基于改进Prony方法的邻近频率正弦信号参数识别  

Parameters Estimation of Intensive Sinusoidal Signals Based on Improved Prony Method

作　　者：高启明[1] 李传江[2,3] 张自强[2] 陈海雄[2] 

机构地区：[1]合肥通用机械研究院,安徽合肥230088 [2]上海师范大学信息与机电工程学院,上海201418 [3]上海大学机电工程与自动化学院,上海200072

出　　处：《机床与液压》2013年第21期168-170,174,共4页Machine Tool & Hydraulics

基　　金：合肥通用机械研究院青年基金资助项目(2012011096)

摘　　要：在旋转机械故障诊断、通信等工程领域中邻近正弦信号参数识别问题较为常见,针对此类问题中非衰减实值信号的特点,采用总体最小二乘改进型Prony方法 (简称IPM)来提取邻近频率分量的频率值,并推导出两个正弦信号频率识别的封闭解。当信噪比较差时,先采用低通或带通滤波器降噪并去除直流偏置,然后采用IPM方法识别参数,大大提高参数估计精度。仿真和实验结果表明:该方法只需少量样本数据即可获得较高的识别精度,且运算实时性较高,具有很好的工程应用价值。Estimation of sinusoidal signal with adjacent frequencies is a common problem in engineering areas,such as fault di-agnosis of rotor system and communication. In view of the undamped real signal in this problem,the total least square improved Prony method was introduced to extract frequencies of signal components with adjacent frequencies. And closed form solution for two sinusoid signals was deduced. When SNR wass low,low-pass or band-pass filter was used to denoise and remove constant bias first,and then IPM was used to estimate parameters,which could improve estimation precision greatly. Simulation and experimental results show that the proposed method has high precision only using little data,and also has good real-time performance and desirable application value in engineering.

关 键 词：密集频谱 参数识别 改进Prony方法 

分 类 号：TH825[机械工程—仪器科学与技术]