中文摘要：

齿轮传动系统是保障机车安全稳定运行的最重要的关键装备之一,其运行状态具有时变性、不可预知性和动态联动性等特点,采用传统故障诊断方法进行故障特征获取仍然存在误诊、漏诊等现象。稀疏盲分离是一种能够在信号传输通道有限的情况下,依据正交基映射将多元非线性信号有效分离的软计算方法。但是在实际工况中,机车齿轮故障数据往往是微弱性和不确定性的,从而导致稀疏分离后的源信号特征无法准确诊断故障。因此,提出一种基于变尺度经验模态分解的自适应时变盲分离方法,利用稀疏化处理和迭代筛选进行分离获取故障源,通过调整时间跨度获取最优本征模态函数,删除冗余因素,有效提高故障特征识别准确率。通过仿真试验数据验证,进一步表明了该方法在低信噪比状态下快速准确获取故障特征的有效性,能够为铁路运输的状态检测和故障诊断提供关键技术。

英文摘要：

Gear transmission system is one of most important key equipment to guarantee safe and stable operation in locomotive. With time variation, unpredictability and dynamic linkage, the problem of fault feature acquisition still be misdiagnosis by traditional fault diagnosis method. Sparse blind source separation is a kind of soft computing method based on orthogonal basis mapping to effectively separate multiple nonlinear signals under signals transmission channel unknown. However, gear fault data in actual working status are weak and uncertainty, thus causing source signal characteristics cannot accurate diagnosis fault after sparse separation. Therefore, a method of adaptive time-varying blind separation based on variable metric empirical mode decomposition（VMEMD） is presented, which is using sparseness and iterative screening to separate fault source and obtain optimal intrinsic mode function by adjusting time span. Redundancy factors are deleted and fault recognition rate is improved. The analysis through simulation experiments shows that fault feature can be obtained quickly and correctly on low signal to noise ratio which provides key technology for state detection and fault diagnosis of railway transportation.