Electromagnetic Interference Diagnosis Based on HPLC Timing Sequence Topology and Machine Learning

Abstract

High power line communication (HPLC) has become a basic communication method in the courts of smart grid areas. Some special electrical equipment in these power grid areas may cause electromagnetic interference, which will undermine the stability of HPLC, and lead the acquisition equipment in the smart grid to miss out or lose control. Consequently, the power supply quality and grid safety will be affected. This kind of influence is temporary, random, and self-recoverable, and it is difficult to be identified and eliminated through on-site inspection. In this paper, the HPLC network topology records of the courts are analyzed, the HPLC timing sequence topology is fitted, and user information and electricity consumption information are correlated. Machine learning and optimization are carried out on the data of the courts by using the association rule learning and decision tree algorithm. It is found that the machine learning method based on the HPLC timing sequence topology can effectively realize the electromagnetic interference positioning of the power grid in the courts, and help the on-site operation and maintenance personnel to conduct troubleshooting.