目的 研究不同感应淬火工艺对重载铁路车轴钢试样梯度结构的影响，阐明不同人工缺陷下感应淬火试样的疲劳断裂行为，并评价损伤容限，为重载铁路车轴感应淬火工艺设计和损伤容限评价提供理论基础和评价方法。方法在AAR-CM车轴钢试样表面采用不同感应淬火工艺参数制备不同深度淬硬层，采用X射线测量淬硬层残余应力分布，通过电火花和空气炮在正火态和感应淬火试样表面制备不同尺寸的电火花或冲击损伤，采用疲劳试验机研究损伤试样的弯曲疲劳性能，借助光学显微镜和扫描电镜分析试样的损伤机理和疲劳断裂行为，并对其疲劳性能进行评价。结果 在3种工艺下感应淬火试样有效淬硬层深度分别为800、1 200、1 400 μm。在电火花缺口尺寸较小时，感应淬火试样的疲劳强度比正火态试样高出50%。在冲击损伤尺寸较小时，感应淬火试样的疲劳强度比正火态试样高出40%。较小的尺寸冲击损伤对正火态和感应淬火处理试样的疲劳强度无影响。结论 感应淬火可提高重载铁路车轴钢试样表面强度，进而提高其耐冲击性能，残余压应力减弱了电火花缺口试样和低速冲击损伤对试样疲劳强度的影响，高速冲击破坏了试样表面完整性，并削弱了残余压应力的作用，试样的疲劳强度急剧降低。文中设计的淬硬层深度对感应淬火试样损伤容限的影响较小。

To provide a theoretical basis and evaluation method for surface induction hardening process designing and damage tolerance evaluation of heavy-duty railway axles, the paper studied the evolution of microstructure, hardness, and residual stress gradient of the hardened layer of heavy-duty railway axle steel specimens under different induction hardening processes, analyzed the damage mechanism of defects introduced by EDM (electrical discharge machining) notch and impact damage, clarified the fatigue fracture behavior of induction hardening specimen under different artificial defects and evaluated its damage tolerance. The AAR-CM heavy-duty railway axle steel was used to prepare testing specimens. The hardened layers with different depths were prepared on the surface of the specimens by different induction hardening process parameters, and the residual stress distribution of the hardened layer was measured by X-ray. Different sizes of EDM notch or impact damage were prepared on the surface of untreated (normalized) and induction hardening specimens by EDM and air gun. The bending fatigue properties of the specimens under different damage were studied with a fatigue testing machine. The damage mechanism and fatigue fracture behavior of the specimens were analyzed with an optical microscope and a scanning electron microscope. Through metallographic observation, hardness test, damage morphology characterization, and fatigue fracture analysis, the fatigue properties of normalized and induction hardening specimens under different damage degrees were evaluated respectively. The results showed that the surface structure of the specimen changed from pearlite and ferrite to martensite during the induction hardening process. The effective hardened layer depths of the induction hardening specimens under the three processes were 800 μm (M1), 1 200 μm (M2), and 1 400 μm (M3), respectively. The thicker the effective hardened layer, the greater the surface hardness and residual stress. The shape of the EDM notch was regular. The impact damage shape of the normalized sample was complete. The impact damage morphology of the induction hardening specimen depended on the impact velocity. There were gaps and microcracks at the edge of the high-speed impact damage. When the size of the EDM notch was small, the fatigue strength of the induction hardening specimen was 50% higher than that of the normalized specimen. As the size of the EDM notch increased, the increase of the fatigue strength of the specimen by induction hardening gradually decreased. When the impact damage size was small, the fatigue strength of the induction hardening specimen was 40% higher than that of the normalized specimen. When the impact damage size increased, the fatigue strength of the induction hardening specimen was equivalent to the fatigue strength of the normalized specimen. Under the same damage parameters, the fatigue strength of the induction hardening specimen was not affected by the depth of the hardened layer. Minor impact damage did not affect the fatigue strength of normalized and induction hardening specimens. Surface induction hardening could improve the surface strength of heavy-duty railway axle steel specimens and enhance their impact resistance. The residual compressive stress would weaken the influence of the EDM notch specimen and low-velocity impact damage on the fatigue strength of the specimen. The high-speed impact would destroy the surface integrity of the specimen and weaken the effect of residual compressive stress, resulting in a sharp decrease in the fatigue strength of the specimen. The depth of the hardened layer designed in this paper had little effect on the damage tolerance of the specimen.