型钢混凝土(steel reinforced concrete， SRC)具有承载力高、刚度大、延性与抗震性良好、防火防腐性能优越、施工速度快等诸多优点，被广泛应用于全世界各地的高层、超高层建筑中。近年来，随着我国高速铁路的飞速发展，与之配套的客运站房、桥梁需要大规模地建设，在这些大型复杂的基础设施中，广泛采用了性能优异的SRC结构。由于高速铁路运输具有动力荷载大以及荷载循环次数较多等特点，因此结构设计时必须考虑高周疲劳破坏这一问题。然而，国内外相关规范均未对SRC梁的疲劳设计作出规定，而且目前国内外关于SRC梁疲劳性能的研究也极少，这无疑限制了SRC结构在疲劳荷载工况下的应用。为此，本文采用试验研究、理论分析和数值计算等手段，对SRC梁的疲劳性能进行了深入系统的研究。 本文的主要研究工作和取得的成果如下： (1)对29根SRC梁与5根纯钢梁开展了疲劳试验，考察了疲劳破坏全过程、变形与混凝土裂缝发展、钢组件疲劳断口形貌等特性。试验结果表明，纯钢梁与SRC梁的主导破坏模式均为受弯疲劳破坏，钢梁裂纹起源于腹板与受拉翼缘之间的角焊缝处。受拉纵筋一般先于内部H形钢梁的受拉翼缘发生完全疲劳断裂，而内部H形钢梁受拉翼缘的完全断裂可作为整个SRC梁疲劳破坏的标志。 SRC梁在疲劳阶段仍满足平截面假定，各组件能保持变形协调、未发生滑移。 SRC梁的混凝土裂缝宽度在静力阶段满足规范要求，在疲劳阶段因受拉纵筋的疲劳断裂以及内部H形钢梁裂纹的扩展而突然增大，最终破坏时形成疲劳主断口。然后，统计了纯钢梁与SRC梁内钢组件的疲劳寿命，为后续数值模型的验证与S-N曲线的回归提供了样本数据。 (2)在SRC梁与纯钢梁疲劳试验结果的基础上，从理论上推导了 SRC梁内各组件的名义应力幅计算公式，公式计算值与试验值吻合较好，并考察了内部H形钢梁含钢率、混凝土强度等级、剪跨比等参数的影响效应。通过比较发现， SRC梁的疲劳性能优于RC梁；受弯破坏的SRC梁的内部H形钢梁与纯钢梁具有相同的破坏特征、断口形貌，但前者在200万次时的平均值疲劳强度比后者提高了20%。分别采用不考虑纵筋断裂影响与考虑纵筋断裂影响两种方法对SRC梁内部H形钢梁回归了S-N曲线，发现后者比前者只提高了4%，而采用前者更加方便。对SRC梁内受拉纵筋也回归了S-N曲线，可供疲劳设计使用。 (3)已有的研究未涉及到SRC梁疲劳破坏模式的讨论，故本文通过一系列的概括与对比，明确了SRC梁的疲劳破坏模式问题。研究结果表明，对于内部H形钢梁表面未焊接连接件的SRC梁，纯弯段的受力更为不利，更易发生受弯疲劳破坏，疲劳破坏模式由受弯破坏主导，应着重考虑其受弯疲劳设计。对SRC梁与纯钢梁的受力机制进行了比较，发现SRC梁比纯钢梁更不易发生弯剪疲劳破坏。对于内部H形钢梁表面焊接连接件的SRC梁，可能发生弯剪疲劳破坏破坏。 (4)内部H形钢梁在整个SRC梁的疲劳抗力中起关键作用，并且SRC梁内部H形钢梁与纯钢梁的破坏特征非常相似。因此，基于断裂力学原理，通过ABAQUS软件及其二次开发程序，对纯钢梁的疲劳裂纹扩展过程进行了参数化数值模拟。围绕纯钢梁裂纹前缘设置楔形奇异单元，并采用J积分法计算SIF。采用大约4段1/4椭圆弧线来模拟裂纹前缘，各段弧线分别由独立的Paris公式进行控制，以裂纹尺寸达到翼缘壁厚或最大SIF达到断裂韧性作为数值模拟的终止条件。5根纯钢梁疲劳寿命有限元值与试验值之间比值的平均值为1.149，表明本文的数值模型是准确可靠的。 (5)在纯钢梁疲劳裂纹扩展模型的基础上，进一步计入钢筋与外部混凝土等组件，基于损伤力学原理提出了受拉纵筋的疲劳损伤演化规律与混凝土的疲劳损伤本构模型，对SRC梁的疲劳破坏过程进行了数值模拟。以内部H形钢梁的裂纹尺寸达到翼缘壁厚或最大SIF达到断裂韧性、或者混凝土最大压应力达到其疲劳剩余受压强度作为数值模拟的终止条件。本文以及本课题组前期研究中共计39根SRC梁内部H形钢梁的疲劳寿命有限元值与试验值之间比值的平均值为1.191，对于受拉纵筋这一数值为1.212，表明本文的数值模型是准确可靠的。对一些关键参数的影响效应开展了深入分析，结果表明：SRC梁内部H形钢梁的疲劳寿命随名义应力幅、受拉纵筋配筋率的增大而降低，随含钢率、受拉纵筋疲劳强度的增大而增大，而剪跨比、混凝土强度等级与配箍率对SRC梁的受弯疲劳性能的影响很小。 (6)结合试验研究与数值模拟的结果，对SRC梁内各组件分别提出了疲劳设计方法。内部H形钢梁是SRC梁中最重要的组件，本文基于大量参数分析的结果，提出了疲劳寿命提高系数，对已有规范中的纯钢梁疲劳S-N曲线进行修正，从而实现该组件的疲劳设计。对于受拉纵筋，采用基于试验数据得到的设计S-N曲线进行疲劳验算，并可考虑应力比的影响。SRC梁外部受压混凝土在试验中未发生疲劳破坏，可直接采用Eurocode 2规范中的设计S-N曲线进行验算。最终，根据试验结果对疲劳设计方法进行了验证，表明本文的设计方法是合理的。 本文明确了 SRC梁的疲劳破坏机制，形成了精细可靠的数值分析模型，并提出了完善合理的疲劳设计方法，为SRC梁的疲劳评估提供了强有力的工具，促进了 SRC结构的进一步推广应用。 关键词：型钢混凝土梁，钢结构，疲劳性能，S-N曲线，疲劳强度，疲劳寿命，破坏模式，试验研究，数值模拟，受弯疲劳破坏，疲劳裂纹扩展，参数影响效应，疲劳破坏机制，疲劳设计方法

 Steel reinforced concrete (SRC) structures have been widely applied in high-rise buildings and super high-rise buildings all over the world, which possesses a lot of advantages, such as greater load carrying capacity, high stiffness, excellent seismic performance and ductility, improved fire and corrosion resistance, short construction period. In recent years, with the rapid development of high-speed railway in China, a large number of railway stations and railroad bridges need to be constructed. In these large-scale and complex infrastructures, SRC structures with excellent performance are used. Due to higher dynamic load and more load cycles of high-speed railway transportation, high cycle fatigue problem must be taken into account in structure design. However, there are no relevant codes and specifications for fatigue design of SRC beams, and studies on fatigue behavior of SRC beams are very limited. The application of SRC structures under fatigue condictions is restricted to a large extent. Therefore, the fatigue behavior of SRC beams was studied comprehensively and systematically in this dissertation by the way of fatigue experiments, theoretical analysis and numerical simulations. The main research work and results are as follows: (1)A total of 29 SRC beams and 5 pure steel beams were tested under fatigue loading in order to investigate the features of fatigue failure process, development of deformation and concrete cracks, as well as fatigue fracture surfaces of steel components, and so on. The fatigue test results indicated that flexural fatigue failure was the dominating failure mode for both pure steel beams and SRC beams, for which the fatigue crack of steel beam originated from the fillt weld at the junction between tensile flange and web. Longitudinal reinforcement usually fractured before the entire tensile flange of internal H-steel, so the final fracture of internal H-steel signifies fatigue failure of the entire SRC beam. During fatigue loading period, all components of SRC beam worked together and deformed compatibly without slip, and plane cross-section assumption was satisfied. The maximum width of concrete cracks during static loading period met the requirements of design code. With fatigue fracture of longitudinal reinforcements and crack growth of internal H-steel during fatigue loading period, the width of concrete crack at a certain section increased significantly. At final fatigue failure, the main crack appeared. Then the fatigue lives of pure steel beams and steel componnets inside SRC beams were recorded, which could provide verification for further numerical models and sample data for regression of S-N curves. (2)Based on the test results of SRC beams and pure steel beams, the formulas calculating the nominal stress ranges of components inside SRC beam were derived theoretically, and the calculated values from the formulas agreed well with the measured values in the tests. The effects on fatigue behavior of SRC beam were investigated, including steel ratio of internal H-steel, shear span-to-beam depth ratio and concrete grade. Moreover, the comparison results suggested that fatigue behavior of SRC beam is superior to that of RC beam. Internal H-steel insude SRC beam with a flexural fatigue failure mode has the same failure feature and fracture surface with pure steel bram, but fatigue strength of the former is 20% higher than the latter. The S-N curves for internal H-steel inside SRC beam were obtained on the basis of two calculation methods of stress range, that is, the methods without and with considering the effect of fatigue fracture of longitudinal reinforcements. It is shown that the difference between the two methods is very small, but the method without considering the effect is more convienent. The S-N curves for longitudinal reinforcements were also derived and could be used for fatigue design. (3)The fatigue failure mode of SRC beam has never been studied or disscussed in the existing researches. In this dissertation, the fatigue failure mode problem of SRC beam was clarified through a series of summaries and comparisons. It is found that SRC beam without any welded attachments on internal H-steel is more prone to flexural failure mode and flexural failure mode is dominant. More attention should be payed to flexural fatigue design of SRC beam. The fatigue mechanism was compared between SRC beam and pure steel beam. The result indicated that SRC beam is less prone to flexural-shear fatigue failure than bare steel beam. However, flexural-shear fatigue failure mode may occur for SRC beam with welded attachments on internal H-steel. (4)Internal H-steel plays a significant role in the fatigue resistance of SRC beam and the fatigue failure characteristics of internal H-steel inside SRC beam and pure steel beam are quite similar. Therefore, the fatigue crack propagation process of pure steel beam was numerically simulated through ABAQUS software and its secondary development program based on the basic principle of fracture mechanics. Wedge singular elements were arranged around the crack front of pure steel beam and J integral method was employed to calculate the stress intensity factor (SIF). About four elliptic arcs were applied to simulate the fatigue crack front and each elliptic arc was controlled by an independent Paris formula. The situation that crack depth reached the wall thickness of tensile flange or the maximum SIF reached material fracture toughness was considered as the terminal condition of the finite element analysis (FEA).Finally, for 5 pure steel beams, the average value of the ratios between fatigue lives obtained from FEA and fatigue tests is 1.149. It is shown that the numerical model for pure steel beam in this dissertation is accurate and reliable. (5)Based on the fatigue crack growth model of pure steel beam, the fatigue failure process of SRC beam was also numerically simulated by adding in reinforcing cage and external concrete. Fatigue damage evolution model for tensile longitudinal reinforcements and fatigue damage constitutive model for concrete were proposed on the basis of the basic principle of damage mechanics. The situation that the maximum compressive stress of concrete reached its post-fatigue residual strength as well as the terminal conditions for pure steel beam were all considered as the terminal conditions of FEA for SRC beam. Finally, the average value of the ratios between fatigue lives obtained from FEA and fatigue tests is 1.191 for internal H-steels inside 39 SRC beams, while that average value for tensile reinforcements is 1.212. It is indicated that the numerical model for SRC beam in this dissertation is accurate and reliable. Then the effects of key parameters on fatigue behavior of SRC beam were further investigated. The results suggested that the fatigue life of internal H-steel inside SRC beam increases with the decrease of nominal stress range or steel ratio of tensile reinforcements, and it increases with the increase of steel ratio of internal H-steel or fatigue strength of tensile reinforcements. However, shear span-to-depth ratio, concrete grade and stirrup ratio have rather small influence on flexural fatigue behavior of SRC beam. (6)Finally,the results of fatigue emperiments and numerical analysis were combined to propose the fatigue design mothods for each component inside SRC beam. An increase coefficient of fatigue life for the most important component, that is, internal H-steel inside SRC beam, was proposed based on the results of a large number of parametric analysis. This increase coefficient could achieve fatigue design for internal H-steel inside SRC beam by modifying the S-N curve for pure steel beam in existing design codes. The design S-N curve for tensile reinforcements was derived based on the fatigue test results. Furthermore, the effect of fatigue stress ratio could be taken into account. In the fatigue test, external concrete in compressive region did not fail, so the design S-N curve in Eurocode 2 was deriectly used for fatigue checking calculation of SRC beam. At last, the fatigue design method for SRC beam was verified by the test results and it is shown that the fatigue design method in this dissertation is reasonable. In summary,fatigue failure mechanism was clarified, accurate and reliable numerical model was built, and reasonable fatigue design method was proposed in this dissertation. These research results provide a powerful tool for fatigue assessment of SRC beam and can promote the application of SRC beam. Keywords:Steel reinforced concrete (SRC) beam, Steel structure, Fatigue behavior, S-N curve, Fatigue strength, Fatigue life, Failur mode, Experimental investigation, Numerical simulation, Flexural fatigue failure, Fatigue crack propagation, Effect of parameter, Fatigue failure mechanism, Fatigue design method