Impact Simulation Benchmarking for the Double Asteroid ...

Impact Simulation Benchmarking for the Double Asteroid Redirection Test (DART)

Angela M. Stickle1, O.S. Barnouin1, M. Bruck Syal2, A. Cheng1, C. El-Mir3, C.M. Ernst1, N. G?ldemeister7, R. Luther7, P. Michel4, N. Oklay5, M. Owen2, M. Price6, E.S.G Rainey1, K.T. Ramesh3, S.R. Schwartz4, J.-B Vincent5, K. W?nnemann, and the AIDA Impact Simulation Working Group

Goddard Space Flight Center Johnson Space Cener Langley Research Center

1JHU-Applied Physics Laboratory, Laurel, MD; 2LLNL; 3Johns Hopkins University HEMI; 4CNRS; 5MPI; 6University of Kent; 7Museum fur Naturkunde, (angela.stickle@jhuapl.edu)

Introduction: Code Benchmarking

In support of the AIDA mission concept, an international working group was formed to better understand the range of possible outcomes of the DART impact. Impact modeling is one of the primary tools to be used to interpret the results of the kinetic impact deflection, to infer the physical properties of the target asteroid, and to advance our understanding of impact processes on asteroids. Several types of numerical methods can be used to model the DART impact, all of which differ in their fundamen-

Basalt Fully dense 20%

tal approach to solving flow equations as well aposrositmy od45e%lpionrosgity of 60% porosity

material properties and responses to impact streDsenssietys[g./ccA] 2.n8 i2n.6itial 1.5 1.2

benchmarking campaign was established to compQSaCormepretsshiveestrrenegtsh ults [MPa] 400 400 400

of simplified test cases across a variety of hydroc40o0 des. NumeriDynamic Compressive

cal schemes used in this campaign are outlined below. Strength [MPa] Th6o00ugh some of these codes were benchmarked against each other using simple test cases of strengthless metal targets by [1], updated implementations and material models are included here.

AIDA/DART

The Asteroid Impact and Deflection Assess- DART (NASA)

Didymos binary system

ment (AIDA) mission is Kinetic Impactor

designed to test the ef-

fectiveness of kinetic

impactors for planetary

defense. AIDA is a joint

effort between NASAand

ESA. The NASA-led portion is the Double Asteroid Redirect Test (DART)

Earth-based observations

AIM (ESA) Rendesvouz/Observer

and is composed of a ~300-kg spacecraft designed to impact the

moon of the binary system 65803 Didymos at ~7 km/s. The de-

flection of the moon will be measured by the ESA-led Asteroid

Impact Mission (AIM) (which will characterize the moon) and

from ground-based observations.

Working Group Members

The Modeling and Simulation of Impact Outcomes Working Group leads the effort to simulate the DART impact and the resulting effects on the Didymos system using a variety of different modeling techniques. The group is composed of international participants from universities and industry (Table, below). The group's two main goals are to:

(1) Determine the expected outcome of the DART impact and its sensitivity to initial impact conditions; (2) Assess the effect of the DART impact on the moon of Didymos, focusing on implications for asteroid deflection

and properties.

Name Erik Asphaug Olivier S. Barnouin Megan Bruck-Syal Andrew Cheng Gareth Collins

Dan Durda Charles El-Mir Carolyn M. Ernst Galen Gisler Nicole G?ldemeister

Affiliation ASU

JHU-APL LLNL

JHU-APL Imperial College

SWRI JHU JHU-APL LANL MfN

Name Keith Holsapple Kevin Housen Robert Luther

Paul Miller Patrick Michel Nilda Oklay J. Michael Owen Cathy Plesko

Mark Price Emma S. G. Rainey

Affiliation UW

Boeing MfN LLNL OCA MPS, MPG LLNL LANL University of Kent JHU-APL

Name

K.T. Ramesh Jim Richardson

Eileen Ryan Peter H. Schultz Stephen R. Schwartz Angela M. Stickle Jessica Sunshine Jean-Baptiste Vincent Kai W?nnemann

Affiliation

JHU Arecibo

NMT Brown OCA JHU-APL Maryland MPS, MPG MfN, Leibniz

Introduction: Code Parameters

To maintain consistency across numerical simulations, all modelers were asked to

use a Tillotson equation of state, with parameters taken from [2]. For basalt, gen-

eral strength parameters were also specified (below). For more complete code de-

scriptions, see the abstract. Though these parameters are given to all teams, each

team was free to choose a strength and damage model, and implementations vary.

This may explain some of the differences in results observed here.

Basalt

Fully dense 20% porosity 45% porosity 60% porosity

Density [g/cc]

2.8

2.6

1.5

1.2

QS Compressive strength [MPa]

400

400

400

400

Dynamic Compressive Strength [MPa]

600

600

600

600

QS Tensile strength [MPa]

30

30

30

30

Dynamic Tensile Strength [MPa]

80

80

80

80

Toughness Mpa (m)

1.6

1.6

1.6

1.6

Youngs Modulus [GPa]

70

70

70

70

Shear Modulus [GPa]

29

29

29

29

Bulk Modulus [GPa]

49

49

49

49

Coefficient of Friction

0.6

0.6

0.6

0.6

Poisson's ratio

0.25

0.25

0.25

0.25

Adaptive SPH

2D and 3D cases using the Adaptive Smooth Particle Hydrodynamics (ASPH) code Spheral are run by LLNL.

Strength Model: Tensor form of the Benz-Asphaug Grady-Kipp damage model [2-3] with von Mises strength.

Porosity Model: Strain-based porosity from Collins et al. [4].

Resolution: Resolution is concentrated within 3-8 cm of the impact point, with a maximum resolution of 0.08 cm (right). On the order of 3M SPH particles used.

CTH

2D & 3D cases using the CTH model are run at the JHU Applied Physics Laboratory.

Strength Model: A pressure-dependent yield surface is used for plasticity, coupled to a Johnson-Cook Fracture damage model.

Porosity model: The p- model, with parameters from Jutzi et al. [5].

Resolution: Adaptive Mesh Refinement is used to track the shock wave and moving interfaces. The maximum resolution is 0.08 cm (8 cppr).

iSALE

Cases using the 2D iSALE shock physics code are run by teams at the Berlin Museum for Naturkunde, the Max Plank Institute, and the Imperial College.

Strength Model: Uses the strength and damage model described in Collin et al. [6].

Porosity Model: Strain-based porosity from W?nnemann et al [7] and Collins et al. [4].

Resolution: For cases 1-3 (half-space cases), resolution is 40 cppr. For the spherical target, the resolution is 15 cppr.

Uintah MPM

3D cases using the MPM model are run by Johns Hopkins University.

Strength Model: Modified Tonge-Ramesh model [8], allowing explicit definition of the sub-scale flaw distribution (right). Includes pressure-dependent granular flow of damaged material and pore compaction through the porosity model.

Porosity Model: The p- model, with parameters from Jutzi et al. [5].

Resolution: 16 material points per 5 mm bin. Simulations utilize on the order of 2 million material points.

Flaw distribution parameters

min. flaw size [m] 5.E-06

max. flaw size [m] 1.E-03

flaw density

2.E+12

flaw dist. Type Pareto

Pareto exponent [9] 3

References/Acknowledgements

[1] Pierazzo E., et al. (2008) Met. & Planet. Sci. 43, 12, 1917-1938; [2] Benz, W. and E. Asphaug (1999) Icarus 142, 5-20; [3] Benz, W & E. Asphaug, (1995) Computers Phys. Comm., 87, 253-265; [4] Collins, G., et al. (2011) Int. J. Imp. Eng., 38:434-439; [5] Jutzi et al. (2009) Icarus 201, 802-813; [6] Collins, G. S., et al. (2004). Met. & Planet. Sci., 39:217--231.; [7] W?nnemann, K. et al. (2006). Icarus, 180:514-527; [8] Tonge, A. L., Ramesh, K. T., & Barnouin, O. (2016) Icarus, 266, 76-87; [9] Housen, K. R., & Holsapple, K. A. (1999) Icarus, 142(1), 21-33.(1999).

Portions of this work were performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52- 07NA27344.

Code Comparisons

To enable comparison between codes, five initial test cases were sent to each team performing simulations. These cases cover impact into both aluminum and

basalt targets and were designed to isolate the effects of geometry (impacts into a half-space versus a sphere), material strength, and porosity. All simulations

utilize a Tillotson EOS, with properties for basalt taken from [2]. Comparisons between models include metrics such as crater size, ejecta velocity, and damage

growth as well as measurements of (the momentum transferred to the target by the impact). Some measurements (e.g., crater size) match well between codes

for certain cases, while some measurements behave very differently, likely due to differing strength and damage model implementation or resolution of model

runs. We are still examining the large differences in calculated values between codes.

Porosity and Strength Model Effects

Effects of Target Geometry

Crater Size

Damage

Crater Size

Width [cm]

Depth [cm]

10

MPM

8

6

4

2

00

50

100

150

200

250

300

Time [usec]

ASPH

5

4

3

2

1

00

50

100

150

200

250

300

Time [usec]

Case 3: 2D CTH, basalt half-space, width, =0% Case 3: 2D iSALE, basalt half-space, width,=0% Case 3: 2D iSALE, basalt half-space, width,=20% Case 3: 2D iSALE, basalt half-space, width,=45% Case 3: 2D iSALE, basalt half-space, width,=60% Case 3: 2D CTH, basalt half-space, depth, =0% Case 3: 2D iSALE, basalt half-space, depth, =0% Case 3: 2D iSALE, basalt half-space, depth, =20% Case 3: 2D iSALE, basalt half-space, depth, =45% Case 3: 2D iSALE, basalt half-space, depth, =60%

iSALE

1.0 0.9 0.8

CTH

0.7

0.6

0.5

0.4

0.3

0.2

0.1

1.0

10

0.9

0.8

0.7

9

0.6

0.5

8

0.4

0.3

0.2

7

0.1

6

(left) Depth and diameter (width) for CTH and iSALE calculations of impacts into half-space targets (bottom) compared

Length [cm]

5

to a 30-cm spherical tar-

4

gets (top, damage con-

3

tours below). Impacts

at 5 km/s by a 0.635-cm

2

basalt sphere. All simu-

1

lations had a strength

0 0

50

100

150

200

250

Time [usec]

300

350

400 and damage model. Cra-

10

ter sizes are similar for

Width (Diameter) [cm]

8

simulations into a half-

6

space compared to a fi-

4

nite sphere for both CTH

2

and iSALE, for both 2D

00

50

100

150

200

250

Time [usec]

300

350

400 and 3D geometry.

10 8 6

Depth [cm]

(top) crater size (depth and diameter) as a function of porosity from CTH and iSALE calculations. As porosity increases, craters become increasingly narrow and deep. Fully dense 2D CTH models predict larger craters than corresponding 2D iSALE models, though the trends in crater size and shape as a function of porosity are consistent between codes for 3D CTH and ASPH models (see crater shapes in the damage contours, right). The fully-dense 2D CTH calculation predicts a much wider and deeper crater than iSALE.

(top) Damage contours from an MPM simulation into a fully-dense basalt half-space. (middle/bottom) Damage contours from a Spheral calculation into a basalt half-space at two different porosities, showing the inner 3-cm around the crater. As porosity increases, the damage patterns change and the craters become narrow and deep.

(left) Damage contours from 2D iSALE calculations at varying target porosities. (right) Damage contours from 3D CTH calculations at varying target porosities, showing large effect on crater shape. The damage model also affects the observed deformation and the ejecta behavior in the simulation, though crater shapes and amount of damage remain similar between codes.

4

2

0

0

50

100

150

200

250

300

350

400

Time [usec]

Damage and Strain

Ejecta Behavior and Momentum Transfer

Ejecta behavior from 2D iSALE models. Comparisons with other numerical codes are in progress. (left) Normalized launch velocity as a function of launch position; (right) Ejection angle as a function of normalized launch position. As the target becomes increasingly porous, the ejection angles increase, but the launch velocity decreases. This suggets that a kinetic impactor hitting a porous asteroid would produce a lower-speed, high-angle ejecta curtain.

(left) Momentum transfer from CTH models. As porosity increases, the momentum enhancement factor of the target decreases significantly, though it remains above 1. There is also a significant difference in calculated beta values between 2- and 3-D calculations.

1.4 1.35 1.3 1.25 1.2 1.15 1.1 1.05

10

1a

1a 1b

1b 23

2

3

10

20

30

40

50

60

70

80

Time (?s)

(top) Momentum transfer calculated for 2D and 3D CTH models for impacts into strengthless halfspace targets (solid and dashed lines) compared to impacts into half-space targets with a plasticity and damage model included in the calculations. 2D strengthless models give the highest beta values because they generate more ejecta than the cases with strength. For 3D models, simulations incorporating strength models provide higher beta values.

In all cases, CTH predicts higher beta values than equivalent ASPH calculations (bottom). CTH calculations for Case 2 are the closest to the values predicted by the ASPH calculations.

50

100 ?sec

0

50

1.0 0.9

200 ?sec

0.8

0.7

0

0.6

0.5

0.4

0.3

0.2

0.1

1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1

Z (mm) Z (mm)

0

100

200

X (mm)

1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1

0

100

200

X (mm)

Impacts into a 30-cm basalt sphere by a 0.635-cm basalt projectile at 5 km/s. Color overlays show damage contours from 0 (blue) to red/orange (damage = 1, fully damaged). All simulations predict roughly the same extent of

damage, though the details are dependent on damage model. (top) 3D ASPH simulations with 20% porosity ~ 100 sec after impact. (topleft) Damage trace contour. Trace = 3 is fully damaged. (top-right) Damage trace contour for the inner 10-cm of the sphere showing details of damage zone. (middle) 3D CTH simulations, showing the sphere at 100 and 200 sec after impact. Grey material is undamaged. (bottom) 2D iSALE simulations at 100 and 200 sec after impact. Backside damage is spallation.

 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download