Final Report

STRENGTHENING OF REINFORCED CONCRETE COLUMNS

FOR EARTHQUAKE RESISTANCE

by

Lawrence F. Kahn

Prepared for

National Science Foundation

Washington, D.C. 20550
under
Research Initiation Grant No. ENG77-06478

June 1979

Georgia Institute of Technology

School of Civil Engineering
Atlanta, Georgia 30332
Report NSF ENG77-06478

STRENGTHENING OF REINFORCED CONCRETE COLUMNS

FOR EARTHQUAKE RESISTANCE

Lawrence F. Kahn
School of Civil Engineering
Georgia Institute of Technology
Atlanta, Georgia 30332

June 1979

Final Report

Prepared for

National Science Foundation

Division of Problem-Focused Research Applications
Washington, D.C. 20550
 ABSTRACT

Four identical, 10-inch square, reinforced concrete columns were

constructed using four No. 7 bars and 6400 psi concrete. Their design
included no special transverse reinforcement for earthquake resistance.
Three of the columns were strengthened externally using various tech-
niques in order to improve their shear resistance and ductility. One
technique used 2-inch wide steel packaging bands which were wrapped
around the column and spaced at 4-inch on center. The space beneath
the bands was packed with grout. A second column was spirally wound
with a 1/4-inch diameter steel bar on a 1.1-inch pitch. The space be-
neath the rectangular spiral was grouted. For the third strengthening
technique, U-shaped clamps were fabricated from 2-inch x 5/16-inch steel
bar and from 3 x 5 x 5/16 inch steel angle. Two U-clamps were bolted
together around the column to form a hoop; these hoops were 4.25-inches
on center.
All four columns were tested under static reversed cycle deflections
of increasing magnitude and with a constant axial load of 80,000 lbs.
The unstrengthened column collapsed when the lateral deflection was
about twice the deflection causing yield of the tension steel. The
three strengthened columns responded nearly identically and resisted
three reversed cycles at four times the yield deflection with little
deterioration. Based on the test results and the ease of construction,
it was concluded that the U--clamp and banding techniques showed great
promise in providing low cost, easy-to-construct methods for greatly
improving the ductility and earthquake resistance of existing reinforced
concrete columns.
 ACKNOWLEDGEMENTS

Financial support was provided by the National Science Foundation,

Research Initiation Grant ENG77-06478, and by the Georgia Institute
of Technology. This support gratefully is acknowledged.
The main steel members for the structural test frame were pro-
vided by the Georgia Department of Transportation through the good
offices of Mr. Vernon Smith. All structural steel fabrication and
miscellaneous steel was donated by Owen of Georgia, Steel Fabricators,
with the generous assistance of Messrs. Hank Cobleigh and Robert Heany.
Mr. Benjamin Suriano accomplished much of the experimentation,
and he was aided by Mr. William Bynum and Ms. Loretta Britsch.
Opinions, findings and conclusions expressed herein are those
of the author and do not reflect necessarily the views of the sponsors.
 TABLE OF CONTENTS

ACKNOWLEDGEMENTS

CHAPTER Page

1. INTRODUCTION 1
1.1 Purpose 1
1.2 Scope 1
1.3 Background

2. EXPERIMENT DESIGN 7
2.1 Design Philosophy 7
2.2 Specimen Design 9
2.3 Specimen Construction 9
2.4 Test Set-up and Instrumentation 21
2.5 Test Procedure 21

3. EXPERIMENTAL RESULTS 29
3.1 Load-Deflection Response 29
3.2 Physical Observations 35
3.3 Moment-Curvature Response 38
3.4 Strain Observations 45
3.5 Energy Dissipation 52

4. DISCUSSION 55
4.1 Quantitative Analysis 55
4.2 QualitativeAnalysis 59
4.3 Limits of Findings 61

CONCLUSIONS AND RECOMMENDATIONS 63

5.

APPENDIX: MATERIAL TESTS 65

REFERENCES 95
 1. INTRODUCTION

1.1 Purpose

The purpose of this research was to experimentally investigate

several methods of strengthening reinfornced concrete columns to improve
their seismic resistance.
Due to increasing alarm over the possible occurrence of earthquakes,
some geographical sections of the United States and other parts of the
world have considered modifying zoning criteria for earthquake resistant
design. Furthermore, some facilities now are desired to survive and
function after severe earthquakes. For these reasons, the designer must
turn his attention to methods of strenghtening existing structures.
Ideally, these methods should be employed before an earthquake occurs;
4 however, they car be used along with the repair of a previously damaged
structure.
The objective of this research was to initiate investigations
aimed at providing the designer with qualitative and quantitative in-
formation on how existing reinforced concrete columns may be economically
strengthened.

1.2 Scope

The scope of this project was limited. Four identical reinforced

concrete columns were cast and three enterior strengthening methods were
employed. These strengthening methods were the only variables in the
experimental program. They were used to provide additional shear cap-
acity for the structural members when compared to an unstrengthed member
and to provide confinement.

1.3 Background

The deterioration of strength of reinforced concrete columns due

to earthquake type loading has been investigated by several researchers
(4,7,9,11,13,19,20) and design recommendations have been proposed in-
cluding that the minimum transverse reinforcement ratio be 0.6 percent
and that the transverse reinforcement be designed to carry the full shear

1
forces. Other studies have investigated the adequacy of the repair of
concrete structures after being damaged by earthquake type loading (5,
7,9,10,13,14).
Results from repaired specimens tend to show that the repaired
structures respond with similar strength and in the same manner as the
original specimen; although, significant differences can occur (13).
The joint ASCE-ACI Task Committee 426 (16) defined the basic
mechanisms of shear transfer and failure criteria for reinforced concrete
columns. The most prevalent mechanisms are shear transfer by concrete
shear stress which occurs in uncracked members. or portions of structural
members, interface shear transfer which is stress along a diagonal ten-
sion crack and is called aggregate interlock, dowel shear which is shear
resisted by longitudinal reinforcement, and shear reinforcement.
Under repeated and reversed loadings the Committee states that
deterioration of the first three mechanisms will occur rapidly and that
only by employing closely spaced stirrups will splitting along the long-
itudinal reinforcement be restrained. Therefore, dowel action and shear
reinforcement will account for the full shear transfer. However, since
dowel action is dependent on shear reinforcement, the stirrups should
provide for the full shear in beams and columns.
Vallenas, Bertero, and Popov (19) investigated column cores confined
by rectangular hoops and loaded axially. A total of 14 reinforced
specimens were tested varying the effects of 3 parameters, concrete
cover, lateral reinforcement, and longitudinal reinforcement.
The first group consisted of two plain 20 in. long by 10 in. square
concrete columns which underwent a relatively brittle type of failure
with a large diagonal crack opening suddenly.
The second group of six specimens tested were confined concrete
columns with no longitudinal reinforcment, 3 with cover and 3 without.
The confinement was obtained by using plain #7 wire at 1.33"c/c spacing.
This lateral reinforcement did not rupture at failure, instead the hooked
ends slipped out of the concrete showing a need to use deformed bars for
better anchorage. The specimens with cover had a slight increase in max-
imum load before cracking than those without cover.
The third group of specimens tested were similar to the 2nd group,
with the addition of 8 4#6 longitudinal bars. As in the above group,

2
3 columns had cover and 3 did not. In both these cases an increase in
concrete strength was obtained with failure occurring from buckling
of the longitudinal reinforcement and rupture of the stirrups.
The confinement of the test specimens produced an increased strength
of 13% when compared to the plain specimens. It was also noted that the
Uniform Building code (UBC, 18) and American Concrete Institute (ACI-318-77,3)
equations for ratio of confining steel can be combined and this equation
can be used for any type of confinement system by varying the confinement
effectiveness ratio for different types of confinement and materials.
Bertero and Popov (4) investigated the hysteric behavior of re-
inforced concrete beams subjected to high and low shear stresses. They
suggested that this deteriorating behavior can be improved by using a
closer spacing of stirrups and increasing the are of compression
reinforcement.
Wight and Sozen (20) investigated the hysteric behavior of 12 re-
inforced concrete columns subjected to large shear reversals. The spec-
imens represented a column between the points of contraflexure above
and below a story level. The principal variables of the test were the
amount of axial load, the transverse reinforcement ratio, and the re-
quired deflection ductility (total deflection divided by yield deflection)
for each cycle. A. comparison was made between specimens with and without
an axial load using the same transverse reinforcement ratio; the specimens
without an axial load suffered a more rapid decrease in strength with
each complete cycle of load reversals. The specimens with axial loads
had higher yield and ultimate shear capacities. Additional results
from the tests indicate that the shear capacity of the member should
be based on the shear capacity of the column core confined with closely
spaced stirrups.
Lee (13) tested beam-column subassemblages subjected to earthquake
type loading. The main variable between two types of models was the
amount of transverse reinforcement in the joint. The first design was
in accordance with the ACI 318-71 code for nonseismic areas. This
was assumed to represent an existing structure which was designed with-
out considering seismic loading. The second design was in accordance
with ACE 318-71 including "Appendix A" for the design of ductile moment-

3
resisting space frames. The testing included virgin & repaired specimens
of both types. Results from these experiments demonstrated that epoxy
injection and removal and replacement techniques of repair can effect-
ively restore the stiffness, strength and energy dissipation capacity
of beam-columns. The repaired specimens were found to be stronger than
the original specimens at the same deflection level due to the strain
hardening of reinforcement and to the higher strength repair materials.
Because of the specimens increased strength, the beam to column joint
is usually stressed to a higher level, thus creating the possibility
of damage moving from the beam to the unrepaired joint. Also, Lee
concluded that stirrups should be designed to carry all of the shear
force at the points of maximum moments.
A comprehensive collection of the most recent literature presented
on the earthquake repair and strengthening of structures was given at
the 6th World Conference on Earthquake Engineering. A few of the ex-
perimental programs pertaining to reinforced concrete members are out-
lined below.
Gulkan (7) tested two three-fourths size beam-column connections
which were subjected to reversed cycle deflections of double curvature
before and after repair. The columns were loaded axially while the
beams were loaded in reversed shear to produce the double curvature.
The only difference between the two specimens was the lack of the beam
stub representing an out-of-plane beam framing into the joint. After
failure of the virgin specimens, the original shell of the column was
chipped off and replaced with more longitudinal and transverse rein-
forcement and concrete cast around the original core. This repair
technique improved the strength of the column considerably but forced
failure into the joint core.
Higashi and Kokusho (10) investigated strengthening methods of
existing reinforced concrete buildings. Three experimental test pro-
cedures were conducted with respect to these methods.
In the first test a comparison was made between a monolithic shear
wall cast with a rigid frame and a rigid frame strengthened by a shear
wall poured under pressure. Results from the above test show remark-
able increases in the strengthened frame to a degree almost equal to the
monolithic shear wall. Rigidity and lateral capacity were increased

4
substantially under large deflections. Although the behavior under
working loads were similar, the mode of failure between the two speci-
mens was different.
The second test procedure consisted of strengthening columns by
the addition of wing walls at the sides of the existing columns. The
most substantial increase in rigidities and strengths occurred when the
wall reinforcements were welded to the hoops in the columns before
the walls were cast. Other methods of fastening the wing walls (steel
anchor pieces, mprilar grouting, etc.) did not show significant strength
increases.
In the third test two specimens were compared; (1) an existing
reinforced concrete column, and (2) the same column in (1) with welded
wire fabric wrapped around the column and mortar poured in place. The
column in (2) had gaps at both ends to prohibit spalling of the column
at the face of the joint. The results indicated that (2) showed def-
inite increase in the ductility and deformation capacities under re-
versed cyclic loading. Due to the additional reinforcement, the con-
finement area of the column was increased and therefore, the ultimate
capacity of the existing column was guaranteed.
Freeman (6) described motification of an existing hospital facility
to satisfy the new Veterans Administration (VA) seismic design criteria.
Since the new criteria was more severe than the original, a response
spectrum modal analysis was made with the aid of a digital computer
program. The output data provided the force distribution to the members
for there modes of vibration.
Several strengthening modification schemes were evaluated for their
feasibility and economic application. The proposed scheme was a comb-
ination of a shear wall and rigid reinforced concrete frame placed
around ther perimeter of the 15-story tower structure. The shear walls
were cast on the existing mat foundation at the corners of the tower and
extend its full height. Peir like columns were then cast at the present
exterior column lines to form the rigid frame. A majority of the lateral
force resisting capacity has been offered by the shear walls, but the
frame system reduced the buildup of overturning moments at the base of
the shear walls to produce a ductile seismic resistant structure.
Strengthening has been accomplished along with the repair of

5
structures. The repair of the Mene Grande Building, Caracas, Venezuela,
after the earthquake of July 29, 1967, included the placement of additional
transverse reinforcement in the columns (8).
The repair of the Holy Cross Hospital after the San Fernando Earth-
quake included the strengthening of some columns (17). In locations
where the columns had failed, damaged concrete was removed, and new ties
were placed. Gunite was then shot in place. Kajfasz (12) found that
concrete beams could be adequately strengthened in shear by epoxy bonding
steel stirrups to the exterior of the beams.
The past research has shown that reinforced concrete columns do
fail in shear under earthquake forces and that rapid deterioration of
strength may occur if insufficient shear reinforcement is present. Yet,
little research has been done to determine what methods may be used to
strengthen the shear resistance of existing columns and what the adequacy
of those methods might be. This experimental program examined some
appropriate strengthening methods to determine their potential for improving
the earthquake resistance of existing columns.

6
 2. EXPERIMENT DESIGN

2.1 Design Philosophy

In order to investigate strengthening techniques, it was necessary

to design experimental models which were weak with regard to earthquake
resistance; yet, the test specimens needed to model actual existing
construction. Because of laboratory considerations, the size of the
models were limited, but the model size needed to be large enough to
represent accurately such behavior as bond of reinforcement and aggregate
interlock for shear.
Therefore, the specimens were designed as two-thirds scale model
columns according to provision in ACI 318-63 ( 2 ) without regard to
earthquake effects or to concepts of ductile concrete which were dev-
eloped during the 1960's (Blume Newmark Corning 1). The overall depth of
the column was chosen as 10 in. to correspond to models tested by
Wight and Sozen (20). The width of the specimen was set at 10 in. (un-
like Ref. 20) because the 'typical reinforced concrete column is square.
Also typical of existing columns are reinforcement ratios (p) between
2 percent and 3 percent, and ratios of core dimension (d - d') to thick-
ness ( h ) of 7 to 8. This ratio is generally termed r. As shown below,
the model speciems had p and r ratios within these ranges.
Past earthquakes have demonstrated that reinforced concrete columns
often fail in the first story just below their connection with the second
floor girders. The stiffness of the second floor and the structure above
is often much greater than that of the first columns; this rapid transition
in building stiffness apparently induces significant shear together with
flexure and axial load in the columns which results in column failure.
Furthermore, large interstory lateral deflections induce a P-A moment
in those column.
In order to model this weak column - stiff girder connection and to
provide a P-L\ effect, a model like that shown in Figure 2.1 was selected.
The large center block represents the stiff girder connection. With
a constant axial load, the specimen was flexed in single curvature by
applying a lateral load at the center block. The P-A moment was gen-
erated by the difference in the line of action of the axial force and

7
 Lateral load -z
Axial Load -----

Center block (beam-column

connection)

Face of joint
.
Simple supports

ti
'-10"x10" square column

Figure 2.1 Column test specimen under axial load and flexed
in single curvature.

8
the deflection at the column-block joint.
The specimen was symmetric about the center line, and in the elas-
tic range the deflections were also symmetric. Therefore, each half
of the specimen represented the upper half of this hypothetical first
story column. The specimen was designed to be typical of reinforced
concrete columns, but it was not conceived to model all possible var-
iables. The purpose of the tests was to develop a qualitative under-
standing of these simple strengthening techniques rather than to explore
the range of parameters affecting reinforced concrete column response.

2.2 Specimen Design

As shown in Figure 2.2, the 10-in. square columns were reinforced

with a No. 7 deformed bar in each corner and with 11 gage (0.22 in. dia-
meter) ties at 10 in. spacing. The original design called for Grade 60
steel and a concrete compressive strength ( f' ) of 4000 psi. As listed
in Table 2.1 below, the actual material strengths were different than
those design strengths. No "special transverse reinforcement" as req-
uired by current standards (ACI 318-77 & SEAOC Code) was included for
concrete confinement or for shear resistance. The columns were purposely
designed so that the shear resistance provided by the concrete under an
80 kip axial load would by about 15 percent less than the lateral load
which wuold cause the ultimate moment at the column-center block joint.
Shear calculations based on ACE 318-77 ( 3 ) were known to be somewhat
conservative ( 16 ), and it was desired that the specimens be weak
in shear so that the strengthening methods would be required to aid
in the shear resistance.
Design for a full size column based on ACI 318-63 would require a
11/2 in. cover as opposed to the 1 in. cover used. But that code does
not allow use of 1/4 in. diameter ties as used in these specimens.
The shear span of 50.75 -in. was chosed to fit existing laboratory
equipment and to model at two-thirds scale one-half the height of an
actual column.

2.3 Specimen Construction

2.3.1 Unstrengthened Columns

9
 Table 2.1 Material Properties

f' f" #7 reinf. Strengthening Strengthening

Sp ecimen c c
fy System fy
(psi) (psi) (psi) (psi)

1 6350 5540 56,800 None -

Packaging
2 6470 6220 56,800 93,700
Bands

No. 2 bar
3 6350 6390 56 , 800 66,900
Spiral

42,600*
4 6470 6130 56,800 U-clamps
34,100**

* 2" x 5/16 bar

** 3 x 5 x 5/16 angle

10
 ' - 0"

anr-11.••• ■
5 @ 10" = 4'-4" 1 "

#4 gage wire ties, typ.

Fi gure 2 .2 C ol umnr ei nfor ci n gb ar d e si gn .

7,
r

A
Elevation

#4

#3 2"-2" 1" chamfer

#4 gage wire ties

#7

1 " clr. typ.

1'-5" "
 The four column specimens were cast in two pours: Specimens 1
and 3 at one time, and Specimens 2 and 4 several weeks later. The four
specimens had identical plywood forms (Figure 2.3).
The main and transverse reinforcing bars were strain gaged as dis-
cussed below, and then they were tied into cages (Figures 2.4,2.5).
The cages were positioned into the forms using precast cement blocks
as chairs.
A nominal 4000 psi concrete with a 3/4 in. maximum sized aggregate
was ordered from a local supplier and was delivered in a ready-mix truck.
For each pour, slump test assured that the slump was greater than 4 in.
for workability. The concrete was poured directly from the truck into
the forms (Figure 2.6). At the same time, nine 6-in. x 12-in. concrete
test cylinders were cast. As shown in Figure 2.6 the columns were cast
horizontally so that no cold joint was formed at the column to center
block joint. Vibration with a spud vibrator assured compaction.
About four hours after casting, the forms and test cylinders were
covered with wet burlap. The burlap was kept moist for one week, after
which the forms were removed and the specimens placed within the lab-
oratory building. Cardboard forms from three of the test cylinders were
removed one day after casting, and the cylinders were stored in a fog
room, 100 percent humidity at 73 ° F, until the cylinders had aged 28 days.
The remaining test cylinders were kept with-the specimens and were cured
under identical conditions.

2.3.2 Strengthening Techniques

Specimens 2,3 and 4 were strengthened using different techniques

after the columns had cured for a minimum of one month. Specimen 2 was
strengthened by hooping the column with 2-in. wide and 0.045-in. thick
steel strapping bands; such bands are used typically for packaging.
Figure 2.7 shows the first band being tightened around the column using
a lever device; one end of the band was fixed in the base of the device
while the other was secured in a spindle. The spindle was rotated by
the lever, and the band, thus, was tensioned. One free end of the
band was lapped over the hoop and was secured with two metal clips
(figure 2.8) by crimping the clips with special pliers. As the lever
device removed, the tension in the band was released. As the band

12
Figure 2.3 Typical form for column specimens.

13
Figure 2.4 Reinforcing bar cage.

Figure 2.5 Reinforcement in forms.

14
Figure 2.6 Casting column specimens. Figure 2.7 Tightening banding around Specimen 2.
tension decreased, a gap ranging from 1/8 in. to 3/8 in. occurred be-
tween the band and the column; although the bands remained tight around
the corners*.
The first band next to the joint was spaced 1 in. clear from the
joint. The other bands were spaced on 4 in. centers over a distance of
40 in. on each side of the joint. The gap beneath each band was packed
with a non-shrink grout (Embeco 636 by Master Builders) to assure con-
finement.(Figure 2.9).
Specimen 3 was strengthened with a rectangular spiral. A plain
1/4-in. diameter steel rod was hammered around the column to form a spiral
with a 1 1/16 in. pitch. A starting loop (zero pitch) was placed within
1/2 in. of the joint. For ease of construction, straight 10 ft. lengths
of the rod were used to form the spiral. After one 10 ft. length had
been wrapped, the next length was lap welded to the end of the existing
spiral (Fiugre 2.10). A lap weld of 4 in. was used; it was calculated
that this length would develop the yield strength of the rod.
Gaps between the spiral and the column were as large as 1/16 in.
These gaps and the space between the rods was filled with a Portland
cement grout made of equal parts sand and cement and sufficient water
to provide and workable mortar. (Figure 2.11). The column was thoroughly
wetted prior to applying the mortar.
Specimen 4 was strengthened with U-shaped clamps. The clamps
were made of 2-in. x 5/16 in. hot rolled steel bar cut 10 in. long
which was fillet welded to A36 steel angle 5 in. x 3 in. x 5/16 which
was cut 2 1/4 in. wide (Figure 2.12). The angle had a lower yield
stress than the bar; so the angle was cut wider than the 2 in. bar so
that the total yield force of the angle and bar would be equal.
Holes of 13/16 in. diameter were drilled in the 3-in. outstanding
legs of the U-clamps for A325 3/4-in. diameter bolts. It was calculated
that the yield force of the bolt was greater than the yield force of
either side of the clamp.

* The author originally desired that the bands remain tensioned, but
the lever device required about a one-half inch gap beneath the band.
As the device was removed, this space permitted the band to slacken.

16
Figure 2.8 Crimping metal clips to secure banding hoop.

Figure 2.9 Specimen 2 banding showing clips and non-shrink grout paCked
beneath the bands. Photograph taken after deflection
sequence was completed.

17
Figure 2.10 Hammering No. 2 bar around Specimen 3. Note lap
welds.

Figure 2.11 Specimen 3, cement grout was mortared around the

rectangular spiral. Photograph was taken after
deflection sequence was completed.

18
 Once fabricated the U-clamps were easily secured around the columns
and bolted together. The clamps fit tight and required gentle hammering
to seat them. The first clamp was spaced 1-in. clear from the joint,
while the remaining clamps were spaced on 4 1/4 in. centers over a
40 in. distance (Figure 2.13).

2.3.3 Materials

Detail description of material properties are given in Appendix A.

General properties of the materials used to construct the specimens
is given below and in Table 2.1
The compressive strength of the 28 day fog cured concrete cylinders
(f') was 6350 psi for Specimens 1 and 3, and it was 6470 psi for Specimens
2 and 4. As stated above, an additional three cylinders were cast with
each column and were cured under identical conditions. These cylinders
were compression tested when each specimen was tested; the average
strength of these field cured concrete cylinders (f") is given in Table 2.1
The tensile stress-strain response for the steel reinforcement is
given in the Appendix A; two tension tests for each type of reinforcement
were conducted. The average yield stress for the No. 7 bar was 56,800
psi, and the average 0.2 percent offset yield stress for the 11 gage
wire was 77,900 psi.
Tension tests of the materials used for the strengthening techniques
gave the yield stress results listed in Table 2.1 The yield stress
given for the banding steel was the average yield stress of two tests
of 3 ft. lengths of material. Two additional tension tests were conducted
with two clips in an identical manner to the connection used on the
column hoops. These tension tests were designed to examine the capacity
of the clip connection. The lapped bands began slipping through the
clips at an average load of 7800 lbs., and the bands freely slipped at
8400 lbs. The initial slip load was divided by the gross area of one
band of 86,700 psi, a value which was 7.5 percent less than the actual
0.2 percent offset yield stress of the banding. Therefore in calculations
regarding the force capacity of the banding hoops, the author believes
that the "apparent yield stress" should be used.

19
 Figure 2.12 U-clamps bolted on Specimen 4.

• 1
- • 4••••••••.

r- I

11, 7-1
- 11,1.1
•

-
•

-00 ?N gra

it-■

till, It 11111111

Figure 2.13 Specimen 4 with U-clamps ready for test.

20
2.4 Test Set-up and Instrumentation

Reinforcing bar strain measurements were taken with standard 1/4 in.
electrical resistance strain gages. The gages were located at the
center of the joint and at approximately one inch from the face of
the joint on the main reinforcement (Figure 2.14). Strain gages also
were placed on the first column tie from the face of the joint. The
gages were bonded to a machined suface of the steel reinforcing bar
with a two part adhesive, M-bond 200. An epoxy resin coating was used
as a final step to protect the gages against impact during casting
(Figure 2.5). Also, strain gages were bonded to the near and far sides
of the first strengthening technique from the face of the joint on
Specimens 2 and 4 (Figure 2.15).
Figure 2.13 and 2.16 illustrate the test set up. The axial load
cell was positioned at the right end of the specimen while the load was
applied by a hydraulic loading ram on the left end of each specimen.
The lateral load was applied at the center of the joint and monitored
by a strain gage load cell which was used for both downward and upward
loading. Lateral deflection measurements were taken at the edge of
the center block adjacent to the column and at 10 in. from the face
of the joint.

2.5 Test Procedure

All test specimens were mounted in the structural test frame as

shown in Figure 2.13; pinned bearings without rollers were provided
for all supports (Figure 2.17). An axial load of 80,000 lbs. was applied
with the hydraulic jack; this load was maintained throughout the test
sequence or until the column failed. This load produced a stress of
800 psi which was considered to represent a typical working axial stress
used by Wight and Sozen (20); so comparison with their results would be
facilitated.
Each specimen then was cycled through the lateral deflection sequence
shown in Figure 2.18. Lateral loads were applied by the vertically
oriented hydraulic jacks located above and below the center of the
specimen. Downward deflection and loads were considered positive. The
lateral deflection at which the main tension reinforcement would first

21
 - Strain Gages on Main Reinforcement
(I) - Strain Gages on No. 4 Ties

Figure 2.14 - Strain Gage Locations within the Specimen

Figure 2.15 Specimen 4, strain gage on U-clamp nearest the joint
face. Photograph was taken after deflection sequence
was completed.

23
 I
Lateral force hydraulic jack
Dial gage atioint face
Dial gage 10" from joint
'Lateral force load-cell
Axial force hydraulic jack

Axial force load-cell

Pin support, see_Fig.. 2.17

4
Lateral force hydraulic jack

Figure 2.16 Test set-up.

E
I
4

Figure 2.17 Close-up view of pin bearings and axial force

hydraulic jack.

II

25
 Cycle

Figure 2.18 Lateral deflection sequence

26
yield (A ) was calculated. The maximum deflection of the first three
cycles was equal to one-half this calculated A value. The maximum
deflection of the fourth cycle was to the actual yield deflection level
of the specimen. During loading the strain on the tension bars was
constantly monitored. When the yield strain was detected, the deflection
was continued to the nearest 0.05 in. level for ease of testing. The
next three cycles (cycles four through six) were to this actual A level.
Thereafter followed three cycles to twice the A level and three cycles
to four times the A level. The maximum deflection in a cycle (A) was
divided by A to give a ductility ratio A/A . Each deflection cycle
required 15 to 30 minutes.

27
 3. EXPERIMENTAL RESULTS

3.1 Load-Deflection Response

The shear force-deflection response for the four specimens is

illustrated in Figures 3.1 through 3.4. Because of the symmetry of
the load and specimen, the shear force at the column-joint face was
equal to one-half the lateral load. The lateral deflections given in
the figures was that found by the dial gage located at the joint face
next to the side of the specimen where the reinforcement was instru-
mented with strain gages (the left side as viewed in the figures).
The centerline of the load was off-set !I; in. toward the side of the
specimens with the strain gages in order to force failure on the instru-
mented side of the models. In the elastic range this slight discrep-
ancy from exact symmetry could not be detected by dial gages placed on
either side of the center block. As plastic hinging occurred in the
column at the face of the joint, the instrumented side deflected more
than the right side.
The four hysteresis curves illustrate that in the cycles to
1/2 Ay (about 0.3 :in.) and to Ay ( about 0.65 in.) all specimens behaved
nearly elastically. The maximum load did not degrade in the cycles
to 1/2- Ay. The maximum load degraded a maximum of 7 percent between the
first and third cycles for the cycles to Ay. The maximum shears to
the 1/2 Ay and the Ay deflection levels are given in Table 3.1, along with
the experimentally determined Ay. The reader will observe on the
hysteresis curves that the specimens were deflected slightly beyond
the actual Ay deflection level during the second three-cycle set in
order to facilitate the experimental procedure. Both the yield deflec-
tions and yield loads for the four specimens were within 14 percent
of each other. The maximum shears observed for the four columns during
the entire deflection sequence were within 7 percent of each other.
Table 3.1 clearly shows that the specimens appeared stronger in the
negative direction (upwards) than in the positive direction (downwards).
Dead load of the specimens accounts for about 0.8 kips of the difference
in shear force between the two directions. With a correction for dead

29
 •

(sdpi) mays

Figure 3.1 Specimen 1, load-deflection response.

30
 (sd!1) Joaqs

Figure 3.2 Specimen 2, load-deflection response.

31
 •C
•••••■•••••

0 O
CN
(Sdp1) Mail S1--

Figure 3.3 Specimen 3, load-deflection response.

32
0 0
111■011M

s d 1.1 ) abaLIS

Figure 3.4 Specimen 4, load-deflection response.

33
 Table 3.1 Load-Deflection Data

Specimen 4y Max. Shear Max. Shear Max. Shear

Experimental (kips) (kips) (kips)
first 3 cycles* Second 3 cycles**

Positive •Postive Negative Postive . Negative Postive Negative

1 6.0 +8.2 -9.6 +14.8 -17.3 +14.8 -17.3

2 6.4 +9.3 -10.1 +14.5 -15.1 +14.5 -16.5

3 +6.0 +9.5 -10.1 +14.5 -15.8 +14.8 -16.3

4 +6.0 +10.5 --11.3 +15.5 -16.9 +15.6 -17.4

* nominally termed 1/2 4y

** nominally termed 4y
load effect, the difference in applied shear is smaller. The remainder
of the difference is believed to have resulted from accidental eccentri-
city of the axial load.
At deflections beyond Ay, the load-deflection response of the un-
strengthened column differed markedly from the responses of the other
columns. During the seventh deflection cycle with a planned maximum
deflection of 1.5 in. (2-Ay), Specimen 1 failed at a deflection of
1.48 in. Figure 3.1 shows an immediate drop in lateral load. Simul-
taneously, the axial load fell to about 50 kips from the original 80
kips. Attempts to return the axial load to 80 kips caused spalling
of the concrete cover and fracturing of the core. At a later deflec-
tion of +2.0 in., the axial load was reduced to zero for Specimen 1.
The column effectively had collapsed. The remainder of the hysteresis
curve was determined for the column under no axial load.
Specimens 2, 3 and 4 demonstrated similar, stable hysteretic res-
ronsesduring deflection cycles to 2-Ay and to 4-Ay. For the three
specimens at deflection cycles to 4-Ay, the load degraded less than 5
percent between the first and third cycles.
For Specimens 2, 3 and 4 as the deflections were increased beyond
approximately 1.2 in (2-Ay), the lateral load decreased slightly with
increased deflection during the first cycle to 4-Ay. This decrease
illustrated the P-A effect whereby the moment created by the eccentri-
city of the axial force due to the lateral deflection had a significant
contribution to the moment at the joint. The second and third cycles
to 4-Ay did not show clearly this P-A effect.

3.2 Physical Observations

After the th..7ee cycles to the yield deflection level, Specimen 1

showed prominent cracks which were highlighted (Figure 3.5). At a
deflection of +1.3 in. the concrete at the top surface crushed and
began to spall (Figure 3.6). At a deflection of +1.5 in. the concrete
cover on the top and bottom spalled. The axial load fell to 50 kips
and was further reduced with increasing deflections. During increased
deflection to +2.0 in., more concrete spalled, the #7 reinforcement
buckled, and a single 4 gage wire tie unraveled. Figure 3.7 shows

35
Figure 3.5 Specimen 1 after three cycles to yield deflection.

Figure 3.6 Specimen 1, initial crushing at +1.3 in. deflection.

36
 Figure 3.7 Specimen 1 after failure.

Figure 3.8
Specimen 2 at +2.5 in. deflection.

37
Specimen 1 as the lateral load was reduced to zero. Reversed cycle
deflection forced much of the fractured concrete to fall from the column.
Specimen 2 with the packaging bands for strengthening showed no
dramatic events during the entire deflection sequence. Only a single
set of diagonal cracks were observed; these were small and crossed
about 6 in. from the joint. Figure 3.8 shows Specimen 2 at the maximum
+2.5 in. deflection.
Specimen 3, reinforced with the No. 2 bar rectangular spiral,
again showed no significant cracking or spalling during the test. Figure
3.9 shows Specimen 3 after the deflection sequence. Some mortar cracked
and spalled from between the spirals; the concrete under the mortar
appeared undamaged. The figure illustrates that some concrete on the
joint face had crushed and spalled.
A set of crossing diagonal cracks was evident in Specimen 4 as
shown in Figure 3.10, when the column was at a deflection of -1.2 in.
The cracks crossed at the center of the second clamp from the joint. As
deflections were increased, these cracks did not widen. Specimen 4
demonstrated no spalling or major cracking during deflections to 4-Ay
(Figure 3.11). Crushing of the concrete was observed under the clamp
nearest the joint during the maximum deflection cycles (Figure 3.12).

3.3 Moment-Curvature Response

A rough experimental measure of column curvature was determined

by the lateral deflection gages at the column-joint face and at 10 in.
from that face. The difference in the deflection measurements were
divided by the 10-in. distance between gages to yield an average rota-
tion over the 10-in. space which included the maximum column moment
and plastic hinge region. The moment at the joint face was calculated
as the sum of the shear times the distance from the support to the
joint face and the axial force times the lateral deflection at the
joint face. Plots of these calculated moment-curvature relations are
shown in Figure 3.13 through 3.16 for Specimens 1 through 4 respectively.
For Specimens 2, 3 and 4 these moment-curvature diagrams closely
resembled the load-deflection curves, and were typical for reinforced
concrete members dominiated by flexural response. The moment-curvature

38
Figure 3.9 Specimen 3 after test sequence.

Figure 3.10 Specimen 4 at -1.2 in. deflection.

Figure 3.11 Specimen 4 at -2.4 in. deflection.

Figure 3.12 Specimen 4 showing crushing under U-clamp after

deflection sequence.

40
 0
O
0

CO
_

Op
1 0

cJ

—0.20 —0.12 —b.04 0.04 0.12 0.20 0'.29

CURVATURE 1RRD,1

Figure 3.13 Specimen 1, moment-curvature response.

41
0

0
C_

0_
Cpl

op_

CJ

— 0 , 04 -0,02 0.00 0.02 0.04 0,06 0.06

CURVATURE RAD

Figure 3.14 Specimen 2, moment-curvature response.

42
O
O
C_7_1

CD
CD

O
(-A

-0.04 -0.02 0.00 0.02 0.04 0 1 .05 0'.00

CURVATURE (RAD,1
Figure 3.15 Specimen 3, moment-curvature response.

43
 O
O
O

0
O
(.7 _

ID
a_ c)
H
v

2:c3

0
0

Di 1

0
CJ

1 _0 08 -0
1 .06 -004 -b.02 OGO 0•02 004
CU:RVRTURE

Figure 3.16 Specimen 4, moment-curvature response.

44
plot for Specimen 1 shows almost immediate failure at curvatures beyond
the elastic range. The most important observation in the plots for
Specimens 2, 3 and 4 is that the moment continued to increase at curv-
atures between 0.02 radian and 0.04 radian during the first cycle to
the maximum deflection. This increasing moment corresponds to the de-
creasing lateral load at large deflections mentioned in Section 3.1
above. Although the P-A effect was evident, the moment capacity of the
columns was stable and slightly increasing.

3.4 Strain Observations

The majority of strain observations have been omitted from this

report because they shed little insight into the behavior of the strength-
ening techniques. Figures 3.17 through 3.20 show the average strain on
the bottom No. 7 bars where the gages were located 1 in. from the joint.
Only data for the first cycles to Ay and 2-Ay were included for
clarity. Figures 3.21 and 3.22 show the strain on the packaging band
and U-clamps respectively.
For all specimens the maximum recorded tensile strain on the #4
gage wire tie located 1 in. from the joint was about 1000 micro-inch/inch.
This represents a stress less than one-half the yield stress of the
ties. Crack patterns in Specimens 1 and 4 indicated that the strain
in the second tie located 11-in. from the joint would have been much
greater than in tie tie next to the joint. It was the second tie that
unraveled when Specimen 1 failed. Unfortunately, the second tie was
not instrumented.
The band next to the joint in Specimen 2 and the U-clamp next to
the joint in Specimen 4 were instrumented with strain gages. For the
band the maximum tensile strain readings was less than 300 micro-inch/inch
in the deflection cycles to Ay, 2-Ay, and 4-Ay. This low strain rep-
resented a stress less than 9000 psi in the band. For the U-clamp the
maximum tensile strain was less than 30 micro-inch/inch, a stress less
than 1000 psi. These stresses represent forces in the band of about
800 lbs and in the U-clamp of about 600 lbs.
The crack pattern shown in Figure 3.10 indicated that the strain
in the second U-clamp from the joint would have been larger than the

45
cc
•-•

cc

CD

-0.10 -0 .02 0.06 01 .14 0 I . 22 0.30 0.38 0.46

STRR I ( IN./IN. ) x10 -2

Figure 3.17 Specimen 1, average strain on bottom No. 7 bars, 1-in. from joint
for deflection cycles 1 and 4.
 CO

O
C
CO

CO

-0.08 1.04 0.00 0.04 0.08 0.12 0.16 0.20

STIN IN ./IN. x10 -1

Figure 3.18 Specimen 2, average strain on bottom No. 7 bars, 1-in. from joint
for deflection cycles 1, 4, and 7.
o

o
o

CO

1
-0.24 -0.18 - 10.08 0.00 0.08 0.16 0.24 0.32
STRAIN (IN./IN.) *10 -2
Figure 3.19 Specimen 3, average strain on bottom No. 7 bars, 1-in. from joint
for deflection cycles 1, 4, and 7.
D
0

CD _
..--i

C)
0

CO

o
D
Cr
c-N 1 1 1 1 1
1 i
-0.08 0.00 0.08 0.16 0.24 0.32 0.40
STRAIN (IN./IN.) x10 -2

Figure 3.20 Specimen 4, average strain on bottom No. 7 bars, 1-in. from joint
for deflection cycles 1, 4, and 7.
 CC1

CO

co

m
D

I
-0.05 0.00 0.05 0.10 0.15 0.20 0.25 0.30
STRAIN (IN./IN.) x10 -3

Figure 3.21 Specimen 2, strain on packaging band next to joint for deflection
cycles 1, 4, and 7.
 co

0 ).
-0.33 -0.25 - .17 -0.09 -0.01 0.07 0 15 0.23
STRAIN (IN./IN.) x10 -4
Figure 3.22 Specimen 4, strain on U-clamp next to joint for deflection
cycles 1, 4, and 7.
strain in the first U-clamp. Observations of the second U-clamp and
bolt did not indicate any yielding. The second clamp had added flex-
ibility across the crack because of the bending of the outstanding legs.
The bending of these legs reduced the stiffness and effective confining
capacity of the clamp.

3.5 Energy Dissipation

The energy dissipated by each column in each deflection cycle was

determined by measuring the area within the hysteresis loops of the
shear force - deflection curves, Figures 3.1 through 3.4. The cummula-
tive dissipated energy for the four specimens is shown in Figure 3.23.
Specimens 2, 3 and 4 absorbed and dissipated similar amounts. The
differences resulted because the deflection magnitudes at 2-Ay and
4-4y for each specimen were slightly different. Specimen 1 dissipated
an order of magnitude less energy than the strengthened columns.

52
o o #1
A A #2
#3

I I

6. 9 12
Cycles
Figure 3.23 Cummulative dissipated energy.

53
 4. DISCUSSION

4.1 Quantitative Analysis

4.1.1 Moment Capacity

Using standard principles of reinforced concrete analysis
without any capacity reduction factors, the ultimate moments (M u )
of the specimens were calculated and were listed in Table 4.1 together
with the maximum observed moments. The observed maximums were all
greater than the calculated M values. For the strengthened specimens,
u
the average of positive and negative maximum moments was 18 percent
greater than the calculated M . The author believes this significant
u
increase in moment capacity resulted from the confining effect of the
bands or spiral. Not only would the strength of the concrete be in-
creased by biaxial effects, but the ultimate strain of the concrete
would also be increased. The increase in concrete strain capacity
allowed roations which brought the tension steel into the strain
hardening region. The rise in steel stress above the yield value
resulted in the higher moments observed.

Table 4.1 Moment Capacities

Calculated M Observed M (inch-kips)

Specimen u u
(inch - kips)
Positive Negative

1 798 + 800 - 940

2 799 + 820 - 1000
3 798 + 870 - 1030
4 799 + 880 - 1060

4.1.2 Confinement and Shear

The ultimate shear capacity provided by the concrete (V ) was
c
calculated by the provisions in ACI 318-77 (3), and the values are
listed in Table 4.2. Also listed are the calculated ultimate capacities

55
 Table 4,2 Shear and Confinement

Specimen/Technique V V A A A actual
c s V max sh sh sh
average required actual A req'd
sh
(kips) (kips) (kips) (sq. in.) (sq. in.)

1. None 13.5 16.0 - - -

2. packaging 13.6 33.4 15.5 0.18 0.09 0.51

Bands

3. #2 Spiral 13.5 49.8 15.6 0.07 0.05 0.78

4. U-clamp 13.6 104.9 16.5 0.39 0.63 1.61

of shear resistance provided by the various strengthening techniques
(V ) and the average of positive and negative observed shear values
s
(V ). As the author had anticipated from previous research (16),
max
Specimen 1 sustained a shear force 18 percent greater than that cal-
culated using code provisions (3). But once shear failure occurred,
Specimen 1 collapsed.
Table 4.2 shows that all the observed V values were about 17
max
percent greater than tha calculated V value. These V values imply
c max
that the shears on Specimens 2, 3 and 4 were about equal to the ultimate
shear capacity of the concrete.
But Specimens 2, 3 and 4 possessed extensive reserve shear strength
provided by the wrappings. The shear resistance of the wrappings was
not utilized. The strain data indicated less than 9000 psi stress in
the packaging bald and 1000 psi in the U-clamps next to the joint. The
minor diagonal cracks beneath the second band and the second U-clamp
indicated higher strains in those wrappings than in the ones nearest
the joint. But the stress in the first band of Sepcimen 2 would have
been substantial, over 43 ksi*, if the bands were required to carry the
total shear force,V That the stress in the band was less than
max .
21 percent of this value showed that the shear forces were resisted
principally by the concrete, even at the 4-Ay deflection level.
The confinement provided by the strengthening techniques allowed
the concrete to resist shear over the large deflection cycles. In
Specimen 1 as the compression concrete spalled (Figure 3.6), the
shear carried by the compression zone (V , Reference 16) deteriorated
cz
rapidly. With the opening of the diagonal crack, shear transmitted by
aggregate interlock (V a ) decreased. The result was rapid loss of all
shear resistance. In Specimens 2,3 and 4 the wrapping nearest the
joint prevented spalling; the concrete in the compression zone resisted
shear forces. The wrapping away form the joint prevented the diagonal
crack from opening so that V a was maintained.
The quantity of confining ties needed in reinforced concrete

* V for Specimen 2
max = 43
(2 bands)(.18 in 2 /band)(29,000 ksi)

57
columns was given in ACI 318-77 Appendix A (3) as the following
l p s
h s h
A =
sh 2
2
where A = area of transverse hoop bar (one leg), in.
sh
1 = maximum unsupported length of rectangular hoop. In
h
this case 1 was taken as 10 in., the column dimension.
h
s = center-to-center spacing of hoops which was 4.0 in. for
h
the bands, 1.1 in. for the No. 2 spiral, and 4.25 in.
for the U-clamp.
Ag ft
p s = 0.45 (AgA - 1) c
ch

but not less than

f'
0.12 c
f

Ag = gross area of section, which was 100 in

A = area of rectangular core of column measured out-to-out
ch
of hoop which was taken as Ag for Specimens 2, 3 and 4.
Because Ag and A ch were the same for the strengthened columns, the
minimum provisions governed for the amount of externally applied hoops.
The calculated values for A are given in Table 4.2, along with the
sh
values of A actually provided. The value of f used for the pack-
sh
aging bands was that determined by the slip of the bands through
the clips rather than the higher value of the actual yield stress
of the steel (Chapter 2).
The ratio of the Aprovided to the A required by ACI 318-77
sh sh
is shown in the last column of Table 4.2. As discussed in Chapter 2,
the author specifically designed the strengthening techniques to provide
a range of A sh from less than to greater than that required. The
exact ratio was tempered by constructability and judgement of appro-
priate hoop spacing that might be used in actual retrofit practice.
Whether the strengthening techniques provided about 50 percent
(Specimen 2) or 160 percent (Specimen 4) of the A required did not
sh
affect the elastic or inelastic response of the columns. As shown by
the load-deflection hysteresis curves and by the dissipated energy plots,
the responses of Specimens 2, 3 and 4 were almost identical. The author
concluded that the ACI provisions (3) for A required for confinement
sh
were too conservative. Less A may be used to provide confinement as
sh

58
idemonstrated by Specimens 2 and 3.
It must be remembered that these strengthening techniques con-
fined the whole column and not just the core area bordered by the main
reinforcement. Typical hoops used for new construction would confine
only the core area. An explanation for the fact that less than the
required Ash satisfactorily confined the concrete is that by confining
the compression zone the wrappings peiiiiitted V to be effective and
cz
increased the failure strain of the concrete. Confining hoops within
a column only act after the cover has spalled and when the column has
a much reduced section. The area of hoops (A required) are then
sh
designed to provide the strength lost by the spalling (16). The
strengthening techniques do not need to provide for lost material.
Therefore, the requirements for confinement related to exterior
strengthening of existing columns is different than the need for
confinement in new designs.
Furthermore, the rectangular spiral of Specimen 3 utilized plain
No. 2 bar. The current code (3) requires that the minimum size hoop
be a No. 3 bar. The No. 2 spiral perfoLmed well; no bulging was noticed.
For columns larger than 10 in. x 10 in., a larger diameter bar may be
required; but for the small size column tested, the 1/4-in. diameter
bar for the spiral was satisfactory.
From the above the author concluded that the requirements
in ACI 318-77 are not directly applicable to retrofit of exising
structures. Modified requirements for repair and strengthening
are needed.

4.2 Qualitative Analysis

The two most significant qualitative results were the following:

(1) the three strengthening techniques greatly improved the ductility
and cyclic resistance of the existing reinforced concrete columns,
and (2) the various types of strengthening used, even though providing
different amounts of steel (A sh), produced the same ductile type of
structural behavior.
The term ductility used herein means the ability of the column
to sustain axial load and a lateral force through increased lateral
deformations and to absorb and dissipate energy over reversed, in-

59
elastic deflection cycles. The strengthened columns demonstrated
much greater ductility than the unstrengthered column. The response
of Specimens 2, 3 and 4 is that desired for earthquake resistance, for
it sustains lateral loads over inelastic deformations and dissipates
seismic energy without severe structural degradation.
Because the type of strengthening did not affect the ductile
response, the choice of strenghtening technique would depend upon
constructability, ease of application in an occuppied building and
cost. This research did not investigate all possible strengthening
techniques nor the cost of large scale application of the three types
studied. But the author did gain an appreciation for the construction
of each.
The U-clamp technique was the easiest to apply to the column
and would be the cleanest to work with in an occuppied building.
Fabricating the U-clamps in the machine shop required considerable
time, and in actual application this fabrication would be more expen-
sive than the banding technique (Specimen 2). But the shop-time was
compensated for by the short time required in the field bolting
application of the clamps.
The packaging bands were applied easily to Specimen 2, but such
banding would be slightly more difficult to a column in a vertical
orientation. Grouting beneath the bands was the time consuming part
of the construction. Such dry-pack work would be somewhat messy in
an existing structure. From the observations made during the test,
the author believes that the grouting under the bands was necessary
to provide confinement.
The rectangular spiral was the most difficult to apply. The
wrapping and plastering would create more disturbance in an occuppied
structure than created by the other two techniques. Use of larger than
the 1/4-in. diameter rod would make fabricating the rectangular spiral
difficult. If provisions of ACI 318-77 were followed, a No. 3 bar
would be the minimum size required.
Use of any of the techniques seems to provide an economic alternative
to placing No. 3 or No. 4 hoops around the column and casting concrete
or applying shotcrete. Shotcrete and cast-in-place concrete disrupt
the use of an occuppied structure and require extensive clean-up.

60
 After application of any strengthening technique, the column
would be covered with an architectural finish or surrounded with
gypsum board or paneling.

4.3 Limits of Findings

While the findings of this research do point to use of low cost

techniques for strengthening existing columns, the results are quite
limited in scope and must be applied with judgement. The three tech-
niques worked well on the 10-in. x 10-in. column with corner reinforce-
ment. For larger size columns with intermediate reinforcement along
the sides, the external confining system would have to resist greater
bulging forces. The requirement (3) that intermediate bars be restrained
by supplementary crossties could not be satisfied by exterior wrappings
only.
The techniques increased the ductility of the columns and,
thereby, provided a strengthening effect at large deflections and over
repeated loading cycles. But the systems did not increase the maximum
lateral load carried by the columns. The author believes that increas-
ing ductility is primary in improving the earthquake resistance of
existing structures, but increasing the lateral load resistance some-
times is required. These techniques would not accomplish the latter.
By strengthening a column with any of the techniques, the failure
zone of the existing structural system may be shifted from the column
into the beam-cclumn connection. Most existing structures outside
California lack the special stirrup-ties in the joint required for
ductile performance. Beams framing into a joint help confine the
joint (7, 13), and column strengthening may not necessarily lead to
joint failure. But the designer must be cautious and not assume that
by strengthening one structural component he has strengthened the en-
tire system.
Each existing structure is unique, and a retrofit technique for
improving earthquake resistance must be engineered especially for
that structure. Application of the concepts experimentally tested in
this research must be applied with careful judgement for each specific
condition.

61
 5. CONCLUSIONS AND RECOMMENDATIONS

Four identical reinforced concrete columns were constructed, and

three were strengthened using various techniques, packaging bands,
No. 2 bar spiral, and U-clamps. Each technique greatly improved the
ductility and, therefore, the earthquake resistance of the existing
columns. The unstrengthened column collapsed at a deflection ductility
ratio less than two, while the strengthened columns resisted three
cycles of lateral deformations to ducitlity ratios of four with little
deterioration. Although the three techniques provided different areas
of confining steel, the responses of the three strengthened columns
were the same.
Both the packaging bands and No. 2 bar spiral provided significantly
less transverse reinforcement than required by ACI 318-77 (3); yet
the columns behaved satisfactorily. It was concluded that for the
retrofit of existing structures, the building code provisions used
for new construction may be too conservative.
The application of any strengthening technique will depend on its
ease of construction within an occuppied structure. The shop-fabricated
U-clamps were easiest to install on the column and would require the
least disturbance to building occupants. The banding simply was
tightened around the column. Grouting beneath the bands was required
for confinement; such grouting would be time consuming and messy but
would not present excessive difficulties. The rectangular spiral was
most difficult to construct. Both the U-clamp and banding techniques
show significant potential for strengthening of existing columns and
tentatively are recommended.
The use of these techniques should be limited to small columns
like those tested for this research. Further research is necessary
to investigate other parameters such as column size, reinforcing bar
location, axial load, hoop size and spacing, and concrete strength.
Nevertheless, this research has demonstrated that simple, low-cost, easy-
to-construct techniques may be used to greatly increase the ductility
and improve the earthquake resistance of existing reinforced concrete
columns.
63
 APPENDIX
MATERIAL TESTS

by William E. Bynum, III

A.1 Reinforcing Bars

Several methods and combinations of methods were employed to

measure and record strain during tensile tests of reinforcing steel
specimens. The techniques used were bonded electrical resistance
strain gages, inscribed gage marks, an LVDT extensometer with auto-
matic graphing. The No. 7, No. 4 and No. 3 bars were loaded with
Tinius-Olsen Universal Testing Machine, while the No. 2 bars were
tested with an Instron Testing Machine. Material properties for
each reinforcing bar size were computed and are listed in Table A.1.
The yield stress for each bar was found by using the ASTM accepted 0.2
percent offset method.
Two No. 7 bar specimens were strain gaged to create a half
Wheatstone bridge resistance unit. Two gages, diametrically opposed,
were oriented along the axis of the bar at a single location. By
electronically averaging the strains, bending strains were canceled.
The bars also were inscribed with two marks, approximately two inches
apart, near the strain gages. The strain gages responded until a
strain of about .012 inches per inch, at which time the bond failed.
Deformation readings then were taken up to fracture with calipers
positioned between the inscribed marks. This two step procedure is
not recognized by the ASTM, but was adopted for No. 7 bar testing
for two reasons: first, the use of strain gages during elastic be-
havior would give a more accurate determination of the yield point,
and second, the No. 7 bars in the column prototypes were strain gaged,
so the material test specimens were gaged similarly for uniformity.
Figure A.1 is the average stress-strain curve plotted with data from
the two tests. The specimens exhibited ductility and large plastic
deformation prior to rupture which is reflected by the cup-cone type
fracture in Figure A.2. The stress and strain results show that the
No. 7 bars meant specifications for a 615 Grade 40 reinforcing steel.

65
 TABLE A.1

Material Properties for Reinforcing Bars

Bar Size Yield Ultimate Ultimate Modulus of

Stress Stress Strain Elasticity
(ksi) (ksi) (in/in) (ksi)

No. 7** 56.80 92.33 .158 29,200

No. 4*** 64.30 95.00 .157 28.100

No. 3* 63.10 94.09 31,000

No. 2** 77.90 82.38 .022 29,800

Average 1 Test
Average 2 Tests
*** Average 3 Tests

66
80.0

oo

.2 % Offset Line

20.0

.10 .15
STRAIN (INCHES/INCH)

Figure A.1 - Average Stress-Strain Curve for No. 7 Reinforcing Steel Bar
 11

Figure A.2 - Cup-Cone Fracture of No. 7 Reinforcing Bar Specimen

 Three No. 4 bar specimens were tested using a Tinius-Olsen auto-
graphic plotter together with an LVDT extensometer of two-inch gage
length to plot load-strain curves up to a strain of approximately .005
inches per inch. After yielding, readings were taken with calipers
positioned between two marks, initially inscribed eight inches apart.
Both techniques used for recording No. 4 bar strain measurements are
ASTM approved methods. The average stress-strain curve derived from
data of each of the three specimens is given in Figure A.3. The
cup-cone fracture of one of the specimens is shown in Figure A.4. The
stress and strain results show that the No. 4 bars meant specifications
for A 615 Grade 60 reinforcing steel.
Two No. 3 bars were tested using one electrical resistance strain
gage oriented along the length of the bar. Strain was recorded as the
bars were loaded to fracture. Inscribed gage marks were not used. One
of the specimens gave unreasonable results; the modulus of elasticity
was found to be 46,000 ksi, an obvious error. Possible explanations
were unsymmetric tensioning of the bar or load indication errors of
the testing machine. Following ASTM recommendations this specimen
was disregarded, and all material properties for the No. 3 bars were
based on a single specimen and were found to meet ASTM specifications
for A 615 Grade 60 reinforcing steel. The resulting stress-strain
curve is shown in Figure A.S. Excellent ductile behavior was exhibited
as illustrated by its fracture shown in Figure A.6.
The Instron Testing Machine, complete with an ASTM accepted
autographic plotter, was employed to test the two No. 2 , reinforcing
bars. An electrical resistance strain gage extensometer (one-inch
gage length) was clipped to the bar and was electrically connected to
the plotter. The load-elongation graphs produced by the plotter were
averaged and transformed into the stress-strain curve presented in
Figure A.7. Extreme necking occurred in the bars prior to rupture
as shown in Figure A.8. From ASTM specifications, the barE were
found to meet A 615 Grade 40 reinforcing steel.

A.2 Concrete Cylinders

ASTM approved 28 day compressive strength and split tensile

69
100.0 -

80.0

.2 % Offset Line

.04 .6 8 .12 . i6
STRAIN (INCHES/INCH)

Figure A.3 - Average Stress-Strain Curve for No. 4 Reinforcing Steel Bar
Figure A.4 - Cup-Cone Fracture of No. 4 Reinforcing Bar Specimen
 Ultimate Stress

80.0 —

1-1 60.0—

40. 0_ .2 % Offset Line

E-1

20.0—

.005 .010 . 01 5 .020

STRAIN (INCHES/INCH)

Figure A.5 - Average Stress-Strain Curve for No. 3 Reinforcing Steel Bar
Figure A.6 - Cup-Cone Fracture of No. 3 Reinforcing Bar Specimen
 80.0

60.0

1-1
1
% Offset Line
40. 0 I
U) 1
111
E-1
I
1
20.0

1
.01 .02 .03
STRAIN (INCHES/INCH)

Figure A.7 - Average Stress-Strain Curve for No. 2 Reinforcing Tie

Figure A.8 - Cup-Cone Fracture of No. 2 Reinforcing Tie Specimen
strength tests were conducted on fog room cured concrete cylinders.
Field cured cylinders were tested for compressive strength when
the actual column prototypes were tested. All cylinders are ASTM
specified 6 inches in diameter and 12 inches in height.
Fifteen cylinders were cast when the first two column specimens
were poured. On the day after casting, six cylinders were moved to
a fog room for 28 day cure at 100 percent relative humidity and 73 ° F
temperature condition. The remaining cylinders were placed next to
the columns to simulate field cure conditions. Fifteen test cylinders
were cast with the second set of two columns; again six were fog cured
while nine were field cured. Since the castings took place at dif-
ferent times, the following 28 day compressive and split tensile test
procedures were preformed twice so material properties for each pouring
could be determined. The properties are listed in Table A.2.

A.2.1 28 Day Compressive Tests

After 28 days had elapsed, three fog cured cylinders were capped
with liquid sulfur and positioned in a standard cylindrical compress-
ometer. This apparatus, was equipped with a ten-inch gage length and
deflection dial accurate to 1/1000 inch.
Stress and strain values for the first set of three cylinders
were averaged to produce the graph in Figure A.9. The modulus of
elasticity was found using the ASTM recommended equation (5):

E = (S -S )/(E -.00005)
2 1 2

where S and E are stress and strain values at 40 percent of the ul-
2 2
timate stress, while S is the stress corresponding to .00005 inches
1
per inch strain.
Figure A.10 is the average stress-strain curve for the three
cylinders cast for the second set of columns, Number 2 and 4. The
modulus of elasticity was computed as previously described.
All six cylinders fractured to produce the usual cone configura-
tion. The nominal compressive strength of 6000 psi was much higher
than the 4000 psi requested of the commercial ready-mix distributor.

76
 TABLE A.2

Material Properties for Concrete

(Properties are Averages of Three Tests)

Column Compressive Tensile Modulus of

Specimen Strength Strength Elasticity
Numbers (psi) (psi) (psi)

1,3 6,350 490 3,700,000

2,4 6,470 445 3,200,000

77
 6.0-

_5.0 -
STRESS(K SI)

3.0-

2.0 —

1.0 T

1
.0005 .0010 .0015 .0020 .0025
STRAIN (INCHES/INCH)

Figure A.9 - Average Stress-Strain Curve for Concrete Cylinders 1, 2 and 3

 6. 0 -

5.0-

4.0 —
ST RE SS ( K SI )

3. 0 -

2.0 —

1. 0 --

1
.00075 .00150 .00225 .00300
STRAIN (INCHES/INCH)

Figure A.10 - Average Stress-Strain Curve for Concrete Cylinders 4, 5 and 6

A.2.2 Split Tension Tests

Three fog cured cylinders of each set were used to determine

the split tensile strength. The ASTM specified method for testing
split tensile specimens was followed. The average splitting load,
computed from the three tensile test results, was used to calculate
the tensile strength following the ASTM equations:

a = (2P)/(TrdL)
t

where a equals the average tensile strength (psi), P equals the

t
average splitting load (lbs), d equals the cylinder diameter (in)
and L equals the cylinder length (in). This is an approximation
because of local stress conditions at the load lines and the presence
of stresses at right angles to the tension stresses.

A.3 Specimen Wrappings

The reinforced concrete column specimens were strengthened by

wrapping various reinforcement materials around the columns. Samples
of these materials were tested in tension to determine their properties.
These strengthening materials included a cold drawn No. 2 reinforcing
bar, 5/16 x 2 inch flat plates, 5/16 x 7/8 inch coupons cut from steel
angles 3 x 5 x 5/16 and 1/20 x 2 inch packaging steel bands. The
fourth wrapping technique using standard U ties will not be employed
inthe immediate project research. Table A.3 lists material properties
calculated from the specimen tests. The ASTM recommended 0.2 percent
offset method was used ofr locating the yield stresses.

A.3.1 No. 2 Reinforcing Bar

One cold drawn No. 2 reinforcing bar was tensioned by an Instron

Testing Machine. A one-inch gage length extensometer which was elec-
trically attached to an autographic plotter was used to register and
graph the load-elongation curve. This curve was transformed into a
stress-strain curve shown in Figure A.11. The reinforcing bar, which
met ASTM specifications for A 615 Grade 40, exhibited ductility prior
to fracture.

80
 TABLE A.3

Material Properties for Specimen Wrappings

Wrapping Material Yield Ultimate Ultimate Modulus of

Stress Stress Strain Elasticity
(ksi) (ksi) (in/in) (ksi)

No. 2 Cold Drawn

Reinforcing Bar* 66.88 72.34 .058 29,000

2" x 5/16" Flat

Coupon** 42.60 69.00 .225 29,600

Angle Coupon** 34.10 55.00 .432 28,900

Packaging Band** 93.70 117.44 27,600

* Average 1 Test
** Average 2 Tests
 7 0. 0

60.0

50.0

.2 % Offset Line

U)

30 .0

20.0

10. 0

.61 .02 .03 .04 .0 5

STRAIN (INCHES/INCH)

Figure k,11 - Average Stress-Strain Curve for Cold Drawn No. 2 Bar
Figure A.12 - Cup-Cone Fracture of Cold Drawn No. 2 Reinforcing Bar
3.3.2 U-Clamps

The second wrapping technique was composed of two separate steel

components welded together as shown in Figure A.13. These components
were 10 x 2 x 5/16 inch steel strips and 2 1/4 inch wide strips of
3 x 5 x 5/16 steel angle. The angle contained bolt holes in the 3 inch
leg for keeping two U-Clamp assemblies together. Coupons were saw cut
from lengths of the 2 x 5/16 bar and from the angle for determining
properties of these steels. Testing was performed in a Tinius-Olsen
Universal Testing Machine with an LVDT extensometer of two-inch gage
length electrically attached to an autographic plotter. Load-strain
curves were automatically plotted unitl yield strains occurred (approx-
imately .002 inches per inch for flat strips and .015 inches per inch
for angle strips). After yielding, readings were taken with calipers
positioned between two marks, initially inscribed two inches apart.
This technique is an ASTM accepted method for testing sheet type specimens.
Two of these coupon specimens were machined to a width and length
of 3/4 inch and 4 inches repsectively, in the center area of the bar.
The machined specimens met ASTM specifications for a sheet type spec-
imen with a single exception; the gage width was cut slightly larger than
recommended. The load-strain curves, along with data from caliper
measurements, were reduced and the resulting average stress-strain
curve is shown in Figure A.14. The ductile fracture of one of the
specimens, which occurred outside the gage length, is illustrated
in Figure A.15.
In order to obtain coupon specimens of adequate testing length
from the steel angle, one leg of an angle was cut into two 10 inch
strips as shown in Figure A.16. The center portions of these two
strips were machined to a width and length of 11/16 inch and 3 7/16
inches respectively. The specimens satisfied ASTM requirements with
a single exception; the overall specimen width was slightly smaller
than recommended. The average stress-strain curve was plotted with
data from the tests of both specimens and is shown in Figure A.17.

3.3.3 Steel Banding

Two packaging band specimens, which represented the third wrapping

84
Figure A.13 - Welded Components of the Second Strengthening Technique
 .2 % Offset Line

.100 .1 15 0
STRAIN (INCHES/INCH)

Figure A.14 - Average Stress-Strain Curve for 10 x 2 x 5/16 inch Flat Plate
Figure A.15 - Ductile Fracture of a Flat Plate Specimen
 Specimens
3"

5"

Figure A.16 - Locations of the Two Angle Test Specimens

 6o.o i
STRES S( KSI )

.2 % Offset Line

.10 .20 .30 .4o .50

STRAIN (INCHES/INCH)

Figure A.17 - Average Stress-Strain Curve for 3 x 5 x 5/16 inch Angle

technique, were tested in a Tinius-Olsen Unversal Testing Machine and
strain was automatically plotted using the identical LVDT arrangement
previously mentioned. Gage marks were not used. The specimens ex-
hibited classical necking configurations shown in Figure A.18. Yield
lines developed diagonally across the necked area. The rupture is
shown in Figure A.19. The average stress-strain curve, plotted from
the load-strain graphs, is shown in Figure A.20.

90
Figure A.18 - Necking of a Packaging Band Specimen
Figure A.19 - Rupture of a Packaging Band Specimen
 120. _
Ultimate Stress

80.0_

.2 % Offset Line
U)

UZ
40. 0_

.010 .020 .0310

STRAIN (INCHES/INCH)

Figure A.20 - Average Stress-Strain Curve for Packaging Band

 REFERENCES

1. Blume, J.A., Newmark, N.M., and Corning, L.H., Design of Multistory

Reinforced Concrete Buildings for Earthquake Motion, Portland Cement
Association, Chicago, Illinois, 1961.

2. "Building Code Requirements for Reinforced Concrete (ACI 318-63)",

ACI Committee 318, American Concrete Institute, Detroit, Michigan, 1963.

3. "Building Code Requirements for Reinforced Concrete (ACI 318-77)", ACI

Committee 318, American Concrete Institute, Detroit, Michigan, 1977.

4. Bertero, V.V., and Popov, E.P., "Hysteretic Behavior of Ductile Moment-

Resisting Reinforced Concrete Frame Components", Earthquake Engineering
Research Center Report EERC 75-16, University of California, Berkeley,
April, 1975.

5. Celebi, M., and Penzien, J., "Hysteretic Behavior of Epoxy-Repaired

Reinforced Concrete Beams", Earthquake Engineering Research Center
Report No. EERC 73-5, University of California, Berkeley, February, 1973.

6. Freeman, S.A., "Modification of Structure to Satisfy New Seismic

Criteria", Preprints, Sixth World Conference on Earthquake Engineering,
New Delhi, India, 1977, pp 7-91, 7-96.

7. Gulkan, P., "The Inelastic Response of Repaired Reinforced Concrete

Beam-Column Connections", Preprints, Sixth World Conference on
Earthquake Engineering, New Delhi, India, 1977, pp. 7-55, 7-60.

8. Hanson,R.D., and Degenkolb, H.J., The Venezuela Earthquake, July 29,

1967, American Iron and Steel Institute, New York, 1969.

9. Hidalgo, P., and Clough, R.W., "Earthquake Simulator Study of a

Reinforced Concrete Frame", Earthquake Engineering Research Center
Report No. EERC 74,13, University of California, Berkeley, December, 1974.

10. Higashi, Y., and Kokusho, S., "The Strengthening Methods of Existing
Reinforced Concrete Buildings", Proceedings of the U.S. - Japan
Cooperative Research Program in Earthquake Engineering with Emphasis on
the Safety of School Buildings, Honolulu, Hawaii, August, 1975.

11. Ikeda, A., "Load-Deformation Characteristics of Reinforced Concrete

Columns Subjected to Alternating Loading", Report of the Training
Institute for Engineering Teachers, Yokohama National University,
March, 1968.

12. Kajfasz, S., "Concrete Beams with External Reinforcement Bonded by

Clueing, Preliminary Investigation", Proceedings of the RILEM Inter-
national Symposium, Paris, France, September, 1967.

95
13. Lee, D.L.N., "Original and Repaired Reinforced Concrete Beam-Column
Subassemblages Subjected to Earthquake Type Loading", Ph.D. Thesis,
University of Michigan, Ann Arbor, April, 1976.

14. Mahin, S., Bertero, V.V., Atalay, M.B., and Rea, D., "Rate of Loading
Effects on Uncracked and Repaired Reinforced Concrete Members", Earth-
quake Engineering Research Center Report No. EERC 72-9, University
of California, Berkeley, December, 1972.

15. Murphy, L.M., Coordinator, San Fernando, California, Earthquake of

February 9, 1971, Volume 1, National Oceanic and Atmospheric
Administration, U.S. Department of Commerce, Washington, D.C., 1973.

16. "Shear Strength of Reinforced Concrete Members", ASCE-ACI Task

Committee 426, Journal of the Structural Division, ASCE, Vol. 99,
No. ST 6, June 1973, pp 1091-1187.

17. Spracklen, R.W., "Repair of Earthquake Damage at Holy Cross Hospital",

ASCE National Structural Engineering Conference, Preprint 1941, San
Francisco, California, April, 1973.

18. Uniform Building Code, International Conference of Building Officials,

Whittier, California, 1973.

19. Vallenas, J., Bertero V.V., and Popov, E., "Concrete Confined by
Rectangular Hoops and Subjected to Axial Loads",UCB/EERC-77/13,
Earthquake Engineering Research Center, University of California
Berkeley, August, 1977 (NTIS Accession No. PB275165).

20. Wight, J.K., and Sozen, M.A., "Strength Decay of RC Columns Under
Shear Reversals", Journal of the Structural Division, ASCE, Vol. 101,
No. ST5, May, 1975, pp. 1053-1065.

96