Repair of Damaged Aluminum Truss Joints of Highway Overhead

Sign Structures Using FRP

Amir Fam a*, Sarah Witt b and Sami Rizkalla c

a Queen’s University, Kingston, Ontario, Canada K7L 3N6 b Fyfe Co. LLC, San Diego, California, USA 92121

c North Carolina State University, Centennial Campus, Raleigh, NC, USA 27695-7533.

ABSTRACT

An innovative repair technique is introduced for aluminum truss-type highway overhead sign

structures, using fibre-reinforced polymer (FRP) sheets. The welded k-joints are typically subjected

to excessive fatigue-induced cracking under the effects of wind and moving traffic. The repair

technique proposed in this paper utilizes longitudinal FRP layers bonded to the diagonals and

wrapped around the main chord to form alternating v-patterns, followed by additional circumferential

layers for anchorage. Eight tests were conducted on four full-scale specimens. Weld lines at the

junction between diagonals and main chord were ground to simulate a 90 percent loss of joint

strength. After repair, diagonals were loaded to failure in tension. The study showed that full strength

of the welded joints was restored using carbon-FRP sheets. Only 70 percent of joint strength was

restored when using glass-FRP. The strengthening technique is particularly sensitive to quality

control during installation. A field application using the proposed technique was successfully

completed by the New York State Department of Transportation for a cracked aluminum truss over

Route 88 in NY State.

Keywords: Aluminum, joint, crack, FRP, repair

______________________________________________________________________________ * Corresponding author. Tel: 613-533-6352; Fax: 613-533-2128; E-mail: fam@civil.queensu.ca

1

1. INTRODUCTION

The use of fibre reinforced polymers (FRP) in restoring the strength of damaged or deteriorated

concrete and masonry structures has been proven to be quite successful [1, 2]. On the other hand, for

metallic structures, including both steel and aluminum, a recent state-of-the-art study [3] has pointed

out that the research on using FRP materials for retrofit of metallic structures is still limited and field

applications have not yet gained a wide potential.

One of the promising retrofit applications of metallic structures, where FRP would be quite

suitable, is the repair of truss-type highway overhead sign structures, shown in Fig. 1(a). These

aluminum overhead sign structures have been in service for the past five decades. A large number of

these structures suffer from fatigue cracking of the welded joints between the diagonal and chord

members due to vibration under wind loading and the effect of moving traffic. The cracking in the

weld could propagate and cause complete failure of the connection as shown in Fig. 1(b). The

problem in these situations is that field welding of aluminum is not allowed because of the

difficulties associated with providing the necessary quality control conditions in the field. On the

other hand, replacing the whole structure could be very costly. The use of FRP flexible sheets to

repair the cracked joints could be a quite promising technique. In addition to their ease of

installation, light weight and durability, the FRP sheets offer additional unique advantage in this case.

The diagonal and chord members are likely to have circular cross sections of different diameters and

are also connected at an angle, which result in a complex and quite irregular surface geometry at the

joint. Therefore, FRP sheets can be effectively used in this situation as they conform quite well to

such a surface.

Care should be taken when bonding carbon-FRP (CFRP) materials to metals, especially

aluminum. Although the CFRP material itself is non-corrosive, as the carbon fibers become in

contact with aluminum, a galvanic corrosion process may be generated. The three requirements

necessary for galvanic corrosion to occur are: electrolyte (such as salt water) bridging between the

2

two materials, electrical connection between the two materials, and a sustained cathodic reaction on

the carbon. Karbhari and Shully [4] suggested that, using a hybrid of glass and carbon, where the

glass fabric is sandwiched between the steel or aluminum surface and the subsequent CFRP layers

adds durability to the system.

2. OBJECTIVES

This study is intended to demonstrate the effectiveness of a proposed repair technique for damaged

truss-type aluminum highway overhead sign structures using FRP wraps through laboratory tests and

actual field application. The welded joints between the diagonal members and the main chords of this

type of structures suffer excessive fatigue-induced cracking, yet welding of these cracked joints in

the field is not feasible. The goal of this research program is to restore the strength of cracked joints,

using FRP sheets. The evaluation is based on testing the joint under axial tension load applied to the

diagonal member of the truss. The cracking in the welded joint has been simulated in the laboratory

by grinding the weld line between the diagonal member and the main chord to simulate a 90 percent

loss of joint strength. Both GFRP and CFRP wraps were investigated individually. While CFRP

wraps were installed directly on the aluminum surface in this short term study, it is recommended to

use a layer of GFRP as a barrier between the aluminum and the CFRP for field applications, where

galvanic corrosion could take place over time.

3. EXPERIMENTAL PROGRAM

The experimental program included eight tests conducted using four different specimens tested at the

Constructed Facilities Laboratory (CFL) of North Carolina State University [5]. Each specimen

consisted of a k-shape portion of the aluminum truss, including two diagonal members connected to

one horizontal chord. For each specimen, two tests were conducted by loading each of the diagonal

members independently in tension up to failure.

3

3.1 Configuration of the Test Specimens

Four full scale k-shape aluminum joints, each including two diagonal members connected to one

chord member, were taken from an actual structure and used in this experimental program. The joints

are referred to as G1, C1, C2 and C3. Specimen G1 was repaired using GFRP sheets whereas

specimens C1, C2 and C3 were repaired using CFRP sheets. Each specimen consisted of two

diagonal members welded to the main chord and oriented at 45 degrees with respect to the main

chord. The two diagonals consisted of round aluminum tubing of 50.8 mm outer diameter and 4.8

mm wall thickness. The main chord also consisted of round tubing of 101.8 mm outer diameter and

8.1 mm wall thickness. The tubing material was aluminum alloy ASTM 6061-T6 with tensile yield

strength of 240 MPa, ultimate tensile strength of 290 MPa and modulus of elasticity of 69.6 GPa [6]

as shown in Table 1(a). The strength of the diagonal members is normally governed by the strength

of the welded joint. The diagonals were connected to the main chord through shop welding using an

inert gas arc-welding process. The allowable tensile stress in the diagonal near the weld is 82.7 MPa

[6].

3.2 Surface Preparation and FRP Installation

In order to simulate the damage of the welded joint, which is typically induced by fatigue cracking,

the weld between the diagonal members and the main chord was completely ground along the

perimeter of the joint, except for two spots on opposite sides, 10 mm long each, left intact in order to

hold the diagonal member in place in the appropriate configuration. This represents about 90 percent

loss of the capacity of the joint. Fig. 2 shows the grinding process and the reduction of the weld

length. A wire brush was then used to roughen the surfaces of the diagonals and main chord in the

area surrounding the joint, where the FRP was to be applied.

Two types of FRP sheets were used in this experimental program, namely GFRP and CFRP

(commercially known as Tyfo SEH-51 and Tyfo SCH-35 respectively), both utilizing epoxy resin for

4

bonding. The GFRP was used to repair specimen G1, whereas CFRP was used to repair specimens

C1, C2 and C3. The GFRP laminate has tensile strength and modulus of 575 MPa and 20.7 GPa

respectively and a thickness of 1.3 mm as shown in Table 1(b), while the CFRP laminate has tensile

strength and modulus of 991 MPa and 78.6 GPa respectively and a thickness of 0.89 mm as shown in

Table 1(c).

The repair scheme and process of installation was identical for all specimens. The surface of

the aluminum tubes was wetted with epoxy and left till the epoxy became tacky. The dry glass (or

carbon) fabrics were saturated with epoxy before being applied to the aluminum joints. Six FRP

layers, 840x100 mm each, were wrapped around the joint in a v-shape pattern in an alternating

fashion such that the fibers are parallel to the diagonal members as shown in Fig. 3(a). Those layers

were notched 13 mm from both sides, at 305 mm from each end as also shown in Fig. 3(a), in order

for the sheets to conform to the curved surface at the joint, where the diameter of the diagonal

member is different from that of the main chord. Four additional 510x100 mm layers were wrapped

transversely around the diagonal members, at two locations each, to confine and anchor the

longitudinal FRP layers. One additional 1020x100 mm layer was also wrapped transversely around

the main chord at the joint as shown in Fig. 3(b). The surface was finally finished using a thick paste

of silica fume mixed with epoxy and left to cure in room temperature.

3.3 Test Setup and Procedure

The main chord of the aluminum truss specimen was anchored to the strong floor of the laboratory as

shown in Fig. 4(a). This was achieved using a solid steel shaft inserted inside the main chord and

held down at each end using a stiff hollow square section (HSS). Each HSS was anchored to the

strong floor using two anchor rods as shown in Fig. 4(c). One diagonal member of the specimen was

connected to a 980 kN MTS hydraulic actuator. The actuator was anchored to the vertical reaction

wall and positioned at 45 degrees angle to line up with the diagonal member of the specimen as

5

shown in Fig. 4(b). The actuator included a hinged connection at each end in order to insure axial

tension loading in the diagonal members.

In order to facilitate the connection between the diagonal members and the hydraulic

actuator, aluminum end plates, were attached at the ends of the diagonals. Three types of connections

were used, “A”, “B” and “C” as shown in Fig. 4(a). In connection “A”, the diagonal member was

directly welded to a square aluminum plate. This connection was successfully used for specimen G1,

however, it was inadequate for the specimens retrofitted with CFRP, due to the higher capacity of the

specimens. In connection “B”, which was used for specimen C1, the end of the diagonal member was

inserted into aluminum ring and welded from both sides. The ring was then positioned to bear on the

square end plate through a grove cut into the plate as shown in Fig. 4(a) “Joint B”. One diagonal

member of specimen C1 failed within the tube, while the other failed within the weld of connection

“B”. It was, therefore, decided to use a stronger connection “C” for specimens C2 and C3.

Connection “C” was similar to “A”, except that four additional gusset plates were welded between

the end square plate and the diagonal member as shown in Fig. 4(a), “Joint C”.

Axial tension load was applied using stroke control at a rate of loading ranging between 0.25

to 0.75 mm/min, until failure occurred. The specimen was then rotated 180 degrees and the second

diagonal member was also tested to failure. The axial tension load and axial displacement at the

loading end of the diagonal were continuously monitored. In specimen G1, the longitudinal strain in

the diagonal member was also measured in the aluminum using electric resistance strain gages.

4. TEST RESULTS AND FAILURE MODES

Specimens C1 and C3 (repaired with CFRP) failed by fracture of the aluminum tube, close to the

loading end, near the weld, and hence, provided an experimental value of 114 kN for the axial tensile

strength of the diagonal member. This value is based on the average strength of the two diagonals of

each of specimens C1 and C3 of 115, 105, 125, 112, and 125 kN respectively. It should be noted that

6

this load corresponds to an induced stress of 164 MPa, which is double the maximum allowable

stress of 82.7 MPa specified by the Aluminum Association Inc. [6], but also lower than the 290 MPa

ultimate strength of aluminum, mainly due to the welding effect. Since the welded length at the

retrofitted joint was reduced by 90 percent, the reduced strength of the member (after simulating the

fatigue cracking) was predicted as 11.4 kN. Summary of the test results of all specimens is provided

in Table 2.

The axial load – displacement behavior of specimen G1 (repaired with GFRP) is shown in

Fig. 5. Both diagonals failed by fracture of the GFRP wraps at the joint as also shown in Fig. 5. The

failure loads for both diagonals were 78.7 and 78.3 kN and the displacements at failure loads were

6.5 and 6.9 mm respectively. These loads are about 31 percent lower than the target strength of 114

kN (strength of a welded aluminum tube), which indicates that the GFRP repair scheme was

inadequate to recover the full strength of the diagonal member. It, however, increased the strength of

the artificially damaged specimen by about 6.9 times. The load – strain behavior of Diagonal 1 is

shown in Fig. 6. The strain was measured on the aluminum tube outside the FRP repair system. It is

noted that failure occurred before yielding of the aluminum as indicated by the linear response.

Based on the ultimate strain of 0.00177 and the elastic modulus of 69.6 GPa, the stress level

developed in the diagonal member at failure is 123 MPa only, which is relatively comparable with

the stress based on the measured ultimate loads (113 MPa).

The load – displacement behavior of specimen C1 (repaired with CFRP) is shown in Fig. 7.

As indicated earlier, both diagonals failed at the loading end within the aluminum tubes, near or at

the weld as also shown in Fig. 7. The ultimate loads and displacements were 115 and 105 kN and

11.4 and 11.8 mm respectively. The CFRP repair system was fully intact at failure and the full

strength of the aluminum diagonal was achieved. The strength of the artificially damaged specimen

was increased 9.6 times.

7

Although specimen C2 (repaired with CFRP) was quite similar to specimen C1, both

diagonals of C2 failed by fracture of the CFRP wraps at the repaired joint as shown in Fig. 8. The

ultimate loads and displacements of both diagonals were 102 and 104 kN and 14.5 and 19.4 mm

respectively as also shown in Fig. 8. This indicates a 7 percent lower strength and 46 percent higher

displacement in comparison with C1. This is, in fact, attributed to the slightly lower level of quality

control encountered in specimen C2. It is believed that the FRP sheets were not wrapped quite tight

around the joint, which led to some slack. This was reflected in the higher axial displacements

measured in C2 compared to C1. The strength of the artificially damaged joint was however

increased 9 times. While flexible FRP sheets are ideal for conforming to curved surfaces in general,

it should be noted that the geometry of the curved surfaces at the joint of the aluminum truss in this

case is quite complex due to the difference in diameter and the angle between the diagonal and main

chord. As such, quality control is crucial for such application in order to produce consistent results,

especially when the FRP is hand applied. Following the testing of specimen C2, it was decided to

fabricate another identical specimen, C3, insuring a better quality control and conduct another testing

to confirm the results of specimen C1.

The load – displacement behavior of specimen C3 (repaired with CFRP) is shown in Fig. 9.

Both diagonals failed at the loading end within the aluminum tubes, near the weld as also shown in

Fig. 9. The ultimate loads and displacements were 115 and 125 kN and 14.1 and 15.1 mm

respectively. Similar to specimen C1, the CFRP repair system was fully intact at failure and the full

strength of the aluminum diagonal was achieved. The strength of the artificially damaged joint was

increased 10.5 times. This test has provided a rather consistent result with specimen C1 in terms of

failure mode and a slightly higher strength.

It should be noted that in all the eight tests conducted on the FRP repaired joints, no slip was

observed between the FRP repair scheme and the diagonal aluminum tube, which suggests sufficient

8

bond strength. In all tests, failure occurred by fracture of the aluminum tube, out side the FRP repair

patch or by fracture of the FRP wraps at the joint.

5. FIELD APPLICATION

The New York State Department of Transportation (NYS DOT) has found in a recent study that 10

percent of the State’s overhead sign structures had some form of structural damage. The most

common damage found was cracking of the joint between internal diagonal trussing and the main

chord. The NYS DOT was approached with the proposed innovative repair method described in this

paper to repair the cracked truss joints. NYS DOT joined partners with the State of Utah DOT to

research this solution to examine the feasibility as well as time and cost savings. After conducting the

laboratory tests, it was concluded that the repair using this technique is promising and can restore the

strength of a fully welded connection.

The NYS DOT decided to conduct a field application using the proposed technique on a

cracked aluminum overhead sign truss located over Route 88 (Westbound Direction) in NY State,

about half a mile east of Exit 2. A Contract was awarded to EMI Guiderail. Tyfo Fibrwrap System

was used in the repair of the cracked structure. Representatives from EMI Guiderail, Inc., R. J.

Watson, Inc. (Local distributor of Tyfo Fibrwrap System), NYS DOT (Materials), and NYS DOT

Region 9 were present during this unique retrofit application on September 10, 2003. The weather

conditions were quite suitable for the installation. Access to the aluminum sign post was provided by

EMI Guiderail, using a man-lift. Similar to the steps followed in the laboratory testing, the two-

component epoxy matrix was mixed for approximately 5 minutes prior to saturating the dry fabric.

The aluminum sign structure was prepared by grinding the surface area where the FRP was to be

applied. A prime coat of epoxy was applied to the areas that have been prepared and left until the

epoxy became tacky. The dry fabric strips were saturated with epoxy and applied in a pattern similar

to that shown in Fig. 3. The repair system was then left to cure.

9

6. SUMMARY AND CONCLUSIONS

An innovative retrofit technique using FRP sheets has been introduced for repair of truss-type

aluminum highway overhead sign structures comprised of welded tubular members. The welded K-

joints between diagonal members and main chords are typically cracked due to fatigue loading. The

repair system utilizes narrow FRP longitudinal strips wrapped around the K-joint and attached to the

diagonal members in an alternating fashion. The longitudinal layers are then wrapped with additional

layers in the circumferential direction for anchorage. Both GFRP and CFRP materials have been

investigated. The cracking at the joints was simulated by grinding the welded perimeter at the

intersection between diagonals and main chord to simulate a 90 percent loss of weld.

Based on eight tests conducted on four specimens, it is concluded that the full strength of the

welded joint can indeed be restored when CFRP layers are applied using the proposed repair

technique along with adequate level of quality control. When similar number of layers of GFRP

were used, only 79 percent of the capacity of the joints was restored, compared to CFRP. While

using additional layers of GFRP could have achieved the desired target strength, this was not

investigated in this study. In all tests conducted, failure was either due to rupture of the FRP sheets

at the joints or fracture of the aluminum tube near or at the weld at the loading end. In no case bond

failure was observed between the FRP wraps and the aluminum tubes. It is also noted that failure of

one diagonal at the repaired joint has insignificant effect on the strength of the second diagonal

attached to the same joint, when tested later.

Based on the test results, the New York State Department of Transportation has decided to

implement this technology in the field. On September 2003, aluminum truss overhead sign structure

with cracked joints has been successfully repaired using the FRP repair scheme and procedure

described in this paper as shown in Fig. 10. The structure is located over Route 88 (Westbound

Direction), about half a mile east of Exit 2. To date, no signs of unsatisfactory performance have

been reported.

10

7. ACKNOWLEDGEMENTS

The authors wish to acknowledge Fyfe Co. LLC for supplying the FRP materials, New York State

Department of Transportation for providing the test specimens, Mr. Jerry Atkinson for his assistance

during the experimental work, and the Constructed Facilities Laboratory of NC State University.

8. REFERENCES

[1] Fam A, Musiker D, Kowalsky M, Rizkalla S. In plane testing of damaged masonry wall repaired

with FRP. Advanced Composites Letters Journal 2002; 11(6):277-283.

[2] Rizkalla S, Hassan T. Design recommendations for the use of FRP for reinforcement and

strengthening of concrete structures. Journal of Progress in Structural Engineering and Materials

2003; 5(1):16-28.

[3] Shaat A, Schnerch D, Fam A, Rizkalla S. Retrofit of steel structures using fiber-reinforced

ploymers (FRP): state-of-the-art. Transportation Research Board (TRB) annual meeting,

Washington, D.C., Jan. 2004, CD-ROM (04-4063).

[4] Karbhari VM, Shulley SB. Use of composites for rehabilitation of steel structures – determination

of bond durability. Journal of Materials in Civil Engineering, ASCE 1995; 7(4):239-245.

[5] Fam A and Rizkalla S. Experimental testing of aluminum truss joints repaired using FRP sheets.

North Carolina State University-CFL, Report submitted to Fyfe Co. LLC 2001.

[6] Aluminum Association Inc. Section 1: Specifications for Aluminum Structures, Aluminum

construction manual, 5th Edition, Washington DC 1986.

[7] State of New York Department of Transportation (NYSDOT) Region 10 and Office of

Engineering Structures Design and Construction Division. Overhead Sign Structures Inventory and

Inspection Manual 1999 p.42.

11

List of Tables

Table 1 Mechanical properties of the aluminum tubes and FRP wraps

Table 2 Summary of test results

12

13

List of Figures

Fig. 1 Typical aluminum truss overhead sign structure and associated problem

Fig. 2 Reducing the length of weld at the joint through grinding

Fig. 3 Installation of FRP repair scheme

Fig. 4 Details of test setup

Fig. 5 Load-displacement behavior and failure mode of specimen G1

Fig. 6 Load-axial strain behavior of Diagonal 1 of specimen G1

Fig. 7 Load-displacement behavior and failure mode of specimen C1

Fig. 8 Load-displacement behavior and failure mode of specimen C2

Fig. 9 Load-displacement behavior and failure mode of specimen C3

Fig. 10 Field repair of aluminum truss over Route 88 in New York State using CFRP sheets

Table 1 Mechanical properties of the aluminum tubes and FRP wraps

(a) Aluminum alloy ASTM 6061-T6 PropertyTensile yield strengthUltimate tensile strength

Compressive yield strength Modulus of elasticity

Value240 MPa290 MPa

240 MPa

69.6 GPa

Allowable tensile stress 82.7 MPa

(b) GFRP Tyfo SEH-51PropertyLongitudinal tensile strengthLongitudinal modulus

Transverse tensile strength Laminate thickness

Value575 MPa26 GPa

20.7 MPa

1.3 mm

Elongation at break 2.2%

(c) CFRP Tyfo SCH-35PropertyLongitudinal tensile strengthLongitudinal modulus

Transverse tensile strength Laminate thickness

Value991 MPa78.6 GPa

0 MPa

0.89 mm

Elongation at break 1.26%

Table 2 Summary of test results

SpecimenI.D.

G1

TestNo.

Ultimate Load (kN)

Displacement at ultimate (mm)

Strength as % age ofDamaged specimen

Intactspecimen

1 78.7 6.5 690 69.02 78.3 6.9 687 68.7

C1 1 115 11.4 1009 100.92 105 11.8 921 92.1

C2 1 102 14.5 895 89.52 104 19.4 912 91.2

C3 1 115 14.1 1009 100.92 125 15.1 1096 109.6

(a) Truss-type overhead sign structure

(b) Damaged joint - adopted from reference [7]

Fractured weld

Fig. 1 Typical aluminum truss overhead sign structure and associated problem

Fig. 2 Reducing the length of weld at the joint through grinding

10 mm

Bottom chord

10 mm Diagonal

(a) Applying longitudinal v-shape FRP layers

Fig. 3 Installation of FRP repair scheme

(b) Complete repair after applying transverse FRP layers

305 mm

13 mm deep

840 mm

100 mm

305 mm

1020 mm

100 mm

510 mm

100 mm

Fig. 4 Details of test setup

Hinge

Hinge

(b) General view of test setup (c) Close-up of test setup

Aluminum tube

Square plate

Ring

4 gusset plates

Aluminum tube

Square plate

Joint B

Joint C

Hydraulic actuator

Load cell

Hinge

HSS

Test specimen

Strong floor

Anchor bars

Solid steel shaft

MTS

FRPwraps

(a) Schematic view of test setup and connections

weld

Aluminum tube

Square plate

Joint A

Fig. 5 Load-displacement behavior and failure mode of specimen G1

0

20

40

60

80

100

120

0 5 10 15 20Displacement (mm)

Load

(kN

)

Failure of Diagonal 1 by fracture of GFRP

Failure of Diagonal 2 by fracture of GFRP

Full strength of the welded diagonal = 114 kN

Reduced strength of diagonal = 11.4 kN

Diagonal 2

Diagonal 1

0

5000

10000

15000

20000

0 500 1000 1500 2000

22

44

66

88

0

Load

(kN

)

Strain x 10-6

Strain gage

Fig. 6 Load-axial strain behavior of Diagonal 1 of specimen G1

0

20

40

60

80

100

120

0 5 10 15 20Failure of Diagonal 1 in

the weld at the loading endFailure of Diagonal 2 by

fracture of aluminum tubeDisplacement (mm)

Load

(kN

)

Full strength of the welded diagonal = 114 kN

Reduced strength of diagonal = 11.4 kN

Diagonal 2

Diagonal 1

Fig. 7 Load-displacement behavior and failure mode of specimen C1

0

20

40

60

80

100

120

0 5 10 15 20

Load

(kN

)

Displacement (mm)

Failure of Diagonal 1 by fracture of CFRP

Failure of Diagonal 2 by fracture of CFRP

Full strength of the welded diagonal = 114 kN

Reduced strength of diagonal = 11.4 kN

Diagonal 1

Diagonal 2

Fig. 8 Load-displacement behavior and failure mode of specimen C2

0

20

40

60

80

100

120

0 5 10 15 20Failure of Diagonal 1 by

fracture of aluminum tubeFailure of Diagonal 2 by

fracture of aluminum tube

Load

(kN

)

Displacement (mm)

Full strength of the welded diagonal = 114 kN

Diagonal 1

Reduced strength of diagonal = 11.4 kN

Diagonal 2

Fig. 9 Load-displacement behavior and failure mode of specimen C3

(a) Aluminum overhead truss over Route 88

Fig. 10 Field repair of aluminum truss over Route 88 in New York State using CFRP sheets

(b) Cracked joint shown after surface preparation

(c) Application of epoxy coating to the surface (d) Application of CFRP wetted sheets

(e) Sheets applied in longitudinal and transverse directions (f) Repaired structure