Rolling bearings, widely used to support heavy rotating machinery and transfer load, generally working in low speed and heavy load conditions, thereby easily resulting in damages and mechanical equipment shutdown. Therefore, it is necessary to propose a method for fault period impact characteristic extraction of rolling bearings based on maximum correlated kurtosis deconvolution. The proposed method takes advantage of the period impact characteristics excited by partial failure during bearing operation as well as maximized correlated kurtosis to select the optimum parameters of finite response filter. Besides, based on iterative convolution operation, noise in vibration signals can be eliminated and period impact characteristics excited by failure of rolling bearing fault can be further extracted, thereby bearing faults location can be diagnosed according to the period of impact characteristics and bearing failure diagnosis can be realized. Simulation and test data of rolling bearing indicates the feasibility of proposed method and it is better than ensemble empirical mode decomposition (EEMD) in bearing failure diagnosis by comparing with the results extracted of ensemble empirical mode decomposition method.%滚动轴承广泛应用于重型旋转机械支撑和传送负载,经常工作在低速、重载等恶劣工况下,特别容易损坏,从而导致机械设备停运停产的事故,因此有必要提出一种基于最大相关谱峭度解卷积的滚动轴承故障周期冲击特征提取方法.该方法利用轴承运行过程中局部故障激发起的周期性冲击特征,通过最大化相关谱峭度选择最佳有限冲击响应滤波器参数;通过迭代卷积运算,消除振动信号中的噪声,提取出滚动轴承故障激发起的周期性冲击特征;依据冲击特征的周期判断轴承故障所在位置,从而实现轴承故障诊断.通过仿真和滚动轴承实验数据验证提出方法的可行性,并与广泛应用的集总经验模式分解方法提取结果进行对比,结果表明该文提出的方法在轴承故障诊断中展现出更好的优势.
展开▼
