基于粒子滤波的齿轮箱故障诊断研究
发布时间：2018-08-11 13:37
【摘要】：本课题来源于国家自然科学基金资助项目“基于粒子群优化和滤波技术的复杂传动装置早期故障诊断研究”(项目编号:50875247)。 齿轮箱作为机械设备中最常用的传动装置,通常处于长期负载运转,由于制造、装配中存在的误差,以及疲劳、老化等效应的存在,齿轮箱在工作中总会发生故障,而它工作的状态又直接关系到整个设备的运行,因此对其进行状态监测和故障诊断有重大的意义。 粒子滤波是一种基于模型的解决非高斯非线性随机系统估计问题的有效方法。将粒子滤波应用到齿轮箱故障诊断中能解决齿轮箱振动信号的非高斯非线性问题。利用粒子滤波进行故障诊断需要知道系统的状态空间模型。文章通过建立齿轮箱振动信号的ARMA模型,用ARMA模型的参数来作为齿轮箱的状态空间模型参数。在建立ARMA模型中采用FPE准则对模型定阶,然后利用最小二乘法对参数进行估计计算。仿真验证了粒子滤波状态估计算法在信号降噪中的应用,用其对实验室采集的齿轮箱正常工况和故障工况的振动加速度信号进行降噪处理,比较分析降噪前后数据的特征值,降噪后的数据特征值都优于前者。 研究了基于粒子滤波优化神经网络的算法,在该算法的基础上建立了粒子滤波优化神经网络模型。从降噪后的齿轮箱振动信号中提取特征参量,对提取的特征参量利用粒子滤波优化神经网络进行故障识别,并取得理想效果。这同时也证明了粒子滤波信号降噪的效果是理想的。
[Abstract]:This paper comes from the project "Research on early Fault diagnosis of complex Transmission device based on Particle Swarm Optimization and filter Technology" (project No.: 50875247) funded by the National Natural Science Foundation of China. The gearbox, as the most commonly used transmission device in mechanical equipment, is usually in long-term load operation. Due to the errors in manufacture and assembly, as well as the effects of fatigue and aging, the gearbox will always fail in its work. The state of its work is directly related to the operation of the whole equipment, so it is of great significance to monitor and diagnose the state of the equipment. Particle filter is an effective method to solve the estimation problem of non-Gao Si nonlinear stochastic systems based on model. The application of particle filter to the gearbox fault diagnosis can solve the non-Gao Si nonlinear problem of the gearbox vibration signal. It is necessary to know the state space model of the system for fault diagnosis by particle filter. In this paper, the ARMA model of gearbox vibration signal is established, and the parameters of ARMA model are used as the parameters of the state space model of the gearbox. In the establishment of ARMA model, the FPE criterion is used to determine the order of the model, and the least square method is used to estimate the parameters. The application of particle filter state estimation algorithm in signal de-noising is verified by simulation. The vibration acceleration signal of gearbox under normal working condition and fault condition is processed by Particle filter, and the eigenvalues of the data before and after noise reduction are compared and analyzed. The data eigenvalues after noise reduction are better than the former. The optimization neural network algorithm based on particle filter is studied, and the particle filter optimization neural network model is established on the basis of the algorithm. The characteristic parameters are extracted from the vibration signal of the gear box after noise reduction, and the fault identification of the extracted characteristic parameters is carried out by using the particle filter to optimize the neural network, and the ideal results are obtained. At the same time, it is proved that the noise reduction effect of particle filter signal is ideal.
【学位授予单位】：中北大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：TH165.3
本文编号：2177148
[Abstract]:This paper comes from the project "Research on early Fault diagnosis of complex Transmission device based on Particle Swarm Optimization and filter Technology" (project No.: 50875247) funded by the National Natural Science Foundation of China. The gearbox, as the most commonly used transmission device in mechanical equipment, is usually in long-term load operation. Due to the errors in manufacture and assembly, as well as the effects of fatigue and aging, the gearbox will always fail in its work. The state of its work is directly related to the operation of the whole equipment, so it is of great significance to monitor and diagnose the state of the equipment. Particle filter is an effective method to solve the estimation problem of non-Gao Si nonlinear stochastic systems based on model. The application of particle filter to the gearbox fault diagnosis can solve the non-Gao Si nonlinear problem of the gearbox vibration signal. It is necessary to know the state space model of the system for fault diagnosis by particle filter. In this paper, the ARMA model of gearbox vibration signal is established, and the parameters of ARMA model are used as the parameters of the state space model of the gearbox. In the establishment of ARMA model, the FPE criterion is used to determine the order of the model, and the least square method is used to estimate the parameters. The application of particle filter state estimation algorithm in signal de-noising is verified by simulation. The vibration acceleration signal of gearbox under normal working condition and fault condition is processed by Particle filter, and the eigenvalues of the data before and after noise reduction are compared and analyzed. The data eigenvalues after noise reduction are better than the former. The optimization neural network algorithm based on particle filter is studied, and the particle filter optimization neural network model is established on the basis of the algorithm. The characteristic parameters are extracted from the vibration signal of the gear box after noise reduction, and the fault identification of the extracted characteristic parameters is carried out by using the particle filter to optimize the neural network, and the ideal results are obtained. At the same time, it is proved that the noise reduction effect of particle filter signal is ideal.
【学位授予单位】：中北大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：TH165.3
【引证文献】
相关硕士学位论文 前2条
1 李桃;基于粒子滤波技术的齿轮箱故障诊断研究[D];中北大学;2012年
2 郭姗姗;基于改进粒子滤波的红外弱小目标检测前跟踪算法[D];哈尔滨工程大学;2012年
，本文编号：2177148
本文链接：https://www.wllwen.com/kejilunwen/jixiegongcheng/2177148.html
