Auckland, Derailed – The Documents

Last month Kiwirail and Auckland Transport surprised us by announcing nearly 3 years of rail closures starting at Christmas in order to fix the foundations that sit under the tracks. Yesterday Kiwirail released a bunch of reports and ministerial briefings from over the last 3 years covering the issue.

I haven’t read all of the documents in detail but here are some of the most interesting things I picked up from skimming through them.

The issue has been known about for a long time

This work to fix the rail network is known as the Rail Network Growth Impact Management project (RNGIM). It’s origins stem from 2019 when Auckland Transport commissioned a business case “to assess the case for change, optioneering and value for money of ‘Catch-up Renewals’” on the Auckland metro network.

A high level infrastructure review in May 2019 found the track was in poor condition and that while Kiwirail knew about the issues, noted:

Whilst KiwiRail has experienced and competent track maintenance personnel, the efficiency of the work undertaken is compromised by the lack of track access time and resources. This has the result of programmed work not being finished within the track time allocated and the deferment of some works.

From evidence gathered, the effectiveness of KiwiRail’s maintenance strategy is arguably sufficient for the existing network operations. However, given the scale of the maintenance task ahead and the limited availability of track time and skilled resource, we believe that there is a high risk of network deterioration to below an acceptable condition that warrants further intervention now.

In our view, it is clear that the existing asset condition, resourcing and maintenance strategies will not be sufficient for the reliable and, ultimately, safe operation of planned future rail services in the Auckland area.

Concerningly, faults on the network were getting worse, averaging close to 120 per month by early 2019.

Temporary Speed Restrictions (TSRs) as a result of faults appear all over the network but appear to be concentrated in a number of locations. By mid-2019 they were averaging about 50 per month and covered over 8% of the network – and both of these figures were increasing.

All up, an analysis of train delay data between July 2016 and June 2019 found 411,000 hours of delays “with a root cause attributable to maintenance and renewals practices“. There’s also a potentially 52k hours that could potentially fall into this category.

That infrastructure review fed into the business case which came up with the option to overhaul the network as well as improvements in technology, equipment, training and processes for ongoing maintenance. They estimated this would cost $183 and had a benefit cost ratio of 5.4, which is better than most big road projects. This work was anticipated to take about four years from 2020 but the project was put on hold when it was discovered just how bad the Rolling Contact Fatigue (RCF) issue was in August 2020.

If sitting on this knowledge about the state of the network for three years sounds bad, worse is it seems most of the issues were known about at least as far back as 2014. In a paper to Kiwirail board in April 2021 – where staff asked for approval of an additional $140 million for the project – in a section talking about the root cause of the RCF issue they say:

‘The closest single root cause could therefore be stated as a missed opportunity during 2014-2017 to implement the recommendations of the 2014 Network Rail Consulting report. Appropriate responses would have been to:

1. Upgrade the track asset OR
2. Modify or operate the units in a way that specifically recognised the track was not upgraded OR
3. Introduce a rigorous grinding regime OR
4. Study and implement the lowest TCO approach that optimised all three issues.’

That 2014 report was commissioned by AT in advance of the introduction of the new EMU fleet

That board paper also highlights an interesting feedback loop issue to do with the RCF. One of the issues identified was the wheel profile on the EMUs causing Wheel Rail Interface issues which contributed to the RCF. However, the paper notes that the wheel profile was modified from Kiwirail’s standard “to avoid anticipated high wheel flange wear” as a result of high vehicle stiffness. But the high stiffness designed into the trains was so that they could better negotiate Kiwirail’s track irregularities safely and comfortably.

The extra $140 million was primarily to cover an increased project scope and more accurate estimated costs. This brought the BCR down to 2.3 – it’s interesting they revised it when Waka Kotahi never seem to with their mega state highway projects, even if they double (or more) in cost.

Ministerial Briefings

It appears the Minister of Transport was told about this issue at least as early as August 2021 and the documents released include briefings in October 2021 as well April, July, August and September 2022.

In the October-21 paper they some scope on the RINGM project, with Kiwirail comparing scale of work as the equivalent to rebuilding a third of the Auckland network. They suggest they would need about 1060 work days, which if they carried delivered it at their current pace of about 42 days per year, would take till 2041.

Kiwirail’s initial plan was to deliver the programme over 5 years from mid-2021 and finishing at the end of 2026

With a lot more planned access and moving progressively around the network, KiwiRail is confident that RNGIM could be completed in 2026, without the need for a full network shutdown. The initial focus would be to prioritise upgrades to the core network which are critical to CRL before it begins operation, then move onto the outer parts of the network.

It seems they don’t count the Western line as part of the core of the network as work on it wouldn’t have started till after the CRL opened. That’s crazy given the Western line benefits the most from the CRL and carries the most passengers, though as a Western line user it also doesn’t surprise me as services to the west are almost always de-prioritised over Southern and Eastern lines if there is a conflict.

To achieve that 2026 timeframe, they were investigating a couple of options alongside AT.

The approach could include a combination of:

1. AT’s initial idea to close lines or sections of line, possibly for months at a time, to allow RINGM and any relevant NZUP work to be completed outright. This would require commuter services to be replaced by buses but would avoid longer lasting disruption and operational changes for commuters. Once reopened, the lines would be up to standard and without speed restrictions, which currently impact AT’s timetables.
2. Closing lines during the day between peaks to carry out work. This would also presumably require bus replacement and is not as efficient as complete closures, given the time needed for staff to set up and decamp. This could result in works taking a lot longer.
3. There is potential on parts of the double-tracked network to close one line for sections, diverting trains onto the other line. This single line running is similar to Stop/Go on the road network and allows one side of the track to be closed for construction or maintenance work while diverting traffic to the other side to keep the network open. This would support the continuation of metro services in some parts of the network, albeit with likely timetable changes. This is the approach taken with crossovers being installed between Papakura-Pukekohe over Christmas, which will allow freight services to continue operating while P2P and Drury Stations work is carried out over the next few years. As part of the existing programmes, additional infrastructure is being installed at strategic locations on the network to allow for more single line running operations.

In April 2021 they note they’re close to an agreement on the timing of the project. One thing that stands out is a comment that they planned to still run some services on the eastern line during its 8-month shutdown but only during the peak and only in the peak direction. This was dropped from the final plan.

The next major update is the September briefing which confirms the plan that was then announced in October. A few interesting paragraphs from the briefing are below:

15. Auckland Transport (AT) will be using a mix of bus replacement and existing bus services to help manage commuter disruption. Given the general inconvenience to the public in stopping regular rail services and potential capacity issues with bus replacements and existing bus services, we do expect to see a drop in public transport use in specific areas as people choose to work from home, cycle or drive instead.
16. However, we expect this will be limited to the period of works and AT will be encouraging people back to trains once work is finished in each stage. Given the renewals will remove existing Temporary Speed Restrictions, we expect more reliable services across the network over the next few years, prior to the start of CRL operations.
17. KiwiRail and AT have agreed the timing of the first two stages of work, with exact timing of the remaining stages yet to be finalised. The focus is on completing the inner and other key parts of the Auckland Metro Network before CRL begins operating in 2025. The Manukau Branch Line, Newmarket Line (which needs to be used for CRL commissioning but will see train frequency drop from 18 to four trains an hour once CRL begins operating) and the Swanson – Henderson section of the Western Line, which has lower commuter use, are expected to be completed by Q3 2025.

It feels like AT are being overly optimistic that people will just come back once the works are finished. The reality is the rail network still hasn’t recovered to the to the same level as buses and ferries since those works with previous users seemingly changing if or how they travel.

It’s also notable that they say the Newmarket Line will only see four trains per hour after CRL opens. We’re yet to see any official post-CRL network plans.

The do include this table which shows how many people will be impacted and some of the alternatives for people for each station.

There are also some comments about some of the other work that will happen around the network while the shutdowns occur.

19. Where possible, work stages are timed to coincide with other projects, to avoid creating further disruption later. For example:
    1. While the Eastern Line is closed, AT is co-ordinating its work to install a new pedestrian underpass near the Meadowbank train station.
    2. Renewals on the Southern Line between Papakura and Pukekohe will be undertaken while the line is already closed to metro commuter services for the electrification project. This means it will already be up to standard when commuter trains again run on the line from late-2024, cementing the improved level of service on this section of line. The building of new motorway overpasses in the area will also be undertaken during the line closure.
    3. The Western Line will be broken into sub-stages and will be aligned with installation of a new power feed at Glen Eden, funded through KiwiRail’s Rail Network Investment Programme (RNIP) – which will help reduce commuter disruption. We will also prioritise the section between Newmarket and Kingsland, so it can be used to support transport to events at Eden Park.
    4. As noted above, Henderson – Swanson and the Newmarket Line may need to be left to post CRL opening. The decision on when to open CRL will include consideration of the RNGIM works. During renewals work in each stage we will also do other maintenance, where possible. This could include, replacing sleepers that may need to be replaced in the next year or two, or remove trees that we can see being a problem later. This is about thinking ahead and trying to avoid the need to come back and cause further disruption in the coming years.

I suspect the underpass mentioned near Meadowbank is for the connection from Gowing Dr to the Eastern Path – I note that the path is below the rail line at that point.

As someone who will be affected by the Henderson to Swanson section being delayed, I find it quite insulting that they just dismiss this section as being low use. In 2018-19 the three stations of Sturges Rd, Ranui and Swanson combined had over 700k boardings. That’s equivalent to being about the 5th busiest station on the network.

The paper also covers the issue of partial services on lines being worked on.

22. KiwiRail and AT have investigated alternatives to full commuter rail closures – for example allowing peak rail services to the CBD in the morning and back in the evening. However, given the overhead electric cables will need to be turned on for EMU services this presented an additional safety risk to track workers. KiwiRail freight trains, Te Huia and the Northern Explorer are diesel and do not require the overhead cables to be switched on. In addition, running peak daily EMU services introduces further operational complexity, which would extend completing renewals to key parts of the network beyond the start of CRL.

Finally, there’s a paper on how the modelling they used to come up with the proposed plan. It seems to suggest that they had options which would have completed the works up to six-months earlier but dismissed them as it would involve working on multiple lines at the same time so had higher costs and customer impact. It also includes this image which shows just how much work is needed for the different types of issues they’re fixing on each section.

There’s a lot of information in these docs so let me know if there’s anything interesting or important I’ve missed.

Share this