异步电机故障诊断方法研究与应用

【作者】 朱建山；

【导师】 叶银忠；

【作者基本信息】 上海海事大学 ， 控制理论与控制工程， 2007， 硕士

【摘要】 三相异步电机以其结构简单、价格低廉、运行可靠而在传动领域占有极其重要的地位。电机的工作状态直接影响了生产的运行。对电机的常见故障进行分析及诊断，可以有效地降低故障率，减少突发事故造成的停产损失，降低维修成本，消除对人员和设备的安全威胁。故障诊断技术虽然进展显著，但误判率高、准确度低。因此，对电机系统故障诊断方法的研究有必要进一步深化和成熟。本文对主要用信号处理的方法对异步电机转子断条故障进行研究，同时对异步电机三种常见故障产生机理进行分析。所做的主要工作概括如下：（1）在综合大量国内外文献的基础上，系统地介绍了电机故障诊断技术的发展历史、研究现状、研究的主要内容以及研究的必要性。对故障诊断方法的前人研究结果进行了认真的回顾、总结，重申了故障诊断研究方法的分类，论述了当前的研究热点。（2）通过对电机常见故障总结，确定了鼠笼式异步电机的三种主要故障为：转子断条、气隙偏心、定子匝间短路，并运用基本电磁关系深入分析了故障特征频率成分的产生机理和原因，研究表明：定子匝间短路故障的最有效判据为负序电流的存在；转子断条故障最有效判据为定子电流中会出现频率为（1±2s）f₁的主要故障特征成分；当存在气隙偏心（包括静态、动态和复合偏心）故障时，会在定子电流中出现频率为f₁±mf_r的故障特征成分；上述工作奠定了本文故障分析的理论基础。（3）详细介绍为了检测定子电流中是否存在故障频率分量而采用的两种种信号处理方法：对应于电机稳态运行时的傅立叶变换频谱分析和对应于电机起动过程时的短时傅立叶变换时频分析。对于每一种信号处理方法作了理论介绍和仿真计算。（4）根据异步电机的非线性多变量动态数学模型，利用LabVIEW基于数学模型对异步电机进行了仿真研究，通过分析转子断条后电机的等效电路图，得到了故障模型。并将故障诊断方法中的信号处理方法应用于电机故障诊断实例中，对正常电机模型和故障电机模型中的定子电流信号进行处理分析，并给出两种信号处理方法的分析结果。分析结果表明，通过傅立叶变换对平稳信号进行分析，和通过短时傅立叶变换对起动过程信号进行分析，都能观察到故障分量的存在，说明这两种信号处理方法有一定的实用价值。（5）介绍了一个电机故障诊断实验系统的结构，功能，以及所包含的主要模块，并提出了各种常见电机故障的在实验系统中的设置方法。更多还原

【Abstract】 The three-phase asynchronous motor, by its simple structure, low cost, reliable operation, occupies extremely important position in the field of drive. The active status of electrical machinery has directly affects the operation of production. By using the analysis and diagnosis of the fault of the common electric motor, can effectively reduce the rate of failure, decrease the loss of Production suspension caused by any contingency matters, lower the cost of maintenance, and eliminate the threats the personnel and equipment. The fault diagnosis technology has had a notable progress, but the rate of misjudgment is high, the degree of accuracy is low. Therefore, the research of deepening and maturity the fault diagnosis of electrical system is necessary.The method, mainly based on Signal processing, of the fault diagnosis of broken rotor bars of induction motor is studied in this paper. The production of three common fault of induction motor is analyzed. The major job summarized as follows:（1）The development, the present situation, the content and the necessity of the electrical fault diagnosis research are systematically introduced on the basis of integrated in large literature. Fault diagnosis of previous research is reviewed and summed up. Reiterated classification of the method of fault diagnosis and discussed the current hotspot.（2）Through the common electrical fault summary, three main faults in the induction machines are determined: broken rotor bars, rotor eccentricity and coils short circuit in stator windings. The cause of the characteristic frequency components of fault is thoroughly analyzed by the use of basic electromagnetic theories. Studies have show when short circuit occurs in the stator, positive sequence current, as well as a certain degree of negative sequence, is the main components; frequency of （1±2ks）f₁ of current of stator can be detected if rotor bars break; frequency of （f₁±mf_r） if rotor eccentricity occurs. This work has laid the theoretical foundation of the fault analysis.（3）Two signal processing methods to test fault frequency components in the stator current are detailed introducedThe Fourier transform analysis are used when the motor run on the steady-state and short-time Fourier transform frequency analysis are adopted in starting period. The theory and simulation are given for each signal processing method.（4）The model of asynchronous motor is found by means of LabVIEW base on the nonlinear, multi-variable, strong-coupled dynamic mathematical model. By means of this model, simulation is done. The fault model of motor are obtained through analysis the electrical circuit diagrams of broken rotor. The signal processing method in fault diagnosis has been applied to the motor fault diagnosis. The stator current in normal motor and in fault motor are analyzed and the results by two signal processing methods are obtained. The results show that can observe the presence of fault component through the Fourier transform to the steady-state signal analysis and the short-time Fourier transform to starting signal analysis. The two signal processing method has practical value（5）An experimental motor fault diagnosis system is introduced. The construction and the functions of such system are reported. And the methods of set up common fault of motor are brought up.更多还原