Computational Elements of Robust Civil Infrastructure

1
Computational Elements of Robust Civil
Infrastructure

• White paper by
• G. Cybenko, K. Fuchs, A. Grama, C. Hoffmann,
• A. Sameh, N. Shroff, M. Sozen, and B.F. Spencer
• September 17, 2002

2
Motivation for Study

• The country has an investment of 20 trillion in
  civil infrastructure.
• Much of this civil infrastructure is
  mission-critical, e.g.,
• bridges
• power plants and power grid towers
• telecommunication centers
• water purification plants

3
Motivation for Study

• Monitoring the health of such infrastructure
  through sensing technology can
• assure timely service,
• detect the onset of catastrophic failure,
• mitigate catastrophic failure, or
• allow for effective contingency plans (crisis
  management).
• Actuation based on sensing infrastructure can
• increase the robustness of such structures very
  significantly,
• enable economical construction of critical
  infrastructure,
• in the event of imminent failure, direct the
  structure to desirable failure modes.

4
Targeted Hazards

• Earthquakes
• Explosions
• Fire
• Rust
• Wind
• Terrorist events

5
State-of-the-art in Controlled Structures -
Passive Control
6
Focus of the study

• Develop the communication, data integration, and
  computational, infrastructure that enables
• Effective design and economical construction of
  highly robust smart structures that sense and
  react to external stimuli and
• Transformation of existing structures into active
  structures that sense, discriminate, and act in
  defense.
• Off-line use of data collected to solve the
  inverse problem determine actual structural
  characteristics and specific stimuli leading to
  failure. This can be done through a series of
  scenario simulations.

7
Research and Development Highlights

• The design and implementation of a low-power/
  low-cost smart sensors-actuators complex (SAC)
  consisting of
• smart sensor networks
• data receptors
• computational elements
• real-time control algorithms

Sensing/Computation/Communication elements -
designed by part of our research team at
Dartmouth. These units cost under 200 and are
the size of a deck of cards. Efforts are on to
develop the next generation of such devices at
Purdue.
8
Research and Development Highlights

• Integrate the SAC with a strut system containing
  controllable dampers (to change the stiffness
  characteristics of the structure).
• a magnetorheological (MR) device, also referred
  to as a smart-strut-device (SSD).

Magnetorheostatic dampers can change their load
bearing characteristics from fully solid to fully
damping in milliseconds when exposed to magnetic
fields.
9
Research and Development Highlights

• Develop distributed strategies for computing
  control vector from sensed signals in real time.
• Develop detailed simulation methodologies for
  validating control strategies and examining a
  variety of what-if scenarios for a range of
  stimuli.

10
Research and Development Highlights

• Detailed methodologies for design of structures,
  including placement and capability of sensors and
  actuators, precise calibration of impact bearing
  capacity of the structure.
• Real-time visual information infrastructure to
  support status checks, and rescue and relief
  efforts.

11
Research and Development Tasks

1. Development of self-configuring, self-calibrating
   wireless sensor networks and low-latency sensor
   data management techniques.
2. Development of algorithms and software for
   continuous real-time testing, diagnosis, and
   maintenance for all communication and
   computational components of the sensor/actuator
   networks.
3. Fault-tolerant operation of the SAC-SSD
   complexes.

12
Research and Development Tasks

• Model reduction of the large-scale dynamical
  system representing the structure (off-line).
• Development of distributed, real-time (on-line)
  algorithms for determining the structures
  response to dynamic impulses using the reduced
  low-order model, together with a real-time
  visualization environment.
• Development of rapid simulation and visualization
  infrastructure for exploring (off-line) a range
  of what-if scenarios for real-time disaster
  management and control strategies.

13
Research and Development Tasks

• Validation of the entire computational paradigm
  on well-instrumented model structures, as well as
  actual instrumented structures in Puerto Rico
  (wind effects), and Japan (earthquakes).

14
Unique Qualifications of Research Team

• Extensive experience building and applying sensor
  networks (Cybenko, Dartmouth, Shroff, Purdue).
• Pioneered the development and use of smart-strut
  devices (Sozen, Purdue, Spencer, Illinois).
• Fundamental work in fault tolerance, testing, and
  system validation (Fuchs, Cornell).
• Experts in geometric modeling, large scale
  simulation, and visualization infrastructure,
  more recently, applied to the Pentagon crash
  simulation (Hoffmann, Purdue),
• Parallel and distributed algorithms for
  structural modeling, model reduction, and control
  (Sameh, Grama, Purdue).

15
Relation of Project to Other Sensor Network
Efforts

• The fundamental goal of this project is to build
  robust civil infrastructure.
• From this point of view, the aim is one of
  integrating a range of existing technologies, and
  where needed, to develop new technologies.
• Its primary aim is not to build a new class of
  sensors or RF communication devices. It is our
  belief that these technologies have matured to a
  point where they can safely be used for solving
  the critical task of securing civil
  infrastructure.