• CN:11-2187/TH
  • ISSN:0577-6686

›› 2009, Vol. 45 ›› Issue (12): 232-236.

• 论文 • 上一篇    下一篇

扫码分享

基于希尔伯特变换的核电站松动部件定位方法

杨将新;郑华文;曹衍龙;程实;方力先;谢永诚   

  1. 浙江大学现代制造工程研究所;杭州电子科技大学自动化学院;上海核工程研究设计院
  • 发布日期:2009-12-15

Estimation Method for Impact Location of Loose Parts Based on Hilbert Transform

YANG Jiangxin;ZHENG Huawen;CAO Yanlong;CHENG Shi;FANG Lixian;XIE Yongcheng   

  1. Institute of Manufacturing Engineering, Zhejiang University School of Automation, Hangzhou Dianzi University Shanghai Nuclear Engineering Research & Design Institute
  • Published:2009-12-15

摘要: 针对当前的核电站松动部件定位方法在估算信号到达时间差时存在起始点难以准确判定、误差大、抗噪声性能差的缺点,提出一种基于希尔伯特变换的时间差估计方法,从而根据时间差估算松动部件位置实现定位。希尔伯特变换可提取信号轮廓线,以轮廓线的第一个峰值到达时间作为能量最大弯曲波到达传感器的时间;分析信号的主频率可以计算出弯曲波的传播速度,根据波速和信号到达不同传感器之间的时间差使用三角形定位法实现松动部件的定位分析。搭建平板试验台,在不同信噪比条件下和不同方法之间进行对比试验,结果显示这种方法的平均估计偏差为34.1 mm,在信噪比为1 dB时仍然能够得到比较准确的结果,具有定位精度高、抗噪声性能好、运算速度快等特点。

关键词: 包络提取, 定位, 松动部件, 希尔伯特变换, 多目标优化, 径向基模型, 抗撞性, 热成形, 梯度性能

Abstract: In view of the shortcomings of the present positioning method of loose parts of nuclear power station when estimating the signal arrival time difference, such as difficult judging the starting point accurately, big error and poor anti-noise performance, a time difference estimation method based on Hilbert transform is proposed, thereby the positioning can be realized by estimating the location of loose parts on the basis of the time difference. The Hilbert transform can extract the signal envelope, and the arrival time of the first peak of envelope is taken as the time of the biggest energy bending wave arrival to the sensor. The propagation speed of bending wave can be calculated by analyzing the main frequency of signal. On the basis of the wave speed and the signal arrival time difference between different sensors, the triangle positioning method is adopted to realize the positioning analysis of loose parts. A flat-plate test bench is built, and comparison tests are carried out under different signal-to-noise ratio (SNR) conditions and between different methods. Test results show that the average estimated deviation of this method is 34.1 mm, and even when the SNR is 1 dB, it still can obtain an accurate result. This method features high positioning accuracy, good anti-noise performance and high operating speed.

Key words: Envelope extraction, Hilbert transform, Loose parts, Positioning, Crashworthiness, Hot Forming, Multi-objective optimization, Radial basis function, Tailored property

中图分类号: