Proceedings of the International Symposium on Bond Behaviour of FRP in Structures (BBFS 2005) Chen and Teng (eds)

© 2005 International Institute for FRP in Construction

531

LONG TERM PROPERTIES OF BOND BETWEEN CONCRETE AND FRP

C. Mazzotti and M. Savoia 1 1 DISTART – Structural Engineering – University of Bologna – Italy

ABSTRACT Preliminary results of an experimental study on FRP plates bonded to concrete and subject to long – term loading are presented. Tests are performed in a climatic room, and constant load during time is assured by adopting a specifically designed test frame. Evolution with time of axial strains in FRP plates is measured. It is shown that a shear stress redistribution along the bonded length occurs, mainly due to increased compliance of interface law due to creep deformation. KEYWORDS FRP, concrete, interfacial stresses, creep, experimental behaviour. INTRODUCTION Durability with time of FRP strengthening interventions on Reinforced Concrete structures strongly depends on interface behaviour between concrete and FRP. Delamination is in fact the most important failure mode governing design criteria against end debonding, bending and shear (Teng et al, 2002, 2003). Several studies can be found on FRP – concrete bonding problem, from experimental, numerical or design points of view (see, for instance, Miller et al 2001; Savoia et al 2003; Lu et al 2005; Ferracuti et al 2005). Effectiveness of strengthening can be remarkably reduced due to rheological properties of materials involved (concrete, adhesive, FRP). An important role is also played by the temperature variation. In this framework, only few information can be found in literature concerning the time behaviour of strengthened civil structures (Plevris and Triantafillou 1994; Savoia et al. 2005). In the present paper, first results of an experimental campaign concerning long-term behaviour of bond between concrete prisms and FRP plates are presented. Three different bonded lengths (from 100 mm to 400 mm) have been adopted. Seven to eleven strain gauges (depending on bonded lengths) along FRP plates have been used to measure longitudinal strains. A mechanical system able to apply a traction force constant in time to the extremity of plates has been designed. In order to eliminate strain thermal drift, long-term tests have been carried out in climatic room with standard ambient conditions. Strains have been measured during time by using an automatic control system. Strain profile evolutions with time along the bonded length have been recorded. At the moment, time duration of tests is about 6 months. It is shown that a significant redistribution of shear stresses along the anchorage occurs due to creep deformations at the interface level. A set of short-term delamination tests on identical specimens has also been performed and considered as reference tests. At the end of tests (loading time about 1 year), specimens subject to long – term loading will be loaded up to failure in identical conditions as reference specimens, in order to estimate if a long – term loading reduces bond strength of FRP anchorage. GEOMETRY AND MECHANICAL PROPERTIES OF SPECIMENS Specimen preparation Long-term behaviour of CFRP plates bonded to concrete surfaces has been investigated by considering three bonded lengths (100, 200 and 400 mm) and two concrete specimens (two anchorages for each specimen). Concrete specimen dimensions were 150×200×600 mm. They were fabricated using normal strength concrete. Concrete was poured into wooden forms, externally vibrated. The top was steel – troweled. Three 15 cm – diameter by 30 cm – height standard cylinders were also poured and used to evaluate mechanical properties of

532

concrete, according to Italian standards (see UNI 1976, 2002). Specimens were demoulded after 24 hours and covered with saturated clothes for 28 days; subsequently, they were stored at room temperature and uncontrolled humidity inside the laboratory. Mean compressive strength fcm = 52.6 MPa from compression tests and mean tensile strength fctm = 3.81 MPa from indirect traction tests have been obtained. Mean value of elastic modulus Ecm = 30700 MPa and Poisson ratio ν = 0.227 have been obtained from preliminary tests. Concrete specimens were sufficiently old before tests (two years) so that shrinkage strains can be neglected. For composite plates, CFRP Sika CarboDur S plates, 80 mm wide and 1.2 mm thick, have been used. According to technical data provided by the producer, plates have carbon fiber volumetric content equal to 70 percent and epoxy matrix. Minimum tensile strength is 2200 MPa and mean elastic modulus is Ep = 165000 MPa. Two opposite surfaces of each concrete block have been grinded with a stone wheel to remove the top layer of mortar, until the aggregate was visible (approximately 1 mm). Plates have been bonded to surfaces by using a 1.5 mm thick layer of two – components Sikadur – 30 epoxy adhesive, having mean compressive strength of 95 MPa and mean elastic modulus Ea = 12800 MPa, according to producer data. No primer before bonding has been used. Two CFRP plates have been bonded to opposite faces of each specimen, adopting different bonded lengths (Block P1: B.L. = 100, 200 mm; Block P2: B.L. = 200, 400 mm), see layout reported in Figure 1a. Curing periods of anchorages of blocks P1 and P2 were, respectively, 15 and 20 days prior to testing. The bond surface Two different positions of bond surface on the concrete specimen were considered in previous experimental tests and FE numerical investigations (Mazzotti et al. 2004), i.e., starting from the loaded end of the specimen or at a given distance from it. These studies showed that if bonding of CFRP plate starts close to the front side of concrete specimen, very high tensile stresses are present in this concrete portion and, typically, an early failure occurs in delamination tests due to concrete cracking of a prism with triangular section. On the contrary, when plate bonded length starts far from the front side, tensile stresses are much smaller and FRP – concrete interface is subject to prevailingly shearing stresses up to delamination failure. Behaviour of the interface is not affected by boundary effects and this test set-up is than more appropriate to obtain data for calibration of interface laws. In the present experimental investigation, plate bonded length starts 100 mm far from the loaded end of specimen (Figure 1a). EXPERIMENTAL SETUP AND INSTRUMENTATION Experimental setup is depicted in Figure 1a. Concrete block was positioned on a rigid frame with a front side reaction element (60 mm thick) in order to prevent longitudinal displacements. Free ends of plates were mechanically clamped within two steel plate systems. A double hinge system allowed for rotations around transverse axes. Traction force was applied to steel plates by using a mechanical frame allowing for magnification of applied force at the opposite side of horizontal arm. Force was given as weight of a number of

(a) (b)

Figure 1: Geometry of specimens and experimental setup for long – term tests.

150

200 CFRP plate

80×1.2 mm Adhesive 50×1.5 mm

7 to 11 strain gauges

CONSTANT LOADS

L1 L2 L9 … 100 mm

7 to 11 strain gauges

533

steel plates (Figure 1b). Amplification factor of mechanical system is about 4. Tests were performed by prescribing a constant value of applied force (about 35% of delamination force, Mazzotti et al. 2004, 2005). Along CFRP plates bonded to concrete, series of seven – to – eleven strain gauges (depending on the plate length) were placed, in the centerline. For each bonded length, spacing between strain gauges are reported in Table 1.

Table 1: Distances (mm) between strain gauges along the FRP plate, for different bonded lengths (B.L.). B.L. L1 L2 L3 L4 L5 L6 L7 L8 L9 L10 L11 100 10 10 10 15 15 15 15 200 10 10 10 20 20 20 30 30 40 400 10 10 20 20 40 40 50 50 50 50 50

RESULTS OF EXPERIMENTAL TESTS Instantaneous loadings Tests have been carried out with a loading phase up to 12.30 kN of traction force for each anchorage (about 35 percent of maximum transmissible load, according to delamination tests on analogous specimens), followed by six months at constant loading. Longitudinal strains along the plate at different loading levels and different time intervals have been recorded by an automatic computer system; behaviour of bonded plates during loading phase has been compared with analogous results previously obtained by the authors adopting analogous specimens and a standard setup for delamination tests (Mazzotti et al. 2005). Recorded strains along FRP plates for 200 and 400 mm bonded length are reported in Figures 2a, b. In both cases, strain values have been reported for the first 160 mm only; longitudinal plate strains recorded beyond that distance from the force application, in fact, are almost zero. Strains at x = 0 are calculated from the value of applied force as ε0 = F / Ep Ap. Good matching between experimental data previously obtained by using a standard delamination set–up and results from the present study is found (see Figures a, b), so assessing the reliability of symmetric double plate system (Figure 1a) adopted in this experimental campaign. Long term behaviour Figures 3a – d show time evolution of longitudinal strains along the plates for 100 – 200 mm (specimen P1) and 200 – 400 mm (P2) bonded lengths. For each strain gauge, distance from the beginning of bonded part is reported in parenthesis. Time is reported adopting a logarithmic scale: behaviour after small time intervals can then be observed. Furthermore, for specimen P2, strains during loading increase are also reported (Figures 3c, d). It is worth noting that, due to stability of climatic condition and having adopted an automatic data acquisition system during the whole test, smooth curves have been obtained, even if strain variation of various gauges was small. In order to verify the reliability of long-term loading system and repeatability of experimental results, two identical plates with 200 mm bonded length have been considered in the experimental campaign (one plate for

0

100

200

300

400

500

600

700

800

0 20 40 60 80 100 120 140 160

Distance (mm)

Long

itudi

nal s

trai

n ( µ

ε)

4.00 kN

4 kN - Mazzotti et al 2004

8.00 kN

8 kN - Mazzotti et al 2004

12.30 kN

12 kN - Mazzotti et al 2004

0

100

200

300

400

500

600

700

800

0 20 40 60 80 100 120 140 160

Distance (mm)

Long

itudi

nal s

trai

n ( µ

ε)

4 kN

4 kN - Mazzotti et al 2004

8 kN

8 kN - Mazzotti et al 2004

12.3 kN

12 kN - Mazzotti et al 2004

(a) (b)

Figure 2: Instantaneous loading: strains in FRP plates along the bonded length at different loading levels, comparison between present and previous experimental results (from Mazzotti et al. 2005); (a) B.L.=200 mm, (b)

B.L. = 400 mm.

534

each specimen). Figures 4a, b show a comparison between strains recorded during constant loading phase. Strains measured by first five strain gauges (starting from the force application side) are reported for both specimens in Figure 4a, whereas strain evolution of other strain gauges are given in Figure 4b. Very close agreement can be seen between data obtained from two specimens. Some differences can be observed only in Figure 4b, i.e. for strain gauges far from the loaded end, where strain values are very small. From Figures 3a – d, it can be observed that at low stress level, plates with different bonded length exhibit a similar behaviour: rate of strain starts decreasing immediately after load application (as can be observed in classical concrete creep tests). Adopting time log scale, stabilization of delayed strain rate is typically considered when a linear increase of strain is attained. This behaviour has been reached after about three months of constant loading and can be considered as the beginning of steady-state increase of delayed strain.

0

100

200

300

400

500

600

700

0.0001 0.001 0.01 0.1 1 10 100 1000

Time (days)

Long

itudi

nal s

trai

n ( µ

ε)

E1E2E3E4E5E6E7

(10)

(20)

(30)

(45)

(60)(75)(90)

0

100

200

300

400

500

600

700

0.0001 0.001 0.01 0.1 1 10 100 1000Time (days)

Long

itudi

nal s

trai

n ( µ

ε)

E1E2E3E4E5E6E7E8E9

(10)

(20)

(30)

(50)

(70)

(150)

(90)(120)

(190)

(a) (b)

-100

0

100

200

300

400

500

600

700

0.0001 0.001 0.01 0.1 1 10 100 1000

Time (days)

Long

itudi

nal s

trai

n ( µ

ε)

E1E2E3E4E5E6E7E8E9

(10)

(20)(30)

(50)(70)

(150)

(90)(120)

(190)-100

0

100

200

300

400

500

600

700

0.0001 0.001 0.01 0.1 1 10 100 1000

Time (days)

Long

itudi

nal s

trai

n ( µ

ε)

E1E2E3E4E5E6E7E8E9E10E11

(10)

(20)

(40)

(60)(100)

(240)

(140)(190)

(290)(340) (390)

(c) (d)

Figure 3: Long – term constant loading. Time evolution of strains in FRP plates at different positions along the anchorage: specimen P1 – (a) B.L.=100 mm, (b) B.L.=200 mm; specimen P2 – (c) B.L.=200 mm, (d) B.L.=400

mm. Number in parenthesis indicates the distance from the initial section of the anchorage.

0

100

200

300

400

500

600

700

0.0001 0.001 0.01 0.1 1 10 100 1000

Time (days)

Long

itudi

nal s

trai

n ( µ

ε)

P2-E1P1-E1P2-E2P1-E2P2-E3P1-E3P2-E4P1-E4P2-E5P1-E5

P2-E1P1-E1

P2-E2

P1-E2

P2-E3

P1-E3

P2-E1P1-E1

P2-E1P1-E1

-50

-25

0

25

50

75

100

125

150

0.0001 0.001 0.01 0.1 1 10 100 1000

Time (days)

long

itudi

nal s

trai

n ( µ

ε)

P2-E6P1-E6P2-E7P1-E7P2-E8P1-E8P2-E9P1-E9

P2-E6

P1-E6

P2-E8

P1-E8P2-E7

P1-E7

P2-E9

P1-E9

(a) (b) Figure 4: Long – term loading. Time evolution of strains in FRP plate at different positions along the anchorage.

Comparison of results from specimens P1, P2, for B.L.= 200 mm.

535

A delayed strain coefficient has been defined as the ratio between delayed strain and instantaneous strain:

i

del tt

εε

=δ)(

)( .

If stress distribution was constant in the anchorage during time, and considering linear creep behaviour for both adhesive and concrete cover, delayed strain coefficient at a given time should be equal for all strain gauge positions.

Table 2 shows values of instantaneous and delayed strains as well as corresponding delayed strain coefficients at various distances from the traction side (after 180 days at loading). Delayed strain coefficient increases significantly far from loaded plate side: delayed deformation is up to 8 times greater than instantaneous deformation. Figure 5 shows, for three bonded lengths, variation of delayed strain coefficient with the distance from the traction side. Two types of behaviour can be observed for each bonded length: close to loaded end, delayed strain coefficient increases almost linearly, and slope increases with bonded length. At a distance from loaded end of about 150 – 200 mm, curves reach a second linear branch with smaller slope up to the end of plates (for 200 mm and 400 mm cases). Finally, Figures 6a – d show the strain profile along four plates at different times. Strain profiles are quite similar, and delayed strains during test are almost constant along the bonded length, even if instantaneous strain is much higher close to loaded end. DISCUSSIONS OF RESULTS

Variation of delayed strain coefficient along FRP plates is the result of different phenomena. 1) CFRP plates may exhibit creep deformation when subject to long – term loading. Nevertheless, according to experimental results reported in the literature (Ascione et al. 2005), CFRP plates adopted in test are practically insensitive to creep strains, being less than 2 per cent of instantaneous strain even for medium level stresses. 2) Adhesive and concrete cover where load is transferred from plate to concrete are subject to creep strains, increasing compliance of the interface. Increasing with time, plate – concrete slip provides for a shear stress redistribution along the interface: high shear stresses close to loaded end decrease and a longer length of interface contributes carry applied force. 3) Due to very high shear stresses at adhesive and concrete cover level, creep phenomenon may be highly non linear close to loaded end. Hence, compliance increase at the beginning of the anchorage is much higher than far from it.

Table 2: Instantaneous strain, delayed strain and delayed strain coefficient, of significant positions of the anchorage, for different bonded lengths (after 180 days of loading).

B.L. = 100 mm B.L. = 200 mm B.L. = 400 mm Distance εi (µε) εdel (µε) δ= εdel /εi εi (µε) εdel (µε) δ= εdel /εi εi (µε) εdel (µε) δ=εdel /εi

10 454 124 0.27 466 148 0.32 403 180 0.44 90 27 53 1.96 34 106 3.12 29 106 3.65 150 12 58 4.83 14 93 6.64 390 5 43 8.10

0

1

2

3

4

5

6

7

8

9

0 50 100 150 200 250 300 350 400

Distance (mm)

δ =

ε del

/ ε i

400 mm200 mm100 mm

Figure 5: Ratio between delayed (at 180 days of loading) and instantaneous strains along the plate.

536

Evolution with time of shear stress redistribution and creep deformation at the interface level (both variable along the anchorage) may explain the experimental results described in the previous section. In fact, far from the loaded section of the plate, normal stresses in the plate increase and, consequently, the same behaviour is exhibited by axial strains even if creep strains in FRP are absent. Moreover, shear stress redistribution is more evident in the case of long bonded lengths, i.e. longer than transmission length (estimated as about 150 – 200 mm). For this reason, the rate of increase of axial strains in FRP plate is higher for the longest length (400 mm), see Figure 5. At distance from loaded end greater than 200 mm, shear stresses increase with a smaller rate after long term loading application, and, consequently, the second branch in Figure 5 has smaller slope. Finally, shear stress redistribution and creep deformation provide for an almost constant strain increase with time along the bonded length.

CONCLUSION

The first results of an experimental campaign concerning long-term loading of FRP plates bonded to concrete are presented. Four anchorages are subject to a load about 50 percent of delamination load for a duration of 200 days. Strain profiles along the plates at different times are measured. It is shown that a shear stress redistribution along the bonded length occurs, due to increasing compliance of interface related to delayed deformation. At the end of tests (about 1 year), FRP – concrete specimens will be subject to standard delamination tests and results compared with those obtained for analogous specimens never loaded before. It will be then estimated if a long-term loading may cause damage of FRP – concrete interface reducing delamination load. ACKNOWLEDGMENTS

The authors would like to thank the Sika Italia S.p.A. for providing CFRP plates and adhesives. The financial supports of (Italian) MIUR (Ministry of Education, University and Research –PRIN 2003 Grant, FIRB 2001 Grant) and C.N.R. (National Council of Research), PAAS Grant 2001, are gratefully acknowledged.

0

100

200

300

400

500

600

700

0 10 20 30 40 50 60 70 80 90 100

Distance (mm)

Long

itudi

nal s

trai

n ( µ

ε)

end of loading7 days32 days70 days192 days

0

100

200

300

400

500

600

700

0 20 40 60 80 100 120 140 160 180 200

Distance (mm)Lo

ngitu

dina

l str

ain

( µε)

end of loading7 days32 days70 days192 days

(a) (b)

-100

0

100

200

300

400

500

600

700

0 20 40 60 80 100 120 140 160 180 200

Distance (mm)

Long

itudi

nal s

trai

n ( µ

ε)

End of loading10 days30 days75 days175 days

-100

0

100

200

300

400

500

600

700

0 40 80 120 160 200 240 280 320 360 400

Distance (mm)

Long

itudi

nal s

trai

n ( µ

ε)

End of loading10 days30 days75 days175 days

(c) (d) Figure 6: Evolution with time of strains along the bonded length, at different times. Specimen P1 – (a) B.L.=100

mm, (b) B.L.=200 mm; specimen P2 – (c) B.L.=200 mm, (d) B.L.=400 mm.

537

REFERENCES

Ascione, F., Berardi, V.P., Feo, L. and Giordano, A. (2005). “Experimental test on long term behaviour of CF pultruded elements” (in italian), Proceedings of XVII AIMETA, Florence, Italy, September, 1-11 (on CD).

Ferracuti, B., Savoia, M. and Mazzotti, C. (2005). “A numerical model for FRP – concrete delamination”, Composites B – Engineering, in publication.

Lu, X.Z., Teng, J.G. Ye, L.P. and Jiang J.G. (2005). “Bond – slip models for FRP sheets/plates bonded to concrete”, Engineering Structures, 27(4), 920-937.

Mazzotti, C., Ferracuti, B. and Savoia, M. (2004). “An experimental study on FRP –concrete delamination”, Proceedings of FraMCoS – 5, Li et al., eds., Vail, Colorado, U.S.A., V. 2, 795-802.

Mazzotti, C., Savoia, M. and Ferracuti, B. (2005). “FRP – concrete delamination results adopting different experimental pure shear test setups”, Proceedings of ICF XI, Turin, Italy, March 2005, 1-6 (on CD).

Miller B., Nanni A, De Lorenzis L. (2001). “Bond of FRP laminates to concrete”, ACI Material J., 98(3), 246-254. Plevris, N. and Triantafillou, T.C. (1994). “Time-dependent behaviour of RC members strengthened with FRP

laminates”, J. Struct. Engineering, ASCE, 120, 1016-1042. Savoia, M., Ferracuti, B. and Mazzotti, C. (2003). “Non linear bond-slip law for FRP-concrete interface”,

Proceedings of FRPRCS-6, Singapore, 1-6 (on CD). Savoia, M., Ferracuti B. and Mazzotti C. (2005). “Long-term creep deformation of FRP-plated r/c tensile

members”, J. of Composites for Construction, ASCE, 9(1),63-72. Teng, J.G., Chen, J.F., Smith, S.T. and Lam, L. (2002). FRP strengthened RC structures. J. Wiley and sons, UK. Teng J.G., Smith S.T., Yao J., Chen J.F. (2003). Intermediate crack-induced debonding in RC beams and slabs.

Construction and Building Materials, 17(6-7), 447-62. UNI 6556 (1976). “Tests of concretes. Determination of static modulus of elasticity in compression.” (in Italian). UNI 12390 (2002). “Testing hardened concrete - Compressive strength.” (in Italian).

538