滚动轴承支承下的转子碰摩故障动力分析、特征提取与智能诊断

针对滚动轴承支承下的转子碰摩故障诊断问题,考虑滚动轴承非线性赫兹接触和轴承径向间隙,建立了含不平衡-碰摩耦合故障的转子-滚动轴承系统动力学模型.应用三维谱图分析碰摩诊断依据和确定频段特征,利用仿真数据获取结构最优的神经网络模型,并对试验碰摩故障数据进行诊断,识别率高达90%以上.计算结果充分表明了建立的动力学模型的有效性,同时,研究结果为有效结合故障机理分析与故障智能诊断提供了新的方法. 

Dynamics analysis,feature extraction and intelligent diagnosis for rubbing fault of rotor supported on bull bearings

This paper aims at diagnosing the rubbing fault of a rotor supported on ball bearings.Firstly,a dynamics model of unbalance-rubbing coupling fault of rotorball bearing system was established by taking into account the nonlinear Hertzian contact and the radial clearance;secondly,3-D spectrum diagram was used to analyze the rubbing fault and determine the characteristics of every frequency zone;thirdly,an optimized self-adaptive neural network was obtained using simulation data,and rubbing test samples were diagnosed with a recognition accuracy over 91%.The results validate the usefulness of this dynamics model,meanwhile the research findings provide a new method for effectively combining malfunction mechanism analysis and intelligent diagnosis.