Modeling the Influence of Disturbances in High-Speed Railway Systems

Abstract

Accurately forecasting the influence of disturbances in High-Speed Railways (HSR) has great significance for improving real-time train dispatching and operation management. In this paper, we show how to use historical train operation records to estimate the influence of high-speed train disturbances (HSTD), including the number of affected trains (NAT) and total delayed time (TDT), considering the timetable and disturbance characteristics. We first extracted data about the disturbances and their affected train groups from historical train operation records of Wuhan-Guangzhou (W-G) HSR in China. Then, in order to recognize the concatenations and differences of disturbances, we used a K-Means clustering algorithm to classify them into four categories. Next, parametric and nonparametric density estimation approaches were applied to fit the distributions of NAT and TDT of each clustered category, and the goodness-of-fit testing results showed that Log-normal and Gamma distribution probability densities are the best functions to approximate the distribution of NAT and TDT of different disturbance clusters. Specifically, the validation results show that the proposed models accurately revealed the characteristics of HSTD and that these models can be used in real-time dispatch to predict the NAT and TDT, once the basic features of disturbances are known.

Prediction of train arrival delays considering route conflicts at multi-line stations

“…Generally, the results show that the delay distribution presents heavy-tailed distribution forms, meaning short delays have extremely high frequencies, whereas long delays have relatively low frequencies. The Weibull distribution (Goverde et al, 2013;Goverde et al, 2001), log-normal distribution (Huang et al, 2019;Wen et al, 2019;Wen et al, 2017), and exponential distribution (Briggs and Beck, 2012) are the most commonly used distribution models for delay durations. Also, clustering models (Cerreto et al, 2018;Huang et al, 2019) can be used to separate delays into different categories, contributing to investigating the delay distribution detailly.…”

“…The Weibull distribution (Goverde et al, 2013;Goverde et al, 2001), log-normal distribution (Huang et al, 2019;Wen et al, 2019;Wen et al, 2017), and exponential distribution (Briggs and Beck, 2012) are the most commonly used distribution models for delay durations. Also, clustering models (Cerreto et al, 2018;Huang et al, 2019) can be used to separate delays into different categories, contributing to investigating the delay distribution detailly. In another application, the probability model was used to infer delay evolutions; given the distributions of departure delays at the previous station and the running time distributions in the previous section, the arrival delays were predicted using a statistical inference method (Lessan et al, 2018).…”

Predictive models for influence of primary delays using high‐speed train operation records

“…In addition, Goverde, Corman, and D’Ariano (2013) found that Weibull distributions can be fitted to the PD distribution using empirical data, whereas Wen et al (2017) noted that PD distributions could be well approximated by log‐normal distributions, while line regression models can be used to approximate NAT distributions. Huang et al (2019) first used a K ‐means clustering algorithm to classify NAT and TTAT due to PD into four categories, and then fitted an optimal distribution model based on historical data from Wuhan–Guangzhou HSR. Results showed that the log‐normal and gamma distribution had better performance.…”

Cascading dominates large-scale disruptions in transport over complex networks

“…Related literature aims to understand and predict the evolution of delays in transport systems, both under regular and disrupted circumstances. Most delay evolution models, however, focus on regular circumstances and predict how delay fluctuations develop using high-resolution statistics obtained from particular incidents or scenarios [20,22], or from particular stations [23,24] or lines [25][26][27]. These models come in various forms, mainly in the context of air and railway transport: analytical [28], agent-based [29], stochastic [30][31][32][33] and purely data-driven [3,34].…”