Plastic deformation of thin copper plates from water hammer

The simulations on this page try to reproduce experimental results provided by V.S. Deshpande, University Cambridge. A strong pressure wave in water is created by firing a steel projectile on a piston at the end of the tube. The wave propagates through the tube and impinges onto a thin copper plate sealing the other end. Depending on the initial projectile velocity the pressure induces plastic deformation of different rupture patterns.

The experimental facility is described in debth in V.S. Deshpande, A. Heaver, N.A. Fleck, An underwater shock simulator. Royal Society of London Proceedings Series A, 462(2067):1021-1041, 2006.

Basic computational setup

Fluid

  • Pressure wave generated by solving equation of motion for piston during entire fluid-structure simulation.
  • Full water shock tube of 1.3m length, 64mm diameter considered in fluid. 1d simulations used to provide proper flow field for coupled 3d simulations.
  • Modeling of water with stiffened gas equation of state with \gamma=7.415, p_\infty=296.2\,\rm MPa
  • Multi-dimensional 2nd order upwind finite volume scheme, negative pressures from cavitation avoided by energy correction
  • AMR base level: 350x20x20, 2 additional levels, refinement factor 2,2.
  • Approx. 1,2M cells used in fluid on average instead of 9M (uniform)

Solid with SFC

  • Copper plate of 0.25mm, J2 plasticity model with hardening, rate sensitivity, and thermal softening
  • Solid mesh: 4675 nodes, 8896 elements
  • 8 nodes 3.4 GHz Intel Xeon dual processor, Gigabit ethernet network, ca. 130h CPU to t=1ms

More details can be found in F. Cirak, R. Deiterding, S.P. Mauch, Large-scale fluid-structure interaction simulation of viscoplastic and fracturing thin shells subjected to shocks and detonations, Computer & Structures 85 (11-14): 1049-1065, 2006.

Plastic material deformation

Comparison of plate for p0=34MPa at the end of the simulation (1ms simulated) and after the experiment:

Plate at the end of the simulation, p0=34MPaPlate after the experiment, p0=34MPa

Parallel performance

Task F=6, S=2 F=12, S=4 F=24, S=8 F=48, S=16 F=96, S=32
s % s % s % s % s %
Integration 1246 27.3 628 22.4 349 16.1 183 12.5 83 8.8
Boundary sync 1884 41.3 1113 39.6 1044 48.1 735 50.1 447 47.3
Recomposition 152 3.3 154 5.5 138 6.4 124 8.5 98 10.4
Interpolation 46 1.0 27 1.0 15 0.7 8 0.5 4 0.4
Regridding 33 0.7 24 0.9 13 0.6 6 0.4 4 0.4
GFM Find cells 144 3.2 76 2.7 43 2.0 23 1.6 11 1.2
GFM Interpolation 458 10.0 242 8.6 133 6.1 69 4.7 33 3.5
GFM Overhead 72 1.6 21 0.7 6 0.3 7 0.5 3 0.3
CPT 22 0.5 17 0.6 15 0.7 13 0.9 12 1.3
Level set sync 119 2.6 206 7.3 187 8.6 134 9.1 107 11.3
ELC 260 5.7 208 7.4 165 7.6 117 8.0 102 10.8
Coupling data calc 25 0.5 9 0.3 6 0.3 4 0.3 2 0.2
Misc 105 2.3 84 3.0 55 2.5 43 2.9 39 4.1
Total 4566   2809   2169   1466   945  

F denotes the number of CPUs for the fluid solver, S the number of CPUs for the solid solver.

-- RalfDeiterding - 12 Feb 2007

Attachment sort Action Size Date Who Comment
PPlateLegend3d.mpg manage 2490.8 K 12 Feb 2007 - 22:02 RalfDeiterding Impact of the pressure wave on the plate, legends for all quantities displayed
PPlate3d.mpg manage 1841.4 K 12 Feb 2007 - 22:03 RalfDeiterding Impact of the pressure wave on the plate without legend displayed
PPlateVelocity3d.mpg manage 890.9 K 12 Feb 2007 - 22:05 RalfDeiterding Plate alone, velocity component in plate normal direction displayed
PPlateLev3d.mpg manage 447.9 K 12 Feb 2007 - 22:05 RalfDeiterding Refinement levels in entire domain

CoupledSimulations > PlateWaterhammer
Copyright © 1997-2024 California Institute of Technology.