Stochastic Process and Simulation of Traction Load for High Speed Railways

Abstract

The traction load of a high-speed railway (HSR) may cause significant voltage fluctuations in the traction power supply system and the utility grid. To evaluate power quality, modeling traction load is an urgent need as well as a fundamental work. This paper aims to build a probabilistic model to describe the pattern of traction load changing with time. It is very challenging due to the randomness of train behavior. This paper first investigates the stochastic process of traction load and points out that the size and frequency of load fluctuation, respectively, rely on the slope and length of the rail ramp. Based on this, train power (TP) and headway (TH) are set as two basic components of the probabilistic model, where TP and TH are described by a normal distribution and an exponential distribution, respectively. In addition, a sliding time-window method is developed for generating time-dependent load samples. The proposed method has low complexity, as well as can consider the randomness, giving more realistic simulation results. It can be used to forecast traction load for a planning railway or to establish a measurement-based model for an operating railway. A case study is performed for parameter identification and model verification, and the rapid voltage change (RVC) caused by the HSR loads has been evaluated in different operation scenarios.