摘要
滚动轴承的早期故障诊断对于设备预测和健康管理具有重要意义,然而受环境噪声、传递路径、信号衰减及源信号本身比较微弱的影响,滚动轴承故障的初期微弱信号特征往往难以提取。为了解决这一问题,提出了一种基于最小熵解卷积(minimum entropy deconvolution,MED)与希尔伯特变换(Hilbert transform,HT)相结合的滚动轴承故障特征提取方法(MED-Hilbert),该方法首先应用MED算法对传感器信号进行处理以提高信号的信噪比,然后通过希尔伯变换提取冲击能量信号,最后用谱分析技术提取故障对应的特征频率,并与理论故障频率比较后成功确定故障。与信号仅仅进行包络分析方法相比,该方法具有很好的降噪效果以及对微弱故障特征的增强作用。计算机仿真与实验验证了该方法在滚动轴承早期故障诊断中的有效性。
Initial fault detection of the rolling element bearing is very important for prognostic and health management. However,it is difficult to extract the initial fault from rolling element bearings for the influence of environmental noise,transmission path,signal attenuation and weakness of source signal. In order to solve this problem,a method based on minimum entropy deconvolution( MED) and Hilbert transform( HT) is proposed.The signals from sensors are firstly proceeded with MED algorithm to improve signal to noise ratio( SNR),then demodulated by Hilbert transform to get the impulse energy signal. Finally,the fault is confirmed by spectrum frequency analysis of the impulse energy signal. The proposed method can detect and enhance the weak fault feature successfully compared with the envelope analysis method only. The effectiveness of the proposed method is verified by the simulations and experiments of the rolling element bearing.
出处
《河南理工大学学报（自然科学版）》
CAS
北大核心
2018年第1期91-96,共6页
Journal of Henan Polytechnic University（Natural Science）
基金
国家自然科学基金资助项目(U1304523
5150050464)
河南理工大学创新型科研团队项目(T2017-3)
