Vous êtes sur la page 1sur 59

SACS / Collapse

Paris, November 29-30

Dr Parvinder Jhita
Dr Jerry Greenough
Engineering Dynamics Inc.
Overview
• Introduction: Collapse Capabilities
• Element Technology + Limit States
• Joint strength, joint flexibility
• Benchmarks
• Demonstrations
Ship impact analysis
Dropped object study
Non-linear blast analysis

• Trouble-shooting
EDI / SACS
Brief History

EDI was established in 1973 by three engineers who


had worked on NASA’s Apollo Project.

Converted aerospace analytical techniques into a


Structural Analysis Computer System
(SACS)

John Fowler
Dr David Garland
James Angehr
SACS Collapse
Brief History

Developed approx. 15 years ago by


John Fowler (EDI President)

Adressed Re-assessment, safety


design work Post Hurricane Andrew

Survivability, Reserve Strength Ratio


(RSR) determination
Gavin Fury (Fury Consult)
Dr Parvinder Jhita
Collapse - Capabilities

• Large deflection, elasto-plastic analysis


• Non-linear geometric and material behavior
• Non-linear pile/soil foundation behavior
• Accounts for member local/global buckling
• Includes joint flexibility effects
• Joint strength/failure (API / Norsok / MSL)
Collapse - Capabilities
Continued

• Non-linear spring elements


• Sequential load stacking (user controlled
incrementation)
• Plastic strain-hardening with unloading
• User-defined ductility limit
• Impact event definitions (+automatic
unloading)
• Non-linear blast analysis
Elasto-plasticity
• Von Mises Yield Function dσ . (df/dσ) = 0
• Prandtl–Reuss Flow Rules dεp = λ(df/dσ)
• Isotropic Strain Hardening H = dσy/dεps
• Strain Hardening ratio, ρ ρ = Ep / E
• H = E / (1-ρ)
• Elastic Unloading
Elasto-plasticity and unloading
Under loading, the plastic strain is monitored
Whilst unloading, the plastic strain is retained,
resulting in permanent deformation.
Elasto-plasticity and unloading
Plasticity Assessment - Members
Sub-segments, Sub-areas

Member is divided length-wise into sub-segments


Cross section is divided into sub-areas (rods)
By default, 8 sub-segments
Tubular section : 12 sub-areas

Sub-area (rod)
Member sub-areas
Various cross sections

Cross-sectional sub-area distributions:


wide-flange, channel, tee, box, angle, prism
Plasticity Assessment - Plates
Layers

Layer is divided into 5 layers

Plasticity is allowed to develop through the plate


thickness

Plastic Layer
Plate Performance
• Improved elastic performance
• Based on Discrete Kirchhoff Theory (DKT)
• Improved out-of-plane deflection profile
• Results available at element corners using ‘DK’
• Triangles recommended for very large out-of-
plane deflections
• Previous plate theory available using ‘ND’
• Benchmarks (inc, patch test) – SAMPLE16
Local Buckling
Three methods

• API LRFD - LR
• Marshall, Gates et al. - MG
• API Bulletin 2U – 2U
Stress checked in each sub-area.
A moment-free hinge is inserted at the
exact location of the local buckling point.
Axial capacity is retained.

Local buckling methods specified using CLPOPT (52:53)


Global Buckling

• Elastic and inelastic instability can occur during


a Collapse run.
• An elastic buckling monitor is available, for
identification of elastic instability.
• Global member buckling is often accompanied
by a limit point.
Interactive buckling monitor ‘BI’ specified using CLPOPT (34:35)
Elastic buckling monitor ‘EB’ specified using CLPOPT (34:35)
Limit Points
Member Buckling

Local limit point is an


indicator of local
structural failure.

In the event of a local


limit point, Collapse
will attempt to jump to
the next stable state.
Joint Flexibility
• Joint Flexibility – Distortion of chord cross
section due to forces in the brace and chord

• Particularly Important
for old structures where joint
cans were not used.

MSL Formulation - MF
Fessler’s Approach - JF
Joint flexibility specified using CLPOPT (34:35)
Joint Flexibility - Fessler
• Linear behavior ( no
material/geometric non-
linear behavior).
• Uncoupled – each brace is
independent of forces in
other braces or chord.
• Method developed in 1986
gives reasonable answers
for elastic joint flexibility.
Joint Flexibility - MSL
• Method accounts for
interaction with chord load.
• Based upon joint
classification (K,Y, X joints).
• Accounts for interaction
coupling between internal
loads and moments.
• Apply ductility limits to
predict tensile joint failure.
Joint Strength / Failure
• Norsok - ND
• MSL – based on the joint flex. Approach - MS
• API LRFD - Punching shear criterion - JS

Joint strength specified using CLPOPT (38:39)


Benchmarks
• Current Collapse Benchmark Assessments

(1) EQE International, Inc. - Independent


Benchmark Study.

(2) BOMEL Engineering - Joint Industry Tubular


Frames Project Phase III
EQE Benchmark
Overview
• 13 Participants in Benchmark Study
• Software used: ASADS
CAP
EDP
KARMA
Micro SAS
RASOS
SAFJAC
StruCAD 3D
USFOS
Original Benchmark described in OTC7779 paper
EQE Benchmark
Description

• Benchmark model – existing platform


installed in 1970 in the Gulf of Mexico

- 4 legs
- 157ft (48m) water depth
- 30ft (9m) distance between legs at WP
- 4 conductors
- 355ft (108m) piles
EQE Benchmark
Loading from 270o from True North
EQE Benchmark
Results for loading from 270o from True North
EQE Benchmark
• Independent Assessment Conclusion:
“ The results indicate Collapse provides a good
estimate of platform ultimate capacity
compared to other nonlinear codes. This
Benchmark has been a standard of
comparison for pushover analysis and
Collapse has been shown to match the
standard.”
Bomel Benchmark
Overview

• 13 Participants in Benchmark Study


• Software used: ABAQUS
APCA
ASAS NL
Offshore DYNA
PALS
RONJA
COLLAPSE
SAFJAC
USFOS
Bomel Benchmark
Description

Test frame
6m x 12m x 12m

Support rig with non-linear


supports (non-linear springs)
Bomel Benchmark
Load Case 2
Bomel Benchmark
Results for Load Case 2
Future Developments
• Improved Dynamics Functionality
• Improved Post-buckling Capability
• Foundation Analysis
Shallow foundations, Spud-cans
Soil Plasticity Models
• Corrugated plates
Improved performance
Elasto-plasticity of Orthotropic Plates
Non-linear Blast Analysis
Front Wall of Quarters Building
Blast Analysis
Front Wall Loading

6.5m

(12.6m)

t = 1 cm
box stiffened

3.4m

Ppeak = 120 kPa


Blast Load Profile
Time history loading
Pressure
(kPa)

P0.015 = 120 kPa (reflected over-pressure)

P0.029 = 48 kPa (clearing pressure)

5.0
0.015 0.029 0.2 time (secs)
Blast Analysis Procedure (i)
1. Loading a) self weight (dead load)
b) blast wave over-pressure
c) live loads
2. Generate mode shapes + mass matrix info.
3. Create load profile in the DYR input file
4. Run DYR and Collapse
front_wall\sacinp.blast
front_wall\dyrinp.blast
front_wall\clpinp.blast
Blast Analysis Procedure (ii)
SACINP

Dynpac

CLPINP Mode shapes


(partial) DYRINP
Mass matrix

Dynamic Response

CLPINA DYROCI
(full) (= SACINP + new loads)

Collapse
Collapse Blast Loading
LDAP line

The LDAP lines enable Collapse to load the


structure incrementally with the blast forcing
plus inertial loads.
Blast Analysis – DIF Options
Implementation of Dynamic Increase Factor
(DIF) in Collapse Input File

• YSFACT Universal yield stress factor


• YSUOVR Universal yield stress override
• YSUMOD Modification of YS by value
• YSMGOV Override YS by member group
Ship Impact Study
5000 MT Ship at 0.5 m/sec.
(operational impact load)

• The impact load is


modeled as a single joint
force.

• Loading progresses until


the total deformation
energy exceeds a pre-
defined kinetic energy.

ship_impact\sacinp.ship_impact
ship_impact\seainp.ship_impact
ship_impact\clpinp.ship_impact
Ship Impact Procedure
IMPACT, ENERGY and SHPIND lines

• Identify impact joint


• Define an impact load condition
• Calculate the ship’s kinetic energy at impact
• Progressively load the structure until the kinetic
energy is absorbed by the deformation, ie:

Estructure + Eship > KE impact

• Optional unloading after energy absorption


Impact Event
Definition of an Impact Event – IMPACT line

• Specify impact joint


• Specify impact load condition
• Calculate or specify kinetic energy at impact
• Optionally specify a ship indentation curve in
order to model the ship indentation energy
• Optionally specify a member denting energy
• Exclude unloading after energy absorption
Member Dent Energy
Two approaches to Force v. dent depth

• Furnes Formula: (use ‘F’ on IMPACT line, col. 38)


Pd = 15 Mp (D/t)0.5 (X/R)0.5
• Ellinas Formula: (use ‘E’ on IMPACT line, col. 38)
Pd = 40 Fy t2 (X/D)0.5

Mp Plastic moment capacity of the tube


X Dent depth
Pd Denting force
Two formulae from API RP 2A-WSD,
C18.9.2C ‘Damage Assessment’
Based on experimental test results
Member Dent Energy
Alternative: Mesh the Tubular

• Mesh the tubular using quads


(Precede)
• Plate plasticity in Collapse
• DKT plate theory
• Advanced out-of-plane
deflection performance
Ship Indentation Curve
Defined using a SHPIND line
Impact
Force
(MN, kips)

Indentation (m, ft)

Collapse uses the ship indentation curve to calculate


the ship indentation energy for a given impact force.
DNV Ship Indentation Curves
Implicitly defined in Collapse

• DNV1 Bow impact


• DNV2 Broad side impact, OD = 1.5 m
• DNV3 Broad side impact, OD = 10 m
• DNV4 Stern impact, OD = 1.5 m
• DNV5 Stern impact, OD = 10 m
DNV Technical Note TN A 202, ‘Impact Loads from Boats’
Ship Indentation Curve
Demonstration curve – MAR1

di(m) F(MN)

1 0.0 2.5
2 1.0 8.0
3 2.0 19.0
4 3.0 27.0
5 4.0 30.0

SHPIND
Dropped Object Study
1 MT mass dropped from 4.5 metres

1T
Dropped Object Study
Impact Procedure

• Identify impact joint (decking mid-point)


• Define an impact load condition
• Calculate the object’s kinetic energy at impact
• Load the structure until the kinetic energy is
absorbed by the deformation, ie:
Estructure > KE impact

dropped_object\sacinp.dropped_object
dropped_impact\clpinp.dropped_object
Mesh Quality
Free Triangular (Delaunay) Mesh
Mesh Quality
Collapse Trouble-shooting
• Check convergence
Increase maximum iterations / member
iterations.
• Use finer increments
• Check for limit points + other events. Use Collapse
View with a high deflection factor.
• Run with fixed pile-heads (check soil data)
• Examine wishbone lengths
• If possible, elevate wishbones away from pile-head
joints
SACS Manual, Section 4.0 ‘Trouble Shooting’
Collapse Trouble-shooting
Wishbones

• The purpose of a wishbone is to constrain a


joint perpendicular to a member, but allow
motion along the member (sliding joint)
• Non-grouted pile / leg connection
• Frame / Conductor connection
• Directional constraint (i.e. non DX, DY, DZ)
• Defined using an offset (aligned with the
reference member) + releases (‘100111’)
Collapse Trouble-shooting
Wishbones

Leg

Wishbone
Unstable
Pile Configuration

Wishbones should have sufficient original


length to avoid the unstable configuration.
Local Buckling API LRFD
Check stress in each sub-area to initiate local buckling
Local Buckling API LRFD
Collapse: for D/t <60 assume no local buckling.
Failure will result from plastic hinge formation.
Local Buckling – Marshall, Gates et al.
Lower bound critical strain criterion
Local Buckling – API Bulletin 2U
Local Buckling assumed for D/t < 134
58

Local Buckling Comparison


Contact Information
Engineering Dynamics, Inc.
2113 38th Street,
Kenner,
LA 70065
USA
Phone: +001 504 443 5481
Fax: +001 504 443 6120
Email: sacs@edi-nola.com
Web: http://www.sacs-edi.com

Vous aimerez peut-être aussi