The MOST Experiment: Earthquake Engineering on the Grid

1
The MOST Experiment Earthquake Engineering on
the Grid

• Lee Liming
• George E. Brown, Jr. Network for Earthquake
  Engineering Simulation (NEES)
• National Science Foundation
• Division of Civil and Mechanical Systems

2
Vision of NEES
NEESgrid
3

• NEESgrid facilitates research capabilities
  previously unavailable
• NEESgrid links earthquake researchers across the
  U.S. with leading-edge computing resources and
  research equipment and allowing collaborative
  teams (including remote participants) to plan,
  perform, and publish their experiments
• NEESgrid is a coordinated and secure
  architecture/environment
• NEESgrid is a modular and extensible environment
  with a customizable user interface
• NEESgrid provides common tools that allow
  leveraging resources and experiences
• Rather than having to worry about the required
  cyber infrastructure, NEESgrid allows researchers
  to focus on the earthquake engineering challenges
  at hand
• The goal of the System Integrator (SI) is to
  develop NEESgrid as the Cyber Infrastructure that
  will facilitate this next generation of
  experimentation/simulation in earthquake
  engineering

4
The Grid in NEESgrid
Experimental Component
Grid Data Repository
Grid Operations Center
Campus Net Component
NEESgrid Component
Hub C
Hub A
Hub B
NEESpop A
Teleobservation Equipment
Experimental Equipment
Telepresence Equipment
Passive co-PI
Video I/O
Active PI
Data Cache
Audio I/O
Data Cache
Site A Experimental Data Producer
Site B Remote Lead Investigator
Site C Passive Collaborator
5
The Main Components of NEESgrid

• Computational Simulation tools and interfaces
• Uniform Telecontrol interface (equipment and
  simulation)
• Uniform Streaming Data interface (equip. and
  simulation)
• Data and Metadata services and Data Visualization
  tools
• Data Acquisition (DAQ) and related services
• Tele-Observation (remote viewing) services
• Electronic Notebook capability
• Remote Collaboration tools and services
• Compatibility with (inter)national Grid
  standards, making NEESgrid potentially part of a
  much broader cyberinfrastructure

6
Status of NEES and NEESgrid

• Three-pronged Construction Approach
• 15 Major Facilities Development Awards
• 1 System Integration (SI) Award (NEESgrid)
• 1 Consortium Development Award to establish a
  team for ongoing maintenance
• The SI Activity (NEESgrid) is both a development
  and a deployment effort not separated
  contractually.
• June 2001 through September 2004
• SI team and NEES Consortium are now engaged in
  transition of operations, to be completed in
  September 2004.

7
MOST ExperimentExperiment-based Development

8
What was the MOST Experiment?

• A partnership between the SI Team, UIUC and
  Colorado Equipment Sites to showcase new (and
  previously developed) NEESgrid capabilities
• A large-scale experiment conducted in multiple
  geographical locations. Combining physical
  experiments with numerical simulation in an
  interchangeable manner
• The first integration of NEESgrid services with
  application software developed by Earthquake
  Engineers (UIUC, Colorado and USC) to support a
  real EE experiment
• Demonstrated how collaborative designs can be
  realized by teams of IT professionals and
  earthquake engineers

9
Scientific Rationale

• Scientists currently use testing, analysis, and
  field observations in their work.
• Each method has strengths and weaknesses.
• Field observations are considered useful mainly
  for validating theories resulting from testing
  and analysis.
• The MOST Experiments combination of distributed
  pseudodynamic testing and finite element analysis
  seems ideal for large-scale experiments.
• Pseudodynamic testing involves online,
  computer-controlled testing of real physical
  structures in laboratory conditions using applied
  forces.
• This class of experiment is generally too large
  to be conducted in any single U.S. laboratory.

10
Test Structure for MOST Experiment
Note for ease of programming, all computational
models are Matlab-based for this exercise.
Why This Experiment?
Because we already knew the answer!
11
Primary Capabilities Demonstrated by MOST

• The NCTP telecontrol service used with both
  physical and simulation components, including an
  interchangeable back-end control interface
• Streaming data (Colorado site) and video
  monitoring
• Collaboration interface (especially chat room and
  data viewer)
• Data collection and storage

12
Telecontrol Using NTCP
UIUC Experimental Model
U. Colorado Experimental Model
ExperimentController(NTCP Client)
Each NTCP service is configured with
asite-specific plug-ininterface to the
localequipment/code.
NCSA Computational Model
13
1. Simple Computation/Control Model
UIUC Computational Model
U. Colorado Computational Model
SIMULATION COORDINATOR
NCSA Computational Model
14
2. Switch to NTCP Protocol
UIUC Computational Model
U. Colorado Computational Model
EXPERIMENT COORDINATOR (NTCP Client)
NCSA Computational Model
15
3. Switch One Site to Physical Mode
UIUC Experimental Model
EXPERIMENT COORDINATOR (NTCP Client)
NCSA Computational Model
16
4. Two Physical Sites (MOST)
UIUC Experimental Model
U. Colorado Experimental Model
EXPERIMENTCOORDINATOR(NTCP Client)
NCSA Computational Model
17
Data Collection

• Local Data Acquisition (DAQ) system collects data
  from sensors and provides time-stamped data.
• NEESgrid components transfer data to local
  repository (archival storage) and provide NSDS
  (streaming data) interface.
• Local system is virtualized by services
  interfaces on NEESPOP.

18
Feedback from Users

• Remote participants (those observing through the
  Web) need explicit orientation.
• Communication items (narration/status, chat)
• Documentation items (for both experiment setup
  and NEESgrid features)
• The Web interface is all most people will see.
• The rest of the system just works.
• Dont expect comments on infrastructure
  experiment observers care mainly about the user
  interface, not the infrastructure.

19
Lessons Learned by the SI Team

• Fault-tolerance is important.
• Both the NTCP and the data management services
  experienced faults during the MOST experiment.
  Automatic recovery is critical.
• Video of work areas should focus on the
  experiment.
• All sites must meet NEES specifications to ensure
  satisfactory performance.
• NTTP (clock syncing), network requirements, etc.

20
Two Important Lessons

• Developing standardized tools and resources
  allows the Earthquake Engineer to focus on the
  engineering problem at hand.
• Using common tools leverages resources and
  experiences.
• While a number of potential improvements have
  been identified, it was clear that the basic
  system worked for the MOST experiment.