We need to talk about availability
MRT reliability isn’t just about avoiding disruptions.
2024 is a year of project delivery; projects that will increase carrying capacity of the rail network across various MRT and LRT lines.
But how effective can these projects be? And even if they are completely delivered, it may yet take much longer than expected.
Not so far apart
What do Pasir Ris and Punggol have in common? They both have inefficient terminal designs, and thus form an overall bottleneck in the entire line.
The Pasir Ris turnback, billed as a means to increase the carrying capacity of the EWL by 20%, by increasing crossover efficiency and thus train throughput.
Is it really necessary? If considered in the context of the TEL taking over the Changi Airport branch, it might. Changi Airport was initially meant to have through service to the EWL, and the current shuttle is only that way to preserve capacity towards Tampines and Pasir Ris. Tampines has the DTL as an alternative today and Pasir Ris will have the CRL come 2030, maybe making this less important. Still, what’s built is built, and it will be possible to see increased service levels in the area. The only issue now is the trains.
As of time of writing, only 16 new R151 trains have been introduced out of 106 ordered. A further 3 are in testing, and 15 more have been delivered to Singapore but have not yet been tested. With 38 previous-generation trains already decommissioned and more not seen in service for a while, this may leave SMRT with an overall shortage of at least 22 train sets compared to the 198 they’re supposed to have.
It makes some sense to get rid of them quickly, seeing as especially the Siemens trains are the most fault-prone. But still, the rollout is not keeping pace; it may mean SMRT engineers are forced to find ways to keep old trains in service past their expiry dates which may not always be possible. Breakdowns also require more trains to be kept in reserve.
A similar story is seen on the NEL. While Punggol Coast has been sold as a rail expansion project to serve SIT and the upcoming Punggol Digital District, I can’t help but think it serves a similar role too. The 6 new NEL trains were delivered in 2021, but only entered service in the middle of 2023 — a 2 year wait at least. And the first refurbished NEL train entered service in February 2022; 2 years later, only 7 trains, less than a third of the fleet, have been refurbished. The project is expected to finish in 2026, and from now to then, 1 or 2 trainsets are out of service at any time for upgrading.
So what should the LTA do? It should prioritize delivery. Completing the R151 rollout by the end of 2026 will require them to work much faster than they are now, launching around 3 trains every month in the remaining 34 months between now and December 2026. With so many trains already delivered but not seen testing, if quality issues are found on one set, the team can move on to another set, hopefully unaffected, and come back later. Software updates will benefit all trainsets, too.
I will go as far as to say that the R151 rollout should only be paused if the kind of catastrophic fault that also merits taking the already in-service trains out of service are seen in testing — similar to the C751B gearbox failures in 2002, shortly after their introduction. Even if there are issues preventing them from running on the NSL, deploying them on the EWL anyway means an EWL train can be transferred to the NSL.
And for the NEL, whilst they may not be able to significantly speed up the project now, the project can possibly move to the SRTC come 2025; either as a second production line in addition to work done in Sengkang Depot, or even to shift the whole operation. Full-day train testing can be done at the SRTC to minimize downtime, also demonstrating capabilities in train overhaul should the LTA be interested in leasing out the SRTC facilities.
Black holed
It is not that we cannot increase service on the DTL. 92 trains were purchased; only 52 are in use every day, the remaining 40 trains remain in the depot. Is this a fair use of the taxpayers’ money spent to purchase these trains? In fact, according to SGTrains Spotters, 25 trains amongst the overall fleet have not been seen in the past month, making up more than a quarter of the overall fleet.
This is not an issue now, but will be sometime soon enough, when works take place to connect the new tracks around Tanah Merah and the new platform there, to the operational EWL. It will be highly disruptive; depending on how they plan it, there may well be periods of no EWL service at all east of the nearest crossover at Eunos.
When this happens, the DTL will have to take over as the main line for residents of Tampines and much of Bedok — though the latter also depends on whether new temporary or permanent bus routes will be introduced to connect the rest of Bedok, as well as Kembangan, to TEL4 stations. Service may need to be significantly increased to cope with large amounts of passengers diverted from EWL trains.
But how did we get here in the first place? I can think of 5 possible causes, and there may be more.
- Is it because SBST refuses to do so because of additional financial risk without more passengers?
- Is it because of parts/staff shortages or other forms of maintenance issues, that forces SBST to have to selectively manage maintenance schedules to keep trains running?
- Is it because SBST cannot find the personnel to staff every train under their existing operating practices?
- Could the LTA and SBST be taking advantage of fleet excess to take some trains out for R&D to facilitate improvements?
- Or is it because the LTA has borrowed trains to conduct testing inside the East Coast Integrated Depot, taking them out of use until the Sungei Bedok extension opens in 2025?
The last two may be reasonable excuses, considering the size of the depot (and one wonders whether the TEL may follow suit). But the first three probably need to be addressed; whether by getting SBST to adapt operating practices to facilitate more cost-effective service within their resource constraints, or to work more closely with suppliers like Alstom in establishing long-term partnerships to provide parts and technical support to keep trains running, like SMRT has.
And once all the pieces are in place, run more trains. At the very least, revert to the kind of service levels seen near opening day in 2017. After all, under NRFF v2, the LTA is supposed to share in losses; perhaps the Expo end may also get busier when Sungei Bedok station opens.
All problems are local
No one’s pretending that changing train control systems is not hard. But it does, ultimately, raise a question on just what has been happening with new LRT trains.
While 2 refitted BPLRT vehicles able to operate on the new CBTC system have re-entered service and a third is currently being refitted, not a single one of the new trains has entered service since the first two were delivered in April 2022. At least it looks like we have at least 4 vehicles in Singapore now. What has happened in the meantime, especially in the past 7 months since CBTC-fitted trains entered service? To be fair, test trains are seen moving in and out of the depot.
Giving Bukit Panjang residents reasons to ride the rails again, by restoring bidirectional service, lets policymakers adjust the amount of public transport service in the area; which means policymakers have an interest in fast project completion. As I’ve said, Bukit Panjang is a 15 minute town with the LRT; a return of bidirectional service with these new vehicles will clearly show LRT superiority over traffic stuck at traffic lights. More immediately, there may be a train shortage like on the NSEWL; new trains must enter service soon to resolve that.
Denver International Airport has also purchased similar vehicles and they hope theirs are serving passengers by summer. Might we be able to beat them to it? And even if we do, will all 19 enter service by the end of 2025, maybe even end of this year? This is necessary in order to decommission the old train control system and replace power rails by 2026, the last chapter of the BPLRT upgrade. Upgrading of existing trainsets must also finish by then; or maybe cancelled/descoped and replaced with more new trains.
In the northeast, the need for capacity expansion is also apparent; at least the LTA expects to start receiving the two-car LRT trains for Sengkang and Punggol within this year. Developments have opened up in the Punggol West Loop and near Fernvale station, creating record ridership across the Sengkang and Punggol LRT networks. There’s also space yet to be developed near Kupang and Thanggam LRT stations.
So it remains to be seen how quickly these new trains can be put into service. While the definition of a train is being changed, with completely new test procedures and other pre-service checks needing to be drawn up, at least the SPLRT is still keeping its current train control system, which should make the introduction process much easier.
There may be concerns on where the trains can be kept in the absence of the depot expansion, but if the LTA dared to commit to a 2024 delivery date, I presume they know what they’re doing. Also, perhaps, the arrangements made here may be a case to buy additional 2-car trains as well, leaving them along the LRT lines overnight should even the expanded depot be unable to fit them.
The work continues
Overall, so far, Singapore has been fortunate. Foresight in previous rolling stock orders has meant that we have a surplus of vehicles that can be used — on the newer lines, we may even be able to avoid having to buy new trains for future extensions to Changi T5 and Sungei Kadut, maybe even getting economies of scale buying them now than later. Even the Circle Line, woe as it may be, no longer has the issue of leaving only 1 spare train in the depot; something which I understand Hong Kong still has to practice.
Taipei, also, is forced to operate short turns on the Blue, Red, and Green lines because of a shortage of trains — this means things like Taipei 101 in Xinyi District getting a 10 minute interval MRT service in the off peak period. I’m not joking.
But it’s important, overall, to make sure we keep up the good work. To keep up with increasing rail usage without buying more trains, we will have to improve vehicle availability; and it seems the biggest obstacle in the short term may be the projects. This gives the operators time to invest in retraining and reskilling for new condition-based maintenance paradigms; and the projects themselves bringing new features that make operations easier.
Higher levels of rail service will also hopefully help people to overcome rail reliability PTSD, by showing them what’s possible. After all, run it, and they will come. Just like how Kuala Lumpur’s RapidKL turned itself around.
Like what you read? Join the Telegram Channel for updates!
