| 基于判别字典学习的电能质量扰动识别方法 |
| |
| 引用本文： | 沈跃,张瀚文,刘国海,刘慧,陈兆岭. 基于判别字典学习的电能质量扰动识别方法[J]. 仪器仪表学报, 2015, 36(10): 2167-2173 |
| |
| 作者姓名： | 沈跃 张瀚文 刘国海 刘慧 陈兆岭 |
| |
| 作者单位： | 江苏大学电气信息工程学院镇江212013 |
| |
| 基金项目： | 国家自然科学基金（61301138）、江苏省博士后科研资助计划（1401053C）、江苏大学高级人才启动基金(10JDG136)项目资助 |
| |
| 摘 要： | 电能质量扰动识别方法通常是先通过数字信号处理工具对信号进行检测和特征提取,再采用人工智能方法对特征进行分类识别,增加了识别过程的复杂性和冗余性。提出一种基于判别字典学习(DDL)的稀疏表示电能质量扰动识别方法,可有效减少识别步骤、降低复杂性,并提高识别率。该方法首先采用主成分分析方法将K类扰动训练样本集降维为扰动降维特征训练样本集,由各类样本分别训练出冗余子字典,然后级联成判别字典。接着基于l0范数算法求解出降维测试信号在该判别字典下的稀疏表示矩阵,最后利用不同的冗余子字典重构测试样本,由冗余残差最小值确定目标归属类,实现对电能质量扰动信号的识别。仿真实验结果表明该方法能有效地对不同电能质量扰动进行识别,过程简单、数据量少、抗噪声鲁棒性好,在信噪比20 d B以上的噪声环境中电能质量扰动识别准确率达到95%以上。
|
| 关 键 词： | 电能质量；识别；稀疏表示；字典学习；重构算法 |
Power quality disturbance identification method based ondiscriminative dictionary learning |
| |
| Shen Yue,Zhang Hanwen,Liu Guohai,Liu Hui,Chen Zhaoling. Power quality disturbance identification method based ondiscriminative dictionary learning[J]. Chinese Journal of Scientific Instrument, 2015, 36(10): 2167-2173 |
| |
| Authors: | Shen Yue Zhang Hanwen Liu Guohai Liu Hui Chen Zhaoling |
| |
| Affiliation: | School of Electrical and Information Engineering, Jiangsu University, Zhenjiang 212013, China |
| |
| Abstract: | Traditional power quality disturbance identification methods usually use digital signal processing tools for signal detection and feature extraction first, and then use artificial intelligence methods to classify the features, which increases the complexity and redundancy of identification process. This paper presents a method based on the sparse representation of discriminative dictionary learning (DDL) for power quality disturbance identification, which can reduce the identification procedures and complexity, and improve the recognition rate effectively. At first, the K class disturbance training sample set is dimension reduced to disturbance feature training sample set with the PCA method, and various kinds of samples are trained to obtain the redundant sub dictionaries, which are cascaded to form the discrimination dictionary. Then, the sparse representation matrix of the disturbance feature training sample set in the discrimination dictionary is solved based on the l0 norm algorithm. Finally, the test samples are reconstructed using different redundant sub dictionaries, the object class is determined through minimizing the residual error between test sample and its sparse representation, and the power quality disturbance identification is achieved. Simulation and experiment results show that the proposed DDL method can effectively identify different power quality disturbances, and features simple identification process, small data and good anti noise robustness; with this method the power quality disturbance recognition rate reaches above 95% in the noisy environment with the SNR above 20dB. |
| |
| Keywords: | power quality identification sparse representation dictionary learning reconstruction algorithm |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《仪器仪表学报》浏览原始摘要信息 |
|
点击此处可从《仪器仪表学报》下载免费的PDF全文 |
|
