基于EMD与Duffing振子的风机轴承早期故障诊断研究  

Research on wind turbines bearing early fault diagnosis based on EMD and Duffing oscillator

作　　者：吕跃刚[1] 李子民[1] 

机构地区：[1]华北电力大学控制与计算机工程学院,北京102206

出　　处：《可再生能源》2016年第5期687-691,共5页Renewable Energy Resources

基　　金：中央高校基本科研业务费专项金(2014ms24)

摘　　要：针对风电机组滚动轴承早期故障诊断,文章提出了一种以改进的Duffing振子与经验模态分解(EMD)相结合的混沌检测系统对早期微弱的故障信号进行有效识别的方法。首先利用EMD将采集到的振动信号分解成几个内蕴模式函数分量IMF,将包含故障特征的IMF作为外加策动力输入混沌系统,通过正逆向检测过程观察相轨迹的变化情况来确定是否捕捉到轴承早期的微弱故障特征信号。Duffing振子不仅能很好地抑制噪声,而且对内部策动力同频的微弱周期信号非常敏感,信噪比可达-45 d B。EMD分解法对信号进行初步筛选进一步提高了检测门限。通过轴承实验数据验证了该方法的有效性。A method has put forward for the wind generator for early fault diagnosis of rolling bearing in this paper, which is based on the improved the Duffing oscillator for the model and combined with empirical mode decomposition（EMD） chaos detection system for early effective identification of weak periodic signal. First, the EMD collected vibration signal is decomposed into several intrinsic mode function（IMF） components. Then the IMF containing the characteristic frequency input the chaotic system and as a driving force, through watching the phase trajectory is the reverse process to determine whether to catch early weak bearing fault characteristic signal. Duffing oscillator can not only well restrain noise but also for internal policy power with the frequency of weak periodic signal is very sensitive, signal to noise ratio can reach to-45 d B. The EMD decomposition method to signal a preliminary screening to further improve the detection threshold. Through the experimental data has proved the effectiveness of the proposed method.

关 键 词：振动信号 DUFFING振子 EMD 风电机组 滚动轴承 

分 类 号：TK83[动力工程及工程热物理—流体机械及工程] TH165.3[机械工程—机械制造及自动化]