Identification Methods for Structural Systems

Prof. Dr. Eleni Chatzi

Lecture 1 - 24 February, 2015

Institute of Structural Engineering

Identification Methods for Structural Systems

Course Information
Instructor
Prof. Dr. Eleni Chatzi, email: [email protected]
Office Hours: HIL E14.3, Wednesday 10:00-12:00 or by email

Assistant
Dr. Vasileios Ntertimanis, email: [email protected]
Office Hours: HIF B21, Tuesday,Thursday 10:00-12:00 or by email

Course Website
Lecture Notes and Homeworks will be posted at:
http://www.chatzi.ibk.ethz.ch/education/identification-methods-forstructural-systems.html
Suggested Reading
T. S
oderstr
om, P.G. Stoica, System identification, Englewood Cliffs,
Prentice-Hall, 1989
available for download from: http://user.it.uu.se/ ts/sysidbook.pdf
Institute of Structural Engineering

Identification Methods for Structural Systems

Grading Policy
Homeworks (30%) - Final Exam (70%)

Homework
3 Homework Problems will be assigned in total
Homeworks are due in class up to 3 weeks after assignment
HWs 2 and 3 will also require some light coding
This maybe done using any coding language (MATLAB, Fortran, C,
MAPLE) - example code will be provided in MATLAB

Final Exam
The final written examination will have a duration of 2 hours
Topics will address examples done in class or in Homework problems
No coding will be necessary

Institute of Structural Engineering

Identification Methods for Structural Systems

Background & Motivation

Operating under continually changing Environments

Wind Induced Vibrations

Wind Turbine

Pedestrian Vibrations
Millennium Bridge

(source: Youtube)
(source: Getty Images)

EVACES15, 19 - 21 October 2015

Monitoring based on Uncertainty

Background & Motivation

Operating under continually changing Environments

Wind Induced Vibrations

Wind Turbine
Pedestrian Vibrations
(source: imgur)

Millennium Bridge
(source: Youtube)

EVACES15, 19 - 21 October 2015

Monitoring based on Uncertainty

Background & Motivation

Deteriorating Infrastructure

Wake-up Call
I-35W highway bridge Collapse (2007)

EVACES15, 19 - 21 October 2015

Monitoring based on Uncertainty

American Society of Civil Engineerings (ASCE) report card that we have all come

much. Since its inception in 1988, there have been 5 complete report cards (at le
Background & Motivation
found) and one trend report. When growing up, I wonder if I brought home a repo

poorly, time and time again, if my parents wouldve awarded me with nice stimulus

Ha. The only stimulus I wouldve received would have been to my backside from t
Deteriorating Infrastructure
wooden spoon!
Take a look:
2003
1988 1998 2001 Trend 2005 2009
CATEGORY
Aviation
Bridges
Dams
Drinking Water
Energy
Hazardous Waste
Navigable Waterways
Public Parks & Rec
Railways
Roads
Schools
Solid Waste
Transit
Wastewater

BC+
N/A
BN/A
D
N/A
N/A
N/A
C+
N/A
CCC

CCD
D
N/A
DN/A
N/A
N/A
DF
CCD+

D
C
D
D
D+
D+
D+
N/A
N/A
D+
DC+
CD

N/A
N/A

D+
C
D
DD
D
DCCD
D
C+
D+
D-

D
C
D
DD+
D
DCCDD
C+
D
D-

(source ASCE)

I know that if my kids came home with these grades one marking period, you woul
you would see improvement the next. And then, as a parent, I would work with th
school to learn the root of problem and then make the necessary changes that wou
term success. Finding short term solutions and putting a band-aid over the situation

Institute of Structural Engineering

Identification Methods for Structural Systems

Background & Motivation

Ageing Infrastructure is a worldwide trend
In Western Europe, where
some infrastructure
systems are far older,
maintenance and repair
issues surfaced long before
they did in the US.

This country desperately needs new

infrastructure. The road network is
deteriorating.- Boris Johnson

European nations have

already drawn up extensive
infrastructure plans.
Bridge - dehumidification project

Institute of Structural Engineering

Identification Methods for Structural Systems

Background & Motivation

Structural Innovation

Institute of Structural Engineering

Identification Methods for Structural Systems

Background & Motivation

Structural Health Monitoring
The strategic and monetary value of the civil infrastructures
worldwide call for the development of SHM systems.
Smart
Materials
Damage

Sensors

SHM
IT
Network

Institute of Structural Engineering

Detection

Life
Cycle

Identification Methods for Structural Systems

Background & Motivation

Structural Health Monitoring
Current data acquisition technology.

Smart
Materials

Damage

Sensors

SHM
IT
Network

Institute of Structural Engineering

Detection

Life
Cycle

Identification Methods for Structural Systems

Background & Motivation

Collection of Sensory Information
Force

Displacement

Acceleration

Meteo

EVACES15, 19 - 21 October 2015

Strain and tilt

Data from GPS

Monitoring based on Uncertainty

Sensor Technology

Fiber Optics
Ultrasonic Sensor
Laser Vibrometer

Institute of Structural Engineering

Identification Methods for Structural Systems

Smart Sensors & Materials

Shape Memory
Alloys

Wireless Smart
Sensor

Institute of Structural Engineering

Bridge Skin
(source: University of
Michigan, Jerome Lynch)

Identification Methods for Structural Systems

Smart Sensors & Materials

Fiber Reinforced Polymer Materials

(courtesy of Dr. Fuggini, DAppolonia S.p.A)

Institute of Structural Engineering

Identification Methods for Structural Systems

Sensor Technology
Interferometric Radar
Generates maps of surface deformation over time

courtesy of A.W. Smyth

Institute of Structural Engineering

Identification Methods for Structural Systems

Sensor Technology
Acoustic Emission Detection Schemes

Strain sensor detects

approaching trains and wake AE
sensors.
Using Artificial Intelligence
Methods fatigue crack growth is
quantified.
Information is transmitted via
wireless network or satellite
transmission.
Decision making Process is
activated.
taken from http://www.tisec.com
Institute of Structural Engineering

Identification Methods for Structural Systems

IT Network Technology
Efficient SHM systems shall completely avoid the issue of wire management
the development of efficient data communication systems and energy
harvesting techniques is required.

Institute of Structural Engineering

Identification Methods for Structural Systems

Link between acquired data & modeling of structural behaviour ?

System Identification
Institute of Structural Engineering

Identification Methods for Structural Systems

Link between acquired data & modeling of structural behaviour ?

System Identification
Institute of Structural Engineering

Identification Methods for Structural Systems

Definition
System Identification
is the process of developing or improving the mathematical
representation of a physical system using experimental data.
analytical or numerical
model

excitation

dynamic response

Model

Forward Problem (analytical/numerical model exists a priori)

ambient
excitation

vibration
response

and/or
known applied forces

Time domain
x[t] = Ax[t-1]+Bu[t]
y[t] = Cx[t]+Du[t]
Structural
identification

Frequency Domain
H(ejt)
Modal space
, ,

Inverse Problem

(structural model is obtained from experimental data)

Institute of Structural Engineering

Identification Methods for Structural Systems

Background & Motivation

Turning Data Into Knowledge - System Identification
is the process of developing or improving the mathematical representation
of a physical system using experimental data.

(Manhattan Bridge Monitoring Project - Columbia University, 2008)

EVACES15, 19 - 21 October 2015

Monitoring based on Uncertainty

Application of Structural ID for Modal Analysis

Cross Section

Institute of Structural Engineering

Identification Methods for Structural Systems

Application of Structural ID for Modal Analysis

Dynamic Differential GPS and Acceleration Sensing

Institute of Structural Engineering

Identification Methods for Structural Systems

Application of Structural ID for Modal Analysis

Dynamic Differential GPS and Acceleration Sensing

Institute of Structural Engineering

Identification Methods for Structural Systems

The Manhattan Bridge - Instrumentation Plan

Courtesy of A. W. Smyth

Institute of Structural Engineering

Identification Methods for Structural Systems

Application of Structural ID for Modal Analysis

Sensor Installation

Institute of Structural Engineering

Identification Methods for Structural Systems

Background & Motivation

The Manhattan Bridge - Identified Modes

Courtesy of Prof. A.W. Smyth

EVACES15, 19 - 21 October 2015

Monitoring based on Uncertainty

Application of Structural ID for Vibration Control

Tuned Mass Damper Installation to the Millennium Bridge, London

8 Horizontal TMD
Weight: 2500 kg
Frequency: 0.45 Hz

50 Vertical TMD
Weight: 1000 - 2000 kg
Frequency: 1.2 - 2.2 Hz
(source: www.gerbusa.com)

Institute of Structural Engineering

Identification Methods for Structural Systems

Background and Motivation

Multi-source Uncertainties

Random (stochastic) excitation: wind, turbulence, traffic, earthquake, combined effects

due to various sources, road or rail irregularities, etc
Operational conditions often assumed as WN.
Random (stochastic) noise: instrument noise, discretization noise, modelling errors
environmental effects and disturbances, interferences.
The structural system: Assumed to be deterministic usually (but not necessarily) and
either linear & time invariant (stationary) or less often nonlinear

EVACES15, 19 - 21 October 2015

Monitoring based on Uncertainty

Background and Motivation

Multi-source Uncertainties

Random (stochastic) excitation: wind, turbulence, traffic, earthquake, combined effects

due to various sources, road or rail irregularities, etc
Operational conditions often assumed as WN.
Random (stochastic) noise: instrument noise, discretization noise, modelling errors
environmental effects and disturbances, interferences.
The structural system: Assumed to be deterministic usually (but not necessarily) and
either linear & time invariant (stationary) or less often nonlinear

EVACES15, 19 - 21 October 2015

Monitoring based on Uncertainty

Damage Detection
Nondestructive testing (NDT): defines methods used to examine
an object, material or system without impairing its future usefulness.
SHM - Damage Detection:
Monitor changes in the dynamic
properties of a structure

Available Techniques: impedance

tomography, radiography,
ultrasounds

Institute of Structural Engineering

Identification Methods for Structural Systems

Application of Structural ID for SHM

Baselinefphasef?modellingfoffthefhealthyfstructuralfdynamics)
non-parametric
estimationfoffPDF

PDF

Temperaturef?oC)

vibration
response

Qo()

Temperaturef?oC)

Timef?h)

ambient
excitation

Structuralfmodel
estimation

PDF

Average
floadfmassf?kg)

Averagefloadfmassf?kg)

Timef?h)

Q: Characteristic statistical quantity that may be used for the

description of the structural state (e.g. natural frequencies)
: Vector of uncertainty parameters

non-parametric
estimationfoffPDF

Inspectionfphasef?damagefdetection)
structure under unknown state
measured conditions 1, 2
ambient
excitation

vibration
response

statisticalfdecisionfmaking

Structural
model
estimation

Qu()

Qo()

YES

healthy
state

Qu()
NO

faulty
state

Institute of Structural Engineering

Identification Methods for Structural Systems

Worldwide Monitoring Projects

Fiber Optic sensing system - Sutong Bridge (China)

Advantage: capability of distributed sensing and measurement
Obstacle: Lack of engineering demonstration of the durability and
long-term performance

Institute of Structural Engineering

Identification Methods for Structural Systems

Worldwide Monitoring Projects

Egnantia Highway - Greece

(source: Ntotsios et al 2009)

Institute of Structural Engineering

Stork Bridge Outdoor

Deployment - Winterthur

(source: EMPA)

Identification Methods for Structural Systems

Monitoring of Tall Structures

SHM System:
Monitoring of strain/stress
and structural temperature,
Acceleration, tilt and
displacement,
Ground motion, wind and
weather conditions
Corrosion
Monitoring Guangzhou TV tower - source:
Hong Kong Polytechnic University

Institute of Structural Engineering

Identification Methods for Structural Systems

Worldwide Monitoring Projects

CSMIP and Caltrans created the bridge strong motion program
after the Loma Prieta EQ (1989).
Includes instrumentation of highway bridges, free-field sites near
major bridges, and subsurface geotechnical arrays.
Goals of instrumentation:
Provide valuable information to improve
seismic design and analysis of bridges.
Near real-time recovery and processing
of recorded data provides significant
information for assessment of the safety
of the bridge structures right after
damaging EQ.

Institute of Structural Engineering

Identification Methods for Structural Systems

Chair of Structural Mechanics, ETH Zurich - Monitoring Activities

The ETH House of Natural Resources: An opportunity to implement,
demonstrate and validate innovative structures in hardwood

EVACES15, 19 - 21 October 2015

Monitoring based on Uncertainty

10

Chair of Structural Mechanics, ETH Zurich - ETH HoNR

SHM - Data Stream

Dynamic field tests during construction
Ambient dynamic recordings during operation
Long-term Static Monitoring (construction & operation)

Snow loads
Wind loads

BIBLIOTHEK / AUFENTHALT
BRO

BRO

SEKRETAR.

SITZUNGSZIMMER

BRO

User behaviour
BRO

ARCHIV

EVACES15, 19 - 21 October 2015

Monitoring based on Uncertainty

11

Chair of Structural Mechanics, ETH Zurich - ETH HoNR

Data Stream: Dynamic Regime

Accelerometers

Accelerometer placement

Amplifier Box

EVACES15, 19 - 21 October 2015

Monitoring based on Uncertainty

12

Chair of Structural Mechanics, ETH Zurich - ETH HoNR

Data Stream: Static Regime

EVACES15, 19 - 21 October 2015

16 Load cells

12 LVDTs (horizontal/
vertical)

12 Temperature and
humidity sensors

40 Moisture content
measurement points

20 FBG Strain sensors

2 LUNA Strain cables

Monitoring based on Uncertainty

13

Chair of Structural Mechanics, ETH Zurich - Monitoring Activities

Strain Monitoring - Elefantenhaus Z
urich Zoo

EVACES15, 19 - 21 October 2015

Monitoring based on Uncertainty

18

Chair of Structural Mechanics, ETH Zurich - Monitoring Activities

Strain Monitoring - Elefantenhaus Z
urich Zoo
Kaeng
Krachan
Elephant Shelter
at Zoo Zrich,
Switzerland

Wooden free-form cupola of 80 m

span.
Roof supported by a post-tensioned
concrete ring.

EVACES15, 19 - 21 October 2015

Monitoring based on Uncertainty

19

Infrastructure Management through SHM

Detect damage and deterioration before it poses a risk
Monitor expected design behavior of structural system not only
at initial stage but also during the life cycle of a structure
Prognosis, i.e., determine remaining life of structures
Trigger, remedial action, passive or active control of structure
Optimize investment in maintenance
Manage risk

Goal:

is not to supersede traditional inspections with monitoring

but to complement and optimize the inspection and assessment
process with modern tools

Institute of Structural Engineering

Identification Methods for Structural Systems

Chair of Structural Mechanics, ETH Zurich

Value of Information for Infrastructure Management

EVACES15, 19 - 21 October 2015

Monitoring based on Uncertainty

Chair of Structural Mechanics, ETH Zurich - Monitoring Activities

Facilitating Maintenance Decisions through Monitoring

EVACES15, 19 - 21 October 2015

Monitoring based on Uncertainty

44

Course Outline

Physics-based modeling
Time-domain (ODE and State-space)
Laplace-domain (Transfer function)
Frequency-domain (Frequency response function)

Data-based modeling
Fundamentals of Signal Processing
Non-parametric methods (Peak-picking, FDD)
Parametric methods (ARMA, Stochastic Subspace ID, KF)

Extension to non-linear and non-stationary structures

Recent advances and applications

Institute of Structural Engineering

Identification Methods for Structural Systems

Thank you for your attention!

We welcome questions/comments/collaboration
[email protected]

EVACES15, 19 - 21 October 2015

Monitoring based on Uncertainty

56