A Method Based on Linear Prediction for Inrush Current Detection

The occurrence of inrush currents in power transformers is directly related to the reduction in its lifetime, as well as safe operation of electric power systems. In this article, the linear prediction technique is used to process three-phase current signals and detect the inrush current. The linear predictors of orders 2 and 4 are implemented, and the linear prediction error is designed to detect the phenomenon. The methods performance is evaluated with several different levels of the inrush current simulated in alternative transients program, allowing the creation of a database containing 285 current signals, and the modeled electric system is based on real distribution network. Aiming to evaluate the reliability of the method, Gaussian noise is added to the signals, and a noise sensitivity analysis is performed. The results show the viability of the linear prediction as a tool able to identify the inrush current, which performed successfully in situations with SNR of 55, 60 and 65 dB.