Information processing and visualization of power system wide area time varying data

This paper presents a methodology for analyzing volumes of time varying data from wide area power systems and intelligently extracting information. The contributions are two-fold. Firstly, algorithms have been developed to identify distinct features of the system dynamic response, flagging outliers in the same process. The second contribution is in the use of advanced visualization tools such as “spark-lines” on geographic overlays to represent this information. The proposed methodology has been applied on 20 sec transient stability data in a system with over 2400 generators. The developed algorithms analyzed generator speeds and identified abnormal response characteristics in 2 generators. These abnormal responses were caused by errors in these generators' exciter models that were previously undetected and were subsequently corrected. This provides evidence of the capability and effectiveness of the proposed methodology as an operational tool.