考虑小波奇异信息与不平衡数据集的输电线路故障识别方法

发布时间：2018-06-15 16:38

本文选题：输电线路 + 故障类型识别　；参考：《中国电机工程学报》2017年11期

【摘要】：鉴于输电线路故障识别中数据集的非均衡性问题,提出一种基于小波奇异信息和改进合成少数类过采样(synthetic minority over-sampling technique,SMOTE)算法的输电线路故障识别方法。首先,通过PSCAD/EMTDC仿真构造输电线路故障不平衡数据集,结合平稳小波变换(stationary wavelet transform,SWT)与奇异值分解(singular value decomposition,SVD)技术提取相电流及零序电流的故障分量的小波奇异值作为特征参数,然后采用改进SMOTE算法在少数类的样本中心邻域进行插值再抽样处理,调整数据集的不平衡度,利用优化后的数据集训练支持向量机(support vector machine,SVM)组合分类器,对不同故障工况下的10种输电线路故障类型进行分类识别。仿真结果表明,该文的方法能有效地提高分类算法在样本数据不平衡的情况下对少数类的识别能力和整体的识别准确率,具有较好的泛化性和较强的鲁棒性,并且对多种分类算法同样适用。
[Abstract]:In view of the disequilibrium of data sets in fault identification of transmission lines, a new method for fault identification of transmission lines based on wavelet singular information and improved synthesis of a few kinds of over-sampling synthetic minority over-sampling technique (SMOTET) algorithm is proposed. Firstly, the unbalanced data set of transmission line fault is constructed by PSCAD / EMTDC simulation, and the wavelet singular value of fault component of phase current and zero sequence current is extracted as characteristic parameter by using stationary wavelet transform and singular value decomposition (SVD) technique. Then, the improved SMOTE algorithm is used to interpolate and resample the sample center neighborhood of a few classes, to adjust the unbalance degree of the data set, and to train the support vector machine (SVM) combined classifier by using the optimized data set. Ten types of transmission line faults under different fault conditions are classified and identified. The simulation results show that the proposed method can effectively improve the recognition ability and the overall recognition accuracy of a few classes when the sample data is unbalanced, and has better generalization and robustness. It is also applicable to many classification algorithms.
【作者单位】：武汉大学电气工程学院;
【基金】：国家自然科学基金项目(51277134)~~
【分类号】：TM75

【相似文献】