论文部分内容阅读
机械设备发生故障时,故障特征的提取是很重要的.为了从观测信号中分离出不同的故障特征源信号,并根据分离信号准确地进行故障诊断,从观测信号样本出发,提出了基于有限支持样本核函数的盲源分离(FSS-kernel BSS)方法.此方法利用有限的观测样本估计信号的概率分布,得到了评价函数,具有很好的自适应能力.仿真试验结果表明:此方法能成功地分离超、亚高斯混合信号,与其他盲源分离方法相比,此方法具有更好的分离性能.将该方法用于转子不平衡和支座松动的复合故障信号的盲分离,分离出了各复合故障的主要频谱.分离结果表明:此方法应用于机械设备复合故障诊断中是可行的.
In the event of a mechanical equipment failure, the extraction of fault features is very important. In order to separate the different fault source signals from the observed signals and to accurately diagnose the fault based on the separated signals, we propose a new algorithm based on the limited support (FSS-kernel BSS) method of sample kernel function, which uses a limited sample of observations to estimate the probability distribution of the signal and obtains the evaluation function with good self-adaptability.The simulation results show that this method can successfully , This method has better separation performance than other blind source separation methods.This method is applied to the blind separation of composite fault signals with unbalanced rotor and loose bearings, The main spectrum of each compound fault.The separation results show that this method is feasible for the complex fault diagnosis of mechanical equipment.