The impacts on freight train operational performance of new rail infrastructure to segregate passenger and freight traffic

TL;DR: In this paper, the authors analyzed the freight impacts of the new chord, focusing on three key operational measures (i.e. train routing, scheduled journey times and train punctuality) during 10-week survey periods before and after the opening of the chord.

About: This article is published in Journal of Transport Geography.The article was published on 2017-01-01 and is currently open access. It has received 6 citations till now. The article focuses on the topics: Rail freight transport & Traffic management.

Summary

1. Introduction

• European transport policy favours a much increased role for rail in meeting the growing requirement for both freight and passenger movement (European Commission, 2011).
• International rail freight corridors are under development as part of the European Rail Network for Competitive Freight, concentrating on improving service quality through a focus on infrastructure capacity and performance (DG MOVE, 2011).
• There is no financial or cost-benefit analysis in the public domain but, importantly for the project’s justification, the chord has removed a bottleneck by providing a freight route independent of the busy East Coast Main Line (ECML), where capacity had been limited by the mix of fast passenger trains and slower freight trains (Network Rail, 2011).
• Section 3 then presents the background to the study, both regarding trends in recent British rail network activity and, specifically, the context for the North Doncaster Chord itself.
• Section 8 ends the paper, setting out the key conclusions and the wider applicability of the research approach.

2. Literature review

• Both cost and service quality are important in determining which mode of transport will be selected for freight flows.
• A survey of existing and potential rail freight customers in Britain (ORR, 2012) identified that on-time delivery was ranked as the second most important service attribute after cost/price.
• Compared with passenger transport, where research regularly investigates the impacts of network changes on people’s trip-making, journey opportunities and travel behaviour (see, for example, Bjarnason, 2014; Shaw et al., 2014), there is little detailed investigation of the impacts of infrastructure enhancements on freight transport activity, particularly within rail.
• Capacity is influenced by infrastructure, traffic and operating characteristics (Abril et al., 2008).
• The research presented in this paper aims to further the understanding of this important subject.

3. Detailed study context

• An increase in rail’s share of both the freight and passenger markets in Britain over the last 20 years was noted in Section 1.
• By contrast, passenger train kilometres have grown considerably but at broadly only half the rate of the increase in journeys and passenger kilometres.
• In response to the growing demand, there is a considerable investment programme in the British rail network.
• Other recently implemented schemes 7 which provide greater flow segregation and fewer conflicting train movements include the Nuneaton North Chord (Network Rail, 2012) and the Ipswich Chord (Network Rail, 2014c).
• The upper diagram in Figure 1 shows the indicative routing prior to the use of the North Doncaster Chord, with coal trains coming in from Immingham through Thorne (to the right of the diagram) and looping round to end up at Drax, the largest of the three Aire Valley power stations.

4. Research methods

• The analysis in this paper is based on a comparison of real-time freight train data1 of the situation ‘before’ and ‘after’ the opening of the North Doncaster Chord.
• For the ‘before’ period, data were gathered for 1,029 trains, though in 25 cases there was missing information relating to departure time, arrival time and/or intermediate routing.
• Unfortunately, this was impractical for a number of reasons relating to the limitations of the real-time data.
• Firstly, a higher proportion of empty trains did not operate directly from power station back to Immingham, unlike the vast majority of loaded trains which did operate directly from port to power station.
• Rakes of empty wagons often travel to an intermediate yard (e.g. at Doncaster or Milford Junction) after discharging their load at the power stations, either for maintenance or for staging until needed for the next loaded trip.

5. Comparative analysis of ‘before’ and ‘after’ survey periods

• This section presents the results from the ‘before’ and ‘after’ survey periods relating to the three key assessment criteria: train routing, scheduled journey times and train punctuality.
• The results are presented for the flows from Immingham to each power station separately and overall for the group of Aire Valley power stations; given the lack of any trains to Ferrybridge power station in the ‘after’ survey period, the latter are indicative only and need to be treated with caution.
• As such, it is more meaningful to separately compare the ‘before’ and ‘after’ survey results for Drax and Eggborough power stations.

5.1 Train routing

• As discussed in Section 3, the freight-related case for the North Doncaster Chord was predicated on the pre-existing route from Immingham to the Aire Valley being that shown in the upper diagram in Figure 1.
• Three of these combinations accounted for 61 per cent of trains and eight of them in total accounted for less than 3 per cent.
• By definition, all trains avoiding the ECML in the ‘after’ period used the North Doncaster Chord, though four of these continued via the Leeds area for weekend stabling and therefore did not travel directly to their destination.
• 14 Table 4 shows the change in weighted average distances between the ‘before’ and ‘after’ survey periods.
• The shortest route in the ‘before’ survey was 62.96 miles and the longest was 112.16 miles, so the longest was almost 80 per cent further than the shortest.

5.2 Scheduled journey times

• 2015a) 15 For both the ‘before’ and ‘after’ survey periods, Table 5 summarises the average scheduled journey time between and each of the power stations as well as the overall weighted average.
• Those trains that were staged at an intermediate location have been excluded from the calculation due to the lengthy layover times, so the sample sizes are slightly reduced.
• Author’s surveys based on data from realtimetrains.co.uk; n = 990, n = 963 Scheduled journey time reductions of the magnitude of 8 to 11 per cent were identified when comparing the ‘’ study to the ‘’, also known as Source.
• Excluding the ‘before’ Ferrybridge services from the analysis has a minimal impact, the overall reduction increasing from 8 to 9 per cent in this case.
• The weighted average journey time reduction was marginally less than the weighted average distance reduction and, in comparing Tables 4 and 5, the scheduled average speed decreased very slightly (and not significantly) from 18.1 mph ‘before’ to 17.9 mph ‘after’; this may have resulted from a reduction in the distance travelled on the section of the higher speed ECML.

5.3 Train punctuality

• In addition to interpreting the train schedules themselves, it is important to assess how well the trains performed in reality (i.e. comparing the actual to the planned) since this is closely linked to network capability and resilience.
• For the ‘before’ survey period, 83.4 per cent of trains met the FPM threshold and even the relatively poorly performing Ferrybridge arrivals slightly exceeded the national 17 FPM.
• When comparing the two periods it must be borne in mind that, broadly speaking, the proportion of trains arriving 10 minutes or more late reduced from one in six to just one in ten .
• Some of these trains appeared to be delayed during stopovers in the Doncaster yards, where driver and locomotive changeovers often take place, rather than on the ECML itself, so the genuine delays incurred on the ECML may well have been even less.

6. ‘After’ survey: analysis of North Doncaster Chord versus alternative routings

• This section further develops the analysis of the ‘after’ survey data.
• In Table 8, the results for the key measures of average scheduled journey time, percentage of trains arriving within 10 mins of schedule and average delay per train incurred en route (for trains experiencing an en route delay) are summarised for those trains routed via the North Doncaster Chord and those not.
• The average scheduled journey time for all trains was 03:42, while it was 03:24 for those using the most direct routing via the North Doncaster Chord (Drax 03:23, Eggborough 03:24).
• While the scheduled average speed for trains using the direct route via the North Doncaster Chord in the ‘after’ survey was 17.9 mph, the same as the overall ‘after’ average, 92 of the trains using the North Doncaster Chord had scheduled times that were quicker than the fastest by an alternative routing.
• The relatively small number of trains arriving 10 minutes or more late combined with the location of initial delay for the different routing options suggests that these figures should be treated with caution.

7. Implications of study findings

• In broad terms, the outcomes of the research are in line with expectations, with clear operational improvements resulting from the new infrastructure.
• The investigation has revealed some interesting findings that deserve more detailed discussion.
• These can be categorised as relating to: 21 The complexity of train routing Meeting customer requirements: implications for cost and train performance .

7.1 Complexity of train routing

• Fundamentally, the complexity of train routing combinations in both the ‘before’ and ‘after’ survey periods is an interesting outcome which challenges the picture set out by Network Rail in its publicity material relating to the justification of the North Doncaster Chord.
• In part, the message portrayed is understandable, since it is not easy to represent the myriad routing combinations in a clear manner, but it does not accurately reflect reality.
• Specifically, it is perhaps surprising that only 64 per cent of trains in the ‘after’ survey made use of the new, direct routing afforded by the North Doncaster Chord.
• The extent of this constraint is not documented formally, but discussions with staff from two of the freight operating companies (FOCs) operating over the route have confirmed that this short single track section does limit operational flexibility.
• In some cases, the alternative routes pass train crew depots, locomotive maintenance facilities and freight yards (e.g. in the Doncaster area) so it can be operationally convenient to use these routes where drivers or locomotives need to be changed or where en route train staging is required to meet customer demands or scheduling constraints.

7.2 Meeting customer requirements: implications for cost and train performance

• As discussed in Section 2, the literature identifies cost and service performance, notably ontime arrival, as key customer requirements.
• While perhaps not as time-critical as other freight flows (e.g. fast-moving consumer goods for supermarkets), on-time rail freight service performance is an important issue in the electricity supply industry.
• According to DfT (2008), the coal is generally unloaded and taken straight into the furnace, although stockpiles are often available to cope with any supply disruption.
• As noted earlier, perhaps the key reason for the investment was improved capacity and performance for passenger trains on the ECML.
• It is not possible to ascertain from published data any specific effect of the North Doncaster Chord on this improved performance and it was beyond the scope of this study to gather real-time data for passenger trains on the ECML.

7.3 The role of exogenous factors

• Longer-term analysis of the operational impacts, particularly with regard to train routing, journey times and service performance, would be worthwhile.
• The longer the time period the more it is likely that exogenous variables will affect the case study flow.
• These specific changes are part of a bigger restructuring of the electricity supply industry in the United Kingdom, which is having a major effect on coal flows.
• On the Immingham to Aire Valley corridor, the focus of this study, by early-2016 just Drax was being served and the number of biomass trains considerably exceeded coal services.
• A final point relates to the potential to use the North Doncaster Chord for traffic other than the key flow upon which the freight element of the project justification was built.

8. Conclusions

• This paper set out the objective to evaluate the key rail freight operational impacts resulting from the implementation of the North Doncaster Chord.
• The use of data at the level of the individual train, with a consistent methodology for both the ‘before’ and ‘after’ survey periods, has allowed an evidence-based evaluation of a number of important criteria relating to operational performance.
• This evaluation has demonstrated improvements in train routing, scheduled journey time and train punctuality.
• It is clear that what appears to be a fairly straightforward bulk rail freight flow actually displays considerable complexity, notwithstanding the wider changes taking place in the electricity supply industry.
• This paper has demonstrated the value of publicly available real-time rail freight data for assessing the operational performance impacts of a new infrastructure initiative, with obvious potential for such data to be used in other studies which require an evidence base.

Figures

Table 4: Weighted average distance (in miles) in ‘before’ and ‘after’ survey from Immingham to power stations

Table 3: Percentage of trains using East Coast Main Line (ECML) in ‘before’ and ‘after’ survey periods

Figure 2: Extract of real-time information for a specific train

Table 6: Average delay per train incurred en route (hh:mm:ss)

Table 7: Location of initial delay for services arriving late (by 10 minutes or more) at power station

Figure 3: Actual vs scheduled arrival times at destination

Table 2: Composition of ‘before’ and ‘after’ surveys, by number of trains to each power station

Figure 1: Indicative rerouting on core corridor from Immingham to Aire Valley Power Stations

Table 5: Weighted average scheduled journey times (in hh:mm) between Immingham and power stations

Frequently asked questions

Q1. What are the contributions in "The impacts on freight train operational performance of new rail infrastructure to segregate passenger and freight traffic" ?

This includes the North Doncaster Chord, opened in June 2014 primarily to provide a more direct route from the port of Immingham to the major Aire Valley power stations ( i. e. Drax, Eggborough and Ferrybridge ). The paper analyses the freight impacts of the new chord, focusing on three key operational measures ( i. e. train routing, scheduled journey times and train punctuality ) during 10-week survey periods before and after the opening of the chord. The use of this empirical method to assess the detailed rail freight operational impacts is an important element in the process of evaluating the effects of network enhancement. The results demonstrate improvements in each of the three operational measures, but also reveal a situation considerably more complex than that suggested by the published material relating to the justification for this new infrastructure. 

Q2. What other attributes were ranked as being of high importance?

Other attributes ranked as being of high importance included network access, overall service quality and flexible service/recovery strategy. 

Q3. What is the importance of comparing the actual and scheduled journey time?

In addition to interpreting the train schedules themselves, it is important to assess how well the trains performed in reality (i.e. comparing the actual to the planned) since this is closely linked to network capability and resilience. 

Q4. What is the reason for the growth in network activity?

The growth in network activity is exacerbating the conflicts that arise from the operation of a mixed traffic railway (i.e. one that caters for both passenger and freight traffic) due, for example, to the incidence of flat junctions between converging or diverging routes and to speed differentials between varied types of train. 

Q5. What is the cost of maintaining and improving the rail network?

Network Rail, the organisation with responsibility for the rail infrastructure, is authorised to spend £38 billion by 2019 to maintain and enhance the network (Network Rail, 2014b). 

Q6. What is the point of delay that was a direct bearing on a delayed arrival of 10?

The point of initial delay that had a direct bearing on a delayed arrival of 10 minutes or more has been allocated to one of five categories: late departure from Immingham; delays en route prior to reaching the ECML;18delays en route in the vicinity of the ECML itself; delays in the section between the ECML and the mainline connections to the power stations; and delays on the final approach to the power stations themselves. 

Q7. What is the impact of the port-hinterland container market?

Woodburn (2013) assessed the impacts on rail freight efficiency of network enhancements targeted at the port-hinterland container market, identifying considerable improvements in on-train capacity and train loads within this specific market. 

Q8. How many trains were missing in the survey?

For the ‘before’ period, data were gathered for 1,029 trains, though in 25 cases there was missing information relating to departure time, arrival time and/or intermediate routing. 

Q9. Why do competitive rail freight markets pose challenges for in-depth research?

Competitive rail freight markets, increasingly common in Europe and elsewhere, pose challenges for in-depth research such as this because of commercial sensitivities. 

Q10. How far was the shortest route from Doncaster to Eggborough?

The shortest route in the ‘before’ survey was 62.96 miles and the longest was 112.16 miles, so the longest was almost 80 per cent further than the shortest. 

Q11. What is the average scheduled journey time for all trains?

The average scheduled journey time for all trains was 03:42, while it was 03:24 for those using the most direct routing via the North Doncaster Chord (Drax 03:23, Eggborough 03:24). 

Q12. How much lower was the coal volume in the last four quarters?

Using the most up-to-date statistics at the time of writing, rail-borne coal volumes were 54 per cent lower in the last four quarters (i.e. 2014/15 Q4 to 2015/16 Q3) than in the same four quarters two years earlier.