柴油机振动信号盲分离组合算法  

Combination algorithm for blind separation of diesel engine vibration signal

作　　者：韩春杨[1] 姚国凤[1] 赵建 张修路[1] 盛鑫[1] 

机构地区：[1]吉林大学机械科学与工程学院工程力学系,长春130022 [2]汽车振动噪声与安全控制综合技术国家重点实验室,长春130011

出　　处：《振动与冲击》2014年第6期44-47,52,共5页Journal of Vibration and Shock

基　　金：吉林省科技发展计划项目(201205021)

摘　　要：由于机械系统的振动信号能有效反映系统特征,对其进行盲解卷积能提供由混合信号中分离出源信号的可能性,为此提出机械系统振动卷积模型;将多通道盲最小均方差与缩减盲源方法结合提出MBLMS-TDS组合算法,且利用该算法对卷积混合信号进行盲解卷积验证算法的合理性;用该算法对柴油机表面混合振动信号进行分离,获得活塞撞击缸体信号与柴油机燃烧信号。The mechanical system vibration signals effectively reflect the characteristics of the system and its blind deconvolution can provide the possibility of separation of source signals from mixed signals.A convolution model of the mechanical system vibration was proposed and the method of Multi-Channel Blind Least Mean Squares was combined with the method of Deflation Source so as to form the MBLMS-TDS algorithm.For verifing the reasonableness of the algorithm, it was applied in the blind deconvolution of some simulated mixed sources.As a practical example,the piston-slap signal and combustion signal of a diesel engine were extracted from the mixed vibration signal on the surface of the diesel engine piston by using the MBLMS-TDS algorithm.

关 键 词：盲源分离 卷积模型 MBLMS-TDS算法 柴油机 

分 类 号：TN911.7[电子电信—通信与信息系统]