Read Microsoft Word - SEAOC_Final_20_July.doc text version

Seismic Rehabilitation ­ Benefits of Component Testing

Sarah Taylor Lange, Graduate Student Alberto Salamanca, Ph.D., Staff Researcher John Wallace, Ph.D., Professor Structural/Earthquake Engineering Research Laboratory Department of Civil & Environmental Engineering University of California, Los Angeles, CA Kutay Orakcal, Ph.D., Assistant Professor Department of Civil Engineering Bogazici University, Istanbul, Turkey John Gavan, Aaron Reynolds, and Luis Toranzo KPFF Consulting Engineers Los Angeles, CA Roger Young, Mostafa Sobaih, Ron Hartman KPFF Consulting Engineers Irvine, CA

Abstract FEMA 356 backbone relations tend to provide conservative estimates of the available strengths and deformation capacities of reinforced concrete components, leading to costly seismic rehabilitation solutions for California hospitals. The conservatism is driven by the lack of test data for older, typically poorly detailed, structural components. Although FEMA 356 does provide for building specific component testing, this approach is not common due to cost concerns and/or schedule constraints. An assortment of large-scale wall segments, columns, and beam-column-joint assemblies were tested until substantial lateral strength degradation and loss of gravity load support was observed. Presented test results are compared with FEMA 356 backbone relations to highlight the advantages associated with building-specific test programs. In general, the test specimens revealed more strength and deformation capacity than assumed by FEMA 356, and more gradual strength deterioration. The test results, when coupled with the FEMA 356 Nonlinear Static Procedure (NSP) and Nonlinear Dynamic Procedure (NDP), enabled the development of more rational and substantially more economical rehabilitation solutions. Introduction Following damage to hospitals in the 1994 Northridge earthquake, California Senate Bill 1953 passed, requiring evaluation of pre-1973 acute care facilities with a timeline for rehabilitation or change in use. SB 1953 requires deficient facilities to be upgraded by 2013 to prevent collapse and loss of life; facilities must be upgraded to provide for continued operation after an earthquake by 2030. A review of existing California hospitals (OSHPD, 2001) reveals that 975 out of 2507 of pre-1973 buildings were rated SPC-1, buildings that pose a significant risk of collapse and must be upgraded by the 2013 deadline. A significant portion of these at-risk buildings are reinforced concrete construction. The FEMA 356 Pre-Standard (FEMA, 2000), recently updated as ASCE/SEI 41 (2006), is commonly used to evaluate and upgrade existing buildings. These documents provide a range of analysis approaches. However, among the most popular is the nonlinear static procedure (NSP). For the NSP, component modeling parameters and acceptance criteria are assigned for the components that contribute the lateral stiffness and strength of the building. For a given building, pushover

analyses (NSP) are performed, target displacements are defined representing the expected displacement demand for the design earthquake event, and acceptance criteria are checked, for each of the structural components. The modeling parameters and acceptance criteria defined in FEMA 356 substantially impact the results of this process. In some cases, relatively little information exists to assign modeling parameters and FEMA 356 provides limited information. A good example of this condition is the one row of modeling parameters that exist for wall segments controlled by shear in FEMA 356 Table 6-19. In other cases, an abundance of information exists. However, the FEMA 356 provisions tend to provide a lower bound estimate to the observed test data (e.g., bond strength of §6.4.4 and the wall shear strength of §6.4.5) therefore, use of these provisions tends to produce rather conservative results. In other instances, the structural details used within the building may not fit into the predefined categories; several specific examples that fall into this category are provided later. In all of these cases, the conservatism built into the modeling parameters is likely to produce evaluation results that indicate existing buildings are excessively deficient and thus require costly and disruptive seismic rehabilitation solutions. As the 2013 deadline for upgrading SPC-1 rated buildings approaches, it is imperative that more economical seismic rehabilitation solutions be identified. This is particularly important given the significant rise in construction costs, which have greatly increased the overall estimated rehabilitation costs. One important attempt to address this issue is the recent update to the concrete provisions of ASCE/SEI 41, referred to as Supplement #1 (Elwood et al, 2007). Supplement #1 provides updates to modeling parameters for beams, columns, slab-column connections, and walls controlled by flexure and shear, as well as a number of other changes. The most significant change in Supplement #1 involves the changes to columns, where the deformation capacity at the collapse prevention limit state is substantially higher for many columns. However, even with the improvements provided by Supplement #1, substantial conservatism still exists. FEMA 356 and ASCE/SEI 41 include the option to derive modeling parameters and acceptance criteria based on testing in §2.8 provided that: (1) test subassemblies are identifiable with a portion of the structure and replicate construction details and boundary conditions, and (2) the test assemblies are subjected to reverse cyclic lateral loading at increasing displacement levels with the number of cycles and displacement levels based on the expected response of the structure to the design earthquake. The limiting strength and deformation capacities are determined from the experimental program using the average values of a minimum of three tests performed for the same design configuration and test

conditions. Conducting building-specific test programs is not common due to the costs associated with the test program as well as scheduling constraints. However, modeling parameters derived from building-specific tests have the potential to substantially reduce the costs associated with seismic rehabilitation. This is particularly true for cases where the construction details are not consistent with predefined categories specified in the FEMA 356 document. Equally important, the tests can often be completed within a reasonable timeframe that does not cause problems for the design team or the client. Building-specific test programs also offer the advantage of providing greater confidence in achieving the desired performance level and help the design team communicate design objectives to the client. Objectives An overview of several recent building-specific test programs conducted within the UCLA Structural/Earthquake Engineering Research Laboratory is provided. Key issues that motivated the tests are discussed, along with the test results, a comparison of the test results with FEMA 356 backbone relations, and the implications of the test results on the seismic rehabilitation process. UCLA Structural/Earthquake Engineering Research Laboratory The UCLA S/EERL consists of a 40 ft x 60 ft strong floor, eight 3ft x 5 ft x 10 ft stackable reaction blocks, and equipment for large-scale component and system testing. The S/EERL houses a range of servo-controlled hydraulic actuators, including four 500kip actuators with 36-inch stroke, as well as ten other actuators with capacities between 20 kips and 750 kips and strokes between +/- 6 to +/-12 inches. Control is provided by MTS Flextest, 407, and 458 units. Hydraulic power supplies (pumps) include a 20, 30, and 60 gpm units; the 20 gpm diesel-powered unit can be deployed for field work, as can the 30 gpm unit. More than 500channels of high-quality data acquisition exist (National Instruments, Kinemetrics), along with a range of strain, displacement, and accelerometer sensors. Real-time data acquisition and control is accomplished using LabView in conjunction with NEES data viewers and collaboration tools that enable remote participation. Equipment and staff for the NSF funded [email protected] Equipment Site are also housed within the S/EERL. Hospital Test Programs Three major building-specific test programs, including tests on wall segments, columns, and beam-column-joint (BCJ) subassemblies, were carried out in the S/EERL between May 2005 and April 2007. The wall segment and column test

specimens were three-quarter scale, whereas the BCJ assemblies were either one-half and two-thirds scale. Typically, the test specimens were subjected to gravity loads, which were held constant for the duration of the test and reversed cyclic lateral loading under either load or displacement control. Details and results for each of the test programs are outlined in the following subsections, along with comparisons with existing modeling parameters. Wall Segment Tests The approximately three-quarter scale specimens were constructed to be representative of a older, lightly-reinforced wall segments of a hospital with perimeter reinforced concrete walls with window openings. Ten specimens were tested, four horizontal wall segments (spandrels, Fig. 1(a)) and six vertical wall segments (piers, Fig. 1(b)). General attributes of the test specimens are given in Table 1. Nominal Grade 40 bars were used in the test; however, stress versus strain relations on coupons indicated yield stresses of 61.5 and 65.5 ksi for the #4 and #5 bars, respectively. Reinforcement quantities were modified modestly to account for the higher reinforcement yield stress. The target strength for the concrete used for the tests was 4 ksi; the actual concrete strength at test data was approximately 4.5 ksi for most tests. The test specimens included several important features of the actual building. For example, a single curtain of web reinforcement was used near the middle of the wall and no hooks we used on the horizontal web reinforcement at the wall boundary (based on field exploration). Given the lack of hooks, the nominal strength is estimated using ACI 318-05 Chapter 21 (21-7) with a -value of 0.0015 as recommended by FEMA 356. It is noted that a ACI 318-05 § requires two curtains of web reinforcement if Vu exceeds

2 Acv f c' ; therefore, strict interpretation of this requirement

interior of the forms as shown in Fig. 2(a) producing the grooves shown in Fig. 2(b). The testing was conducted to address how these features impacted the load deformation behavior and to enable comparison with the FEMA 356 backbone curves of Table 6-19. Table 1. Test Series Overview

Specimen ID (1) WP1-1-00 WP1-1-05 WP1-1-10 WH1-1-0 WH2-1-0

1 2

Geometry (inches) Height Length Thickness (2) 48 48 48 60 60 (3) 54 54 54 60 60 (4) 6 6 6 6 6 Edge1 (5) 2 - #5 2 - #5 2 - #5 2 - #4 4 - #5

Reinforcement Vert. Web2 Horiz. Web2 (6) #[email protected] #[email protected] #[email protected] #[email protected] #[email protected]

3 4

P/Agf'c3 (kips) (8) 0.00 0.05 0.10 0.0 0.0

VN4 (kips) (9) 125.0 125.0 125.0 135.0 135.0

Specimens (#) (9) 2 2 2 2 2

(7) #[email protected] #[email protected] #[email protected] #[email protected] #[email protected]

# of edge (boundary) bars - bar designation bar designation @ bar spacing in inches

Applied axial Load = P/Ag f'c Estiamte of the Vn = 6(f'c)Acv

l = 15"

lb = 60"

hb = 60"

1 - #4 & 1 - #5 or 4 - #5 Jamb bars WPJ tb = 6" 5 - #4 6 - #4

(a) Spandrel

lp = 54"

would limit the nominal shear strength to 2 Acv

f c' given the

single curtain of web reinforcement used in the wall segments. Test results summarized in (Elwood, et al. 2007) indicate that this interpretation is overly conservative, and the tests help shed more light on this issue for poorly detailed walls. Although hooks were used on spandrel stirrups (vertical web bars), a single curtain of web reinforcement was used and the horizontal web reinforcement (Fig. 1(a)) was not sufficiently anchored into the piers. Spandrels also included a "weakened plan joint" (WPJ) at mid-span, where the thickness of the cross section was reduced (Fig. 2) and two-thirds of the web reinforcement was cut (Fig. 1). The groove at mid-span was created by using wood strips, at ¾-scale, attached to the

hp = 48"

v = ~0.25% h = 0.35%

tp = 6"

(b) Pier Figure 1. Schematic drawings of test specimens

(a) Spandrel WP

(b) Constructed spandrel ­ WPJ Figure 2. Spandrel Weakened Plane Joint (WPJ). The specimens were tested using the test setup depicted in Fig. 3. The vertical actuators (400 kip peak capacity) were used to enforce a boundary condition of zero rotation at the top of the specimen as well as to control the level of axial load in the tests to the specified value (e.g., zero). The horizontal actuator was used to produce a linear moment gradient over the height of the test specimen with zero moment at the mid-height of the specimen. It is noted that, given the test setup, spandrels tests were rotated 90 degrees from the actual orientation. Cyclic loading was applied to the test specimens by controlling the load or displacement applied by the horizontal actuator. Given the very high stiffness of the test specimens, the first two levels of loading were applied using load control to approximately one-half of the predicted cracking load and equal to the cracking load. Three cycles were applied at each of these load levels.

steel testing frame

The load controlled levels were followed by displacement controlled loading to displacement levels equal to 0.5, 0.75, 1.0, 1.5, 2.0, 3.0, 4.0, and 6.0 times the predicted yield displacement of 0.004 times the specimen height (approximately 0.2 inches for both spandrels and piers). Approximately 100 sensors were used on each test specimen to measure applied loads (vertical and horizontal), reinforcement strains, and average concrete strains. In addition, uplift and sliding of the base pedestal were monitored so that any displacement contributions resulting from rigid body rotation could be accounted for during data processing. The degree of instrumentation provided exceeded that needed to determine backbone relations for the tests; however, the additional instrumentation was provided to enable detailed analytical modeling studies (Massone et al., 2004, 2006). A primary objective of the test program was to produce backbone relations for the shear force versus deformation behavior for both typical spandrels and piers. The approach used to develop the backbone relations is generally consistent with the approach defined in the ASCE/SEI 41 Supplement #1 (Elwood et al., 2007). Based on a review of the test results for wall piers, a force versus deformation relation based on the following points was determined:

P /A f' u g c Vcr = 4 f c' 1 + 4 f c' Vy = Vn = Acv c

1/ 2

< 0.6Vn

cr =

Vcr 0.4 Ec



f ' c + n f y


y = 0.004

where Pu is the axial compressive load and Ag is the gross concrete area of the cross section. Other terms are as defined in ACI 318-05. As well, for the lightly reinforced wall piers, shear strength degradation occurred at a deformation of approximately 0.0075 times the specimen height ( 0.0075hw ) , which is the same as prescribed in FEMA 356 (2000). However, a deformation of approximately ( 0.01hw ) was more representative for spandrels and piers tested with zero axial load. Residual strength for the piers tended to be lower than described in FEMA 356, apparently due to the relatively light reinforcement and generally poor detailing. ASCE/SEI Supplement #1 proposed backbone curves and experimental results are compared for several tests (Fig. 4). The comparisons indicate that the proposed backbone relation provides a reasonable good fit for lightly-reinforced tests with zero axial load, although the strength degradation observed in the tests was less pronounced than implied for the spandrels, and more pronounced for piers with axial load., as well as rapid strength loss and relatively low residual strength. The low residual strength was probably impacted by the poor detailing (i.e., lack of hooks on horizontal web reinforcement and WPJ for spandrels).

strong wall

specimen horizontal actuator vertical actuator vertical actuator

strong floor

Figure 3. Schematic of the test setup.

For the wall pier with modest axial load P = 0.05 Ag f c' , the ratio of the peak load in the test to the nominal capacity using (1.1) is roughly 1.5 and the deformation associated with the yield point is less than 0.004, indicating that axial load produces a noticeably stiffer and stronger response. However, test data for wall piers with axial load are limited; therefore, a more detailed assessment of the yield deformation and shear strength for wall piers with axial load is not possible and use of (1.1) is recommended.

Drift (%)

-0.02 100 -0.01 0.00 0.01 0.02



The specimen callout of 1.0GS11M refers to: 1.0 = longitudinal reinforcement ratio (%), GS = G-Loc Splice, 11 = axial load as a percent of Ag f c' , and M = Moderate shear or H= High shear. The column with 4 - #8 bars represents a perimeter column whereas the column with 4 - #10 bars represents an interior column. Transverse reinforcement for both columns consisted of #4 hoops spaced at 2 in. on center over the 18-inch length at both ends of the column (top and bottom) and #3 hoops at 8.5-inch spacing for the remaining 5ft height (Fig. 5). One of the defining features of the columns is the use of G-Loc devices near the middle of the column. The G-Loc device is a relatively short sleeve used to align and support the vertical column bars during construction. It is not a mechanical coupler and cannot be relied upon to transfer tension forces. However, the locations of the G-Loc devices were staggered as shown in Fig. 5; therefore, some rebar continuity is provided at all locations along the column height. The primary objectives of the test program were to assess the influence of the G-loc devices, the poor detailing over the mid-height segment on the lateral-load response, and axial load collapse of the column.

Test Data Proposed FEMA 356

Shear Force (kips)












Lateral Displacement (in.)

1- 6"


Drift (%)

-0.020 150 -0.010 0.000 0.010 0.020

5'- 5.25" 3'- 11.25"

1'- 6"


Shear Force (kips)

Test Data Proposed FEMA 356

2'- 5.25"


Bar discontinuity

0'- 11.25" (1'-3" FS)



0'- 11.25"


Column Symmetry

Bar discontinuity


-150 -1.00 -0.75 -0.50 -0.25 0.00 0.25 0.50 0.75 1.00

Lateral Displacement (in.)

Figure 5. Column Dimensions and Reinforcement Concrete cylinder tests and rebar coupon tests were conducted to determine material properties for the construction materials. Average concrete compressive strength based on five cylinder breaks for 1.6GS11H and 1.0GS11M were 4.94 and 5.38 ksi, respectively, which was less than the target strength of 6 ksi, but not expected to impact interpretation of the results. The test set-up used for the columns was identical to that used for the wall segments (Fig. 3), consisted of two vertical actuators placed on either side of the specimen to apply a constant column axial load of 222 kips as well as to enforce a zero rotation boundary condition at the top of the column.

(b) Pier: P = 0.05 Ag f c'

Figure 4. Wall segment force versus deformation relations. Column Tests The two column specimens were tested, each three-quarter scale replicas of columns in a moment frame building constructed in 1968 in southern California. The specimens were 8-ft tall with an 18-inch square cross section reinforced with either 4 - #10 (1.6GS11H) or 4 - #8 (1.0GS11M) (Fig. 5).

Lateral Load (kips)

Lateral loading was applied using an actuator with a line of action passing through the column mid-height; therefore, the column was subjected to reverse curvature bending with zero moment at column mid-height. For each test, the axial load was applied and held constant, and then reversed cyclic lateral loading was applied. For displacement controlled testing, any slip and rotation of the bottom support block were monitored such that the relative lateral displacement over the column height was computed and used for control. A total of 36 Linear Variable Differential Transducers (LVDTs) were mounted on the front and back of the column faces to measure deformations due to flexure and shear, anchorage slip from the end blocks, and rotation and slip of the bottom support block. Additional LVDTs were attached between the column specimen and a rigid external reference frame to measure the relative lateral displacement over the column height. Strain gauges were affixed to the reinforcement at 17 locations on both longitudinal and transverse reinforcement. The column tests were completed on June 20th and 28th, 2006 for 1.6GS11H and 1.0GS11M, respectively; therefore, only one week was required to setup and complete the second of the two tests. The test results for column 1.6GS11H show that the primary failure mechanism was shear and bond over the mid-height of the column (Fig. 6, 7). The test results indicate that the column could reach approximately 2% lateral drift prior to significant strength degradation. FEMA 356 and ASCE/SEI 41 Supplement #1 backbone relations are also shown on Fig. 6. Results shown in Fig. 6 indicate that the FEMA 356 backbone relation based on 0.5Ig is too stiff, whereas the initial stiffness based on Supplement #1 (Ieff = 0.3Ig) provides a much better fit to the effective elastic stiffness. However, both relations underestimate the shear strength (Vmax / Vn = 97 / 78 = 1.24 ) , substantially underestimate the lateral drift capacity at the initiation of strength degradation ( 2% versus 0.3% for FEMA 356), and overestimate the rate of strength degradation. The test results for 1.0GS11M reveal that the column had substantial deformation capacity, reaching a lateral drift of more than 4% prior to significant lateral strength degradation (Fig. 8). Lateral load versus top displacement relations derived using FEMA 356 and ASCE/SEI 41 Supplement #1 modeling parameters also are shown on Fig. 8. It is observed that the FEMA 356 relation based on 0.5Ig is too stiff and that the Supplement #1 effective stiffness of 0.3Ig provides a better fit. It is noted that FEMA 356 Collapse Prevention (CP) limit is less than 1% drift whereas the ASCE Supplement #1 CP drift limit is 2%; however, the test results indicate a CP drift limit of 4%.

Lateral Drift (%) -5 100












75 50 25 0 -25 -50 -75 -100 -5 -4

FEMA 356 Ieff =0.5Ig FEMA 356 Ieff =0.25Ig ASCE 41 Supplement 1

FEMA 356 Table 6-8: Vn=Vc+Vs=54+24=78 kips My = 4300 in-kips Vy=2My/h=91 kips > Vn Shear Controlled


-2 -1 0 1 2 Lateral Displacement (in.)




Figure 6. 1.6GS11H Load Displacement Test Results

Figure 7. Column 1.6GS11H at 4% drift. .

-5 100 75 50 Lateral Load (kips) 25 0 -25 -50 -75 -100 -5





Lateral Drift (%) -1 0 1





FEMA 356 Ieff =0.5Ig

ASCE 41 Supplement 1 (ii)

mid-height as well as the modest capacity achieved by staggering the locations where the G-Loc devices were used (Fig. 5). Although results are available for only a single test at this time, the potential benefits of conducting building-specific tests are apparent. Beam-Column Joint Tests Three beam-column-joint (BCJ) assemblies were tested, two at one-half scale and one at two-thirds scale. The BCJ specimens were companion specimens to the column specimens discussed previously, i.e., individual column specimens were tested separately and the BCJ tests were conducted to assess the behavior of the beam and joint region. For the BCJ test, the column strength and stiffness were properly modeled; however, the column was well detailed to preclude brittle column failure even though a weak beam response was anticipated. Due to space limitations, only results for the two-thirds scale test are presented. The overall geometry for the two-thirds scale BCJ test is shown in Fig.10. Pin-to-pin dimensions for the beam and column were 16 ft and 8 ft ­ 4 in., respectively. Both the beam and column cross sections were 16-inches square. Beam longitudinal reinforcement consisted of 3 - #7 bottom bars, with one of the beam bottom bars extending only 5 ft from the column face, and 4 - #9 top bars, two extending 5ft and two extending 2ft ­ 8 in. from the column face (Fig. 10). A 33db length splice was provided for the remaining 2 - #7 beam bottom bars at 5 ft from the column face as shown in Fig. 10. Two concrete mixes were used, a lightweight concrete mix for the beam and joint region and a hardrock mix for the column. The target column and beam test-day strengths were 6.6 and 4.4 ksi, respectively. Actually test-day strengths, based on the average of three cylinder tests, were 5.97 and 5.11 ksi, respectively. Rebar yield stresses were 65.6 and 66.0 ksi for the #9 and #7 beam longitudinal bars, respectively. Beam transverse reinforcement consisted of closed hoops adjacent to the joint and U-shaped stirrups elsewhere, as shown in Fig. 10. Additional closed-hoop stirrups were provided over approximately one beam total depth (16") at the rigid link attachment points near the end of the beam to address load transfer issues. ACI 318-05 development and splice lengths for the top and bottom bars based on as-tested material properties are:

3 fy t e s 3 55,500 t (1.0)(1.0)(1.3) d = ld = = 40.8 t db2 40 f ' ( (cb + K tr ) d b ) b 40 4, 400 ( (1.75"+ 0.25) d b ) c

a = 0.006h b = 0.015h f'c = 6000 psi assumed

-4 -3 -2 -1 0 1 2 Lateral Displacement (in.) 3 4

FEMA 356 Table 6-8: My=3100 in-k; Vy=2My/h=66 k kVn=54k+17k=71k > Vy Flexure Controlled Vu = 3.2f'c P = 0.11Agf'c


Figure 8. 1.0GS11M Load Displacement Test Results


Figure 9. 1.0GS11M at 4.0" displacement. For both tests, the G-Loc devices appeared to have no impact of the observed column behavior and failure mode. This result is likely due to the relatively low moment demand at column



Full-scale span = 25 ft 2/3 - scale = 16.65 ft Use 16 ft for test to simplify

7'- 4" 5'- 0" 2-#9 (2-#14)* 2'- 8" 2-#9 (2-#14)*

1.5" ID PVC Column base support anchor location Spacing between pvc to be determined

#2 @ 2.75" 1st hoop @ 1.5" 1.5" + 11 @ 2.75" 2'- 10" #2 @ 2.75"

#2 U's #3 U's 4 @ 6.75 = 27" 4 @ 6.75 = 27" 2- 6" #2 @ 6.75

2'-3" #3 @ 6.75"

2-#7 (2-#10)* 3-#7 (2-#11, 1-#10)*

90 degree Standard hooks On beam bars

1'- 4" 33D lap* *33D Full-scale

#4 @ 3" Hoops First hoop @ 1.33"


4'- 2"

A schematic of the test setup is provided in Fig. 11. The beam gravity load was applied using two cylinders, one for each beam, attached to a rigid linkage system connected to the beam with thru-bolts at four locations. The linkage system was designed to provide a load distribution of 4.5:4.0:3.5:3.0. (Fig. 12) to achieve the gravity moment and shear at the beam-joint interface calculated for he actual building. The test sequence consisted of connecting the horizontal actuator at the top of the column and applying the column axial load equal to Pcol 0.11Ag f c' = 165 kips using a hydraulic jack and two post-tensioning bars connected to steel supports at the top and bottom of the column. The horizontal actuator was then used to ensure the column was vertical (not leaning). Next, the beam gravity load of 14 kips for each actuator was applied at a resultant distance of 37.33 in. from the beam-joint interface (under load control). Finally, the rigid links at the beam ends were connected and cyclic lateral loading was applied at the top of the column. Supports were provided near the beam ends to restrain out-of-plane motion (Fig. 12). Load cells were used to measure the column axial load, the beam gravity load, and the loads in the rigid links at the beam ends. Lateral displacements were measured with a pair of LVDTs and a pair of wire potentiometers at the lateral load application point, as well as at the bottom of the column to measure potential slip (fig. 11). LVDTs also were used at the base column support to measure potential rotation.


Beam bar cover: 1" clear top & bottom beam bars C L


16" x 16" 4 - #9 w/1.25" clear cover With additional #4 bar through joint


Threaded Rebar bolted To clevis Joint


Add 3 bars at lap per S20 Detail 2


Figure 10. Specimen Geometry and Reinforcement. For the #7 beam bottom bars and #9 beam top bars, a development/splice length of 35.7db (31.2 in.) and 59.8db (5'7.5") are computed, respectively. Therefore, the splice on the beam bottom bars is close to adequate (33/35.7 = 0.92); however, the development length on the 2 - #9 bars terminated 2' ­ 8" from the column face, is substantially less than required to developed bar yield (32"/60" = 0.53). Therefore, it is anticipated that negative moment capacity at the beam-joint interface will modestly exceed the value calculated for the 2#9 bars terminated 60" (equal to the calculated development length) from the joint-beam interface. As well, a crack would be expected to open where the 2-#9 top bars are terminated. The expected joint shear demand, calculated assuming the 2#9 top bars and the 3-#7 bottom bars, to reach yield is:

vu , joint = ( 3.8 in (65 ksi) - (Vcol 30 kips ) ) (16" x16") = 850 psi


Load cell for axial load Cyclic lateral load

LC 200 kip Jack for axial load

column, top

PT bar for axial load Beam



Reaction blocks


vu , joint

f c' = 850

4500 = 12.6


Load cell

Gravity load

Gravity load

column , bottom

Load cell

For the test configuration, the nominal joint shear strength is

vn f c' = 15 ; therefore, joint shear failure was not expected

16 ft

assuming the 2-#9 top bars do not contribute significantly to moment strength. The primary objectives of the test were to assess beam behavior given the inadequate development length on the beam top bars and the splice of the beam bottom bars, as well as to verify adequate joint performance. An additional important objective was to assess the ability of the beams to transfer gravity load to the column given the anticipated beam behavior (inadequate development).

Figure 11. Test Set-up Schematic Instrumentation on the beam-column joint consisted of a total of 26 LVDTs placed on the east and west faces of the beam and joint regions: 10 LVDTs placed diagonally to measure beam and joint shear, 12 horizontal LVDTs on the beam east side to measure curvature and slip at the beam-joint interface, and two horizontal and two vertical LVDTs within the joint region (Fig. 13). Strain gauges were affixed to beam

longitudinal, beam transverse, and joint transverse reinforcement at 18 locations to aid in assessing the behavior of the assembly.

60 40

LO A D (kips)

Linkage system for gravity load & vertical actuator

20 0 -20 -40 -60

Rigid beam tip links for lateral load (connected after application of gravity load)

Gravity load actuator Mg/Vg constant (load control)







Figure 14. Beam-Column Joint Cyclic Test Results.

Figure 12. Beam Gravity Load Linkage system


MN+ = 1500"-k Mn- = 2800"-k

MOMENT (in-kips.)


Mn+ = 750 "-K Mn- = 1400 "-k




-3000 -0.03 -0.02 -0.01 0 0.01 0.02 0.03

CURVATURE (radians/inch)

Figure. 15 Beam Column Joint Moment vs. Curvature Figure 13. Beam-Column Joint Experimental The loading protocol consisted of applying lateral drift cycles equal to 0.125, 0.25, 0.5, 0.75, 1.0, 1.5, 2.0 times the estimated yield displacement of 1.5 inches (which proved to be quite accurate). Subsequent cycles were applied to achieve 4%, 6%, and 8.25% drift. Given the column pin-to-pin length of 100", drift ratio is equal to the applied lateral displacement. The stroke of the wire potentiometers for positive loading was exceeded for the 8.25% cycle; therefore, displacement values are clipped for positive load in Fig. 14. The results plotted in Fig. 14 reveal that the test assembly reached a lateral drift ratio of 4% prior to the observed lateral strength degradation. The moment versus curvature relation adjacent to the beamjoint interface is shown in Fig. 15. The relation was derived using the horizontal LVDTs affixed to threaded rods anchored through the beam near the top and bottom faces of the beam as:

( (


- bottom ) h* , where h* is the average vertical


distance between sensors. The positive nominal moment + capacity of M n = 1,500 in.-kips was reached, indicating that the bottom bars reached yield, however, all 4-#9 tops bars did not reach yield as the negative nominal did not reach the - nominal moment capacity of M n = 2,800 for 4 - #9 bars. As noted previously, 2-#9 bars were terminated prematurely, at about one-half the computed development length; therefore,

this result is not unexpected. The negative moment reached about 2,300 in.-kips, or substantially more than the nominal moment associated with the 2-#9 top bars that continue (Fig. 10), or approximately 2,000/2 = 1,400 in-kips. Therefore, the test results indicate approximately 2300/1400 = 1.64 times the negative moment capacity expected if FEMA 356 recommendations were used. The larger negative moment strength achieved within the beam may have significant negative impacts. For example, the higher moment demand increases beam shear demands, joint shear demands, and column demands (moment and shear). For this case, the higher demands do not lead to other brittle failure modes, which is an important finding of the test program. Figure 16 illustrates the sum of the beam gravity forces resisted by the column during the test. For reference, the target gravity load (28 kips) and the applied drift also are plotted. As yielding and damage occur adjacent to the beam-joint interface (Fig. 13), the beam gravity load is redistributed to the rigid links at the beam ends (an artifact of the test setup). At a lateral drift ratio of 3%, the beam gravity load resisted by the column was approximately 18 kips, or about 65% of the original beam gravity load resisted by the column. The test results indicate that, despite significant beam damage, the beam can still transfer a significant portion of its gravity load to the column.

30 10 8 P (kips) 6


little information was available to determine the rate of strength degradation (instantaneous for most FEMA 356 backbone relations) or the deformation capacity associated with loss of axial load capacity (residual strength). Given this lack of information, conducting building-specific test programs may provide valuable information that substantially reduces the costs associated with seismic rehabilitation. Results obtained from large-scale component tests on wall segments, columns, and beam-column-joints are presented. The findings confirm the general conservatism associated with the FEMA 356 backbone relations. In general, the buildingspecific tests indicate greater strength, lower effective stiffness, substantially more deformation capacity, and less pronounced lateral strength degradation. All of these findings have the potential to significantly reduce the costs associated with seismic rehabilitation Acknowledgements The work presented in this paper was supported by funds from KPFF Consulting Engineers and conducted in collaboration with two southern California hospitals. The assistance of Dr. Thomas Kang, Assistant Professor at the University of Oklahoma, as well as UCLA graduate students Mr. David Naish, Ms. Anne Lemnitzer, and Mr. Eric Ahlberg are greatly appreciated. Undergraduate laboratory assistants Joy Park and Nolan Lenahan, PEER summer inter Ms. N. N. Lam, and Senior Development Engineer Mr. Steve Keowen, also helped with the conduct of the experimental program. Any opinions, findings, and conclusions or recommendations expressed in this paper are those of the authors and do not necessarily reflect those of the supporting organization or other people acknowledged herein. References Elwood, K. J. et al.; "Update to ASCE/SEI 41 Concrete Provisions," Earthquake Spectra, August 2007, Vol. 23, No. 3, 30 pp. CA FEMA 356, Global Topics report on the prestandard and commentary for the seismic rehabilitation of buildings, November 2000, Federal Emergency Management Agency, Washington D.C. Massone, L. M., "RC Wall Shear ­ Flexure Interaction: Analytical and Experimental Responses," Ph.D. Dissertation, June 2006, 398 pp., Department of Civil & Environmental Engineering, University of California, Los Angeles. Massone, L. M., Orakcal, K., Wallace, J. W., "Modeling flexural/shear interaction in RC walls", ACI-SP-236, Deformation Capacity and Shear Strength of Reinforced



Column Gravity Load Measured Gravity Applied drift

4 2 0

0 0 4000 8000 TIME (seconds) 12000


Figure 16. Beam-Column Joint column gravity load and % drift during the experimental test. Conclusions FEMA 356 backbone relations used to define the nonlinear force versus deformation response of typical reinforced concrete structural elements tend to be quite conservative. Component strengths are often based on ACI 318 relations, which are commonly are close to lower-bound values. Deformation capacity at the initiation of lateral strength degradation also tends to be quite conservative, and relatively

Concrete Members under Cyclic Loadings, May 2006, Paper 7, pp. 127 ­ 150, American Concrete Institute, Farmington Hills, MI. Office of Statewide Health Planning and Development, "Summary of Hospital Seismic Performance Ratings," April 2001, 27 pp., Public Affairs Office, CA


Microsoft Word - SEAOC_Final_20_July.doc

11 pages

Report File (DMCA)

Our content is added by our users. We aim to remove reported files within 1 working day. Please use this link to notify us:

Report this file as copyright or inappropriate


You might also be interested in

Pushover Analysis of a 19 Story Concrete Shear Wall Building
080702-2008 SEAOC Convention Model Paper - TBI Task 7-Malley
Test Solenoid Design