Harmonic modelling and prediction of high speed electric train based on non-parametric confidence interval estimation method

Abstract

The harmonic currents injected by high speed electric trains cause not only the pollution of power quality (PQ), but can also excite harmonic resonance in the traction power supply system (TPSS) of high speed railway (HSR). It is essential to build a proper harmonic distribution and prediction model for high speed electric trains. This paper presents an approach to obtain the probability density of harmonic current based on a non-parametric Kernel density estimation method, which can describe the probability distribution characteristics using a data-driven mechanism. A prediction algorithm for harmonic current based on confidence intervals, which can simulate harmonic current injection using a continuous sampling method, is presented in this paper. The new high speed CRH380AL electric multiple units (EMU) are used to illustrate the method. The results show that the proposed approach and algorithm can be useful for harmonic source simulation in harmonic power flow calculation and PQ assessment.