| 引用本文: | 闵永智,郝大宇,王 果,等.基于深度自适应K-means++算法的电抗器声纹聚类方法[J].电力系统保护与控制,2025,53(8):1-13.[点击复制] |
| MIN Yongzhi,HAO Dayu,WANG Guo,et al.Reactor voiceprint clustering method based on deep adaptive K-means++ algorithm[J].Power System Protection and Control,2025,53(8):1-13[点击复制] |
|
| 摘要: |
| 在高压并联电抗器声纹信号监测系统中，长时海量无标签声纹的高维非平稳性导致特征提取困难、无监督聚类适应性差。由此提出了一种基于深度自适应K-means++算法(deep adaptive K-means++ clustering algorithm, DAKCA)的750 kV电抗器声纹聚类方法。首先通过采用两阶段无监督策略微调的改进堆叠稀疏自编码器(stacked sparse autoencoder, SSAE)，对快速傅里叶变换后的归一化频域数据提取电抗器原始声纹32维深度特征。进一步提出了依据最近邻聚类有效性指标(clustering validation index based on nearest neighbors, CVNN)的自适应K-means++聚类算法，构建了能自适应确定最优聚类个数的电抗器声纹聚类模型。最后通过西北地区某750 kV电抗器实测声纹数据集进行了验证。结果表明，DAKCA算法对无标签声纹数据在不同样本均衡程度下能够稳定提取32维深度特征，并实现最优聚类，为直接高效利用电抗器无标签声纹数据提供了参考。 |
| 关键词: 750 kV电抗器 声纹聚类 自适应聚类算法 稀疏自编码器 深度自适应K-means++算法 |
| DOI：10.19783/j.cnki.pspc.240502 |
| 投稿时间：2024-04-20修订日期：2024-06-04 |
| 基金项目:国家自然科学基金项目资助(62066024)；甘肃省联合基金项目资助(24JRRA852) |
|
| Reactor voiceprint clustering method based on deep adaptive K-means++ algorithm |
| MIN Yongzhi1,HAO Dayu1,WANG Guo1,HE Yigang2,HE Jianshan1 |
| (1. School of Automation and Electrical Engineering, Lanzhou Jiaotong University, Lanzhou 730070, China;
2. School of Electrical Engineering and Automation, Wuhan University, Wuhan 430072, China) |
| Abstract: |
| In high-voltage shunt reactor voiceprint signal monitoring systems, the high-dimensional non-stationarity of long-term, large-scale unlabeled voiceprint data make feature extraction difficult and reduce the adaptability of unsupervised clustering. To address this, a 750 kV reactor voiceprint clustering method based on deep adaptive K-means++ clustering algorithm (DAKCA) is proposed. First, the improved stacked sparse autoencoder (SSAE), fine-tuned using a two-stage unsupervised strategy, is used to extract the 32-dimensional depth features from the normalized frequency domain data obtained via fast Fourier transform. Then, an adaptive K-means++ clustering algorithm is developed using clustering validation index based on the nearest neighbor (CVNN), and a reactor voiceprint clustering model which can adaptively determine the optimal number of clusters is constructed. Finally, the method is validated using real measured voiceprint data from a 750 kV reactor in Northwest China. The results demonstrate that the DAKCA algorithm can stably extract 32-dimensional depth features from unlabeled voiceprint data under varying sample balance conditions and achieve optimal clustering, providing a reference for the direct and efficient use of unlabeled reactor voiceprint data. |
| Key words: 750 kV reactor voiceprint clustering adaptive clustering algorithm sparse autoencoder DAKCA |
