A real-time intelligent abnormity diagnosis platform in electric power system

With the rapid development of smart grid, intelligent electric meters can be seen in most of the households, and the volume of electric energy data is in a rapid growth. This paper mainly aims at introducing an abnormity diagnosis platform in electric power system. It is used to distinguish the abnormal point according to the historical data and expert experience, and put forward some resolving scheme to ensure the high reliability and stability of power grid. In our approach, we use distributed technologies to process big electric energy data. Specifically, distributed fie system (HDFS) and distributed database (HBase) are applied to data storage, and distributed computing technology (MapReduce) is applied to constructing knowledge base and computing. In the inference engine, we use Hidden Semi-Markov Model. This model can auto-get and modify knowledge in knowledge base, achieve a better real time phenomenon, through self-learning function and machine as well as interacting between human. The results show that this abnormity intelligent diagnoses platform is effective and faster.