Rate Effects in Non-Ductile R/C Columns A Prelude to Real Time - - PowerPoint PPT Presentation

Rate Effects in Non-Ductile R/C Columns

A Prelude to Real Time hybrid Simulation

• W. Ghannoum*
• V. Saouma
• G. Haussmann
• K. Polkinghorne
• M. Eck

Dae-Hung Kang

University of Colorado, Boulder *University of Texas, Austin

September 2, 2010

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 1/30

Outline I

1

Introduction

Motivation

2

Literature Review

Materials Observations Components Shake Table Tests

3

Test Description

Reinforcement Actuators Load Vibrations Mixed Control Controller Performance Inertia Force Removal

4

Test Results

Failure Modes Slow Cyclic Test Behavior

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 2/30

Outline II Rate Effects; Results Rate Effects; Observations Others

5

Video

Monotonic Real Time Monotonic, Slow Motion Cyclic, Real Time Cyclic, Slow Motion

6

Conclusions

7

Credit

8

Bibliography

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 3/30

Introduction Motivation

Motivation Rate Effects in Concrete material: well established. Structural systems: indirectly investigated through shake table tests. Structural components: not sufficiently investigated. Real Time Hybrid Simulation (RTHS)

• f reinforced concrete frames seldom/never investigated.

Prior to its undertaking one must master control techniques for multiple degrees of freedom systems Objective Undertake a series of 10 tests of a non-ductile reinforced concrete columns in order to Sharpen our skills prior to the undertaking of complex RTHS. Determine whether there is indeed a rate effect in reinforced concrete columns (currently not considered in ASCE 31/41)

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 4/30

Literature Review Materials

Materials

Concrete Drastic increase in strain rate effect at 1.0 /sec (well beyond rates induced by earthquakes); Malvar (1998-1) Reinforcement Noticeable increase; Malvar (1998-b)

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 5/30

Literature Review Observations

Observations Several experimental investigations have been performed on RC beams and columns at seismically representative loading rates State of knowledge on loading rate effects on RC members is still in its infancy. Specimens need to be tested under more realistic loading protocols. To date all dynamic tests have used single-degree-of-freedom actuation (i.e., a single actuator or slaved actuators applying load in the same degree of freedom). Difficulties in controlling several actuators at high loading rates have so far hindered efforts to apply more realistic loading protocols. Our tests will address some of these limitations.

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 6/30

Literature Review Components

Background Interest in tested column is three fold:

1

Extensively tested in past investigations:

Varying pseudo static loading protocols (i.e., cyclic and monotonic) Lynn et al. (1996), Sezen and Moehle (2006), Shin (2007) Frame sub-assembly (Elwood and Moehle (2008)) Full frame setup (Ghannoum (2007)

2

Has a shear capacity that is only slightly larger than its flexural one making it an interesting candidate for studying failure mode shifts.

3

Of interest to seismic collapse hazard mitigation as it belongs to a family of lightly confined RC columns that is vulnerable to collapse in earthquakes. Yields in flexure prior to failing in shear at lateral drifts only slightly larger than those causing flexural yielding. Flexure-shear critical (ASCE 41 (2007)). Dynamic loading effects observed in all dynamic tests: increased flexural yielding strength (up to 25%) and associated increases in shear demand. No shifts in failure modes were recorded due to dynamic loading.

Presented column tests were devised to shed more light into the dynamic behavior of this column type.

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 7/30

Literature Review Components

Test Setups: Components Sezen (2000) Large scale static tests Shin (2007) Shake table tests to investigate the dynamic response of ductile and non-ductile reinforced concrete columns.

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 8/30

Literature Review Components

Shake Table Tests of Ghannoum Ghannoum (2007) tested R/C frames with non-ductile columns Column dimensions identical to those of Shin (2007), we will use same column design

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 9/30

Literature Review Components

Shake Table Tests at San Diego

d

Panagiotu, Restrepo and Conte tested a slice of a 7 storey reinforced concrete building Shake table tests represent actual structural performance quite well, however they do not lend themselves to parametric studies of RC element behavior at seismic loading rates (lack of causality), and are relatively expensive to conduct.

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 10/30

Test Description Reinforcement

Reinforcement 8#3 bars longitudinally and 1/8" ties spaced 4" on center, identical to the one tested by Shin (2007) and Ghannoum (2007)

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 11/30

Test Description Actuators

Actuators Dynamic Actuators: Vertical (X2 and X3): MTS 244.41S; 110 kip, 10 in. stroke and 20 in/sec. velocity); horizontal (X1): MTS 244.22 (22 kip, 24 in. stroke and 100 in/sec. velocity)

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 12/30

Test Description Load

Load

Axial load (X2, X3): 10% Paxial

f

: 17 kips Peak displacement amplitude imposed through increasing fractions of first yield drift ( 1%). Care exercised in controlling velocity, but minimize acceleration (not not exceed load cell capacity) Maximum velocity 54 in/sec.

−100 −50 50 100 f = 0.76911 Hz; # cycle ramp = 12; # of cst. cycle = 23; Delta t= 0.00097656 Displacemnt [mm] −400 −200 200 400 velocity [mm/sec] 10 20 30 40 50 60 −3 −2 −1 1 2 3 Time [sec.] Acceleration [g]

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 13/30

Test Description Vibrations

Unwarranted vibrations observed around 16 Hz

5 10 15 20 25 30 −200 200 400 600 800 1000 1200 1400 1600 Frequency [Hz.]

Remedied by “stiffening the setup” (remove vertical leg of beam-rig. Implemented an on-line filtering scheme

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 14/30

Test Description Mixed Control

Mixed Control System

Three degrees of freedom ∆h, ∆v and θ: Imposed displacement by X1 Zero rotation θ = 0 Sum of forces in X2 and X3 constant. MTS STS control system can not handle this mixed control Use a Simulink based controller running

• n an xPC computer to drive through STS

X2 and X3 Lateral displacement directly controlled by STS

XPC STS Servo-Valve Actuator Specimen Transducers Displacement Command Forces and Displacements Forces Displacement Feedback Force Feedback Desired Axial Force Flow rate Command Actuator Command Actuator Feedback

Proportional Gain Kp Integral Gain Ki/p Derivative Gain p.Kd

Flow rate command PID Controller (one for each actuator) Position Feedback Force Feedback Displacements SCRAMNet + Desired Axial Force + X2 Force Feedback + X3 Force Feedback

• Axial Force

FeedBack Integral Gain K/p Unit Delay 1/z + + Displacement Command X2/X3 Displacement Command Simulink +- +-

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 15/30

Test Description Controller Performance

Controller Performance

−40 −20 20 40 X2 and X3 forces for slow cyclic test (test 05)

Displacement (mm)

200 400 600 800 1000 1200 1400 1600 1800 2000 2200 −80 −70 −60 −50 −40 −30 −20

Time (sec) Forces (kN)

X2+X3 force X2 force X3 force −50 50 X2 and X3 forces for fast cyclic test at 20in/sec (test 07)

Displacement (mm)

18 18.5 19 19.5 20 20.5 21 21.5 22 22.5 −140 −120 −100 −80 −60 −40 −20

Time (sec) Forces (kN)

X2+X3 force X2 force X3 force

Axial load curve oscillated before failure (related to the horizontal displacements) Upon failure near end of test, axial load drops rapidly even though the PID controller continually pushes vertically downward in increasing amounts. For slow test, axial loading is maintained with less than 13 kN [3 kips] error until initiation of axial failure. During the fast cyclic test the PID controller does a less satisfactory job of regulating axial load.

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 16/30

Test Description Inertia Force Removal

Inertia Force Removal Effective mass estimated from a linear regression analysis of force vs accelerations Implemented inside Simulink code to eliminate it

−5 −4 −3 −2 −1 1 2 3 4 5 −40 −20 20 40 Unfiltered fast test data Lateral Force (kN) −5 −4 −3 −2 −1 1 2 3 4 5 −40 −20 20 40 Filtered fast test data −5 −4 −3 −2 −1 1 2 3 4 5 −20 −10 10 20 Inertial forces from test equipment Lateral Drift Ratio (%) Lateral Force (kN) −5 −4 −3 −2 −1 1 2 3 4 5 −40 −20 20 40 60 Filtered data minus inertial forces Lateral Drift Ratio (%)

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 17/30

Test Results Failure Modes

Failure Modes

Shear failure (one or two orthogonal inclined cracks close to flange), followed by compression failure (buckling of the longitudinal reinforcement)

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 18/30

Test Results Slow Cyclic Test Behavior

Slow Cyclic Test Behavior Columns tested under a pseudo-static “slow” cylic protocol behaved as anticipated and compared well with full scale tests conducted by Sezen and Moehle (2006). A typical flexure-shear failure mode was observed whereas the column yielded in flexure at about 1% lateral drift ratio and subsequently sustained shear cracking that resulted in shear failure initiating at a lateral drift ratio

• f 1.75%.

−5 −4 −3 −2 −1 1 2 3 4 5 −40 −30 −20 −10 10 20 30 40 Lateral Drift Ratio (%) Shear Force (kN) Sezen and Moehle (scaled) slow cyclic test

Axial failure followed at larger lateral drifts.

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 19/30

Test Results Rate Effects; Results

Rate Effects; Results 10 tests performed, 5 to “fine-tune” experimental setup; and 5 yielded proper results

Maximum Minimum Drift ratio Drift ratio Increase in Increase in Test shear (kN) shear (kN) at max.

• min. shear
• max. shear
• min. shear

[kips] [kips] shear (%) (%) from PS PS Pseudo- static slow cyclic (Test 5) 32.0 [7.2]

• 32.0 [-7.2]

1.43

• 1.41

N/A N/A Cyclic @ 254 mm/sec peak velocity (Test 9) 42.7 [9.6]

• 38.3 [-8.6]

2.26

• 2.24

33% 19% Cyclic @ 508 mm/sec peak velocity (Test 7) 42.2 [9.5]

• 39.1 [-8.8]

1.95

• 1.69

32% 22% Cyclic @ 1016 mm/sec peak velocity (Test 8) 40.9 [9.2] 40.0 [-9.0] 2.44

• 2.31

27% 25% Cyclic @ 1016 mm/sec peak velocity (Test 10) 40.9 [9.2]

• 42.2 [-9.5]

2.32

• 2.30

27% 32%

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 20/30

Test Results Rate Effects; Observations

Rate Effects; Observations Failure Modes In all cases a flexure-shear failure mode was observed, irrespective of load rate. Flexural yielding was followed by shear failure (inclined cracks) and then by axial failure (buckling of longitudinal reinforcement). Maximum Shear Demand Faster loading rates increased maximum shears and moment in test columns (consistent with other experimental observations). Surprisingly, increase appears to be independent of the tested velocities. Since flexural yielding influences maximum shear demand in this type of column, an increase in longitudinal steel yield strength is a likely cause of increased shear values. Recorded max. longitudinal bar strain about 0.35/sec. (corresponding to a yield strength increase of ≃35% (Malvar, 1998).

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 21/30

Test Results Others

Others Column Stiffness Higher loading rates did not significantly affect column lateral stiffness prior to yielding. Cyclic Loading Protocols Loading history prior to longitudinal bar yield does not seem to affect shear failure response. This

• bservation is consistent

with pseudo-static cyclic loading observations.

−5 −4 −3 −2 −1 1 2 3 4 5 −50 −40 −30 −20 −10 10 20 30 40 50

Shear Force (kN)

(a) Slow Test Test 5 (254 mm/sec) −5 −4 −3 −2 −1 1 2 3 4 5 −50 −40 −30 −20 −10 10 20 30 40 50 (b) Slow Test Test 9 (508 mm/sec) −5 −4 −3 −2 −1 1 2 3 4 5 −50 −40 −30 −20 −10 10 20 30 40 50

Lateral Drift Ratio (%) Shear Force (kN)

(c) Slow Test Test 8 (1016 mm/sec) Test 10 (1016 mm/sec) −5 −4 −3 −2 −1 1 2 3 4 5 −50 −40 −30 −20 −10 10 20 30 40 50

Lateral Drift Ratio (%)

(d) Test 8 (1016 mm/sec) Test 10 (1016 mm/sec)

Response Envelopes Higher loading rates seem to produce larger response envelopes than pseudo-static loading. However, higher loading rates appear to increase the rate of cyclic damage in columns as observed in larger drops in shear capacity per cycle in faster tests. At low damage levels, this indicates that

• nce a shear crack forms, it degrades much faster at higher loading rates.
• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 22/30

Video Monotonic Real Time

Monotonic, Real Time

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 23/30

Video Monotonic, Slow Motion

Monotonic, Slow Motion

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 24/30

Video Cyclic, Real Time

Cyclic, Real Time

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 25/30

Video Cyclic, Slow Motion

Cyclic, Slow Motion

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 26/30

Conclusions

Conclusions I Numerous pseudo-static tests and shake table tests are reported. Extraction of rate effects from shake table tests may be problematic as causality (direct assessment of causes on effects) is challenging. Few tests on reinforced concrete columns have been conducted at seismically representative loading rates; with all reported tests using single-degree-of-freedom actuation. This investigation

Was a prelude to real time hybrid simulation Sought to quantify the effect of the load rate on the structural response of a non-ductile column. was based on three independently controlled actuators.

A number of experimental challenges had to be overcomed (in particular data filtering and elimination of inertia forces) Indications are that an increase in lateral load capacity of up to 33% may result if load rates are accounted for.

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 27/30

Conclusions

Conclusions II A significantly larger shear force versus lateral drift envelope can be attributed to higher loading rates. There are indications that cyclic damage and cyclic shear strength degradation can increase at higher loading rates. The observed strength increase is not insignificant, and if further validated by future tests, could have a profound influence on our seismic assessment of non-ductile reinforced concrete frames. By ignoring such an increase, codes are currently erring on the conservative side in strength assessment, but then costly repais/rehabilitation may not be justified by the outcome of this research. Further investigations are needed to corroborate conclusions.

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 28/30

Credit

Credit

• Prof. Wassim Ghannoum

CO-PI, Data Analysis

• Prof. Victor Saouma

CO-PI, Testing

• Dr. Gary Haussmann

Development of Control Algorithms

• Dr. Dae-Hang Kang

Development of Mercury Matlab, c++ Kent Polkinghorne FHT Operator and DAQ Michael Eck FHT Operator Etienne Burdet Data Analysis Casey Champion Lab Assistant

• Dr. Eric Stauffer

Technical consultant Thomas Bowen Laboratory Manager NSF/PEER/UC Berkeley Shake table tests CU-NEES/NEESinc/NSF FHT Equipment State of Colorado Financial support

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 29/30

Bibliography

Ghannoum, W.: 2007, Experimental and Analytical Dynamic Collapse Study of a Reinforced Concrete Frame with Light Transverse Reinforcement, Phd dissertation, Department of Civil and Environmental Engineering, University of California, Berkeley. Malvar, L. and Crawford, J.: 1998a, Dynamic increase factors for concrete, Twenty-Eighth DDESB Seminar, Orlando, FL. Malvar, L. and Crawford, J.: 1998b, Dynamic increase factors for steel reinforcing bars, Twenty-Eighth DDESB Seminar, Orlando, FL. Sezen, H.: 2000, Seismic Behavior and Modeling of Reinforced Concrete Building Columns, Phd dissertation, Department of Civil and Environmental Engineering, University

• f California, Berkeley.

Shin, Y.: 2007, Dynamic Response of Ductile and Non-Ductile Reinforced Concrete Columns, Phd dissertation, Department of Civil and Environmental Engineering, University

• f California, Berkeley.

Saouma, V., Haussmann, G., Kang, D.H., Ghannoum, W., Polkinghorne, K. and Eck M.: 2010, Real Time Hybrid vs Shake Table Simulation, Submitted to Earthquake Engineering and Structural Dynamics. Ghannoum, W., Saouma, V., Haussmann, G., Polkinghorne, K. and Eck, M.: 2010, Rate effects in reinforced concrete columns, Submitted to ASCE J. of Structural Division. Saouma, V., Kang, D. and Haussmann, G.: 2010, Mercury: A computational finite-element program for hybrid simulation, Submitted to Earthquake Engineering and Structural Dynamics.

• W. Ghannoum et al.; Univ. of Texas

Rate Effects in R/C Columns 30/30