强噪声背景下行星轮系微弱特征信息提取及故障诊断研究

发布时间：2018-04-18 02:03

本文选题：行星轮系 + 动力学分析　；参考：《中国矿业大学》2017年硕士论文

【摘要】：像大功率采煤机这一类大型旋转机械通常工作在低速重载、强噪声背景环境下,从而导致在故障诊断过程中获取的振动信号是被强噪声背景深度污染的信噪比极低的信号,严重影响诊断的精确性。而对于大功率采煤机摇臂传动系统中的低速级行星轮系而言,其拥有更低的转频,当其发生故障时,有用的特征信息显得更加微弱。因此,如何提高极端工况下微弱故障信号信噪比就成了故障诊断领域关键课题之一。为此,本文以行星轮系为研究对象,采用随机共振方法对强噪声背景下行星轮系微弱特征信息提取及故障诊断做出了研究,具体如下所示:1)对正常以及故障状态下的行星轮系动力学特性进行了对比分析。包括采用UG软件对两级行星轮系进行三维建模,采用ADAMS软件对两级行星轮系进行动力学特性分析。分析结果表明,故障状态下时域波形中将产生冲击现象,而频域波形中边频带增多幅值增加,且故障越严重越明显。2)基于双稳态随机共振理论,提出了自适应双稳态随机共振方法。该方法采用移频变尺度对大参数信号进行预处理,使其满足随机共振要求;采用改进鱼群算法对系统参数进行同步优化,寻找全局最优值,并以改进信噪比为优化目标。采用余弦仿真信号以及行星轮系动力学仿真数据对算法的有效性进行验证。分析结果表明,该算法可以将噪声能量转移到微弱特征信号上,提高改进信噪比,且相对于EEMD以及小波阈值降噪方法而言更加的优越。3)为了研究势函数对随机共振系统输出的影响,寻找更加高效的微弱特征提取方法,基于多稳态随机共振理论,提出了自适应多稳态随机共振方法。在只改变势函数的基础上,同样采用余弦仿真信号以及行星轮系动力学仿真数据对算法的有效性进行验证。分析结果表明,多稳态系统明显优于双稳态系统,更有助于微弱特征信号的提取。4)通过实验数据对所提出的方法进行验证。这里采集了井下正常状态下的数据,用于确定井下的工作环境。在实验室采集实验台故障数据,用于验证所提出的算法。分析结果表明,所提出的方法在工程应用方面具有一定的能力。对于极低信噪比环境而言,相对于双稳态系统,多稳态系统拥有更强的处理能力。5)文章对所做工作进行了总结,并对相关的研究技术进行了展望。
[Abstract]:Large rotating machinery, such as high power shearer, usually works in low speed and heavy load, strong noise background environment, so the vibration signal obtained in the process of fault diagnosis is a signal with very low signal-to-noise ratio (SNR) polluted by the depth of strong noise background.The accuracy of diagnosis is seriously affected.For the low-speed planetary gear train in the rocker arm drive system of high-power shearer, it has lower rotation frequency, and the useful characteristic information is weaker when it breaks down.Therefore, how to improve the signal-to-noise ratio of weak fault signals under extreme conditions has become one of the key problems in fault diagnosis field.Therefore, taking planetary gear train as the research object, using stochastic resonance method, the weak characteristic information extraction and fault diagnosis of planetary gear train under strong noise background are studied.The dynamic characteristics of planetary gear trains under normal and fault conditions are compared and analyzed.UG software is used to model the two-stage planetary gear train, and ADAMS software is used to analyze the dynamic characteristics of the two-stage planetary gear train.The results show that impulse will occur in the time-domain waveform under the fault state, while the amplitude of the side band increases in the frequency-domain waveform, and the more serious the fault is, the more obvious the fault is. (2) based on the bistable stochastic resonance theory,An adaptive bistable stochastic resonance method is proposed.In this method, the large parameter signal is preprocessed with frequency shift and variable scale to satisfy the stochastic resonance requirement, and the system parameters are synchronously optimized by improved fish swarm algorithm to find the global optimal value, and the improved signal-to-noise ratio (SNR) is taken as the optimization target.The validity of the algorithm is verified by using cosine simulation signal and planetary gear train dynamics simulation data.The analysis results show that the proposed algorithm can transfer the noise energy to the weak characteristic signal and improve the signal-to-noise ratio (SNR).In order to study the effect of potential function on the output of stochastic resonance system and find a more efficient method of weak feature extraction, based on the theory of multi-steady state stochastic resonance, it is more superior than EEMD and wavelet threshold denoising method in order to study the effect of potential function on the output of stochastic resonance system.An adaptive multistable stochastic resonance method is proposed.On the basis of only changing the potential function, the validity of the algorithm is verified by using the cosine simulation signal and the planetary gear train dynamics simulation data.The analysis results show that the multistable system is better than the bistable system, and it is more helpful to extract the weak characteristic signal. 4) the proposed method is verified by the experimental data.Here the normal state of the acquisition of underground data, used to determine the working environment underground.The fault data were collected in the laboratory to verify the proposed algorithm.The analysis results show that the proposed method has certain ability in engineering application.For very low SNR environment, compared with bistable systems, multistable systems have stronger processing power. 5) in this paper, the work done is summarized, and the related research techniques are prospected.
【学位授予单位】：中国矿业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TD421.6

【参考文献】