【正文】 erent types of fibers has been implemented. Three fibers have been selected with a different magnitude of elongations at failure. Figure 1 shows the stressstrain curves in tension for the selected posite fibers, and Table 1 shows their mechanical properties.The technique is based on bining these fibers together and controlling the mixture ratio so that when they arc loaded together in tension, the fibers with the lowest elongation (LE) fail first, allowing a strain relaxation (an increase in strain without an increase in load for the hybrid). The remaining highelongation (HE) fibers are proportioned to sustain the total load up to failure. The strain value at failure of the LE fibers presents the value of the yieldequivalent strain of the hybrid, while the HE fiber strain at failure presents the value of ultimate strain. The load corresponding to failure of LE fibers presents the yieldequivalent load value, and the maximum load carried by the HE fibers is the ultimate load value. Ultrahighmodulus carbon fibers (Carbon No. 1) have been used as LE fibers to have as low a strain as possible, but not less than the yield strain of steel (approximately % for Grade 60 steel). On the other hand. Eglass fibers were used as HE fibers to provide as high a strain as possible to produce a highductility index (the ratio between deformation at failure and deformation at yield). Highmodulus carbon fibers (Carbon No. 2) were selected as mediumelongation (ME) fibers to minimize the possible load drop during the strain relaxation that occurs after failure of the LE fibers, and also to provide a gradual load transition from the LE fibers to the HE fibers. Based on this concept, a uniaxial fabric was fabricated and tested to pare its behavior in tension with the theoretical predicted loading behavior. The theoretical behavior is based on the rule of mixtures, in which the axial stiffness of the hybrid is calculated by a summation of the relative stiffness of each of its ponents. The fabric was manufactured by bining different fibers as adjacent yarns and impregnating them inside a mold by an epoxy resin. Figure 2 shows a photo of one of the fabricated samples. Woven glass fiber tabs were provided at both ends of the test coupons to eliminate stress concentrations at end fixtures during testing. The coupons had a thickness of 2 mm ( in.) and a width of mmACI StructuralJournal/SeptemberOctober 2002 (1 in.) and were tested in tension according to ASTM D 3039 specifications. The average loadstrain curve for four tested samples is shown in Fig. 3 together with the theoretical prediction. It should be noted that the behavior is linear up to a strain of %, when the LE fibers started to fail. At this point, the strain increased at a faster rate than the load. When the strain reached ME fibers started to fail, resulting in an additional increase in strain without a significant increase in load, up to the total failure of the coupon by failure of the HE fibers. A yieldequivalent load (the first point on the loadstrain curve where the behavior bees nonlinear) of kN/mm width ( kips/in.) and an ultimate load of kN/mm ( kips/in.) arc observed.BEAM TESTSBeam detailsThirteen reinforced concrete beams with crosssectional dimensions of 152 x 254 mm (6x10 in.) and lengths of 2744 mm (108 in.) were cast. The flexure reinforcement of the beams consisted of two No. 5(16 mm) tension bars near the bottom, and two No. 3 ( mm) pression bars near the top. To avoid shear failure, the beams were over reinforced for shear with No. 3 ( mm) closed stirrups spaced at 102 mm ( in.). Five beams were formed with rounded corners of 25 mm (1 in.) radius to facilitate the installation of the strengthening material on their sides and bottom faces without stress concentrations. Figure 4 shows the beam dimensions, reinforcement details, support locations, and location of loading points. The steel used was Grade 60 with a yield strength of 415 MPa ( psi), while the concrete pressive strength at the time of testing the beams was MPa (8000 psi).Strengthening materialsThe developed hybrid fabric was used lo strengthen eight beams. Two different thicknesses of fabric were used. The first (Hsystcm, t = mm) had a thickness of mm ( in.), and the second (Hsystcm, / = mm) had a thickncss of mm ( in.). Four other beams were strengthened with three currently available carbon fiber strengthening materials: 1) a uniaxial carbon fiber sheet with an ultimate load of kN/mm ( kips/in.)。 and 2) strengthening material on the bottom face and extended up 152 mm (6 in.) on both sides to cover approximately all the flcxural tension portions of the beam (Beam Group B). The strengthening was installed for m (88 in.), centered along the length of the beam. The cpoxy was allowed to cure for at least 2 weeks before the beams were tested. For the beams strengthened with the developed hybrid fabric (Hsystcm), two beams were fabricated and tested for each configuration to verify the results. Table 4 summarizes the test beams.InstrumentationThe FRP strain at midspan was measured by three strain gages loeated at the bottom face of the beam. The steel tensile strain was measured by monitoring the strain on the side surfacc of ihc beam at reinforcing bar level using a DEMEC (detachable mechanical gage) with gage points for Beam Group A, while strain gages were used for Beam Group B. The midspan deflection was measured using a string potentiometer. The beams were loaded using a hydraulic actuator. The load was measured by means of a load cell. All the sensors were connected lo a data acquisition system to scan and record the readings.TEST RESULTS AND DISCUSSION Control beamThe control beam had a yield load of kN ( kips) and an ultimate load of kN ( kips). The beam failed by the yielding of steel, followed by pression failure of concrete at the midspan. Test results for the control beam arc