【作者】 刘烔；

【导师】 黄进；

【作者基本信息】 浙江大学 ， 电气工程， 2007， 博士

【摘要】 感应电机是工农业生产中广泛应用的驱动设备。随着现代工业系统的飞速发展，电机的单机容量不断增加，所驱动的负载也越来越复杂。电机故障不仅会损坏电机本身，严重时还会使电机突然停机、生产线崩溃，造成巨大的经济损失。统计数据表明，定子绕组匝间短路是感应电机最常见的故障，占其全部故障种类的15％左右。因此，研究定子故障诊断技术以在电机故障早期发现故障并及时进行维修，具有重大的理论意义和社会经济效益。通过理论分析和有限元仿真，对感应电机定子绕组匝间短路故障诊断方法进行了研究。基于零序电压、特征阻抗等对电机定子不对称十分敏感的故障特征量的分析，提出了几种有效的定子故障诊断方法。通过对感应电机断电时定子绕组上的残余电压的分析，指出由于避开了电源的影响，断电后的基波零序电压直接反映了电机定子的对称度，并由此提出了基于断电后基波零序电压的定子故障诊断方法。利用Hilbert-Huang变换分析非平稳信号的优势，对断电后的零序电压进行了经验模态分解(EMD)，提取出基波分量。针对EMD分解存在的端部效应，采用了对信号两端进行镜像延拓的方法，对其进行了有效的抑制。定义基波零序电压与相应的相电压之比为故障特征，有效的排除了断电前负载等级不同造成的影响，具有较高的诊断精度。研究了基于零序电压的定子故障在线诊断方法。对定子绕组匝间短路引起的磁动势的变化进行了分析，指出短路环电流产生的三次谐波磁动势，和正负序电流各自在三相不对称绕组里产生的三次谐波磁动势，是定子绕组里基波零序电压的主要来源。为了分析正负序分量对零序电压的影响，定义了等效电源电压，并利用Hilbert变换的特点，将等效电源电压进行了正负序分离。通过检测电机正常时正负序电压分别与零序电压的关系，以及负载对零序电压的影响，获得一个消除了负序和负载的影响、仅与故障有关的零序电压，并将它定义为故障特征。实现了感应电机定子故障的在线诊断。分析了坐标变换在电机故障诊断中的应用，用正序电压及其Hilbert变换替代同步速坐标变换矩阵里的元素，提出了改进的同步速变换方法。将电流进行坐标变换后可以准确的提取出正负序分量。以此为基础，将特征阻抗和负序电流应用到感应电机定子故障诊断中来。在电源不对称时，采用鲁棒性较强的特征阻抗检测电机故障，电源对称时则采用负序电流作为检测手段。测得正常电机的特征阻抗和负序电流并设定相应的阈值，然后根据实测值与阈值相比较来实现故障的诊断。本文提出的感应电机定子绕组匝间短路故障诊断方法均经过了实验验证，具有较高的工程实用价值。更多还原

【Abstract】 Induction motors are important drive equipments widely used in industrial and agricultural production. With the rapid development of modern industrial system, the capacity of a single motor keeps increasing and the load also becomes more complicated. The motor failures would not only result in damage to the motor, but also result in unscheduled machine downtime and the shutdown of a production line, which will cause huge financial losses. Statistical studies have shown that the stator winding interturn short circuit fault, which accounts for nearly 15% of total failures, is the most frequent fault for induction motors. Consequently, it’s of significant social and economic benefits to do the research on the stator fault diagnosis technology which can detect stator fault at an incipient stage and also allow for carefully planed repair actions.Through the theoretical analysis and finite-element simulation, this paper mainly focuses on the diagnostic method of stator winding interturn short circuit fault in induction motors. Based on the analysis of some parameters which are sensitive to the stator asymmetry, such as zero sequence voltage, feature impedance, several effective methods are proposed.As the residual voltage in the stator winding after motor switch-off avoids the influence of supply, its fundamental zero sequence component could reflect the asymmetry of the stator directly. A diagnostic method based on the residual voltage after switch off is presented in this paper. Using the superiority of Hilbert-Huang Transformation in the non-steady signal analysis, the fundamental zero sequence voltage after switch off is extracted by empirical mode decomposition (EMD). For the sake of reducing the ending effect of EMD, the mirror extending approach is adopted in the signal process. The ratio between fundamental zero sequence voltage and corresponding phase voltage after switch-off is selected as the stator fault feature, that eliminate the influence of different load level effectively, has an impressive diagnosis precision.Online diagnostic method of stator fault based on the zero sequence voltage is researched. Through the analysis of the magnetic motive force (MMF) in the induction motor with the interturn short circuit fault, a conclusion is drawn that fundamental zero sequence voltage mainly comes from the three times harmonic MMF produced by the current in the short circuit cycle and asymmetrical windings. The equivalent supply voltage is defined for analyzing the influence on zero sequence voltage by the positive and negative sequence components respectively. The equivalent supply voltage is separated into positive and negative sequence voltage by use of the characteristic of Hilbert transform. Then, through detecting the relationship between the zero sequence voltage and other sequence voltage and load, a fault feature, which removes the influence of supply and load, is defined. And a method of online diagnostics on stator faults is presented.The applications of multiple reference frames theory in the motor fault diagnosis are discussed. The positive sequence voltage and their Hilbert transformation substituted for the elements of synchronous reference frame matrix, and the improved synchronous reference frame method is proposed. The current could be decomposed to positive and negative sequence components by the improved synchronous reference frame exactly. Based on these results, characteristic impedance and negative sequence current are used to diagnose the stator fault. The robust feature impedance is selected to detect the fault of unbalanced supply, and the negative sequence current is used for detecting the condition that the supply is balanced. Through the comparison between the actual data and the thresholds, the diagnosis can be achieved.The experimental results demonstrate the effectiveness and great engineering value of stator fault diagnosis techniques proposed in this paper.更多还原