HMX Detonation in a Tantalum Cylinder

The simulation is a simplified version of the following setup:
Principle sketch of VTF

AMR base mesh 60x60x120, two additional levels with factors 2, 4 to resolve detonation wave in HMX, Roe+HLL fluid solver
Fluid simulation uses <10M cells instead of 221M (uniform)
Detonation propagation in HMX
Coupling at level 1
Solid mesh 56k elements, Tantalum material model with J2 flow theory with Vinet’s thermal equation of state
The boundary conditions on the solid are varied in the pictures below and in the attached movies
~500h CPU with 14 fluid and 2 solid processes on ALC
Source code: [fluid codes] [solid codes]

Coupled 3 level run, exterior in part fully constrained, Pic 1

Weak scalability study: fractions of time spent in different parts of the coupled applications

Solid	S1	S2	S4	S3
Iterations	1080	1240	1480	1400
Processes	4	4	4	4
Elements (M)	0.443	0.443	0.443	0.443
Init	3207	3195	3188	3229
Main	4553	4446	4473	5209
ELC	7742	7398	7479	9882
Misc	491	499	589	493
Scaled	1524	1296	1092	1344

Fluid	S1	S2	S4	S3
AMR Iterations	27	31	37	35
Processes	26	52	104	78
Cells (M)	6.18	9.77	16.1	13.0
Solver	1149	991	965	1109
Synchronization	7648	7320	6923	10092
GFM Overhead	678	575	547	634
CPT Algorithm	264	278	333	326
ELC	4456	4481	4558	4517
Misc	1300	1294	1367	1284
Scaled	1518	1573	1531	2085

-- RalfDeiterding - 06 Oct 2005

I	Attachment	Size	Date	Who	Comment
	canister_test1R.mpg	6881.3 K	06 Oct 2005 - 03:49	RalfDeiterding	Movie of 3 level AMR run, exterior of cylinder constrained in normal direction
	canister_levels.mpg	3739.4 K	06 Oct 2005 - 04:03	RalfDeiterding	Movie of 3 level AMR run, exterior of cylinder constrained in normal direction, AMR levels in mid plane

You are here: CoupledSimulations > TaCanisterHMX

to top