航空发动机轴承钢滚动接触疲劳模拟及寿命预测

Rolling contact fatigue modelling and life prediction for aeroengine bearing steels

主轴轴承是航空发动机中重要的安全部件,由综合性能优良的轴承钢制备,材料须具备高表面硬度、高断裂韧度、耐高温、抗疲劳、抗腐蚀等特点。然而,严苛的运行条件使轴承钢因滚动接触疲劳(RCF)而失效,严重影响飞行安全,因此,准确预测轴承钢的RCF寿命是保证航空发动机可靠性的关键。本文综述了航空发动机轴承钢RCF和寿命预测方面的重要研究成果和进展,并且展望了该领域未来的研究方向。文章首先介绍了轴承滚动体和滚道间赫兹接触导致的特殊应力场,其中剪切应力分量在次表面达到峰值,这解释了轴承钢次表面RCF复杂机理的原因,提出在理想条件下次表面起源的RCF是轴承钢的重要失效模式,同时,随着接触应力的增加,材料的响应方式从弹性向塑性演进;此外,由于航空发动机轴承实际服役环境恶劣,表面起源的RCF也会发生,因此存在两者之间的竞争。接着,总结对比了三种对RCF寿命的理论预测思路,即概率模型、机理模型和数值模型,并分析了这三种模型各自的优缺点:概率模型发展成熟,工业界应用广泛,但本质是一类统计学模型,缺少RCF机理,科学性较低;决定性模型通过对物理过程的描述预测RCF寿命,科学性高但模型过于简化,精确性不高;数值模型兼顾了工程实际和科学性,是针对RCF寿命预测问题的有力手段,但精确性有待进一步提升。最后,基于当前的研究现状,建议在未来从解决RCF过程中关键科学问题、通过嵌入RCF机理优化寿命预测模型和发展人工智能在RCF寿命预测的应用这三个方面进行研究。

Main-shaft bearings are vital safety components in aeroengines made of bearing steels with outstanding combinational properties.The material is supposed to exhibit high surface hardness,high fracture toughness,high temperature resistance,fatigue resistance and corrosion resistance.However,under harsh operation conditions bearing steels are prone to rolling contact fatigue(RCF)that leads to bearing failure,severely endangering flight.Therefore,accurately predicting the RCF lives of bearing steels is key to the reliability of aeroengine.This manuscript reviews important results and progress in the RCF and life prediction of aeroengine bearing steels and suggests future research directions.Firstly,special stress state as a result of the Hertzian contact between rolling element and raceway is introduced,where shear stress components peak at the subsurface.This explains the origin of complexsubsurface-originated RCF mechanisms in bearing steels and indicates subsurface-originated RCF to be an important failure mode of bearing steels under ideal conditions.Meanwhile,with increasing contact pressure,the response of material evolves from elastic mode to plastic mode.Besides,due to the harsh service environment of aero-engine bearings,surface-originated also takes place and hence the competition between these two mechanisms is present.Next,three methodologies for RCF life prediction are summarized,being probabilistic models,mechanistic models and numerical models,with their advantages and limitations analyzed Probabilistic models are well developed and widely employed by industry,but they are in nature a kind of statistical models without accounting for RCF mechanisms and are hence lacking scientificity;Deterministic models predict RCF life via describing physical processes,which are rich in science but poor in accuracy due to their simplification;numerical models balancing both engineering practice and scientific characteristic is a powerful tool to tackle the problem of RCF life prediction,although their accuracy requires further improvement.Finally,it is suggested that future research may focus on solving key scientific problems in RCF,modifying life prediction models via inserting RCF mechanisms and applying artificial intelligence in RCF life prediction.