本文選題：故障診斷 + 盲源提取��； 參考：《振動工程學報》2014年05期

【摘要】：利用滾動軸承發(fā)生故障時其故障信號往往呈現(xiàn)出一定周期性的特點,首先計算出故障信號的理論基本周期τ,將τ作為待提取源信號的基本周期用所述方法的相關(guān)步驟計算出大致目標源信號及權(quán)重分離矩陣^W。然后將^W作為初始權(quán)重分離矩陣,將基于高階統(tǒng)計量的固定點算法用于原始觀測信號提取出更為精確的目標故障信號。通過仿真信號和實驗信號驗證了所述方法相對于約束獨立成分分析(Constrained independent component analysis,CICA)方法具有以下優(yōu)點:不需要精確估計目標源信號的周期及不需要構(gòu)建精確的參考信號。此外,通過仿真還驗證了所述方法相對于其他較新的盲源提取方法具較高的提取精度等優(yōu)點。
[Abstract]:When the rolling bearing is malfunction, the fault signal often presents a certain periodic characteristic. First, the theoretical basic period of the fault signal is calculated. As the basic period of the source signal to be extracted, the target source signal and weight separation matrix ^W. are calculated by the relevant steps of the method described, and then the ^W is used as the initial weight division. The fixed point algorithm based on high order statistics is used to extract more accurate target fault signals from the original observation signal. The following advantages are verified by the simulation signal and the experimental signal. The Constrained independent component analysis (CICA) method has the following advantages: no need to be accurately estimated. The cycle of the target source signal and the construction of an accurate reference signal are not needed. In addition, the advantages of the proposed method are proved to be higher than other new blind source extraction methods.

【作者單位】： 上海交通大學機械系統(tǒng)與振動國家重點實驗室;
【基金】：國家自然科學基金資助項目(51035007,51105243)
【分類號】：TH165.3;TN911.7

【共引文獻】

相關(guān)期刊論文 前10條

1 蔣夕平;吳鳳凰;蔣昱;修連存;;基于FastICA算法的高光譜礦物豐度反演[J];吉林大學學報(地球科學版);2013年05期

2 張云嬌;雷斌;朱婷婷;;針對電磁干擾的濾波技術(shù)[J];電聲技術(shù);2013年12期

3 孫通;許文麗;胡田;劉木華;;基于UVE-ICA和支持向量機的南豐蜜桔可溶性固形物可見-近紅外檢測[J];光譜學與光譜分析;2013年12期

4 李春蔚;;一種改進的獨立成分分析算法的計算機仿真研究[J];電腦知識與技術(shù);2013年32期

5 納躍躍;謝，

本文編號：1808978