Fault location algorithm based on probe in Electric Power Data Network

Abstract

Fault location plays a crucial role in Electric Power Data Network (EPDNet). It usually realized by analyzing the connection between failures and symptoms. However, many symptoms may not synchronize with faulty nodes and potential failures with unknown types remain undetected in passive network management. In this paper, we propose a Facing-the-Impact-Factor (FIF) algorithm using active approach based on probes. We use the complex network theory to calculate the impact factor of each node according to all services in EPDNet. And probes are applied to detect failures on initiative, thus a real time network state can be showed in the form of matrix, then apply the impact factor to Bayesian network to determine the faulty nodes. Simulation results demonstrate the high accuracy and low false positive rate of FIF algorithm.