-
针对电力设备局部放电(PD)超声波检测中存在的时间-空域特征解耦、硬件成本高及计算量大的技术瓶颈, 提出基于核主成分分析(KPCA)伪白化的改进型非圆FastICA (mnc-FastICA)提取TDOA/DOA参数的混合定位方法. 该方法通过时间-空域特征联合提取与智能优化机制, 实现了小规模传感器阵列下的高精度定位. 本文首先构建KPCA伪白化预处理框架, 利用多项式核函数映射信号非线性升维再降维, 在保留TDOA与DOA特征关联性的同时抑制环境噪声; 其次通过mnc-FastICA算法盲分离超声信号后, 联合广义互相关法GCC与阵列流型解析技术同步提取TDOA/DOA参数; 最后建立融合双参数的最大似然估计模型, 并引入非洲秃鹫优化算法实现全局最优解快速收敛. 实验表明, 在仅配置2个正交阵列(共8个传感器)的小规模硬件架构下, 本方法TDOA估计误差降至2.34%, DOA估计精度优于2°, 定位误差达1.54 cm. 该方法有效解决了PD检测中时间-空域特征联合、硬件成本与定位精度的矛盾, 为电力设备状态监测提供新方案 .
-
关键词:
- TDOA/DOA混合定位 /
- KPCA-mnc-FastICA /
- 局部放电 /
- 超声定位
To address the technical bottleneck of decoupling spatiotemporal feature, high hardware costs, and high computational complexity in ultrasonic detection of partial discharge (PD) in electrical equipment, this paper proposes a TDOA/DOA hybrid localization method based on kernel principal component analysis (KPCA) and modified noncircular FastICA (mnc-FastICA). By integrating spatiotemporal feature extraction with intelligent optimization mechanisms, this method achieves high-precision localization by using a small-scale sensor array. The key innovations are as follows. First, a KPCA-assisted pseudo-whitening preprocessing framework is constructed by using polynomial kernel mapping for nonlinear signal dimensionality reduction, which preserves the correlation between time delay (TDOA) and direction-of-arrival (DOA) features while suppressing environmental noise. Second, after the blind separation of ultrasonic signals via the mnc-FastICA algorithm, TDOA/DOA parameters are synchronously extracted through a combination of the generalized cross-correlation (GCC) method and array manifold analysis. Finally, a maximum likelihood estimation model integrating dual parameters is established, and the African vulture optimization algorithm (AVOA) is introduced to accelerate global optimal solution convergence. Experimental results demonstrate that with a compact hardware configuration of two orthogonal arrays (8 sensors in total), the proposed method achieves a TDOA estimation error of 2.34%, DOA estimation accuracy better than 2°, and localization errors as low as 1.54 cm. This approach effectively resolves the discrepancies among spatiotemporal feature coupling, hardware cost, and localization accuracy in PD detection, providing a novel solution for condition monitoring of electrical equipment.-
Keywords:
- hybrid TDOA/DOA /
- KPCA-mnc-FastICA /
- partial discharge /
- ultrasonic location
-
图 4 盲信号分离算法提取信号波形对比
Fig. 4. Waveform extracted by KPCA-mnc-FastICA compared with original signal.
图 6 NCC与分离信号SNR随输入SNR变化情况
Fig. 6. Variation of NCC and SNR of separated signal with different input SNR.
图 7 不同SNR的TDOA提取方法性能对比
Fig. 7. Performance of different TDOA estimation methods with different input SNR.
图 9 不同信噪比的DOA估计精度 (a) 方位角; (b) 俯仰角
Fig. 9. DOA estimation accuracy versus SNR: (a) Azimuth; (b) pitch.
图 12 不同阵列数目下的各种方法定位估计精度
Fig. 12. RMSE of positioning methods versus number of sensor array.
图 14 试验平台 (a) 示意图; (b) 装置图
Fig. 14. Schematic diagram of test platform: (a) Schematic diagram; (b) device diagram.
图 15 3个传感器阵列接收信号的盲信号分离信号 (a) 阵列1; (b) 阵列2; (c) 阵列3
Fig. 15. The extracted signal obtained by blind signal separation of the received signal from the three sensor arrays: (a) Array 1; (b) array 2; (c) array 3.
表 1 PD源和3路传感器阵列坐标
Table 1. Coordinates of Needle-plate model and three sensor array.
设备类型 坐标/cm PD源 (20, 10, 10) 传感器阵列1 (0, 30, 15) 传感器阵列2 (40, 0, 5) 传感器阵列3 (60, 20, 20) 
下载: 导出CSV
表 2 TDOA估计结果
Table 2. Results of TDOA estimation of the test.
时间差 理论时间差/ms 试验时间差/ms 误差/μs t21 0.1568 0.1530 –3.8 t31 0.2275 0.2221 –5.4
下载: 导出CSV
表 3 DOA估计结果
Table 3. Results of DOA estimation of the test.
阵列 理论方位角/( º) 试验方位角/( º) 方位角误差/( º) 理论俯仰角/( º) 试验俯仰角/( º) 俯仰角误差/( º) 1 –45 –44 1 –10 –10 0 2 153 155 2 12 11 1 3 –165 –166 1 –13 –14 1
下载: 导出CSV
表 4 局放定位结果
Table 4. Localization results of PD source.
阵列组合 阵列序号 估计坐标/cm 均方根误差/cm 1 1和2 (20.95, 11.21, 10.09) 1.54 2 1和3 (20.55, 11.79, 9.89) 1.88 3 2和3 (21.25, 12.11, 9.95) 2.45 4 1, 2和3 (21.13, 10.12, 10.14) 1.14
下载: 导出CSV
-
[1] Md-Rashid H, Refaat S, Abu-Rub H 2021 IEEE Ace 9 64587
Google Scholar
[2] 牛勃, 马飞越, 周秀 2019 高压电器 55 108
Google Scholar
Niu B, Ma F Y, Zhou X 2019 High Volt. Appr. 55 108
Google Scholar
[3] Yan B W, Chang D G, Fan Y H, Zhang G J, Zhan J Y, Shuo X J 2017 IEEE Trans. Dielect. El. In. 24 3647
Google Scholar
[4] 唐炬, 陈娇, 张晓星, 许中荣 2009 中国电机工程学报 29 125
Google Scholar
Tang J, Chen J, Zhang X X, Xu Z R 2009 Procee. CSEE 29 125
Google Scholar
[5] 李保伟, 张兴敢 2020 南京大学学报 56 917
Google Scholar
Li B W, Zhang X G 2020 J. Nanjing Univ. 56 917
Google Scholar
[6] 李易, 任明, 王彬, 席英杰, 刘阳, 董明 2022 电网技术 47 4351
Google Scholar
Li Y, Ren M, Wang B, Xi Y J, Liu Y, Dong M 2022 Power Syst. Tech. 47 4351
Google Scholar
[7] Ghosh G, Chatterjee B, Dalai S 2017 IEEE Trans. Dielect. El. In. 24 237
Google Scholar
[8] 关羽, 董明, 席英杰, 李易, 张崇兴 2023 电网技术 48 1721
Google Scholar
Guan Y, Dong M, Xi Y J, Li Y, Zhang C X 2023 Power Syst. Tech. 48 1721
Google Scholar
[9] 王玉龙, 张晓虹, 李丽丽, 高俊国, 郭宁, 程成 2021 物理学报 70 095209
Google Scholar
Wang Y L, Zhang X H, Li L L, Gao J G, Guo N, Cheng C 2021 Acta Phys. Sin. 70 095209
Google Scholar
[10] Chang Y T, Wu C L, Cheng H 2014 2014 International Symposium on Computer, Consumer and Control Taichung, China, June 10—12,2014 p1
Google Scholar
[11] Thu L N N, Tuan D V, Yoana S 2019 Sensers 19 2121
Google Scholar
[12] Jia T Y, Wang H Y, Shen X H, Gao J J, Liu X 2017 OCEANS 2017 -Aberdeen Aberdeen, UK, June 19—22 p1
Google Scholar
[13] Muhammad U L, Hafiz S M, Amna R, Zakria Q, Abbas Z K, Mahmud M A P 2021 Engrgies 14 3910
Google Scholar
[14] 王国利, 杜非, 潘伟, 王婷婷, 罗勇芬 2019 高电压技术 45 2509
Google Scholar
Wang G L, Du F, Pan W, Wang T T, Luo Y F 2019 High Vol. Eng. 45 2509
Google Scholar
[15] Xu S 2020 IEEE Commun. Lett. 24 1966
Google Scholar
[16] 谢庆, 张建涛, 陈艺丹, 刘怡, 张莹 2019 高压电器 55 1
Google Scholar
Xie Q, Zhang J T, Chen Y D, Liu Y, Zhang Y 2019 High Volt Appr. 55 1
Google Scholar
[17] Visalakshi A, Deepika R S, Chinthaginjala V R, Bagadi K, Mohammad A, Ayman A A 2022 IEEE Access 10 132875
Google Scholar
[18] Gulia S, Prasad P, Goyal S K, Raakesh K 2020 Atmop. Poll. 99 1588
Google Scholar
[19] Ning S, He Y G, Ali F, Wu Y T, Tong J 2021 IEEE Sensors J. 21 6741
Google Scholar
[20] Fokin G 2019 2019 21st International Conference on Advanced Communication Technology (ICACT) PyeongChang, Korea (South), February 17—20, 2019 p1
[21] Yue Y G, Cao L, Hu J, Cai S T, Hang B, Wu H 2019 IEEE Access 7 58541
Google Scholar
[22] Chen T, Wang M X, Huang X S 2018 2018 14th IEEE International Conference on Signal Processing (ICSP) Beijing, China, August 12—16, 2018 p1
[23] Li X L, Yi X, Zhang Z K 2021 Int. J. Antenn. Propag. 1 5512395
Google Scholar
[24] Wu P, SU S J, ZUO Z, Guo X J, Sun B, Wen X D 2019 Sensors 19 2554
Google Scholar
[25] Jiang F, Zhang Z K 2021 IET Commun. 15 802
Google Scholar
[26] Liu F, Wang R, Zhao Y J 2018 IET Commun. 15 802
Google Scholar
[27] 阮宗利, 魏平, 钱国兵, 袁晓垒 2017 电子科技大学学报 46 505
Google Scholar
Ruan Z L, Wei P, Qian G B, Yuan X L 2017 J. Univ. Electr. Sci. Tech. 46 505
Google Scholar
[28] 肖剑, 刘经纬, 胡欣, 齐小刚 2023 吉林大学学报(工学版) 54 3558
Google Scholar
Xiao J, Liu J W, Hu X, Qi X G 2023 J. Jilin Univ. Eng. Tech. 54 3558
Google Scholar
[29] Bingham E, Hyvärinen A 2000 Int. J. Neural Syst. 10 1
Google Scholar
[30] Novey M, Adali T 2008 IEEE Trans. Sig. Proc. 56 2148
Google Scholar
[31] Zhou L J, Cai J Y, Hu J J, Guo L, Liao W 2021 Sensors 18 3747
Google Scholar
[32] Pei S T, Liu D W, Ye Z J, Yang J J, Liu Y P 2023 IET Sci. Meas. Technol. 17 11
Google Scholar
下载:
计量
- 文章访问数: 267
- PDF下载量: 3
- 被引次数: 0
