滑动轴承摩擦故障趋势预测的系统自记忆模型  

System self-memory model for predicting friction fault trend of sliding bearings

作　　者：张峻宁[1] 张培林[1] 华春蓉[2] 吴定海[1] ZHANG Jun-ning ZHANG Pei-lin CHUA Chun-rong Wu Ding-hai(Department 7st Ordnance Engineering College, Shijiazhuang Hebei 050003, China School of Mechanical Engineering, Southwest Jiaotong University, Chengdu 610031, China)

机构地区：[1]军械工程学院七系,石家庄050003 [2]西南交通大学机械工程学院,成都610031

出　　处：《振动与冲击》2017年第11期20-26,47,共8页Journal of Vibration and Shock

基　　金：国家自然科学基金(51205405;51305454)

摘　　要：针对滑动轴承时间序列非线性引起的接触摩擦故障趋势难预测的问题,提出一种基于系统自记忆预测模型的滑动轴承接触摩擦故障趋势预测方法。该方法首先根据信号激励源不同的特点,将采集信号分离为冲击声和随机声,然后采用函数拟合、求导和灰色理论分别反演出冲击声和随机声的系统微分方程,并运用双向差分求取不同微分方程对轴承接触摩擦故障信号系统动力核的影响系数。通过引入自记忆函数,将滑动轴承摩擦故障系统动力核反演成一个微分-差分方程,由此得到滑动轴承的自记忆预测模型。应用到静载荷和动载荷的滑动轴承接触摩擦故障实例中,验证了所提方法的有效性,为滑动轴承磨损退化趋势预测提供了一种新的途径。The sliding bearing’s wear trend is hard to predict because its time series is nonlinear.To solve this problem,a wear trend prediction method for sliding bearing based on System self memory prediction model is put forward in this text.First,according to the different signal excitation source,the acquisition signal is separated into impact sound and random noise.Then the function fitting, derivation and grey theory are respectively inversed to get system differential equations of Impact sound and random sound.And Bidirectional Difference is used to obtain the influence coefficient of differential equation on the dynamic core of the bearing wear signal system.By introducing self memory function,the dynamic core of the bearing wear signal system is inversed into a Differential-Difference Equation.Then a self memory prediction model of sliding bearing is obtained.By applicating the proposed method in the case of sliding bearing wear degradation of static and dynamic loads,the validity of the method is verified.And it provides a new way to predict the degradation trend of sliding bearing.

关 键 词：滑动轴承 摩擦故障 发展趋势 非线性动力系统 信号分离 自记忆模型 

分 类 号：TH117.2[机械工程—机械设计及理论]