基于小波和神经网络的火花塞间隙识别  

Spark Plug Gap Identification Based on Wavelet and Neural Network

作　　者：张琦[1] 蒋淑霞[2] 李翔晟[2] 薛行健[2] ZHANG Qi JIANG Shu - xia LI Xiang - sheng Xue Xing - jian(School of Mechanical Engineering, Central South University of Forestry and Technology, Changsha Hunan 410004, China School of Transportation and Logistics, Central South University of Forestry and Technology, Changsha Hunan 410004, China)

机构地区：[1]中南林业科技大学机电工程学院,湖南长沙410004 [2]中南林业科技大学交通运输与物流学院,湖南长沙410004

出　　处：《计算机仿真》2017年第4期176-181,共6页Computer Simulation

基　　金：国家自然科学基金青年项目(51408616);湖南省教育厅科学研究优秀青年项目(101-4334;14B186)

摘　　要：火花塞间隙异常是造成发动机积碳、动力不足的重要原因,由于其处于气缸内部,工作存在很强的干扰,现有的技术很难实现对其准确监测。应用小波阈值分析对发动机次级点火波形降噪,将提取到的特征曲线作为BP神经网络的输入,反复训练神经网络以获得最佳网络参数,利用该网络结构分析发动机次级点火波形与火花塞间隙的对应关系。通过实验证明该研究方案对预设的火花塞间隙分类区分达到了较准确的识别效果,可实现不拆机监测火花塞间隙区间的目标,并为实现发动机不解体在线故障诊断提供理论方法。Spark plug works in the cylinder with strong noise. The unexpected gap of spark plug is an important cause of engine carbon deposition and power insufficiency. Monitoring method based on traditional technology is inaccurate. Wavelet threshold was used for the engine ignition secondary waveform signal de - noising, and the extract characteristic curve was used as the input of BP neural network. The neural network was trained repeatedly to obtain the optimal network parameters. The corresponding relation between the characteristic curve and the gap of spark plug was identified by the neural network. The experiment proves that this method has relatively higher recognition for the preset section of spark plug gap. The spark gap range can be monitored without disassemble. A theoretical method for the on -line engine fault diagnosis without disassemble is provided.

关 键 词：发动机 点火波形 小波阈值去噪 神经网络 特征提取 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] U472.9[自动化与计算机技术—控制科学与工程]