Research on Fault Diagnosis Technology for Power Grid Equipment Based on Spark

Abstract

With the long-term operation of the power grid equipment, the reliable operation state of the equipment is declining, and the probability of failure of the power grid equipment will gradually increase. In order to ensure the safe and stable operation of the power system, it is necessary to timely diagnose and analyze the power grid equipment. This paper put forward the method of fault diagnosis for power grid equipment based on Spark, and realizes the parallelization of support vector machine and K-means algorithm, which are the traditional fault diagnosis algorithm. Then, this paper takes the transformer as the empirical object, uses the transformer oil dissolved gas data to carry on the transformer fault diagnosis. Experiments show that with the increasing of data, compared with the single environment, parallel fault diagnosis method and can remain a similar accuracy rate, and has a lower running time. It has been proved that in the face of big grid data, the power grid equipment fault diagnosis technology based on Spark has better ability of equipment fault diagnosis.