Analysis of delays and delay mitigation on a metropolitan rail network using event based simulation

https://doi.org/10.1016/j.simpat.2015.01.002Get rights and content

Highlights

  • Analysis of the quality of metro service using event based simulations.

  • Development of delay mitigation tactics.

  • Development of an event based simulation model using Simu8.

  • Improvement of delay response and overall punctuality, hence increasing customer satisfaction.

Abstract

For a number of years, the quality of metro service provided by the Tyne and Wear Metro has been impacted by delayed services and non-punctual running. This study aims to look to combat these issues by first developing an event based simulation using Simu8 software, in order to analyse the current system’s performance and its response to delays. A number of potential delay mitigation tactics are introduced and evaluated to assess their potential worth in mitigating delays that may arise within the system. They are then analysed using the simulation model that has been developed, and compared to a system with no delay mitigation tactics input. The potential benefits and the feasibility of such tactics are then evaluated, alongside an appraisal of the scale of delay that can reasonably be mitigated through the proposed tactics alone. It is hoped that the results presented show high enough potential to be considered as a future development to the metro, which could improve both delay response and overall punctuality, in turn increasing customer satisfaction with the network.

Introduction

The Tyne and Wear Metro Network is the second largest, after London, metro system in the United Kingdom [28]. It first opened in phases in the early 1980’s around Newcastle and North and South Tyneside. It was then expanded, firstly in 1991 with a line to Newcastle International Airport, followed by a 2002 extension to include Wearside [19]. It now provides 37 million journeys a year for passengers [12], with a daily ridership of over 100,000 people [16]. It is owned by Nexus, but since 2010 has been operated by DB Regio Tyne And Wear Limited, to a specification set by Nexus [10].

Nexus conduct reviews over 4 weeks periods appraising both metro punctuality and customer satisfaction of the service provided. These reviews identify a major issue with the system, namely the punctuality and reliability of service provided. It was found that for 2012/13, metro punctuality was just 86% with customer satisfaction of this punctuality being 76% over the same period [9].

There has been minimal independent review into delays and delay impact though, and remarkably little research into how the impact of these delay can be minimised. Furthermore, there is no literature that examines how information about delays is conveyed to passengers, or how an improvement in this provision of information could improve the quality of service received by passengers. This lack of communication with passengers during a delay situation has also been identified by a number of customers, who have raised their concerns at Nexus’ ‘Meet the Manager’ sessions [17], [18].

By analysing the impact, frequency and scope of delays, and attempting to mitigate them, a major obstacle to providing the optimum quality of service for the passenger could be removed.

The first objective of this paper is to create a workable simulation of a section of metro network, which will allow the current performance of the system to be analysed. Then, possible delay mitigation strategies can also be analysed and evaluated using this simulation. Secondly, research into the frequency and types of delay that can reasonably be expected on the system, to identify delay scenarios that can be input into a simulation model to evaluate proposed mitigation tactics against.

This leads into the third objective of the study, to identify and test measures by which the impact of delays can be reduced, and investigate whether these could be practically applied on the Metro network. From this work, an analysis of the operational issues and constraints presented for any possible mitigation strategies is drawn, along with a comparison to how Nexus currently responds to delays on the system.

Firstly, background research is conducted into metropolitan railway networks and railway delay mitigation in Section 2. Then, the categories of delay are identified, along with an appraisal of their prevalence on the Tyne & Wear Metro in Section 3. Section 4 initially identifies a suitable case study area of the network to examine, before presenting the rationale behind, and results for, an observation of the Metro. Section 5 builds on the findings from Section 4 to propose effective mitigation tactics for a variety of delay scenarios. Section 6 describes how the system can be realistically modelled using simulation software, before focusing explicitly on how to model the proposed mitigation tactics in Section 7. The results found from the various simulation set ups are presented in Section 8, with a detailed comparison between an unmitigated system, and the modelled mitigation tactics also shown here. Finally, Section 9 illustrates the conclusions drawn from the previous sections, including an evaluation of the simulation model and the mitigation tactics proposed, with potential future development work also identified.

Section snippets

Mitigation of railway delays

The role of a rail network operator in responding to delays was described by Froloff as the process of returning a schedule that began in an optimal state, but was then delayed to induce a disrupted state, back closer to its original state through a series of corrective actions [7]. Carrel further expanded this sequence of corrective actions, or “interventions”, with the operators first identifying a “target state” during disruptions, before taking measures such as holding trains, cancelling

Forms of delay

For the purposes of this paper, the delay situations that occur will be separated into two categories; primary delays and knock-on secondary delays as defined by Vromans and Kroon [29].

Rationale

A physical observation of the chosen section of the metro line was required, firstly to provide data concerning the length of wait at each station and the mean travel time between stations. This data is required in order to construct an accurate simulation model, where both wait and travel times are needed to accurately replicate the physical system.

Secondly, an observation will provide the opportunity to validate the punctuality figures that Nexus reports each month. This will alleviate any

Tactics to address primary delays

As was previously identified in Section 3.1, primary delay will be considered to be unpreventable from a scheduling perspective, and lack of reliability in both the system and rolling stock will have to be accepted. However, the potential scope and impact of these delays will need to be analysed, and as such; input data for both major primary delays, such as signal failure, and minor delays, such as high passenger volume, will be required. This data will then be input into the simulation model,

SIMUL8 fundamentals

The SIMUL8 software is founded on the basis of a flow of units through a network. In the application, the units represent the individual Metrocars, and the simulation network is the previously identified case study area of the Tyne & Wear Metro System.

The simulation concept behind SIMUL8 includes objects and work items. Objects represent the physical structure of the system being simulated. Work Items represent the clients served by the system being simulated. Work items move through the

Delay situations to model

The simulation model assesses the performance of both the current metro system, and any possible mitigation strategies. Initial delays are input into the model, using three delay formats, and then the secondary delays caused are examined. Firstly, they are analysed with no delays input into the system, so any variability in the time spent in the system is caused solely by the average distribution of station wait times.

Secondly, the system will be subject to reasonably frequent minor delays,

Collection of results

The results for the simulation were collected and analysed using two unique methods; firstly, to record the average time that each Metrocar spent within the system, and to compare this value to the timings of the metro schedule. Secondly, each individual time recorded by a Metrocar travelling through the system can be recorded, and then evaluated using the Nexus definition of non-punctual arrival, not more than 30 s early, or three minutes late. The first strategy will provide a mean travel time

Conclusions

The study produced an accurate and realistic simulation model of the case study section of the Tyne & Wear Metro. The system was validated through the input of observed wait and travel times, and by comparing the mean times spent in the simulation system against the Nexus timetable. The simulation was also able to be programmed to model all the desired potential delay mitigation tactics, and provide clear, repeatable results to assess each of the mitigation tactics modelled.

The best delay

References (29)

  • M. Dessouky et al.

    A Simulation model for analyzing large complex rail network

    Simulation

    (1995)
  • M.M. Dessouky, Quan Lu, R.C. Leachman, Using simulation modeling to assess rail track infrastructure in densely...
  • E. Froloff, M. Rizzi, A. Saporito, Bases et pratiques de la régulation. RAPT, direction du Réseau Routier RC/MSE,...
  • T.W. ITA, Tyne and wear integrated transport authority – metro subcommittee meeting, committee room, Civic Centre...
  • Cited by (0)

    View full text