旋转机械故障诊断与现场动平衡系统研发

发布时间：2018-04-29 16:10

本文选题：故障特征提取 + BP神经网络　；参考：《浙江大学》2015年硕士论文

【摘要】：浙江大学化工机械研究所高速旋转机械实验受南通某公司委托开发一套旋转机械故障诊断和现场动平衡系统,用于监测大中型风机转子的运行状态,提前预测与判定设备故障类型,避免故障带来的严重损失,确保设备安全可靠运行；并对不平衡转子进行现场动平衡以减小振动。本文的主要研究工作如下：(1)研究了振动信号故障特征提取与故障模式识别的方法。针对小波包能量特征参数不具有时间特性和自适应性的局限性,引入小波包能量距作为特征参数,给出了根据转子转速调整采样频率的关系式,从而使提取的故障特征不仅具有时频特性,而且对不同转速下的同种故障具有自适应性,并利用仿真信号和不同转速的不平衡故障数据进行了验证,验证结果表明,该方法简化了小波包选择过程,很好的统一了不同转速故障信号的特征参数。实验模拟突加不平衡、不平衡、不对中、碰磨等故障,对振动数据进行时域、频域和小波包分析,提取不同故障的能量距特征参数,用于训练和测试BP神经网络,结果表明本文设计的BP神经网络具有较高的识别率。(2)研究了刚性转子现场动平衡方法,建立了不平衡振动的轴心轨迹和振动高点的计算模型,将振动高点影响系数法应用到现场动平衡中,解决了传统动平衡方法寻找振动高点的难题,单次动平衡减振率明显提高,大大缩短了现场动平衡时间。(3)研发了一套旋转机械故障诊断与现场动平衡系统。采用工控机+PLC+采集卡+传感器组+仪表组搭建系统硬件,并编写了相应的故障诊断与现场动平衡软件,实现旋转机械的状态监测、故障诊断与现场动平衡等功能。转子试验台模拟测试和现场风机转子测试结果表明,本文研发的系统可靠性高、满足设计要求、动平衡效率高达90%以上。
[Abstract]:The high-speed rotating machinery experiment of the Institute of Chemical Machinery of Zhejiang University was commissioned by a company in Nantong to develop a rotating machinery fault diagnosis and on-site dynamic balancing system for monitoring the running status of large and medium-sized fan rotors. In order to avoid the serious loss caused by the fault and ensure the safe and reliable operation of the equipment, the dynamic balancing of the unbalanced rotor is carried out in the field to reduce the vibration. The main work of this paper is as follows: 1) the methods of fault feature extraction and fault pattern recognition for vibration signal are studied. In view of the limitation of wavelet packet energy characteristic parameter which has no time characteristic and adaptability, the wavelet packet energy distance is introduced as the characteristic parameter, and the relation of adjusting sampling frequency according to rotor speed is given. Thus, the extracted fault features not only have time-frequency characteristics, but also have self-adaptability to the same fault at different rotational speeds. The simulation signals and unbalanced fault data of different rotational speeds are used to verify the proposed method. This method simplifies the selection process of wavelet packet and unifies the characteristic parameters of different speed fault signals. The experiment simulates the faults such as sudden unbalance, misalignment and rubbing. The vibration data are analyzed in time domain, frequency domain and wavelet packet, and the characteristic parameters of energy distance of different faults are extracted for training and testing BP neural network. The results show that the BP neural network designed in this paper has a high recognition rate. The field dynamic balancing method of rigid rotor is studied, and the calculation model of the axis track and vibration height of the unbalanced vibration is established. The influence coefficient method of vibration high point is applied to the field dynamic balance, which solves the problem of finding the vibration high point by the traditional dynamic balance method, and the damping rate of the single dynamic balance is obviously increased. A rotating machinery fault diagnosis and field dynamic balancing system is developed. The hardware of the system is set up by using PLC data acquisition card of industrial control computer, and the corresponding software of fault diagnosis and field dynamic balance is compiled to realize the functions of condition monitoring, fault diagnosis and field dynamic balance of rotating machinery. The results of rotor test and field fan rotor test show that the system developed in this paper has high reliability and meets the design requirements, and the dynamic balance efficiency is over 90%.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TQ050.7;TP183

【参考文献】