Multiple ultrasonic partial discharge DOA estimation performance of KPCA Pseudo-Whitening mnc-FastICA

Zeyu Zhang,Xiaojun Tang, Chongzhi Liu,Xiaoshan Li, Shuangzan Ren

Measurement(2024)

引用 0|浏览0
暂无评分
摘要
Empirical studies into partial discharge (PD) localization face the difficulty of signal contamination from external noise and interference. To extract relevant information from PD signals collected by acoustic sensor array, Kernel Principal Component Analysis (KPCA) pseudo-whitening with modified noncircular Fast Independent Component Analysis (mnc-FastICA) were used. By this approach, separated matrix and observed signals can be achieved. The reconstructed array manifold matrix created from the separated matrix generated by KPCA-mnc-FastICA may be used in the Direction of Arrival (DOA) estimation approach. An angle error correction matrix is designed to compensate for the sensor's phase inaccuracy. The simulation results show that angle identification accuracy is outstanding, with mistakes limited to a small margin of 2°. Furthermore, the results of trials show that this technique can efficiently isolate individual target signals even when they are polluted with significant noise and interference. The average errors in azimuth and pitch angle are less than 2° and 1.1°, respectively. These data support the effectiveness of the approach for effectively separating signals and estimating DOA of numerous signal sources.
更多
查看译文
关键词
Partial discharge,Ultrasonic measurement,modified noncircular FastICA,KPCA pseudo-whitening,DOA estimation
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要