NEESR-CR: Design of soil and structure compatible yielding
to improve system performance[CoSSY]
Team Meeting 14 October 2009
Kutter, Hutchinson, Aschheim, Kunnath, co-PI’s
Hakhamaneshi, Liu, Vargas, grad students
Moore, Martin, Mejia, Mar, Comartin, Browning, TTT
OUTLINE• Background – system concepts, design
concepts for bridges. Modeling footing rocking• NEESR project – Goals and approach,
Fundamental research questions• EOT• Centrifuge Modeling (Task 1)• Discuss goals of meeting: Consensus and
confidence in direction on goals, approach and resesarch questions.
(a) fixed-fixed (b) fixed-hinged (c) fixed-rocking; (d) hinged-rocking
Work for Caltrans: Idealized failure mechanisms
(b)
Seismic load
Plastic hinges Plastic hinges
(d)Plastic hinges
Column is protected by rocking isolation in case (d)
(a)
(c)
Caltrans SDC: “foundation components shall be designed to remain essentially elastic when resisting the plastic hinging moments”.
Destroyed columns from 1995 Kobe earthquake
Inspectable, controllable with proper reinforcing, but catastrophic results if ductility capacity is exceeded.
• There is a critical (minimum) contact length, Lc, required to support the vertical load, V.
• Moment capacity (from equilibrium) is
• Define Ac = Lc B
where B = footing width• note Ac /A = Lc /L for 1-D loading
• Lc/L << 1 for typical bridge foundations.
• Mo,ult is insensitive to Lc /L
Definitions and basic concepts
L
LLVM c
ulto 12,
Container and Test Setup (Slow
Cyclic)
LPLP
LPLoad Cell
Load Cell
Load Cell
Station A Station B Station C
beams continue for length of box
Z
y
ACTUATOR
ACTUATOR
200
Nevada Sand
(Vertical Load Test) (Horizontal Load, Standard Ht)
Transverse Teflon Supports
Model Shear Wall
Moment-rotation-settlement behavior of rocking foundation from slow cyclic tests
L/Lc = 2.2 L/Lc = 3
L/Lc = 3.8 L/Lc = 12
9
Numerical simulation
Black: qz gapping spring; Green: non-gapping spring (control settlement); Red: Elastic perfect plastic springs (side friction)
• Phase 1: A 2-D beam-on-nonlinear-Winkler foundation (BNWF) model was developed to verify JAU01 tests.
n1=6; n2=5
-40000
-30000
-20000
-10000
0
10000
20000
30000
40000
-0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.06
Rotation (rad)
Mom
ent
(kN
.m)
NumericalExperimental
-0.1
-0.08
-0.06
-0.04
-0.02
0
0.02
0.04
-0.0800 -0.0600 -0.0400 -0.0200 0.0000 0.0200 0.0400 0.0600
Rotation (rad)
Set
tle
men
t (m
)
Experimental
Numerical
Numerical calibration of JAU01 tests
Bridge System
ConceptsUnits: mmIn prototype setup
Column elements: beamWithHingesTop pin elements: nonlinear rot. springsFoundation elements: Qzsimple1, EPP...
X100911
12
13
500921
22
23
31
32
33
Y
Z
311
312313
333332
331
(11)
(12)
(13)
(21)
(22)
(23)
(33) (34)
(311)
(313)(312)
(333)
(331)
(332)
(120) (220)120 220
313131
42
3131 41
43
44
45
46
47
48
49
50
(411)(412)
(413)(414)
(415)(416)
(417)(418)
(419)(420)
(421)(422)
LJD02_15 event: Gazli 2.0
2 0 21.5 10
4
1 104
5000
0
5000
1 104
1.5 104
Centrifuge tes tsStatic loading tests
Bending moment vs rotation, SN column
Column rotation (%)
Co
lum
n m
om
ent
(kN
*m
)
2 0 21.5 10
4
1 104
5000
0
5000
1 104
1.5 104
Centrifuge tes tsStatic loading tests
Bending moment vs rotation, SS column
Column rotation (%)
Co
lum
n m
om
ent
(kN
*m
)
0 10 20 30 401
0.5
0
0.5
1Vertical acceleration of bridge deck
Time (s)
Acc
eler
atio
n (g
)
1.5 1 0.5 0 0.5
1 104
0
1 104
Moment vs rotationTheo ftg moment capTheo ftg moment cap
Moment vs rotation, large south footing
Footing rotation (%)
Foo
tin
g m
om
ent
(kN
*m
)
2 0 21.5 10
4
1 104
5000
0
5000
1 104
1.5 104
Centrifuge tes tsStatic loading tests
Bending moment vs rotation, LN column
Column rotation (%)
Co
lum
n m
om
ent
(kN
*m
)
2 0 21.5 10
4
1 104
5000
0
5000
1 104
1.5 104
Centrifuge tes tsStatic loading tests
Bending moment vs rotation, LS column
Column rotation (%)
Co
lum
n m
om
ent
(kN
*m
)
0 10 20 30 400.5
0
0.5
1Vertical acceleration of bridge deck
Time (s)
Acc
eler
atio
n (g
)
Footing Rotation (%) Column Rotation (%)
Small-footing bridge
Large-footing bridge
0.01 0.1 1 100
1
2
3
Container baseSmall-ftg deckLarge-ftg deckSN footingLN footing
Period (s )
Acc
el r
espo
nse
sp
ectr
a (g
)
1 .5 1 0.5 0 0.51.8 10
4
1.2 104
6000
0
6000
1.2 104
1.8 104
Moment vs rotationTheo ftg moment capTheo ftg moment cap
Moment vs rotation, small north footing
Footing rotation (%)
Foo
tin
g m
om
ent
(kN
*m
)
1 .5 1 0.5 0 0.51.8 10
4
1.2 104
6000
0
6000
1.2 104
1.8 104
Moment vs rotationTheo ftg moment capTheo ftg moment cap
Moment vs rotation, small south footing
Footing rotation (%)
Foo
tin
g m
om
ent
(kN
*m
)
1.5 1 0.5 0 0.51.8 10
4
1.2 104
6000
0
6000
1.2 104
1.8 104
Moment vs rotationTheo ftg moment capTheo ftg moment cap
Moment vs rotation, large north footing
Footing rotation (%)
Foo
tin
g m
om
ent
(kN
*m
)
1 .5 1 0.5 0 0.51.8 10
4
1.2 104
6000
0
6000
1.2 104
1.8 104
Moment vs rotationTheo ftg moment capTheo ftg moment cap
Moment vs rotation, large south footing
Footing rotation (%)
Foo
tin
g m
om
ent
(kN
*m
)
2 0 21.5 10
4
1 104
5000
0
5000
1 104
1.5 104
Centrifuge tes tsStatic loading tests
Bending moment vs rotation, LN column
Column rotation (%)
Co
lum
n m
om
ent
(kN
*m
)
2 0 21.5 10
4
1 104
5000
0
5000
1 104
1.5 104
Centrifuge tes tsStatic loading tests
Bending moment vs rotation, LS column
Column rotation (%)
Co
lum
n m
om
ent
(kN
*m
)
Critical plots of LJD02_15 event: Gazli 2.0
10 15 20 25 30
2
0
2
SN columnSS columnSN footingSS footing
Column and footing rotation time histori
Time (s)
Ro
tati
on
(%)
10 15 20 25 30
2
0
2
LN columnLS columnLN footingLS footing
Column and footing rotation time histori
Time (s)
Ro
tati
on
(%)
10 15 20 25 301
2
3
4
5Settlement of SN footing
Time (sec)
Set
tlem
ent (
cm)
10 15 20 25 30
2
0
2
SN columnSS columnSN footingSS footing
Column and footing rotation time histori
Time (s)
Ro
tati
on
(%)
10 15 20 25 30
2
0
2
LN columnLS columnLN footingLS footing
Column and footing rotation time histori
Time (s)
Ro
tati
on
(%)
10 15 20 25 301
2
3
4
5
6Settlement of LN footing
Time (sec)
Set
tlem
ent (
cm)
Small-footing bridge
Large-footing bridge
• Systems with small footings may perform better than systems with large footings– drift, ductility demand on columns
• Rocking foundations provide– Self-centering tendency– Non-degrading moment capacity– Isolation mechanism– Energy dissipation
• Difficult aspect of the problem: How to evaluate settlement (or uplift) associated with rocking.
Learned from experiments
Draft Design Procedure for Bridges with Rocking Foundations
1. Determine design ground motions, site conditions, design spectra.
2. Determine superstructure information, geometry, dead loads and live loads, abutment constraints.
3. Estimate distribution of dead load on footings.
4. Size footings based on settlement considerations. FS(bearing failure) >~10 and “yield acceleration”.
5. Preliminary column design: sized to make their moment capacity greater than the footing moment capacity.
6. Confirm that drift and settlement do not violate serviceability limits in Functional Evaluation Earthquake. If drift is too large increase “yield acceleration”
7. No collapse in Maximum Considered Earthquake.
8. Check distribution of dead load on the footings (assumption in step 3).
9. Final design of columns
NEESR-CR: Design of soil and structure compatible yielding
to improve system performance[CoSSY ???]
And finally back to our NSF NEESR project
(1) Technical Goals and ApproachGoal: Learn how to (account for/take advantage of)
complementary yielding and energy dissipation of geotechnical and structural components of building systems with due consideration of practicality, constructability, and life-cycle costs.
Team: A multidisciplinary team of structural, geotechnical, academic and practicing engineers is needed to perform this work.
Academic Approach: Numerical simulation of varying complexity will be performed. Experiments will be specifically designed to test assumptions and uncertainties in numerical simulations of yielding soil-foundation-structure systems.
Research will be driven by specific “Research Questions”
(2) Procedures for analysis of complementary soil-structure yielding
• Opensees 1 – dual systems (wall-frame-rocking foundation)– Fiber models for beams, columns, walls……..– Winkler models of footings.
• Opensees 2- hypothetical simplistic building systems (few DOF)– Beams with hinges– Winkler models of footings.
• Nonlinear spectral approaches – 1 or 2 DOF– Bilinear, trilinear, Bispec, or MatLab
• Design codes/guidelines
(3.1) Research Questions/Goals1 Compare simple system with rocking footing to one with
yielding column– Show that the magnitude of drift demand depends primarily on
the “yield acceleration” for both systems • Overcome unreasonable fear of tip-over
– Show that rocking foundations are superior in some respects (re-centering, energy dissipation, ductility, ….)
– Show that pushover behavior of rocking system is as good as system with yielding column.
2 How should we characterize the “period” of a rocking system for spectral analysis methods?– Equivalent linear (secant stiffness)– Tangent stiffness just before yield occurs– Trilinear system (K1 and K2)
(3.2) Research Questions/Goals3. Is the existing information sufficient for development of
procedures to predict settlement of rocking footings.
Small FSvLarge FSv
Set
tlem
ent
per
cycl
e, s
/L Uplift, -s/L
Half Amplitude of Rotation (radians)
(3.3) Research Questions/Goals4. If settlement or drift is excessive, can we
– mediate settlement with ground improvement while still maintaining benefits of rocking system. – Design structure to accommodate more settlement
5. What are acceptable performance limits? Do we need to rethink performance criteria for CoSSY systems?
(3.4) Research Questions/Goals6. Is it better to force yielding in one system or
another (footing rocking vs ductile beam hinges), or is it better to take advantage of ductility capacity of both?– Answer may depend on whether systems are parallel or
series.
Parallel system Series system
(3.5) Research Questions/Goals
7. Does vertical acceleration help or hurt?8. How to quantify and parameterize the
problems of series and parallel nonlinear MDOF systems?
9. Does the system become chaotic under some conditions? Can these circumstances be avoided?
(3.6) Research Questions/Goals10. Which of the above can be answered by
numerical simulations, and which require experimental validation?
11. Which of the above questions is the most fundamental?
12. Is “CoSSY” the best acronym? Send me alternatives soon!
(4) EOTE1. Recruit new engineers
– Invite MESA student from local community colleges to “Intro to Civil Engr Day” at Davis; Summer support to one attendee from community college.
– Participate in MSE Transfer Day at UC Davis– Leverage UCSD MESA programs
E2. Small workshop at end of 2011 – refocus final year efforts.
E3. Webinars: One per year, lunchtime– TTT to help recruit participants, announced through NEES, EERI – SEAOC meeting presentations (video tape the presentations
and archive with power point slides)
Centrifuge Testing (Task 1)
6 tests proposed – approx 4 to 6 months per experiment(a)conceptual design, (b) detailed design, (c) construction, (d) testing, (e) data analysis, (f) documentation and data archiving, (g) theoretical analysis, (h) applied analysis, and (j) synthesis tasks.
Isolated footing tests (2 expts?)
• Identify limits on applicability of concepts presented in proposal. • Softer soil than we have used in past experiments• Ground remediation to control settlement and re-centering• Water table and cyclic softening of the soil?• Each experiment might have 6 “SDOF” structures
Centrifuge Models of Soil-Foundation-Structure Sytems (4 expts?)
• Data on system behavior is totally lacking• TTT to help select the prototype scenarios• Each “experiment” (model container) may have 2 simple and one
complex structural system• Vertical excitation in last system test, perhaps.
Possible model container 1 – several simple 2DOF - series systems
w
K1, ay1
K2, ay2
h1
h2
LK1, ay1
K2, ay2
h1
h2
L
K1, ay1
K2, ay2
h1
h2
LK1, ay1
K2, ay2
h1
h2
L
w
ww
Decide which parameters to vary through numerical analysis. But initial idea is to vary the critical parameters (ay1/ay2 or K1/K2 and m1/m2 could be systematically varied by numerical analysis,and any peculiar or counterintuitive results could be tested in experiments.
Need consensus of PI’s and TTT(1) Consensus/motivation on Technical Goals
and Approach(2) Can we (and if so how?) distribute
responsibility for the four (or more) “Procedures for analysis of complementary soil-structure yielding” amongst different members of the team?
(3) Are there other “research questions”? Can they be stated better? Can we start strategizing on how to answer the question(s)?
Literature Review
• What two papers/documents do you (TTT, PI’s, students) highly recommend for others on the team to read? – one by you and – one by others
• We will post them on NEEScentral and make available to whole team.
Technology Transfer Team (TTT)– from proposal
• The team will: – brainstorm and develop concepts of prototype buildings. – synthesize the results from each series of tests and guide
the evolution of plans for subsequent tests – review and contribute to reports and documents
describing the simulations, experiments, results, and conclusions,
– participate in a 1-day workshop in year 3, and – use the information learned during this collaboration to
help guide their on-going efforts to revise and update building codes and guidelines such as the aforementioned IT 3 guidelines for NEHRP provisions.
Agenda1:00 Bruce: Overall Goals, Task 1 (centrifuge testing) (30 min)• Tara: concepts for designing small scale realistic bldgs and
description of task 2 (15 min). Pretest modeling in opensees. • Mark: Task 3 (Development and verification of simplified
design tools). Inelastic spectra for rocking systems. New way of characterizing ground motion. (15 min)
• Sashi: ground motion selection for sim and exp., significance of vertical shaking in last system test, (15 min)
• Discussion (15 min)2:30 Break and Tour of facility (30 min)
3:00 Technology Transfer team input (5 min each)– priorities, cautions, what can project accomplish to help code
development progress– TT members describe how they could help project be effective
4:00 Bruce: Project Schedule and next meetings
Lelio led discussion• Dry sand does not degrade, imortant to evaluate degrading
soils, e.g., clay with pore pressure build up.• Determining the effect of the shape of the hysteresis on
response. • Critical mechanism is p-delta. • Question 6. Is it better to force yielding in one system or the
other? Lelio – it depends on which system we can control. • It is important that we do not over- “sell” the research. Let it
speak for itself, but make sure to clearly and objectively communicate results to people that might use it.
David led discussion• Moment frames with lots of hinges are expensive• Need to counter concern that uncertainty in moment
capacity of a footing is small (M=VL/2)• Grade beams etc - accidental overstrength needs to
be considered. • Danger is weak soil where you get a compression
failure; analogy with reinforced concrete, small L/Lc soil is like overreinforced concrete.
David led discussion• How do I get design parameters? • Response characteristics – consider ground displacement. • Strength is the main factor that controls response at large
deformation. The mechanism (rocking vs plastic hinge) is not so important.
• Size of footing is advantage over columns • Discussion of flag vs full hysteresis. Suggested that flag is preferred
due to recentering characteristics, even though full loops dissipate more energy. Would trade energy absorption for centering. Perhaps this is a good topic for a numerical study.
• Radiation damping will also contribute, to energy dissipation and it is most significant for smaller deformations and large L/Lc ratios.
Craig led discussion
• Recommended reading– FEMA 440A ATC 62 degrading response.– FEMA 440 – chapter from Jon Stewart on kinematic interaction,
embedment effects and radiation Damping
• Vertical Excitation is something we should look into, but gut feeling is that it is not the most important paramter.
• We cannot always think that our research speaks for itself. • Practice is competitive, and trying new things takes time. It is
important to present the results in accessible terms that practitioners are used to seeing. Use the vernacular of practitioners to make it easier for them to adopt in the competitive engineering environment.
Mark led discussion• General perception of engineers is that we know
structures better than we actually do. – A hinge is a hinge is a hinge – system thinking naturally follows. – Rocking is “poor man’s base isolation”
• People are now embracing rocking– Recognized at NEHRP – Recognizing it in concert with structural yield. – There is move afoot; it is saving columns BART.
• Think on component level first, it is important to demonstrate that we have an good understanding of component behavior.
Mark led discussion• Maybe we are making the job harder by worrying too
much about settlement. – Need to package the method of calculating the settlement. – Possibly we are worrying about a second order effect;
engineers do not now explicitly consider second order damage due to column hinging
– Don’t make it too hard. People hang hat on the permanent displacement to avoid use of rocking
• Allowable settlements/movements– Design Basis EQ: a couple inches (may be different for
rocking structures?)– Serviceability EQ: half an inch (new structures are meant to
protect contents)
Mark led discussion• Overstrength of rocking foundations needs to be
considered for rocking systems.– overpour engaging another footing, grade beams, slab on grade
may result in rocking footing to be stronger than desired for optimal performance
• It is getting to the point that we can form a team to contribute to ASCE 41. – But ASCE is very prescriptive.– ATC has this on this list.
• Terminology is important – “foundation hinge” may be preferred over “rocking foundation” or
“foundation yielding”