电缆故障信号特征提取方法的研究.pdfVIP

Subject : Research on Feature Extraction of Cable Fault Signal Specialty : Pattern Recognition and Intelligent System Name : Xu Jian (signature) Instructor : Wang Mei (signature) ABSTRACT With the cost reduction and the power grid upgrade both in urban and rural areas, power supply cables are widely used. Power supply cables play an important role in national economy and social development. However, a variety of faults have occurred in power supply cables in the running process, due to the issues like poor insulation and physical damage. As those cables break down, the society and economy wi suffer tremendously. Feature extraction, as a process of the fault identification, affects greatly the performance of the classifier and the accuracy of the recognition results. Therefore, the study on the feature extraction of the fault signals of power cables has important theoretical significance and practical value for the fault recognition. The fault signal characteristics of the power supply cable are firstly discussed in this thesis, and then the fault causes and the common types of power cables are introduced. In addition, the wavelet decomposition method of cable fault is analyzed，and a new method has been proposed based on wavelet packet and debris for feature extraction of cable fault signal. Secondly, Fisher criterion is adopted here to evaluate the new features generated by the above mentioned feature extraction algorithm while the configuration of the fragment threshold in the process of the fault feature extraction is discussed. Thirdly, the feat