QQ客服 官方微博 反馈留言

题名	复杂运行条件下鼠笼电机转子断条故障诊断方法研究
作者	祝洪宇
学位类别	博士
答辩日期	2014-05-16
授予单位	中国科学院沈阳自动化研究所
导师	胡静涛
关键词	鼠笼电机 故障诊断 转子断条 傅里叶变换 小波变换
其他题名	Fault diagnosis of broken rotor bars for squirrel cage motor oriented to complicated working conditions
学位专业	机械电子工程
中文摘要	鼠笼电机是重要的能源转换装置和量大面广的终端耗能设备。截至2012年，以鼠笼电机为核心部件的变频驱动系统已经占据了电气传动系统80%左右的份额。鼠笼电机的大量使用，使其成为重要的能源消耗大户：从全球范围看，工业生产中70%的用电量都与电机使用有关，电机用电量更是占据到了世界发电总量的50%以上。总体看来，电机带故障运行是造成运行效率偏低，能源浪费严重的主要原因之一。因此，以电机本体为核心的变频驱动系统状态监测和故障诊断对确保安全生产、降低维护费用、减少能源浪费具有重要意义。 　　　转子断条故障是鼠笼电机特有的故障形式，一直是电机故障诊断领域热点研究内容之一。长期以来，与转子断条故障诊断有关的研究主要集中在工频供电的鼠笼电机，变频器供电条件下的故障诊断方法研究并不多见。本文依托中国科学院重点部署项目（KGZD-EW-302）展开技术攻关，主要以变频器供电鼠笼电机转子断条故障为研究对象，针对变频驱动系统工况复杂多变的特点，提出了基于电机不同运行状态研究相应诊断方法的思路。在研究过程中，继承了工频供电条件下经典电机电流信号特征分析（Motor Current Signature Analysis, MCSA）方法的精髓，着力解决制约变频驱动系统转子断条故障诊断效果的方法和关键技术问题，从理论和实践两个方面进行了创新性研究。 　　　1）变频驱动系统电机电流信号模型及特征分析。将转移函数和调制理论运用到电机电流信号特征分析当中，建立了变频驱动系统电机电流信号特征分析模型。研究结果表明：在变频驱动系统中，电机定子电流信号中蕴含着丰富的谐波，但是转子断条故障特征频率分布仍然有规律可循；当电机发生转子断条故障时，反映故障特征的频率成分还会从变频器输出侧向输入侧传递，采用MCSA方法诊断转子断条故障仍然是可行的。作为对电流信号模型的验证，研究了变频驱动系统稳态和暂态运行条件下的转子断条故障诊断问题，结果表明了该模型的有效性，从而进一步丰富和发展了既有的MCSA理论。 　　　2）稳态运行条件下转子断条故障诊断方法研究。提出了基于电机电流信号解调制分析的诊断方案，解决了变频驱动系统转子断条故障诊断效果严重受制于基频频谱泄露和负载大小的瓶颈问题。区别变频驱动系统不同设定频率和电机不同负载等多种情况，对包括空载在内、不同稳态运行时的鼠笼电机转子断条故障诊断问题进行了系统研究。研究结果表明，解调制分析方法将转子断条故障诊断转化为对特征频率2sfs的分析，因而能有效消除基频频谱泄露和噪声干扰，取得明显优于经典MCSA方法的诊断效果，更适合于变频驱动系统稳态运行时的转子断条故障诊断。 　　　3）电机负载连续变化时转子断条故障诊断方法研究。将适用于电机稳态运行时的经典MCSA方法进行创新性应用，解决了负载连续变化时早期转子断条故障诊断难题。理论分析和实验结果均表明：通过对故障特征频率fb=(1-2s)fs 或2sfs在负载连续变化期间的演变过程进行分析，可以有效克服经典MCSA方法无法分析非平稳信号的弊端，从而在电机暂态运行时依然能够诊断出转子断条故障。本文提出定子电流直接离散小波变换和定子电流解析信号包络线离散小波变换两种解决方案，实验对比表明，定子电流解析信号包络线离散小波变换能够获得更加优越的诊断效果。 　　　4) 工频启动暂态运行状态转子断条故障诊断方法研究。作为对既有研究方法的改进和完善，提出了简易诊断和精细诊断相结合的转子断条故障两阶段诊断方法。所提出的方法充分利用FIR低通滤波器的消噪功能和离散小波变换的特征提取功能，根据鼠笼电机启动阶段转子断条故障特征频率fb=(1-2s)f随时间演变波形图和小波能量变化确定转子断条故障发生与否。实验结果表明，所提出的诊断方法可以有效规避电机负载变化对转子断条故障诊断效果的影响，具有节约诊断资源、工程实用性强的特点。与公开发表的国内外同类研究结果相比，所提出方法的表现出明显的性能优势。 　　　5) 变频驱动系统暂态运行条件下转子断条故障诊断方法研究。将解调制理论强大的信号预处理功能和离散小波变换优越的时频表述功能有机结合起来，提出了基于Hilbert变换和离散小波分析的电机暂态运行条件下转子断条故障诊断方法。理论分析和实验结果均表明：由于在信号预处理阶段消除了设定频率fs的影响，事实上实现了故障特征频率的完全分离，从而在变频启动、变频升速、变频降速等典型暂态运行状态下，所提出的方法都能有效诊断出转子断条故障。该方法是对变频驱动系统暂态运行条件下转子断条故障诊断方法的突破，也是本论文最主要创新点和贡献之一，具有较强的理论意义和实用价值。 本论文所做的研究基本上是围绕变频驱动开环控制系统展开的。在工业生产中，变频驱动闭环控制系统的应用领域正日益增多。与开环控制相比，闭环控制下的转子断条故障诊断面临着新的困难，主要表现在已形成的故障特征指标往往被闭环控制算法所屏蔽，因此，开环控制条件下形成的诊断方法在闭环控制中是不可靠的，甚至是无效的。在本文研究基础上，开展变频驱动闭环控制下的转子断条故障诊断方法研究，是未来一段时期主要工作内容。
索取号	TP277/Z85/2014
英文摘要	Squirrel cage induction motors are the main electro-mechanical energy conversion device of industry and most important energy-consuming equipment. By 2012, squirrel-cage motors have dominated the field of electric drive automation, featuring 80% of the motors in use. The fact that squirrel-cage motors are widely used in industrial applications make it important energy consuming device.From the world, over 50% the electric energy in the world are consumed by electric machines, which means that they consume 70% the industry electric energy or so. In general, online monitoring and fault diagnosis for squirrel cage motor is significant for assuring the continuity of the process, reducing the maintenance costs and energy waste resulting from faults. 　　　Broken rotor bar is peculiar to squirrel-cage motor, which has been the hotspot in motor fault diagnosis field for long time. The methods mentioned earlier are widely used in main-fed motors. Bibliography relative to inverter-fed motors in this topic seems to be poor. Supported by the Chinese academy of sciences key deployment project（KGZD-EW-302）, the dissertation has focused on diagnosising broken rotor bars of inverter-fed squirrel cage motor. According to the character of complicated operating conditions, the customized diagnosis methods are presented to solve specific problem. During the study, the ideas and key techniques are seriously seek to breakthough the bottleneck which restrict the diagnosis effectiveness, the dual innovations of diagnosis theory and practice are carried out based on the essence of traditional MCSA. 　　　 1) Analytic model of stator current signature for adjustable speed drived squirrel-cage motor. A simplified fault current signal model is iniatively established using switching function concept and frequency modulation theory. The theory analysis results demonstrate that broken rotor bars fault characteristic frequencies still appear in spite of the remarkable noise included into the inverter-fed motor line current, what is more, the fault characteristic frequencies will transmit from input of the inverter to supply-side so it is feasible to diagnosis broken rotor bars faults using MCSA. Following explicitly derivation from a simplified fault signal model, new online fault diagnosis techniques based on steady and transient MCSA for inverter-fed squirrel-cage induction motors is present. Experimental results show that the model and the proposed techniques are effective and accurate. 　　　2) Fault diagnosis techniques of broken rotor bars under steady condition. The diagnosis techniques based on demodulated stator current are proposed for the purpose of eliminating spectrum leakage and variable loads which prevents diagnosis effectiveness strongly. Having benefited from the elimination of spectrum leakages of fundamental-frequency component, the proposed techniques transform the characteristic frequency as 2sfs. The experiment results on main-fed squirrel cage motor under different load conditions and inverter-fed motor with different reference frequency indicate that the presented method can steadily extract fault signature frequency components and realize the incipient fault diagnosis. The theoretical and experimental results prove that the proposed method is superior to traditional MCSA method and suitable for inverter-fed motors. 　　　3) Diagnosis of broken rotor bars for squirrel-cage motors operating under continuous variable load conditions. Inspired by traditional MCSA suitable for steady analysis, creative solutions are proposed to resolve fault diagnosis of incipient broken rotor bars with squirrel-cage induction motor. The theoretical and experimental results demonstrates that the conclusion of failure or not could be drawn depending on the evolution of pattern of the characteristic frequencies of fb=(1-2s)fs or 2sfs due to variouble loads. An evident effect on diagnosising broken rotor bars under transient regime has been achieved by using proposed solutions. Experimental comparisons show that wavelet decomposition to envelop obtained from analytic signal of the stator current has better fusion performance than some traditional methods. 　　　4) Diagnosis of broken rotor bars for main-fed squirrel-cage motors operating under start-up transient regime. A two-stage method for detecting rotor bars cracks is proposed which involves pre-diagnosis and validation stage in order to improve the existing relative method. Benefiting from FIR low pass filtering and time-frequency feature extraction of DWT, the conclusion of failure or not could be drawn depending on the wavelet energy analysis and the pattern evolution of the characteristic frequencies of fb=(1-2s)f obtained from the wavelet decomposition. The proposed method, characterized by sustainable diagnostic resources and practical convenience, can avoid load fluctuations for the diagnosis of mains-fed squirrel-cage motor with broken rotor bars. Compared with that of existing relative method, the performance advantages of presented method is apparent. 　　　5）Diagnosis of broken rotor bars for adjustable speed drived squirrel-cage motors operating under transient regime. A diagnosis method based on Hilbert transform and DWT is presented, which combined powerful signal preprocessing function of demodulation and excellent time-frequency characteristics of wavelet transform. The theory and experiments show the effectiveness of presented method in variable frequency starting, acceleration or deceleration regime, due to the facts that fault characteristic frequency are completely separated from reference frequency fs which is performed during signal preprocessing period. The presented method contribute to research of broken rotor bars for adjustable speed drived motors operating under transient regime, which is the most significant breakthroughs in this dissertation. 　　　The methods presented in this dissertation are mainly used in open-loop inverter-fed motors. However, in many industrial applications, the use of closed-loop inverter-fed squirrel-cage motors increases day by day. Closed-loop control in IM drives introduces new difficulties in rotor broken bar detection, as traditional fault indicators tend to be masked by a control algorithm. Therefore, the methods developed for an open-loop operation are not able to produce adequate and reliable information on the extent of the fault and have to be adapted. Novel techniques in the near future will be oriented to diagnose machine faults under these conditions.
语种	中文
产权排序	1
页码	118页
分类号	TP277
内容类型	学位论文
源URL	[http://ir.sia.ac.cn/handle/173321/14801]
专题	沈阳自动化研究所_信息服务与智能控制技术研究室
推荐引用方式 GB/T 7714	祝洪宇. 复杂运行条件下鼠笼电机转子断条故障诊断方法研究[D]. 中国科学院沈阳自动化研究所. 2014.

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。