【作者】 张丽娟；

【导师】 杨新华；

【作者基本信息】 兰州理工大学 ， 电力电子与电力传动， 2009， 硕士

【摘要】 电机是现代工业生产和日常生活最主要的原动力和驱动装置。电机一旦发生故障,会造成不同程度的经济损失和社会影响。因此研究不同场合、不同运行状态下电机故障诊断理论和相关技术具有很高的实用价值。电机出现故障时,故障信号中往往含有大量的时变、短时突发性质的成分。因此可以通过检测、分析故障信号,获得电机的故障信息。传统的信号分析方法,如傅立叶变换,是一种纯频域分析,缺乏空间局部性,不能满足故障信号分析的要求。而小波分析和小波包分析法具有良好的时频局部性,能够将信号在任意频段进行划分,从而使在不同频段的各种故障特征信号更加容易被识别和提取。基于小波包分析处理非平稳信号的优越性,本文选用小波包分析对电机故障信号进行分析检测。本文在研究了异步电机常见故障类型和诊断方法的基础上,详细分析了电机滚动轴承异常、转子断条、气隙偏心等故障原因,采用基于信号分析法中的振动诊断法和定子电流检测法,对电机滚动轴承故障、转子断条故障进行诊断。对于存在已知轴承故障的电机,在故障状态下采集到振动信号,利用峭度值计算和小波包分析相结合的方法,选用db3作为小波基,进行小波包分析,对包含有故障特征频率信息的信号进行重构,获得轴承故障特征频率,根据故障特征频率的数值和能量,确定出轴承故障的类型。应用小波包分析和FFT相结合的方法,选用Coif5为小波包基,检测转子断条故障特征频率。在此基础上,采集故障电机的振动信号和电流信号,并分别应用上述方法进行了仿真模拟实验,结果表明这些方法是准确可行的。论文以DSP为核心,完成了电机故障诊断系统的硬件电路的设计,包括信号检测电路、调理电路,A/D转换电路等,并给出了主要的软件流程图。更多还原

【Abstract】 Motor is the most important motive power and drive in modern industrial production and daily life.When electric machinery is in fault,power cut will happen in some places so that there is huge economy loss and important influence on society.So the study on the fault diagnosis theory and technology of electric machinery under all kinds of instance is a pledge,which makes equipments work normally.As the fault signals of electric machine are non-stationary transient ones,the traditional signal analysis methods,such as Fourier Transform,are not efficient enough or useful for the fault signal extraction.However,wavelet analysis and wavelet packet analysis have the excellent time-frequency local performance,it can detect the different frequency components of the fault signals by its adjustable time-frequency window.Considering the superiority of wavelet packet analysis to non-stationary signals,this paper focuses on how to analyze and extract the fault signals by wavelet packet analysis.Based on the works previously,the type of the fault and the method of diagnosis in induction motors were summarized。Then a discussion was made on the fault causes of the motor bearing abnormal、rotor bar breaking、airgap eccentricity and so on.At last,adopting vibration diagnosis and current monitoring method,bearing fault and broken bars fault were diagnosed.In order to simulate true fault,the fault of the motor bearing was made by man.At first,author made use of kurtosis index to judge a bearing is in normal or not,then vibration signals were decomposed by db3 wavelet packet at sixth levels,and then the frequency band signals containing characteristic frequencies were reconstructed by the wavelet packet,thus bearing fault characteristic frequencies were obtained.According to its value and energy, the type of the bearing fault could be confirmed accurately.Meanwhile,method of wavelet packets and FFT to detect the broken bars fault signal of motor were put forward,it was proved that bearing fault and broken bars fault of asynchronous motor could be accurately detected by using these methods.In this paper,the author have designed the main hardware circuit of the induction motor fault diagnosis system based on DSP,including signal detection circuit、signal processing circuit,communication module and so on.The main flowchart of software was given finally.更多还原