- 學位論文
結合小波轉換與類神經網路辨別電力開關之切換
Identification of Power Switching Using Wavelet Transform and Neural Network
摘要
近年來由於高科技產業與電腦設備等精密儀器之迅速發展，相對的對電力品質的要求也愈來愈高。不良之電力品質除對製造工廠會造成重大損失外，嚴重時亦會使機器設備損毀，因此如何改善電力品質是電力公司與用戶皆要認真面對的課題。由電力開關切換所造成之電力系統暫態現象，亦是電力品質中重要的一環。因此，瞭解此暫態現象之特性及對電力系統之影響，是值得重視與研究。 本文是針對負載開關及電容器組在系統中切換時所產生之暫態現象進行分析研究。藉由電力系統模擬軟體產生不同之暫態干擾訊號，如負載開關之投入或切離、電容器組之投入或切離、電力開關發生動作時電壓之不同相角及其位於量測點位置之上游或下游，並結合小波轉換與自我組織映射網路作為分析之工具，對上述各種暫態訊號做辨識之工作。 本論文利用兩個模擬之電力系統作為測試對象，由測試結果找出最佳之小波係數組合，並由結果顯示本論文使用之方法的可行性。
並列摘要
Recently, power quality is becoming an increasingly important topic due to the high technology industry and precision instruments developed rapidly. The poor power quality will make the serious loss of the manufactory and sometimes will damage the equipments. Hence, improvement of power quality is an important task for utility companies and their customers. Power system transients caused by switching is one of the important problem of power quality. Hence, it is worth to understand the property of transients and the influence upon power system. This thesis focuses on the transients, which are caused by the load breaker switching and the capacitor bank switching. Several kinds of transients disturbance signals, such as load breaker switching on/off, capacitor bank switching on/off, the different phase angles of voltages while the switch actions, and the position of the switch, are simulated by the power system software EDSA. This thesis identifies these transient disturbance signals, by the other two analysis tools, which are the wavelet transform and the SOM network. This thesis studies two different power systems, which were simulated by EDSA, and finds the wavelet coefficients for identification. The study results show that the proposed method is applicable.