中文核心期刊
CSCD来源期刊
中国科技核心期刊
RCCSE中国核心学术期刊
重庆交通大学学报（自然科学版）

重庆交通大学学报（自然科学版） ›› 2014, Vol. 33 ›› Issue (5): 111-115.DOI: 10.3969/j.issn.1674-0696.2014.05.25

• • 上一篇    下一篇

基于数据驱动的数据故障诊断模型

陆百川1,2，张凯1,2，马庆禄1,2，邓捷1,2，刘权富1,2   

  1. 1.重庆交通大学 交通运输学院，重庆 400074；2.重庆交通大学 重庆山地城市交通系统与安全实验室，重庆 400074
  • 收稿日期:2013-06-04 修回日期:2013-08-29 出版日期:2014-10-15 发布日期:2015-03-10
  • 作者简介:陆百川（1961—），男，江苏南通人，教授，博士，博士生导师，主要从事交通信息工程及控制方面的研究。E-mail:hljzk1258@126.com。
  • 基金资助:
    重庆交通大学研究生教育创新基金项目（20120110）；山区桥梁结构与材料教育部工程研究中心开放基金项目 (QL2X-2012-6)

Data Fault Diagnosis Model Based on Data-Driven

Lu Baichuan1,2, Zhang Kai1,2, Ma Qinglu1,2, Deng Jie1,2, Liu Quanfu1,2   

  1. 1. School of Traffic & Transportation, Chongqing Jiaotong University, Chongqing 400074, China;2. Key Laboratory of Traffic System & Safety in Mountain Cities, Chongqing Jiaotong University, Chongqing 400074, China
  • Received:2013-06-04 Revised:2013-08-29 Online:2014-10-15 Published:2015-03-10

PDF

摘要： 针对交通和桥梁监测的数据诊断过程中，噪声掩盖了部分故障信息以及故障信息分布的多尺度性，提出了一种基于数据驱动的数据故障诊断模型。以故障检测为目的，加入了一种改进的小波阈值除噪方法，去除大部分随机高频噪声，提高了数据置信度；将重构信号进行了多尺度小波包分解，结合小波包能量分析法和主元分析法完成了故障检测与故障分离。模型实际应用于桥梁挠度监测数据故障诊断，结果表明，该模型可以减小错报率和漏报率，抗噪能力更强。

关键词: 交通工程, 数据驱动, 小波阈值除噪, 交通信息, 主元分析, 故障诊断

Abstract: During the process of fault diagnosis of traffic and bridge monitoring data, it was found that the fault information had a property of multi-scale, and sometimes part of it was covered by noise, so a data fault diagnosis model based on data-driven was proposed. In order to diagnose the fault, firstly, an improved wavelet threshold method was joined to remove most of random high frequency noises, which improved the data reliability. Secondly, the reconstructed signals were decomposed by multi-scale wavelet packet. And then, the model finished the mission of the fault detection and isolation by combining Wavelet Packet Energy Analysis and Principal Component Analysis. Finally, the proposed model was applied in a case study of bridge deflection data fault diagnosis. The results show that the model has many advantages such as lower fault and fail rate, and stronger anti-noise ability.

Key words: traffic engineering;data-driven, wavelet threshold denoising, traffic information, Principal Component Analysis (PCA), fault diagnosis

中图分类号: