基于运行状态的鼠笼电机转子导条断裂故障诊断方法研究

发布时间：2018-03-12 14:36

本文选题：鼠笼电机　切入点：转子断条　出处：《沈阳工业大学》2017年博士论文　论文类型：学位论文

【摘要】：鼠笼电机是国民经济领域应用最广泛的机电能量转换装置和终端耗能设备。电机本体故障除了会造成电机自身损坏、驱动系统突然停机之外,还会导致生产线完全崩溃,严重时甚至造成巨大的经济损失,酿成无法挽回的灾难性后果。电机带故障运行也是造成电机运行效率偏低,能源浪费严重的主要原因之一。开展以预防维护为核心的电机早期故障诊断研究对于提高电力拖动系统运行质量、保障设备工作效率、减少电机维护费用以及降低能源消耗具有重要的理论意和经济意义。转子断条故障是鼠笼电机特有的故障形式。早期转子断条故障不会造成电机立即失效,但是会导致电机运行效率降低、电能隐性流失,甚至诱发偏心及定转子刮擦等次生故障;若发现不及时,还会酿成重大人身伤亡事故。因此,该故障一直是电机故障诊断领域的研究热点之一。本论文依托中国科学院重点部署项目(KGZD-EW-302)和辽宁省科学技术计划项目(2015020140),主要以工频供电鼠笼电机转子断条故障为研究对象,针对电机驱动系统工况复杂多变的特点,采用电机电流信号特征分析(Motor Current Signature Analysis,MCSA)方法,围绕鼠笼电机早期转子断条故障诊断理论及应用技术问题展开相关研究。主要工作包括:(1)针对经典谱分析诊断转子断条故障时存在基频频谱泄漏影响诊断效果的问题,提出了基于单相定子电流信号平方解调制分析的诊断方案。区别工频和变频供电两种情况,对包括空载在内的不同负载条件下稳态运行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,采用硬件方式进行的平方解调制预处理能有效消除基频频谱泄漏和噪声干扰影响,使得用于故障诊断分析的频谱内容清晰丰富,从而增加了故障诊断确定性。相比于经典MCSA方法,所提出的诊断方法可实现鼠笼电机从空载到额定负载全负载运行范围内的转子断条故障诊断,诊断性能明显优于经典MCSA方法,更适合于稳态运行时的转子断条故障诊断。(2)针对以快速傅里叶变换为基础的非参数谱估计方法诊断转子断条故障时存在的频率分辨率低影响诊断精度的问题,提出了基于参数模型谱估计的鼠笼电机早期转子断条故障诊断方法。该方法借鉴数学变换和选频细化思想,通过对单相定子电流信号实施希尔伯特变换,消除基频频谱泄漏影响并将故障特征转移到低频段,继而通过基于离散小波滤波的选频细化功能削弱噪声影响、压缩数据处理成本,最后利用扩展Prony方法短时数据处理能力计算频谱,并对稳态运行状态下的电机转子断条故障做出定性/定量判断。采用所提出的方法对不同故障严重程度、不同负载条件下稳定运行的转子断条故障鼠笼电机进行诊断分析。研究结果表明,采用本文方法可实现短时数据条件下的转子断条故障诊断,有效克服非平稳因素可能造成的不利影响,诊断效果明显优于传统傅里叶变换。(3)鼠笼电机时变转速运行时,MCSA方法的应用条件将遭致彻底破坏并由此造成该方法完全失效。将适用于电机稳态运行时的经典MCSA方法创新性应用,解决了时变转速运行状态下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在时变转速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Park矢量模平方信号离散小波变换两种解决方案。实验对比表明,基于定子电流Park矢量模平方信号离散小波变换的诊断方法能够获得更加优越的诊断效果。(4)针对稳态MCSA方法诊断转子断条故障时诊断性能受制于电机运行状态的问题,提出基于起动电流左边频带谐波瞬时频率分析的转子断条故障诊断方法。利用FIR低通滤波器的消噪功能消除定子电流信号中的基频谐波成分并提取包含故障特征频率的单分量信号,根据左边频带谐波瞬时频率在"频率-滑差"平面上的"V"形曲线变化过程诊断转子断条故障。实验结果表明,起动运行状态下的转子断条故障诊断既可作为独立的诊断手段存在,也可作为稳态MCSA方法的必要补充。相比于稳态MCSA分析中根据单一谱线识别故障的做法,根据故障特征频率-滑差演变过程进行故障发生与否判断更具可靠性。(5)针对转子断条故障特征频率时间-尺度分析和频率-滑差分析中存在的谱图复杂、不易识别的问题,提出基于起动电流Prak矢量模平方信号Wigner-Ville分布的诊断方法。该方法利用离散小波变换的低通滤波性能滤除定子电流Park矢量模平方信号中的谐波干扰,然后通过Wigner-Ville分布描述该信号的能量-滑差变化趋势,并结合能量计算做出诊断决策。实验结果表明,采用"能量-滑差"图形比较和能量计算相结合的方式进行诊断决策具有故障特征清晰直观、可靠性好、便于工程实际应用的明显优势。
[Abstract]:The squirrel cage motor is the electromechanical energy conversion device and the terminal energy consuming equipment widely used in national economy. The fault can cause motor motor drive system itself is damaged, suddenly shut down, will lead to the production line of complete collapse, even cause huge economic losses and disastrous consequences of irreparable. Motor fault belt the operation is caused by the low efficiency of the motor, one of the main reasons of the serious waste of energy. In order to carry out research on early fault diagnosis of motor preventive maintenance as the core to improve the quality of electric power drag system, ensuring the working efficiency of the equipment, reduce motor maintenance costs and reduce energy consumption has important theoretical significance and economic significance. The rotor broken bar fault is form unique rat cage motor. The early fault of rotor broken bar will not cause immediate failure but will cause the motor, the motor is running Efficiency, power recessive loss, and even induce eccentric rotor and scraping and other secondary faults; if that is not timely, will lead to serious personal casualty accidents. Therefore, the fault is one of the hotspots in the field of fault diagnosis of motor. This paper is based on the China academy focus deployment project (KGZD-EW-302) and science technology program project of Liaoning province (2015020140), mainly in the power supply of squirrel cage motor rotor fault as the research object, according to the characteristics of motor drive system in complex conditions, analysis of the characteristics of the motor current signal (Motor Current Signature Analysis, MCSA), around the early squirrel cage motor rotor broken bar fault diagnosis theory and application technology issues related research. The main work includes: (1) according to the classic spectrum analysis frequency spectrum leakage diagnosis diagnosis of rotor effect problems of fault, put forward The stator current signal square modulation scheme based on the analysis of the solution of diagnosis. The difference between the frequency and frequency of power supply two, a fault diagnosis problem of different load conditions including no-load, steady operation of the squirrel cage motor rotor were studied. The results show that, using hardware way of square demodulation pretreatment can effectively to eliminate the frequency spectrum leakage and noise interference, the spectrum analysis for fault diagnosis of the content clear and rich, thus increasing the uncertainty of fault diagnosis. Compared with the classical MCSA method, the diagnosis method proposed can achieve the squirrel cage motor from no load to full load rotor rated load operation within the scope of the fault diagnosis, the diagnostic performance significantly better than the classical MCSA method, fault diagnosis is more suitable for steady state operation. (2) according to the number of non parametric based on fast Fourier transformation Spectrum estimation in the presence of fault diagnosis method of rotor when the low frequency resolution effect of diagnostic accuracy, the fault diagnosis method of early rotor squirrel cage motor parameter model based on spectrum estimation. The method using mathematical transform and frequency selective refinement of ideas, through the implementation of the Hilbert transform of the stator current signal, eliminate the frequency spectrum leakage and the fault feature shifted to low frequency, and then through the discrete wavelet filter based on frequency selective refinement function weaken the noise effect, the compressed data processing cost, finally calculate the spectrum using the extended Prony method of short term data processing ability, and the steady state of the motor rotor fault to make qualitative / quantitative judgment. The proposed method of different fault severity of rotor under different load conditions and stable operation of the fault diagnosis of squirrel cage motor analysis research. The results show that this method can achieve short data under the conditions of rotor fault diagnosis, effectively overcome the adverse effects of non-stationary factors may cause, diagnosis is obviously superior to the traditional Fourier transform. (3) when the squirrel cage motor with variable speed, the MCSA method should be stricken completely destroyed and thus the method complete failure conditions. The classical MCSA method will be applied to the motor steady state operation of innovative applications, to solve the fault diagnosis problem of time-varying fault early rotor speed running. Theoretical analysis and experimental results show that: through the analysis of evolution during the fault feature frequency in time-varying speed, can effectively overcome the classic the MCSA method can not analyze defects of non-stationary signal, which is still a diagnosis of broken rotor bar fault in motor transient operation. In this paper the stator current direct discrete small wave Park transform and the stator current vector modulus square signal discrete wavelet transform two solutions. Experimental results show that the diagnostic method of stator current vector modulus square Park signal based on discrete wavelet transform can obtain better diagnosis effect. (4) according to the fault diagnosis method of rotor steady-state MCSA diagnostic performance depends on the motor running state the problem, put forward the rotor current analysis of left band harmonic instantaneous frequency starting the broken bar fault diagnosis method based on noise cancellation. To eliminate the harmonic component of stator current signal using FIR low-pass filter and extract single component signal containing the fault characteristic frequency, according to the instantaneous frequency in the left band harmonic frequency slip plane. "The" V "curve of rotor faults. The experimental results show that the starting condition of rotor broken bar fault diagnosis can be used as a single There are diagnostic means of legislation, but also can be used as a necessary supplement to the steady-state MCSA method. Compared to the steady-state analysis of MCSA based on a single line fault identification approach, according to the characteristic frequency of fault slip evolution process of failure to determine whether or not it is more reliable. (5) according to the characteristics of broken rotor bar fault frequency time scale analysis slip and frequency spectrum analysis in the presence of complex, difficult to identify the problems, put forward the starting current of Prak vector modulus square signal Wigner-Ville distributed diagnosis method based on harmonic interference of discrete wavelet transform low-pass filter filter performance of stator current Park vector modulus square in the signal by using this method, and then through the description of the Wigner-Ville distribution the signal energy slip trend, combined with the calculation of the energy to make a diagnosis decision. The experimental results show that compared with the energy calculation combining the "energy - slip" graphics The diagnosis and decision making has obvious advantages of clear and intuitive fault features, good reliability and easy to use in engineering.

【学位授予单位】：沈阳工业大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：TM343.3

【相似文献】