This specimen was built of the concrete slab with a compressive
strength of 29.1 MPa. The complete loss of the tension flange
prevented the possibility of measurement of the average tensile web at midspan were used to obtain hypothetical values of strain
in the bottom of the flange. The load-deflection behavior of the
repaired girder is shown in Fig. 12. The girder started carrying the
load with a fairly linear response. The elastic stiffness of the
girder was 19.9 kN/mm and, during the elastic unloading ~not
shown !, there were no signs of permanent deformation or stiffness
change. At 133 kN, the epoxy that filled the cut separated from
the steel. The linear response continued up to the 162.2 kN load
level, at which an extrapolated prediction of the strain at the bot-
tom face of the tension flange reached the yielding strain. The
CFRP sheet stress was 194.4 MPa, equal to 9% of its ultimate
strength. Continuing loading, the longitudinal cracks appeared at
287 kN and a complete unloading and reloading was performed at
300 kN. There was minimal hysteresis in the loop and the loop
closed out at the point that unloading had started. The reloading
stiffness was 20.1 kN/mm, very similar to the stiffness at the
initial loading. After reaching the load level of 311 kN, the plate
started to debond from the steel girder at the cut location in the
midspan. Approaching the load level of 380 kN, the failure signs
of the CFRP ends were noticed. The girder failed at the load level
of 434 kN, due to sudden and complete debonding of the CFRP
sheets ~initiated from the end! and subsequently rupture of the
steel web and failure of the concrete. At the peak load, the CFRP
laminate was carrying an 840.1 MPa tensile stress, equal to 40%
of its strength, and the concrete strain was 0.0016, well below its
failure point. The compression flange and the web did not show
any sign of instability due to the support provided by the slab and
the wooden blocks in the entire experiment.