东篱科研大数据发现系统（DRDS）

位置：成果数据库 > 期刊 > 期刊详情页

一种盲源提取方法及其在滚动轴承故障特征提取中的应用

ISSN号：1004-4523
期刊名称：振动工程学报
时间：2014.10.15
页码：755-762
分类：TH165.3[机械工程—机械制造及自动化] TN911.7[电子电信—通信与信息系统;电子电信—信息与通信工程]
作者机构：[1]上海交通大学机械系统与振动国家重点实验室,上海200240
相关基金：国家自然科学基金资助项目（51035007,51105243）
相关项目：关键设备故障预示与运行安全保障的新理论和新技术

关键词：故障诊断, 盲源提取, 滚动轴承, 约束独立成分分析, fault diagnosis, blind source extraction, rolling bearings, constrained independent component analysis

中文摘要：

利用滚动轴承发生故障时其故障信号往往呈现出一定周期性的特点,首先计算出故障信号的理论基本周期τ,将τ作为待提取源信号的基本周期用所述方法的相关步骤计算出大致目标源信号及权重分离矩阵^W。然后将^W作为初始权重分离矩阵,将基于高阶统计量的固定点算法用于原始观测信号提取出更为精确的目标故障信号。通过仿真信号和实验信号验证了所述方法相对于约束独立成分分析（Constrained independent component analysis,CICA）方法具有以下优点：不需要精确估计目标源信号的周期及不需要构建精确的参考信号。此外,通过仿真还验证了所述方法相对于其他较新的盲源提取方法具较高的提取精度等优点。

英文摘要：

When fault arises in the rolling bearing it usually takes on cyclical characteristics.Firstly,the desired source signal is extracted coarsely and the weight separation matrix ^W is obtained by the proposed algorithm using the estimated fundamental period of the desired fault signal.Then,the fixed-point algorithm is applied to the observed signals using ^Was its initial weight matrix and more accurate desired source signal is derived.The proposed method has the following advantages over the constrained independent component analysis（CICA）：it neither need to estimate the fundamental period τ of the desired source accurately nor need to construct the reference signal precisely which is otherwise for the CICA method.Besides,the proposed method has a higher accuracy over the recent BSE methods.At last,the above stated advantages of the proposed method are verified through simulation and the fault diagnosis test of rolling bearings.