|Alternative Title||Fault diagnosis of rolling element bearing for induction motor based on MCSA method|
|Keyword||异步电机 滚动 轴承 故障诊断 特征提取 Mcsa|
|Place of Conferral||沈阳|
电机电流信号特征分析方法（Motor Current Signature Analysis, MCSA）具有信号获取方便的优势，易于构成非侵入式电机状态监测与故障诊断系统，已成为异步电机故障诊断技术的研究热点。论文主要以异步电机轴承外圈故障为研究对象，基于定子电流信号特征分析，深入研究了强噪声背景下的轴承微弱故障特征提取方法，取得了具有创新性的成果。（1）基于电机结构特征分析的无速度传感器转速估计方法。基于电机结构特征的齿谐波分析是广泛采用的一种无速度传感器转速估计方法，一直是众多学者的研究热点。然而近些年的研究实践表明，在电机转子槽数和极对数的某些组合情况下齿谐波不会出现，导致这种转速估计方法失效。针对这一问题，论文提出引入电机偏心谐波分量进行估计转速。针对工频供电情况下的转速估计问题，提出了一种基于定子电流Teage–Kaiser能量算子（Teage–Kaiser Energe Operator, TKEO）解调的转速估计方法。所提出的方法能够将定子电流频谱中的偏心谐波分量转化为直接与转速相关的谐波分量，有效削弱了基频频谱泄露对提取转子速度谐波的影响，可快速实现转速的准确估计。针对变频驱动情况下的转速估计问题，利用调制理论和转移函数建立了变频器供电侧电流理论模型，提出了一种基于变频器供电侧电流信号电机偏心谐波分量分析的转速估计方法，解决了利用定子电流信号估计转速易受到电力电子器件噪声影响的问题。不同负载运行状态下实验结果表明，依据变频器供电侧电流信号估计电机转速具有更高的可靠性和鲁棒性。（2）基于平方包络线和瞬时频率的滚动轴承故障诊断方法。论文针对电机低负载运行状态下通过提取定子电流信号包络线诊断轴承故障时故障特征不易提取的问题，提出了一种构造定子电流平方包络线信号代替包络线信号的轴承故障诊断方法；利用轴承故障在定子电流信号中引起的相角调制作用相比较于幅值调制作用更加明显的特点，提出了一种新的基于定子电流信号瞬时频率分析的轴承故障诊断方法。实验结果表明上述两种方法均可以将要检测的故障特征频率从传统的边频带成分f1±fof转化为轴承外圈故障振动特征频率fof，削弱了基频频谱泄露和供电系统噪声干扰的影响，能够在不同负载运行条件下有效诊断出轴承外圈故障。（3）结合模态分解和统计分析的滚动轴承故障诊断方法。异步电机复杂工作环境中存在负载波动、电网波动和电网谐波等干扰因素，导致定子电流信号并非严格意义的平稳信号。利用自适应白噪声的完整集成经验模态分解方法（Complete Ensemble Empirical Mode Decomposition with Adaptive Noise, CEEMDAN）对信号中微弱瞬态分量的良好分解性能，论文提出了一种结合CEEMDAN和统计分析的轴承外圈故障诊断方法，对定子电流信号进行时频分解不仅能够清晰地捕捉到定子电流基频分量，而且能够有效提取出包含轴承外圈故障特征的模态分量。为了量化不同轴承外圈故障严重程度水平，提出采用该模态分量的方差、均方根和峰峰值这三个统计特征作为故障严重程度评价指标。实验结果表明，上述方法增强了滚动轴承微弱故障特征的识别能力，并能够对轴承故障严重程度做出有效评价。（4）齿谐波对轴承外圈故障异步电机定子电流影响。现有定子电流诊断轴承外圈故障的磁场特性研究中，均假设定转子光滑，忽略了齿谐波对电机内部磁场产生的影响。论文开展了考虑齿谐波和固有偏心谐波共同作用下的轴承外圈故障电机磁场特性研究，给出了与齿谐波相关的轴承外圈故障特征频率分量在电机定子电流信号中的表达形式。实验分析表明，将齿谐波影响考虑在内，可在定子电流频谱中找出更多与轴承外圈故障有关的边频带故障特征频率，有助于更有效地进行轴承故障诊断。
Motor current signature analysis (MCSA) has the advantages of easy signal acquisition, effective reduction of monitoring cost and easy formation of the non-intrusive motor continuous condition monitoring system. It has become a research focus for diagnosis of IM bearing faults in recent years. The paper mainly takes the IM bearing outer race fault as the research object. Using the stator current signal analysis, the method of extracting the weak fault characteristics of the bearing under the background of strong noise is deeply studied, and innovative results have been achieved in both theory and practice. (1) Research on the IM sensorless speed estimation methods based on machine saliency. The sensorless speed estimation methods based on machine saliency have been widely used. But recent studies have demonstrated that the rotor slot harmonics do no exist for some of the pole pairs and rotor slots, as it would result in the invalid of these methods. To address this issue, the speed estimation methods using the eccentricity harmonics have been introduced in this paper. For the speed estimation of the mains supplied IMs, a method using the stator current based on the TKEO (Teager-Kaiser Energy Operator, TKEO) has been proposed. Using the proposed demodulation method, the fundamental frequency component is eliminated, thereby reducing the adverse effect of its spectral leakage in the detection of the rotor speed harmonic, and improving the rotor speed accurate estimation. For the speed estimation of the inverter-fed IMs, a method using the supply side current based on Hilbert transform has been proposed. Thus, the proposed method can effectively solve the problems that the speed estimation using the stator current usually suffers serious noise pollution, which is caused by the power electronics. The experimental results show that the speed estimated using the supply side current can achieve higher reliability and robustness. (2) Research on the fault diagnosis methods of the rolling bearing based on the square envelope and instantaneous frequency. In the existing research, the bearing fault diagnosis is determined by analyzing the stator current signal envelope, which is difficult to extract the fault features when the IM operates under low load conditions. The paper proposes a bearing fault diagnosis method for constructing the stator current square envelope signal instead of the envelope signal. With the use of the phase modulations is more obvious in the stator current than the amplitude modulations, a bearing fault diagnosis method based on instantaneous frequency analysis is proposed. The proposed method can be used to detect the characteristic fault frequency fof directly instead of detecting the sideband components around the supply frequency f1±fof as used in traditional MCSA. Thus, the proposed approach can significantly reduce the negative influence of the spectral leakage of the main frequency component spectral and the power supply noise. Experimental results show that the proposed methods are effective for bearing outer raceway fault diagnosis under different load conditions. (3) Research on fault diagnosis method of rolling bearing combined with mode decomposition and statistical analysis. For the disturbance factors such as load fluctuation, grid fluctuation and grid harmonics in the complex working environment of IM, the stator current is not a strictly stationary signal. By taking advantage of Complete Ensemble Empirical Mode Decomposition with adaptive noise (CEEMDAN) to extract the small components from the original signal, this paper has presented a bearing fault diagnosis method combined with CEEMDAN and statistical analysis. Using CEEMDAN algorithm to decompose the stator current not only can capture the fundamental frequency component clearly, but also effectively extract the mode function including the bearing outer race fault characteristics. In order to quantify the severity level of different bearing outer ring faults, the three statistical values of the mode function, such as variance, root mean square and peak-to-peak value, are further obtained. The experimental results show that the proposed method not only enhances the recognition ability of the bearing weak fault characteristics, but also provides an effective evaluation of bearing faults with different severity. (4) Research on the effects of the slot harmonics on the stator current in an induction motor with bearing fault. For the existing magnetic field characteristic research regarding bearing fault diagnosis based on MCSA, the smooth stator and rotor are assumed, while the slot harmonics generated by the stator and rotor slots can have an impact on the motor internal magnetic field. In this paper, the magnetic field characteristic has been studied when considering not only inherent eccentricity but also the slot harmonics in an induction motor with bearing outer race fault. The influence of the slot harmonics on the theoretical expression of sideband series to characteristic fault frequencies is given. Some experimental investigations have proved that some new characteristic fault frequencies can be founded around the rotor slot harmonics in the stator current spectrum, these new characteristic frequencies expression can be used as effective assessment indicators for bearing fault detection.
|宋向金. 基于电流信号特征分析的异步电机轴承故障诊断方法研究[D]. 沈阳. 中国科学院沈阳自动化研究所,2018.|
|Files in This Item:|
|基于电流信号特征分析的异步电机轴承故障诊（3901KB）||学位论文||开放获取||CC BY-NC-SA||Application Full Text|
|Recommend this item|
|Export to Endnote|
|Similar articles in Google Scholar|
|Similar articles in Baidu academic|
|Similar articles in Bing Scholar|
Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.