADING NORTH *TCASE

CASE

DAF METHOD

HEIGHT

*** FTCASE

1

FTCOMB

1

FTCASE

1

FTCOMB

5

FTCASE

1

FTCOMB

9

FTCASE

1

FTCOMB

13

FTCASE

1

FTCOMB

17

FTCASE

1

FTCOMB

21

FTCASE

1

FTCOMB

25

FTCASE

1

FTCOMB

29

1.60 MMN 1.0

2

1.0

1.0

6

1.0

3

0.3

1.0

4

1.37 MMN 7

0.91

1.0

8

1.30 MMN 1.0 10

1.0 11

1.0 15

1.0 18

1.0 19

1.0 12

1.0 16

2.74

1.0 27

1.0 24

1.0 31

1.0 34

1.0 35

1.0 3.96

1.0 28

1.14 MMN 1.0 30

1.0 3.35

1.16 MMN 1.0 26

1.0

1.0 20

1.18 MMN 1.0 23

1.0 2.13

1.21 MMN

1.0 22

1.0 1.52

1.25 MMN 1.0 14

1.0

1.0 4.57

1.0 32

1.13 MMN

1.0

FTCASE

1

FTCOMB

33

5.18

FTCASE

1

FTCOMB

37

WVFREQ

1

0.3

239963

0.91

63992

1.52

16295

2.13

4149

2.74

1057

WVFREQ

1

3.35

269

3.96

69

4.57

17

5.18

4

5.79

2

1.0 36

1.11 MMN 1.0 38

1.0 39

1.0 5.79

1.0 40

1.0

*** *** WAVE LOADING NORTH-WEST *** FTCASE

2

FTCOMB

41

FTCASE

2

FTCOMB

45

1.69 MMN 1.0 42

1.0 43

1.0 46

1.0 47

1.0 44

1.41 MMN 1.0 48

0.3 1.0 0.91 1.0

*** and so on for the remainder of North-West, and six other directions...

The options line specifies a forty year design life for fatigue, with yearly wave frequencies, a default SN curve of API X’, no fatigue checks for non-tubular members, Eftthymiou SCF formulation with load path dependence, and so on. The 3 in column seven represents the number of additional Common


Solution Files (CSFs) that will be used in the analysis (i.e. a total of four, equal to the number of wave definition files). The JNTOVR lines are used to override the default values for various joint properties, such as S-N curve and chord length. (The Fatigue manual gives details on the override hierarchy). Chord length overrides are particularly important because the chord length impacts the SCF of the joint. The analyst should carefully examine the options available in columns 74 and 75 of the FTOPT line, as well as columns 17 and 18-23 in the JNTOVR line. For example, the T option could be utilised in column 75 of FTOPT (leaving column 74 blank), and JNTOVR cards used for any joints where the T option would give the incorrect total chord length. Incorrectly short cord lengths (such as those generated by the T option when one of the chord members has a mid-span node) are NOT

conservative. Generally, only tubular jacket members are checked for fatigue. The JNTSEL and GRPSEL lines can be used to exclude tubular deck members, whilst SK in columns 44-45 of FTOPT excludes all nontubular sections. Non-standard SCFs can be specified for groups, members or joints using the various ###SCF lines. For example, at the kink in the main legs of jacket, a SCF of 2.5 is specified in Chevron Specification GS22.01 – hence the JNTSCF line in the example above. The FTCASE, FTCOMB and WVFREQ lines are used to combine the various wave heights and directions, with the correct number of occurrences for each. Each fatigue case (FTCASE) corresponds to one wave height from one direction. Recall that each wave height-direction combination was defined by four positions of the same wave (maximum bases shear, minimum base shear, maximum uplift and maximum downforce) in Section 2.3 – these cases are referenced by the corresponding FTCOMB line so that a stress cycle from the wave can be generated. For example, the first FTCASE / FTCOMB pairing in the example above uses a DAF of 1.60 on a 0.3 metre high wave whose four wave positions are defined by cases 1, 2, 3 and 4 (each with an individual load factor of 1.0). The MMN option in columns 32-34 of the FTCASE line stipulates that the maximum and minimum stresses at each joint from each of the load cases in the corresponding FTCOMB line should be used to generate the worst possible stress cycle for that wave. Common Solution File load cases are referenced in the fatigue run input file in monotonically increasing order. As mentioned in Section 2.3, each wave definition file (and hence each CSF) consists of load cases numbered 1, 2, 3 … however when the CSFs are combined in the fatigue run, the load case numbers are considered to be cumulative. For example:

Within Wave Definition Files Wave Definition File

Load Case Numbers

Load Case Numbers Within Fatigue Run Input File

1

1-40

1-40

2

1-44

41-84

3

1-36

85-120

4

1-44

121-164


Note: 



The DAFs used in the FTCASE lines are those calculated in Section 2.4.4. The example above uses the “Interpolated Deterministic Fatigue Analysis” technique, where wave frequencies are defined by the WVFREQ line and wave heights in this line may be interpolated between those explicitly studied. A “Direct Deterministic Fatigue Analysis” would define the frequency of each wave in the FTCASE line.



The frequency values entered in FTCASE or WVFREQ should correspond to the number of occurrences over the period entered in columns 15-21 of the FTOPT line.


3.

RUNNING THE FATIGUE ANALYSES

3.1

Create CSF Files

Each wave data file must be run with the fatigue model file in order to determine the member stresses at each joint (stored in a binary Common Solution File), before these stresses can be combined and the damage evaluated with the Fatigue module. If there are four wave definition files, four Seastate runs will be necessary. Using the Runfile Wizard in SACS 5.1, select a Static Analysis with Pile Soil Interaction, and choose the first wave definition file as the input file. Select the pile soil file appropriately, and leave element checks, Postvue database etc unchecked since only the binary CSF file is needed. Press OK and then the wizard will read the FILE line in the wave definition file, and prompt you for a Model Data File. Choose the file you created in Section 2.2. Run the analysis and rename the resulting PSICSF file to WAVE_1.CSF or similar. Change the runfile to use the second wave definition file, run the analysis again, and rename the resulting PSICSF file to WAVE_2.CSF, and so on until all wave definition files (usually four) have been used to create a CSF file.

Note: If the structure is dynamically sensitive, the Wave Response package (rather than Seastate) should be used to generate the cyclic wave forces (CSF files), since Wave Response considers the structure’s dynamic characteristics. In such a case, equivalent static loads should be generated so that they may be solved statically.

3.2

Combine SCF Files for Fatigue Life Determination

In the Post Processing section (“Post”) of the Runfile Wizard, select Fatigue. Start the wizard and choose the fatigue run input file you created in Section 2.5. The wizard will prompt you to select the appropriate number (as specified in the FTOPT line) of CSF files (those you created above). Run the analysis, and look at the FTGLST file for a listing of the joints in order of decreasing damage (search the file for the string DAMAGE ORDER ). A damage value greater than one indicates a failure to reach the desired design life. Note that the time units (usually years) are defined by the values you input in columns 8-28 the FTOPT line, and your wave frequency data.

84661385 Deter Minis Tic Fatigue Analysis How To

Recommend Documents