Upload
gilbert-eaton
View
229
Download
0
Tags:
Embed Size (px)
Citation preview
Workshop on Very Large Floating Structures for the Future
Trondheim -- 28-29 October 2004
Efficient Hydrodynamic Analysis of VLFS
by J. N. Newman
Theoretical formulation and restrictions
• Linear (and 2nd order) potential theory (small amplitude waves and motions)
• Frequency-domain analysis (can be transformed to time domain if necessary)
• No considerations of current, slamming, viscous loads
Computational approach
• Boundary Integral Equation Method (BIEM) a.k.a. `Panel Method’ with free-surface Green function (source potential)
• General-purpose codes are applicable to a wide variety of structures and applications (different configurations, air-cushions, hydroelastic effects, multiple bodies)
• All relevant hydrodynamic parameters can be computed (loads, RAO’s, pressures, drift forces, free-surface elevations, etc.)
Computational tasks and restrictions
• Geometric representation of submerged surface
• Structural representation (if hydroelastic)
• Confirmation of results requires convergence tests
• Computing cost (time and memory) increases rapidly with number of unknowns
• Short wavelengths aggravate everything!
recent developments
• Higher-order panel method with B-spline basis functions for the solution
• pFFT Accelerated solver O(N log N)
• Exact geometry
• CAD coupling
Examples of Applications and Results
• Rigid barge 4km x 1km x 5m
• Same barge with hydroelasticity
• Hinged barges with hydroelasticity
• Hinged semi-subs with hydroelasticity
• Large arrays of cylinders
Drift forces on 4km x 1km x 5m rigid bargeWave incidence angle 45o -- Water depth = 50m
Red: Surge drift forceBlack: Sway drift force• higher-order method (727 unknowns) Solid lines: Accelerated pFFT (`precorrected Fast Fourier Transform’) method with 43,000 low-order panels (Order N log N)Both methods take about 0.4 hour/period on a 2GHz PCDashed lines: Asymptotic approximation for short wavelengths
Period10 15 20 25 30
0
500
1000
1500
2000
Methodology for hydroelastic analysis
• Structural deflections represented by generalized modes
• Uniform stiffness ratio EI/(mass*L^2) assumed for simplicity
• Extended vector of RAO’s gives the modal response
• Cf. `direct’ method where the structural and hydrodynamic analyses are integrated
4km x 1km x 5m barge, 27*16 generalized modesL/wavelength = 5,10,20 – depth = 20m – Period = 57,28,15 seconds
Head seas 45 degrees Head seas 45 degrees
5
10
20
5
From C.-H. Lee, `Wave interactions with huge floating structure’, BOSS `97
DEFLECTION PRESSURE
Symmetric/antisymmetric modes for a hinged structure with five elements
Modes j=7-11 and j=12-16 are 1st and 2nd Fourier modes j=17-96 are 3rd to 18th Fourier modes (not shown)
j=1,2 j=3-6 j=7-11 j=12-16
Array of five barges with stiffness SMotion and shear force at the upwave hinge
Period6 8 10 12 14 16 18 20 22 24 26 28 30
0
0.5
1
1.5
2
2.5
3
3.5
4
rigid1E-31E-41E-6S=0
Amplitude of vertical motion
Period6 8 10 12 14 16 18 20 22 24 26 28 30
0
1000
2000
3000
4000
rigid1E-21E-31E-41E-6
Vertical shear force
Array of five semi-subs with stiffness SMotion and shear force at the upwave hinge
Period6 8 10 12 14 16 18 20 22 24 26 28 30
0
500
1000
1500
2000
rigid1E-11E-21E-3
Vertical shear force
Period6 8 10 12 14 16 18 20 22 24 26 28 30
0
2
4
6
8
10
rigid1E-11E-21E-3
Amplitude of vertical motion
Mean drift forces on Array of 100 x 4 cylinders Wave incidence angle 45o -- Water depth = 50m
Based on Higher-order method with exact geometry.
NEQN=4200
2.3 hours per Wave periodOn 2GHz PC
Period8 12 16 20 24 28 32
0
50
100
150
200SwaySurge
Drift forces on the 100x25 Cylinder Array Wave incidence angle 45o -- Water depth = 50m
240,000 low-order panelswere used here (96 on eachCylinder).
These computations wereperformed on a 2.6GHz PC using the accelerated pFFT method. The CPU time averaged about one hourper period.
The peak magnitudes aremuch larger below 8 seconds,where near-trapping occursbetween adjacent cylinders.
Period10 15 20 25 30
-20
0
20
40
60
80
100
120
140
SurgeSway
Conclusions
• Contemporary hardware/software can be used to analyze wave effects for many VLFS configurations with length scales of 1-4 km
• These capabilities will increase in the future
• Short wavelengths, bottom topography, breakwaters, etc, present special challenges
• Analytic/asymptotic theories are still required for >>1km length scales, e.g. ice fields, and to validate computations in the short-wavelength regime