Demec measurements(Whitmore gauge measurements) were taken on the east face ofthe wall according to the pattern shown in Fig. 6a.The tests were quasi-static and the test units were subjectedto a fully-reversed cyclic loading history with step-wise increasinghorizontal top displacement. The first four cycles were run inforce control up to respectively 25, 50, 75 and 100% of the forcecorresponding to the onset of yielding at the extreme flexuralreinforcing bars of the section (first yield). Afterwards, the nominalyield displacement  y – corresponding to displacement ductility   D 1 – was defined and additional cycles in displacementcontrol were carried out. The displacement ductility was thenincreased in steps of 1 up to    D 4 and afterwards in steps of2 up to failure. Due to practical difficulties that were encounteredduring testing, Wall W1 was loaded to slightly different targetductilities. While presenting the results relevant to Wall W1 thecorrect ductilities determined at the end of the test are used.  3. Test resultsIn the following, the test results are presented in termsof force–displacement hystereses and displacement componentswith a discussion of the relevant failuremechanisms. Due to spacelimitations it is not possible to present all data collected during thetests and for further details the reader is referred to [4].The hysteretic behaviour of all test units as well as thedisplacement components at the peak top displacement during thefirst cycle of every ductility level are depicted in Fig. 7. The threecomponents (1) shear displacements, (2) flexural displacementsand (3) fixed-end displacementswere computed fromthe readingsof the Linear Variable Differential Transformers (LVDTs) mountedon the test units (Fig. 6b). The shear displacementswere computed  using the procedure originally presented in [14], the flexuraldisplacements were computed by integration of the curvaturesobtained from the LVDTs W-V-2 to 7-North and South. Thefixed-end displacement is the component which accounts for therotation of the wall at the construction joint and it was computedby integrating the rotation due to the opening of the crack at thejoint measured by means of the LVDTs W-V-1-North and W-V-1-South over the height of the test unit. If the summation of allthree displacement components is compared to the measured topdisplacement, the error is always less than 10%.Test UnitW1 failed during the second cycle to ductility 8.5 (4.1%drift) due to tensile fracture of several D5.2 and D12 reinforcingbars. However, significant damage occurred well before failure.Spalling of the concrete cover in the compression zone startedat ductility 4 (2% drift) and by the next cycle at ductility 5.7(2.5% drift) the region affected by the spalling was so large thatthe concrete cover was no longer able to prevent buckling of theflexural reinforcement. This insufficient behaviour was caused bythe inappropriate downscaling of the prototype’s dimensions tothe test unit’s dimensions. Fig. 3a shows that the thickness of thecover concrete was only about 10 mm. Considering the length ofHFC’smediumand long fibres (12 and 30mm), it is straightforwardto conclude that the cover concrete failed prematurely becauseit was too thin for the fibres to reach a proper distribution. InTest Units W2 and W3 this problem was solved by increasing theamount of fibres in themix design and especially by increasing thethickness of the concrete cover (Fig. 3b and c). These improvementswere extremely successful totally eliminating spalling and provingthe ability of HFC to prevent buckling of the flexural reinforcement.A direct consequence of the spalling which could be observed inthe hysteresis curvewas that the peak load reached during the firstcycle at a ductility of 4 was smaller (push direction) or just slightlyhigher (pull direction) than the peak load that was reached duringthe first cycle at a ductility of 2.5. The spalling of the concretecaused a reduction of the inner lever arm, and hence a reductionof the bending strength which could not be totally compensatedby the hardening of the reinforcing steel.Test Unit W1 also experienced noticeable sliding at theconstruction joint. The axial load acting on the wall was relativelysmall allowing the wall to grow vertically. Additionally, theconcentration of the deformations at the base of the wall and theminor roughness of the crack at the joint led to a situation whereduring large portions of a loading cycle, the entire base shear hadto be carried by dowel action of the flexural reinforcement and thereinforcing bars kinked across the still open joint crack. The kinkingimposed large local inelastic deformations on the bars leading totheir failure. At the same time the sliding also affected the overallshape of the hysteretic loops leading to significant pinching asshown in Fig. 7a. In Test UnitsW2 andW3 this problemwas solvedby using steel sleeves instead of plastic sleeves and by partiallyembedding them into the footing (Fig. 3b and c). The dowel actionexercised by the steel sleeves was enough to transfer the totalityof the base shear across the crack at the joint between wall andfooting.Test Unit W2 was able to complete a full cycle at ductility8 before failure occurred during the second cycle (see Fig. 7c).However, already during the first cycle a D12 reinforcing baron the south side of the wall fractured causing the base shearto drop to about 84% of the measured peak value of 185 kNwhich corresponded to a nominal peak shear stress of  max D  Vmax=.0:8bwlw/ D 2:4 MPa (0:20pf0cMPa). The bar fracturedbecause it experienced a high tensile strain. The strain in the barwas not directly measured. However, a preliminary estimation,based on the maximum crack width of about 31 mm measured atthe joint, shows that the strain had to be around 10%–12%. Thisstrain has the same order of magnitude as the total elongation atmaximum force Agt measured during the material tests (Table 3).The steel sleeves embedded in the footing prevented sliding ofthe wall and the relevant pinching of the hysteresis loops. Thebehaviour of Wall W2 was significantly better than the behaviourof Wall W1 even if the latter was able to reach a slightlylarger maximum deformation. This is due to the fact that theunbonded length of the flexural reinforcement was more thantwice the unbonded length of the flexural reinforcement of UnitW2 (Fig. 3). The use of a longer unbonded length in Unit W2as well would have led to a larger displacement capacity. Thiswas not deemed to be necessary; however it is important torecognize that the maximum displacement capacity of the wallscan be largely influenced by the choice of the unbonded length ofthe flexural reinforcement. Fig. 7d shows that shear deformationswere very limited, which is consistent with the crack patterndepicted in Fig. 8b. Flexural displacements are due to crackingalong the wall and Fig. 7d shows that they are proportional tothe base shear. On the other hand, fixed-end displacements aredue to yielding of the flexural reinforcement and are proportionalto the measured top displacement. At displacement ductility 1,flexural displacements are responsible for more than 50% of themeasured top displacement of Test Unit W2. 非线性循环行为混杂纤维混凝土结构墙英文文献和中文翻译(4):http://www.751com.cn/fanyi/lunwen_57270.html