ESyS_Crustal: A Software Infrastructure for Crustal Dynamics Huilin Xing xingesscc'uq'edu'au Earth S

1
ESyS_Crustal A Software Infrastructure for
Crustal DynamicsHuilin Xingxing_at_esscc.uq.edu.a
uEarth Systems Science Computational
Centre/ACcESS Major National Research Facility
2
Outlines

• Introduction
• ACES, ACcESS, iSERVO
• ESyS_Crustal
• Applications to Earthquake Simulation
• Towards Tsunami Source Modeling
• Conclusions and Future Work

3
ACES, ACcESS and iSERVO
Peter Mora Director, Earth Systems Science
Computational Centre QUAKES, Univ. of
Queensland Chair Science Committee, ACcESS Major
National Research Facility Executive Director,
APEC Cooperation for Earthquake Simulation (ACES)
4
ESSCCEarth Systems ScienceComputational Centre

• www.esscc.uq.edu.au
• Centre Director Peter Mora
• 3 research groups
• QUAKES, Peter Mora
• Geodynamics, Hans Muhlhaus
• Computational, Lutz Gross
• HQ of ACES
• APEC Cooperation for Earthquake Simulation
• HQ of ACcESS MNRF
• Australian Computational Earth Systems
• Simulator Major National Research Facility
• 1.1 GFlops ACcESS supercomputer

5
The APEC Cooperation forEarthquake Simulation
(ACES)

• To develop realistic numerical simulation models
• 2. To foster collaboration
• 3. To foster development of infrastructure
  programs

Develop a unified simulation model for earthquake
generation and earthquake cycles
6
China

• Fracture physics, mesoscopic damage models,
  intraplate observations

CAP, LNM, Peking Univ, Beijing U. Courtesy of
Yin and Xia
7
Australia

• Micro-models
• Earthquake physics
• and dynamics
• Crustal mantle models

ESSCC/QUAKES CSIRO, UWA
Courtesy of Mora, Place, Moresi, Muhlhaus
8
Japan
Earth Simulator Worlds fastest supercomputer
GeoFEM Large-scale finite-element software
platform for solid earth simulation

• Solid Earth Simulator Project
• Forefront macro-scale research
• Subduction zone dynamics,
• crustal activity and strong motion
• Mantle and core dynamics

TU, RIST, ERI, GSJ, BRI, NRIESNDP,... Courtesy
of Matsuura, Okuda, Matsui
9
USASimulation of the CA interacting fault system
NASA/JPL SCEC USC, UCLA, LDEO, ... GEM UC
Davis, Indiana, ...
Courtesy of Rundle, Donnellan and Olsen
10
History program of activities

• 1994 Discussions
• ACES Proposal endorsed
• Planning meeting
• 1999 Workshop (AUS)
• 2000 WG Mtg Workshop (JPN) Visitors
• 2001 WG Mtg (USA)
• 2002 Workshop (USA)
• 2003 WG Mtg (AUS)
• Workshop (CHINA)
• 2006 Workshop (USA) ? iSERVO

11
Outcomes and developments

• Simulation models
• 3 x 2 volume journal issues 3 proceedings
• Collaboration
• Earthquake physics catastrophic failure
• Simulation models and software
• Australia, China, Japan, USA visitors collab
  programs (30 visits, joint publications)
• Infrastructure
• GEM, ServoGrid, QuakeSim
• Earth Simulator, GeoFEM
• Key national program for catastrophic failure
• ACcESS MNRF

12
The Australian Computational Earth Systems
Simulator (ACcESS)Major National Research
Facility
A multi-scale multi-physics ESS

• Achieve a holistic virtual earth simulation
  capability

• Provide a computational virtual earth serving
• Australias national needs

• One of two science Major National Research
  Facilities
• being established in Australia

• Develop software models, and establish
• thematic supercomputer needed for research
  outcomes

ACcESS is a member of AuScope (2007.4- )
(http//www.auscope.org)
13
ACcESS grand challenge drivers
Drive construction of the simulator system by
grand challenge science, societal, or economic
problems

• Particle simulation
• Earthquake nucleation
• Mining innovations/block caving
• Global geodynamics
• Implications to formation of large scale
  mineralisation
• Crustal fault system physics
• Earthquake science
• Energy engineering
• Himalayas
• Understanding earth evolution

ESyS_Crustal
build on PANDAS
14
Earthquakes - Interacting Fault System

• Most of the tectonic earthquakes occur by a
  sudden stickslip frictional instability along
  pre-existing faults or plate interfaces. The
  earthquake is the slip (several seconds), and
  the 'stick' is the interseismic period of elastic
  strain accumulation (tens - hundreds years).
• A Frictional Instability
  Problem
• Earthquakes Mechanism and Forecast Time,
  Location and Magnitude?
• Interior Stress Evolution

Adaptive Finite Element Software for Simulation
of Large Scale Interacting Fault System
Friction nonlinearity Multiple material
properties Complex geometry nonlinearity
Multiple time scales
15

• PANDA
• Parallel Adaptive Nonlinear Deformation Analysis
  Software

(http//www.hit.edu.cn)

• RIKEN, Japan

Univ of Queensland, Australia
1990-1995
1995-2002
2002-
16
PANDAS

• In Chinese
• ? ? ??
• Bear Cat Panda
• In English
• Bear Cat BearCat ? Panda

17
SUMARRY OF CODES FOR STATIC FRICTIONAL CONTACT
18
PANDAS Adaptive
Static/Dynamic FEM

• Equilibrium Equation
• Finite Element Formulation
• Various Time Integration Algorithms
• Quasistatic-explicit algorithm (without
  iteration)
• or
• Implicit algorithm (Static a0 Dynamic a1)
  (with iterations)
• Dynamic-explicit algorithm
• .

19
PANDAS
Nonlinearity

• Discontinuity fault/plate boundaries
• Node-to-point arbitrarily contact element
• Friction
• Unified friction description
• Deformation (finite deformation)
• Materials
• Coupling Effects

20
PANDAS
Discontinuity

• Bodies in contact

• Contact element strategy
• Contact between deformable bodies
• Contact between a slave node s and a point c on a
  master segment
• Node-to-point contact algorithm

slave-master concept
Contact Pair
Node-to-node contact algorithm
21
PANDAS
Nonlinear friction
MECHANICAL FACTORS Imposed velocity Types of
motions Imposed displacements histories
Imposed loads and histories Geometry of contact
Component design
THERMAL FACTORS Frictional heating Thermal
softening/melting Thermal shock effects Solid
state phase transformation

• MATERIAL FACTORS
• Materials pairing
• Mechanical properties
• Thermophysical properties
• Microstructure

Nonlinear FRICTION Response
THIRD-BODY EFFECTS Transfer particles Wear debris
concentration Chemical reactivity of
particles Fluid flow
LUBRICATION FACTORS Quality/method of
supply Regime of lubrication Lubricant properties
TRIBOCHEMICAL FACTORS Relative humidity Surface
reactivity/catalysis Environmental composition
22
PANDAS Adaptive Time Step
Control

• Practical requirements
• Qualitative Quantitative Change
• Accuracy Efficiency
• Qualitative changes
• Relative motion state
• stick slip
• Contact state
• close open
• Material state
• elastic plastic
• Numerical Stability (DE)

• Quantitative changes
• Friction force variation in both stick and
  slip state
• Friction coefficient variation
• Normal contact force variation
• Relative slip displacement variation
• Rotation variation
• Stress variation in plastic state
• Strain variation in plastic state

Rmin
23
PANDAS Software
Flowchart
PyFEM
Pre-Processing Patran, I-Deas Chikaku Mesh
PANDAS
.out or .unv file meshboundary conditions
FEM SOLVER To do adaptive static/ dynamic FE
calculation output the data for
post-processing and/or further calculation
.sve, .svb
CONVERTER To convert all the data to the
standard data for FEM analysis
.ctr
TRANSFORMER To modify the prescribed boundary
conditions
??.unv Or ??.node
.sve, .svb
.tex .lst
Database material and friction
Word Excel et al
Post-Processing Patran or I-Deas
PyFEM
ACcESS Software System - ESyS
24
PANDAS
Applications

• Computational Material/Mechanical Engineering
• Computational Biomechanics
• Computational Solid Earth Science

25

• PANDAS
  Applications

• Computational Material/Mechanical Engineering
• Sheet Superplastic Forming (Aerospace Industry et
  al)
• Plastics Blowing Molding
• Tube Hydroforming
• Sheet Metal Forming (Stamping)
• Structure Assembling and Disassembling
• Assessment of the Integrated Performance of
  Assembled Structure
• Welding
• Thermal-Mechanical Coupled Deformation

26
PANDAS Applications
to Earth Sciences

• Sandwich Frictional Model
• Xing, H.L., Makinouchi, A. (2002), Finite
  element analysis of sandwich friction
  experimental model of rocks, Pure and Applied
  Geophysics, 159, 1985-2009.
• Single Fault Bend Model (Intraplate and
  Interplate Cases)
• Xing, H. L., Mora, P., Makinouchi, A. (2004).
  Finite element analysis of fault bend influence
  on stick-slip instability along an intra-plate
  fault, Pure and Applied Geophysics, 161,
  2091-2102
• Xing, H. L., and Makinouchi, A. (2003). Finite
  element modelling of frictional instability
  between deformable rocks. International Journal
  for Numerical and Analytical Methods in
  Geomechanics, 27 1005-1025
• Multiple Fault Bends Model
• South Australia Intraplate Fault Model
• South California Interacting Fault Model
• Entire Earth Model (Tidal Deformation)
• Wave Propagation in Non-Continuum Media
• Subduction Fault Model (Sumatra)
• Frictional Sliding Induced Heat and Deformation
• Hot Fractured Rock Geothermal Reservoir System
• Tsunami Generation induced by Earthquakes

27
Multiple Fault Bents - Mesh and Boundary
Conditions

• Specimen Dimensions

23890 nodes 18432 elements
Py
1000

• Material Parameters

B
72

• Loading Conditions
• PxPy10 MPa
• The previous hydraulic
• pressures remain 10MPa,
• while the nodes on the surface
• Px move along the x axial at

Px
60
1000
Px
36
Vx
12

• Friction Parameters

0
A
28
Equivalent Stress Variation Rate
29
Relative Velocity Variation (1)
30
Relative Velocity Variation (2)
31
Time for Nodes to Change Initial States
32
South California Fault Model
(Data from Professor John Rundle of University of
California at Davis)
Fault Geometry
33
Equivalent Stress Rate
34
South Australia Fault Model
iSERVO Institute Seed Project (iSERVOInternation
al Solid Earth Research Virtual
Observatory) (with Collaboration of Professor
Mike Sandiford of Melbourne University)
Surfaces of SA Faults
35
Southern Australia Fault Model (SA) - Mesh
iSERVO Institute Seed Project
36
SA Model - Equivalent Stress Rate
iSERVO Institute Seed Project
37
Towards Tsunami Generation Modeling

• Survey
• Simulation of Subduction Fault Model
• Sumatra Area
• Proposal/Future Work for Tsunami Source Modeling

38
Killer Tsunamis in Historical Times
39
Different Faults within a Subduction Zone
(sumatra 2004)
Modified from Geist, E. L. (1999). Local
tsunamis and earthquake source parameters.
Advances in Geophyscis, 39. 116-209
40
Earthquake Source Parameters Affecting Tsunamis

• Static Source Parameters
• Seismic moment
• Fault geometry (source depth, width and length,
  constant dip angle)
• Fault slip (average uniform slip constant slip
  direction)
• Linear physical properties
• deformation (vertical) for simulating
• far-field tsunamis and local-field tsunamis

Analytical equations
41
Problems of Existing Source Models (1)

• The tsunami generation/nucleation process from
  earthquakes is critical for the tsunami
  simulation problem in its entirety. Earthquake
  generation is a very complicated dynamic problem
  and the tsunami theory applied in practice has
  oversimplified the source too much to yield
  accurate and reliable results that would be
  desirable for the real-time tsunamis modeling
  system. For examples
• In the earthquake rupture process, the
  faults/plates at the earthquake source normally
  have a very complicated geometry (not only flat),
  and non-uniform slip with different angles and
  may be highly deformed or even fractured (i.e.
  not a rigid medium)
• Real interacting surface direction - sea floor
  Different from faults
• It is dynamic process, i.e. earthquake itself and
  its interaction with water take a rather long
  period of time rather than in an instant.

42
Problems of Existing Source Models (2)

• Submarine earthquakes may induce a landslide and
  related effects resulting in highly variable
  pressure and viscous drag forces (i.e. frictional
  contact forces), and even decoupling between the
  deformable sliding land/rock and water.
• The actual seafloor deformation resulting from
  the earthquake source process with all its
  complexity as well as any ensuing submarine
  landslide is the real source of tsunamis, and
  this total process is quite complicated.
• This combined complex tsunami generation process
  belongs to a solid (rock/sea floor) and liquid
  (water) coupled non-linear interacting/frictional
  contact dynamic problem in 3-dimensions.
• The total source process is not modelled in
  present tsunami simulations despite the fact that
  their reliability as operational tsunami models
  depends strongly on the source dynamics itself
  and highly coupled solid-liquid nonlinear
  interactions at the source.
• The source data for further tsunami simulation
  will be the real one, such as velocity (including
  wave), force and deformation (not only the
  displacement) if the coupled model applied

43
CASE STUDY Sumatra Subduction
(With Dr S Zhu C Yin)
44
Case 1 Geometry mesh
45
CASE 2
80820 nodes 72000 elements
Geometry and mesh
46
Distribution of nodal velocities
47
(No Transcript)
48
(No Transcript)
49
(No Transcript)
50
(No Transcript)
51
(No Transcript)
52
(No Transcript)
53
(No Transcript)
54
(No Transcript)
55
(No Transcript)
56
(No Transcript)
57
Distribution of nodal velocities
58
Distribution of nodal velocities
59
Relative Velocity along Faults
60
(No Transcript)
61
(No Transcript)
62
(No Transcript)
63
(No Transcript)
64
(No Transcript)
65
(No Transcript)
66
(No Transcript)
67
(No Transcript)
68
(No Transcript)
69
(No Transcript)
70
(No Transcript)
71
(No Transcript)
72
(No Transcript)
73
(No Transcript)
74
(No Transcript)
75
(No Transcript)
76
(No Transcript)
77
(No Transcript)
78
(No Transcript)
79
(No Transcript)
80
(No Transcript)
81
(No Transcript)
82
(No Transcript)
83
Tectonic setting and ruptures of major earthquakes
Seismicity (1964-2000)

• The yellow patches are estimated rupture areas of
  known large subduction events between 1797 and
  2004. Orange patches depict the 2004
  SumatraAndaman rupture where slip was 5 m or
  more. The boundary between Australia and India is
  a diffuse plate boundary between 5 S and 8 N
  (ref. 44). Plate velocities of Australia (black
  arrows) and India (red arrows) relative to Sunda
  were computed from a regional kinematic model.
  Dashed lines are contours of sediment thickness
  at intervals of 2,000 m. The inset shows that the
  age of the sea floor increases northwards, from
  50 Myr in the epicentral area to 80120 Myr at
  the latitude of the Andaman islands.

Next gt 7.0 ?
Subarya et al, Nature 440, 46(Mar 2006)
84
LURR

• LURR - Load/Unload Response Ratio theory is a
  method for earthquake prediction proposed on the
  basis of the constitutive relationship of rock
  deformation.
• Reference
• Yin, X. C. Yin, C., 1991, The precursor of
  instability for nonlinear systems and its
  application to earthquake prediction, Science in
  China, 34, 977-986.
• Yin, X. C., Chen, X. Z., Song, Z. P., Yin, C.,
  1995, A new approach to earthquake prediction
  the Load/Unload Response Ratio (LURR) theory,
  PAGEOPH 145, 701-715.

85
LURR for Sumatra region before 26/12/2004
86
LURR for Sumatra region up to 25/12/2005
Submitted to GJI in June, 2006
87

• How to Extend for Tsunami Source Modeling?
• (ARC Discovery Project 2006-2008)

88
Sumatra Earthquake Vertical Velocity
Proposal (1)
Meltzner et al,JGR, 111(2006) B02407

• Provide such deformation ( including velocity)
• results for the further tsunami simulation

C
C
C
C
89
Slip Pattern Determines Local Tsunami?

• Sumatra Pair
• Dec. 26, 2006
• Mar. 08, 2005
• Tokachi-Oki Pair (Hokkaido, Japan)
• 1952 earthquake
• 2003 earthquake

90
Deeper Earthquake
91
Shallow Earthquakes
92
(No Transcript)
93
Further Applications

• Energy Engineering Field
• Hot Fractured Rock (HFR) Geothermal Reservoir
  Simulation (ARC Linkage Project 2005-2008 with
  Geodynamics Ltd)
• Stress Evolution and Fracture Propagation in
  Yufutsu Gas Reservoir (with JAPEX - Japan
  Petroleum Exploration Co., Ltd. )

94
Schematic Diagram HFR Geothermal Energy
Geodynamics Ltd

• Reservoir construction
• (Hydraulic stimulation)
• Heat extraction
• Evaluation-sustainability

Thermo-Hydro-Geomechanical Coupled Problem
95
Geodynamics Ltd Co. Cooper Basin
Highly Coupled Thermo-Hydro-Mechanical Problem
AGU Fall Meeting 2006 NG08 Geothermal Reservoir
System Conveners Xing Wynborn
Extrapolated basement (3.7 to 4km) temperature
contours in Cooper Basin (Temperatures measured
in existing wells)
Heat flow out of the Habanero 2
96
JAPEX Yufutsu Gas Reservoir
Yufutsu
Yufutsu
97

• ACcESS Beta release (March 2006)
• http//www.access.edu.au

Thanks
98
JAPEX Yufutsu Gas Reservoir - mesh
99
Preliminary Simulation Results Stress Rate
blue lines show the fault geometry
The snapshots of the equivalent stress variation
rate at the different stages (the blues lines
show the fault geometry)
100
Preliminary Simulation Results Velocity
blue lines show the fault geometry
The snapshots of the equivalent stress variation
rate at the different stages (the blues lines
show the fault geometry)
101
(No Transcript)
102
Sumatra Earthquake
103
2. Simulate the solid-water interacting
dynamics in 3D as a whole to provide such
simulation results (water) required for the
further tsunami simulation
104
Proposal (3)

• 3. Simulate the long time scale interacting fault
  system with water in 3D as a whole to

Identify the locations of the high tsunami risk
regions
105
Entire Earth Model
Moon
Earth
106
Tidal Deformation
107
Tidal Deformation
108
Southern Australia Fault Model (SA)
iSERVO Institute Seed Project
109
Grand Challenge Non-Continuum Crust of Entire
Earth
Plate Boundary
Faults