The beam analysiswas based on a dead-weight of 65.7 kN (5.39 kN/m), asupport spacing of 11.79 m, and a load-point spacing of2.50 m.As discussed in Section 3.3, three of the eight pres-tressing tendons broke before the concrete cured inBeam 1. The pattern of breakage was such that thecentroid of the ®ve remaining prestressed tendons waswithin 5 mm of the beam centerline, so Beam 1 wasanalyzed and tested with the assumption that the brokentendons behaved simply as unpretensioned tendons.This assumption was justi®ed by the fact that the ten-dons failed outside of the beam. The jacking tension was142 kN per tendon, or 2000 MPa (0.77 fpu). The mea-sured value of f 0c was 86.3 MPa for Beam 1, providingthe modulus of rupture of 5.8 MPa. Based on thesevalues, an initial cracking load of 240 kN was predicted,and an ultimate load of 294 kN was predicted at a centerde¯ection of 152 mm., limited by failure of the tendons.Beam 2 featured tendons with a reduced jackingtension of 84.5 kN per tendon, or 1220 MPa (0.45 fpu).The concrete was reformulated, resulting on a measuredvalue of f 0c ˆ 71:1 MPa and a computed modulus ofrupture of 5.2 MPa. Based on these values, an initialcracking live load of 218 kN was predicted, and an 3.3. Fabrication and testingThe two test beams were fabricated by TecspanConcrete Structures of Grove City, Ohio. The assembledreinforcements and tendons are shown in Fig. 4. Lead-line wedge grips were used to grip the CFRP tendons atboth ends. Grip couplers were used to link the live endof each CFRP tendon to a steel cable, as shown in Fig. 5.The less expensive steel cable spanned the unused por-tion of the prestressing bed.For Beam 1, tendons were pretensioned to 142 kN, or77% of the nominal strength. Due to a calculation error,this pretension level was closer to the nominal ultimatestrength than intended; nonetheless, based on Mitsubi-shiÕs claim that the wedge grips provide 100% of thetendon strength, tendon failure would not be expected.However, three of the eight tendons failed before curingof the concrete, and possible contributing factors areconsidered here. First, upon tensioning of the steeltendons used to complete the span of the prestressingbed, there was a signi®cant twist induced in the Leadline(one turn every 0.6 m) due to the tendency of the steelcables to untwist. While this would not necessarily re-duce the strength of a uniformly stretched tendon, thetwist could be expected to cause stress concentrationswhere the tendons exit the grips. Second, there werehandling practices used during fabrication that couldhave caused, or did cause, damage when used withCFRP tendons. For example, the end plates for theforms were made of thick steel, and the vertical slots that accommodate the tendons had very irregular sur-faces. The end plates were driven down over the ten-dons, visibly abrading and damaging at least onetendon. The ®rst of the three tendon failures occurredduring vibration of the poured concrete, and may haveoccurred when the vibrator contacted the tendon.Greater care in handling was used during fabrication ofBeam 2.Both Beams 1 and 2 were steam-heated overnightfollowing pouring, and tendons were cut the followingdays after the concrete strength had surpassed 60% of the target compressive strength of 69 MPa. The steelstrands were cut at the mid-length of the prestressingbed using a torch. As each cable was cut, the corre-sponding composite tendon was cut at the oppositebeam end using a cutting wheel. Beam 1 was poured on26 August 1997. Beam 2 was poured on 30 September1997.The two test beams were tested to ultimate failure infour-point bending at the Aerospace Structures TestFacility of the US. Air Force Research Laboratory(AFRL), Wright Patterson Air Force Base, OH. Thetest setup is shown in Fig. 6 with Beam 2 in place fol-lowing testing. The applied loads and support loads weretransmitted from the test ®xtures to the test beamthrough solid steel cylinders, used to create transverseline loads at the appropriate beam stations. Steel bearingplates served to distribute the line loads at the contactlocations. Four hydraulic jacks were used in two pairs toapply the load at the two load stations. A load cell wasmounted in-line with each jack to accurately measure theapplied load. A common hydraulic pressure sourcepowered all four jacks to assure uniform loading. Theapplied loads and support reactions were transmitted tothe test beam through a massive steel beam located underthe test beam. Linear voltage displacement transducers(LVDTs) were used with string sensors running from the¯oor to points on the bottom of the test beam. LVDTswere located at the 1/4, 1/2 and 3/4 length positions on thebeam. Beam1 was tested on 19 February 1998, and Beam2 was tested on 24 February 1998.3.4. Results and discussion 复合材料高强度混凝土的试验研究桥梁英文文献和中文翻译(4):http://www.751com.cn/fanyi/lunwen_33192.html