Next Generation Structural Connections:

shootperchΠολεοδομικά Έργα

26 Νοε 2013 (πριν από 3 χρόνια και 9 μήνες)

105 εμφανίσεις

Next Generation Structural Connections:


HF2V devices versus Hinges and Sacrifice

‘Energy Dissipation Without Damage’

Geoff Chase, Professor in Mechanical Engineering

Christchurch, NZ

Taiwan

Case 1: Long long ago …


Lisbon Earthquake of 1755
:


Lisbon 1754 = major economic and political power


Lisbon 1755 = major earthquake and tsunami on All Saints Day


Lisbon 1756


= never recovered its stature as a major power



10,000
-
40,000 dead (or 5
-
20% of population)


2 years later population was only 114,000 of an original 200,000 as
many survivors left



Economic loss of 32
-
48% total Portuguese GDP


Much through lost treasure which we don’t have stored today!


~20% hit to GDP (very approximate) without that


Case 2: The recent past …


Ji
-
J
i, Taiwan, 1999


Widespread power losses marked with significant structural
damage


2400 deaths, 600,000+ displaced from homes


US$8
-
10B in economic costs & damage = ~5% of real GDP


Measured M = 7.2 on the Richter scale


Relatively small

for
region!



Caused severe disruption to the silicon chip and flat panel supply

o
PC prices actually rose! Waiting times of 3
-
4 months for laptops

o
Billions in losses for major players like Intel, Dell, …


Mfg’ers now require Taiwan companies to insure losses with the
price appearing in the cost to consumer



Taiwanese economy still hasn’t fully recovered by some accounts


Case 3: The near/distant future …


Tokyo, Japan

highly seismic & “due” for a major event


The “Great Tokyo/Kanto Quake” of 1923 measured M = 8.3


A maximum likely event is estimated at up to M = 9.0



Estimated economic and damage costs of US$1T for M = 7.5


Yes, that’s a “T” as in Trillion


~20% of GDP (same as Lisbon)



Estimated 11,000 deaths



Worldwide economic disruption

A Simple Story


For relatively
large design
events



Lives lost has
gone down to
nearly zero



Cost and
economic
damage has
remained
constant

Lisbon

1755

Ji
-
Ji

1999

Cost as %GDP = 15
-
20%

X
-

30
-
50% of population lost

X
-

0.1% of population lost

The Problem?


Structural Damage
:


Stops business from operating (at full capacity or at all)


Displaces people from homes


Economies slow further as the displaced seek opportunity
elsewhere



As a result
:


People seek opportunity elsewhere due to lost jobs and homes


Regions lose talented skilled trades and professionals


Physical reconstruction slows


Business further damaged by loss of skilled employees


Economic recovery drags on even more slowly (over 1 decade)



Life Safety, Sacrificial Design & Damage

Seismic
Damage

Structural
Damage

Occupant Safety
& Internal
Damage

How to mitigate all these effects?!?

Residual Permanent
Deformation &
Yielding?

Joint Working &
Hysteretic
Energy?

Transient Inter
-
story Drift?

Impulse &
Structural Jerk?

Structural
Acceleration?


Current structural design procedures focus on life safety through
a sacrificial design approach


buildings are lost but lives are
saved


Hence, you see lives lost going well down, but economic impact little
changed over time.



A sacrificial design approach is no longer “good enough”, …



… but, what comes next?


Damage Avoidance Design


an emerging discipline


Smart retrofit for existing buildings


Next Generation Structures?

Damage to
concrete and
yielding of re
-
bar

Significantly reduce yielding of the critical reinforcing bars.

This also absorbs more energy


particularly after 1
-
2 cycles.

Problem:

Solution:

Requires very
expensive repair

Does not require
expensive repair

Armoured rocking
interface

Device fixed to re
-
bar cage
and embedded in concrete
for containment

Axial
Prestress

Possible Applications

Seismically Vulnerable Bridge Piers


Steel joints


Reinforced concrete joints


Bridge decks


Tuned mass dampers


Base isolation


Single or
double bulge
extrusion
damper fixed
to column

Steel Beam

Steel
Beam

Steel
Column

Dissipative
rocking with
no damage

Gap transmits
joint rotation to
damper instead
of damage

Direct Placement into Steel Joints

Main Goals


High force capacity

= High dissipation


Only 3
-
10 large response cycles per big earthquake



Small device volume


Tight constraints for typical structural connections
-

Universal
column sections nominally 350mm deep


W14 in American Codes



Maximum energy dissipation per cycle


“Square” hysteresis loop



Goal
:

Dissipate energy in the device every cycle rather
than by damage to structural connections

Outline


Devices


Design


Experimental characterisation, modeling and validation



Large
-
Scale Proof of Concept Experiments


Concrete


post
-
tensioned DAD + external devices


Full joint


Corner joint


Concrete


post
-
tensioned DAD + internal devices


Steel


hinged + device



Spectral Analysis and Design Guidelines


Takes solutions to full scale performance based structural design

Device Mechanics

Plastic extrusion of working material through an annular restriction

Bulged
-
shaft design chosen for low manufacturing cost and
repeatable results

Bulged
-
shaft type extrusion damper

Constricted
-
tube type extrusion damper

Source: Robinson and
Greenbank

120
-
250 kN Devices + Comparison

120 kN device (left) is approximately

33% the size of 250 kN device (right)


Most of size is mounting brackets!


Similar devices are normally 1.7+m!



700 kN




100 kN



1.7m (appr.)

Sinusoidal
Input



1000kN compression testing device



Quasi
-
Static



DARTEC 10,000kN testing device



Psuedo
-
dynamic testing


16
-
20mm/sec

Lead

Fixed in
Place

Testing Method



40mm bulge



Quasi
-
static test



100kN peak force, 50kN average



Not repeatable



Void formation


40 mm diameter bulge with un-pre-stressed lead
-150
-100
-50
0
50
100
150
-20
-10
0
10
20
30
Displacement (mm)
Force (kN)
Results without pre
-
stress

Device coring trailing void

Peak force only in
“new” material

Results with Pre
-
Stress



Air voids decreased by 3
-
5x



‘Squarer’ hysteresis loops


for max energy dissipation



Repeatable results










Small decrease in force still


due to air voids, but much less



Air voids same approx volume



40mm bulge has longer void as


smaller bulge than 50mm as it


is about void
volume




40 mm bulge

50 mm bulge

Experimental Relationships



Relationships do not fit standard extrusion models




linear and polynomial



Linear relationship

in tested region

33%

difference

Experimental Relationships



Ratio of bulge area to force


makes relationship

device


independent




Linear

relationship




Wider variety of devices




Relationship used to estimate


design force


Can be readily used for
new device designs

0
50
100
150
200
250
300
350
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
Face Area bulge/Area cylinder
Force (kN)
Relationship between ratio of bulge area

to cylinder area and Force

Mohr
-
Coulomb Stress Based Model

e
bu
o
shaft
D
D
A
A
D
lg




Combination of direct stress applied to the bulge face area and shear
stress on device shaft.

Much better agreement with the experimental results than classic extrusion models

Re
-
Centring Devices

Change in volume of working material as shaft is displaced through
the cylinder

Increased pressure inside device acts on net face area and forces
shaft out of the device, providing a re
-
centring ability

Applications for Re
-
Centring Devices

Recentring
Device

Precast
Reinforced
Concrete
Beam

Armoured
Rocking
Interface

Post
-
Tensioned
Connecting
Bolts

Reinforced
Concrete
Column

Recentring devices developed into beams carry post
-
tensioning in
place of unbonded tendons.

Main Accomplishments
-

Device



Importance of pre
-
stress in repeatable device behaviour and
maximum energy dissipation identified



Full
-
scale 300kN+ prototype created and tested



Final device design underwent proof
-
of
-
concept testing over entire
speed range



Analytical design space characterised and device independent


Results not same compared to studies reported in the literature



Comprehensive theoretical simulation of device placement in a
structure shows significant impact on response



Future work requires some minor modifications but the concept is
proven for multiple devices



Structural Implementation

Placement of lead extrusion dampers into the rocking interface of a
reinforced concrete joint.

Structural Implementation

Recent research has focused on damage
-
free jointed pre
-
cast
connections

Beams and columns are completely separate entities, held together
by post
-
tensioned tendons running through the joint region

Structural Implementation

Non
-
linear response is achieved by joint
-
opening at the connection
in place of structural damage and the formation of a plastic hinge

Gap opening elongates tendon,
increasing the tendon force that
pulls the joint closed, providing
an elastic resistance, with
hysteretic energy absorption
due to friction alone.

Bare
Joint

Bare
Joint

Damper Hysteresis

+

Overall
Joint

=

Large amounts of energy
absorbed in a completely
damage
-
free manner, with
overall recentring ability

Experimental Joint Testing

A near full
-
scale 3D joint configuration was constructed using
Damage
-
Avoidance Design principles.


This joint utilised a bend tendon profile

Tendon Profile

Externally Mounted Dampers

Damage
-
Avoidance
-
Design (DAD) joint have very low inherent
damping, so another form of energy dissipation is provided.


By mounting lead extrusion dampers across the joint region, joint
opening leads to large amounts of energy being absorbed directly in
the joint.

Experimental Results

Response is a combination of initial linear elastic deflection of
members, followed by non
-
linear response due to gap opening.


Hysteretic area is a combination of extrusion damping and friction of
the tendons within the duct.

Initial linear
stiffness

(elastic member
deflection)

Post gap
-
opening
stiffness

(further elastic
member deflection
and gap
-
opening)

Rocking Edge

Gap Opening

Corner Joint

The original joint has only most amounts of added damping, kept low
to ensure ready recentring of the overall joint.


In reality, much more damping could be provided without the loss of
recentring.


Therefore, one of the seismic beams was removed, and both
extrusion dampers mounted onto one beam to double the amount of
damping provided

Corner Joint Results

Much larger hysteretic area


much larger dissipation


Asymmetry of force due to eccentric tendon profile in corner


Asymmetry of inherent (no dampers) hysteresis due to tendon profile


cancelled over whole floor by the opposing corner(s)

Embedded Joint Design

Another near full
-
scale 3D experiment, this time utilising internally
mounted extrusion dampers and a straight tendon profile


Three 250kN lead extrusion dampers cast directly into the beam ends


a fully internal system

Extrusion Damper with items for scale

Extrusion Damper internally mounted
in beam before casting

0
4
8
12
16
-300
-200
-100
0
100
200
300
Displacement (mm)
Load (kN)
Device Hysteresis

Finished Joint

Fully internal dissipation system



a design that should prove popular with the architects.

Internal Joint Results

Large energy dissipation with any notable damage or
strength degradation

Two fully reversed cycles
at 1, 2, 3 and 4% drift

Internal Joint Results

Comparison with and without dampers

Marked difference between two results


Inherent hysteretic absorption much
lower than previous joint due straight tendon profile and less friction in ducts

Hysteretic Modelling

Compound, rate dependent Menegotto
-
Pinto equation

Good agreement to experimental results


with and without dampers

Steel Joint



120kN devices on
both sides



Could be 1 underneath (stroke limited)



Device can be either in column or beam



Hinge connection at bracket



Cut
-
out allows rotation

Steel Joint

+



Significant added damping with
no

damage



Far better hysteresis loop than sacrificial connection



Consistent and repeatable results over several days



Easy to assemble or retrofit inside connection

Energy Absorption

Energy absorbed


䔠㴠慲e愠睩瑨楮 hystere獩s 潰


E晦楣ien捹Ⱐ
η

= area ratio to perfect hysteresis loop/elasto
-
plastic response

-40
-30
-20
-10
0
10
20
30
40
-3.5
-3
-2.5
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
2.5
3
3.5
Lateral Load (kN)

Lateral Drift (%)



h

㴠=㘥



h

㴠=〥景爠物杩搠d瑥tl



1.33x increase in best case




No Damage!



Rigid steel joint is damaged




x

㴠=㘥敱畩vv楳i潵猠摡浰d湧

Steel Joint Summary



Energy absorption needs to be compared to sacrificial damage approach



Compare enclosed hysteretic area to that enclosed by a perfectly


elasto
-
plastic response with the same initial stiffness and strength.





This approach yields an energy absorption efficiency factor of
h

㴠=㘥



Assumes no slip conditions can be met (not shown)





For 3% drift the dampers provide
36% equivalent viscous damping





Previous research has
h

=

㘰6

浡硩x畭

景爠獴敥氠晲f浥獴s畣瑵牥u





Therefore, incorporating extrusion damping into well
-
detailed joints


results in superior energy dissipation than a conventional welded


connection, and can do so without any damage to the structural system


Spectral Analysis


Spectral Response decreases with increasing damper size



Reduced magnitude of structural response and damage



Multiple earthquake suites (3x20=60) of probabilistically scaled


earthquake records


near field and far field ground motions

Reduction in response with increasing damper size

Reduction in response with increasing damper size

El Centro Response Spectra

Spectral Analysis

Reduction Factors



Design spectra divided by the spectra with added extrusion damping



Indicates the magnitude of the achieved reduction in response


Reduction
factors
increase

with increasing
damper size

Structural Design Impact




Spl
it the spectra into 3 regions based upon existing bifurcation points



Use linear regression analysis and linear interpolation



Obtain equations to estimate damping reduction factors



Enable use in structural design analyses

Multiple Equation Regression Analysis


ADRS for design

The final step in an effective overall analysis method to
relate new damping devices to performance based
design methods.

Damper Tests to ADRS Guidelines

The complete procedure relate additional extrusion damping to
Performance Based Design Methods

3. Empirical Reduction Factor Equations

4. ADRS

1. Full Scale Damper Tests

2. Response Spectra

Resulting in design spectra for

structural designers in practice

High Force
-
to
-
Volume Experiment
and Overall Summary


Significant added force capacity



Weak joint with device can outperform sacrificial
designs in first and every cycle


by design!


Easily fit inside of existing typical structural connections



Two application tests shown


concrete and steel



Design equations derived for devices and for application
(spectral analysis)


we can make any “size” for any
application


The Bottom Line … It’s all about $


HF2V devices provide significant dissipation and ZERO damage




HF2V Dampers



$100NZ

each (or less in volume), no maintenance,
but requires rethinking connection designs




Expected Annualised Losses (EAL) reduced 5x


$1240/$1M of building cost


␲㘰⼤/䴠
ㄮ㈥


〮0㔥 潲敳猡)


Major impact on insurance, depreciation costs, and overall building
economics for end
-
users





Conclusions


Novel devices enable significant new DAD applications


Open several new design avenues not previously available



Spectral analysis


䑥獩杮⁇畩摥汩n敳⽓灥捴牡c捲敡瑥t愠
bridge to the profession and application



Damage can be significantly reduced or eliminated


Save significant annualised costs at the same time



Full or large scale validation of all concepts


Several further applications show promise in analysis (using
validated models) and/or hybrid experimental testing

Thanks to....



Thierry Alnot, V. Novello & M. Miguelgorry (INSA Rouen)



Samuel Mermet, Jeremie Garnier (SUPMECA)



Alexandre Gue’ (ESIA)



Tobias Bacht (Karlsruhe Univ)



Lt Stephen Hunt RNZAF, Kevin Solberg, Cameron Ewing,


and Ishan Singh
-
Levett (UoC)

Questions?