Solving the polycrisis of rail

The Rail Reliability Taskforce’s report is a breath of fresh air, but yet predictable.

In the wake of a series of rail disruptions last year, a taskforce was convened to develop solutions and approaches to improving rail reliability. Which is the same tired trope many of us are accustomed to hearing. After all, has the advice of the last task force been implemented?

Something may seem different this time. But it will only make sense if the solutions proposed can reasonably be carried out, in a reasonable amount of time. And whether they actually turn out to be effective. The practical implications of what then happens after the task force’s recommendations are implemented, must also be discussed.

I sense in the report that they, too, agree that the current issues before us is indeed a crisis of integration; not only in asset management and system design, but also where the knowledge lies in our rail workforce. Recommendations are also made to improve “commuter management”, basically disruption handling. Additionally, action plans were prepared to push forward system renewals and how to do it, which is also something that must be looked at.

The task force’s report can be read here.

Work smarter, not harder

The task force calls for the prioritization and acceleration of renewal projects for core rail systems such as trains, control systems, and power supply.

I say they should tread carefully. Precedent has shown that they just might try to run hordes of buses past closed MRT stations, at great logistical cost, ensuring that they will have a hard time building the resilience we’ll need to drive modal shift to rail.

[embed]The logistics of railway closures There is a social cost every time they want to close the MRT.medium.com

To be fair, so is London, as I saw during my recent trip. They had to close the Metropolitan Line for four whole days (including two weekdays!) in order to conduct critical signalling system testing. And the DLR was also partially closed that weekend. Replacement buses are also offered on such works in the outer sections of London, where there aren’t as many lines as in Zone 1 and thus people can’t reroute around closures easily.

Planned closures are also systematically handled, too, with weekly digests network-wide of works happening everywhere — even sent out by email if you register with TfL. They also have a more aggressive schedule of early closures and late openings, so work can be done without having to fully interrupt the line.

Meanwhile in Singapore? I daresay early closures and late openings for maintenance work will almost never happen, because they don’t run engineering trains between revenue service trains. Perhaps generational trauma from the Clementi train crash of 1993?

[embed]

It also seems clear to me, especially after the Circle Line works, that as much as possible, a priority should be put on finding where and how they can make better use of off peak hours predominantly. It may be worth asking why the LTA panicked and needs to close tunnels for six weeks straight at a time, instead of batching the work to be done across a longer period, and keeping the peak hour and weekend event service (at Stadium station) up in the meantime.

[embed]And the work continues Fresh off the heels of Singapore’s longest MRT shutdown, there are more works yet incoming.medium.com

This episode also demonstrates a need for rail resilience, learning how to get new projects done faster too. Was it feasible to open CCL6 first before undertaking the tunnel works, giving some badly needed resilience to the network by providing additional connections, instead of making things worse on the north half of the Circle?

And when they do this, more service should also be run on other lines, if we’re going to expect them to take over the load. Make this a standard procedure with early and consistent information, and that’s how to make a series of rolling weekend works more palatable to the public.

The consequences of these decisions, today, reverbrate through the rest of the rail network, weakening rail resilience overall. The calculus behind such decisions need to be studied, and lessons should be learnt so that we won’t do this again. It seems unfortunate that the task force hasn’t mentioned any of this, but that may be an implementation detail that needs to be worked out during project planning.

The Combine awakens

The devil, as always, is in the details. While the task force calls for more integration, how can that actually happen?

[embed]A crisis of integration, continued When the status quo of rail projects and operations is no longer relevant.medium.com

One main recommendation of the task force is to “establish a comprehensive and standardised practice of condition monitoring across the MRT network”, and in particular, it also calls for the building of “common and integrated data management systems”.

The report is clear that this should not be at a line level, but at a network level. Which I welcome. With modern solutions allowing for big-data storage and easier trend-spotting with algorithms and integration of data from multiple sources, and clearly plenty of talent in Singapore that can help build such a system, there is no reason not to do so. The question is whether the LTA knows where to find these people.

Perhaps we may all be a bit more circumspect about rail reliability, if the LTA, alongside rail operators and suppliers, somehow hired half the NUS Computing graduating cohort to work on various digitalization and digital transformation projects, be it in the Rail Asset Operation and Maintenance department, or elsewhere. I feel, though, that before that happens, there must be a cultural change — that the insular notion of “railway engineering” being someone on the track, workshop, or relay room can no longer be true, and that they have to attract all kinds of talent to work on the railways.

[embed]How can we progress together? We need to build capabilities, and quickly. How?medium.com

In particular, the big-data insights derived from such a system can help inform lifecycle planning decisions. These also include how much spares to stock, and whether to renew or to replace a given subsystem. The task force even goes on to say that this could even be applied across lines, where different rail lines using similar systems can identify common trends, share lessons learnt, consolidate purchases, and such.

As the taskforce points out, it’s also important to equip workers in the rail sector with the skills to make the most out of the insights generated by them too. Of course, something can probably be said about just throwing some AI into the system or something, and then telling technicians to just do as the machine says. But a longer-range effort to professionalize the workforce will not succeed with certificates useful only for one purpose.

If they want to get people into the rail sector, they may have to think also about how to improve employability outside it, with modern careers prioritizing mobility of industries over keeping the same job for decades. They’ll need data analysts, systems architects and integrators, and other such hitherto non-railway engineering roles in our rail industry.

They want to teach today’s rail engineers these skills too, instead of relying on just “intuition” and “expertise” which there isn’t the time to build in the next generation, not when technical debt is piling up. And that may also no longer be so relevant in an age of digital railways, where good software is as much a part of effective operations as functional hardware.

If anything, some of the work of building this was already done. What happened to the Rail Enterprise Asset Management System?

[embed]A crisis of integration We now know what went wrong last September.medium.com

Don’t think they’re any different

One other recommendation of the task force is for “design standardization” and “modular architecture” in new rail systems. How much of this, though, is in line with the modern rail industry, where increasing integration seems to be the trend? I say no, not when Siemens is showing off a new train control system that fits in a single cupboard.

[embed]To renew is not always to rebuild With plenty of talk about needing to upgrade ageing systems on our rail network, what actually is to be done?medium.com

Perhaps LTA might try to get Siemens or Alstom or someone else to implement ETCS/ERTMS Level 3 as the safety-critical portion of the train control system, with other addons to facilitate fully automated, driverless train operation. But those addons will still perhaps be vendor-specific. Or they can also study things like NYC Subway’s I2S Specification or the Paris Metro’s OCTYS.

Trains can also be designed in a modular way. The TEL infamously has three different types of car for a mere 4-car train — slight differences in the electrical setup (some cars are motored, some not) may mean parts can be different and extra stocks need to be kept overall. When the LTA plans system design of new trains, they can consider how to standardize systems across different cars of a train (like what Berlin does) in order to maximize maintainability and system robustness.

The taskforce also recommends bulk buys, partnering with other foreign systems if there isn’t enough scale locally. To which I ask, who? In terms of trains, the only really similar system is Taipei — but it ends there, where Taipei upgraded their old C301, C321, C341, but they scrapped our equivalent trains.

Maybe the taskforce may have in mind cases like the Barcelona Metro 9000 series, which was also sold to other metro operators in Latin America, or Paris’ MP 89, also operating in Santiago, Chile, and Lausanne, Switzerland. But we’re not going to achieve that today. There might be some commonality with Indian metro operators, where some metro lines have trains of similar specifications, but that’s about it.

They can make this work, though, aligning system build-outs and renewals to scale designs and procurement across different rail lines. That also helps centralize maintenance operations, reduce the overload of independent operations, and make it easier to maintain high standards of rail reliability.

And not only operators — when systems suppliers have the responsibility of maintaining multiple lines, they have reasons to invest further in local teams and capabilities, such as repair (and even manufacturing) localization, and supporting parts substitution if necessary.

We can apply the task force recommendations to new lines too. Building the Seletar and Tengah lines quickly and maintaining them effectively may require integrating them with the Cross Island Line — using the CRL’s depots, workshops, and control systems to get the lines up and running quickly. To further sweeten the deal, train orders for the NEL can also be thrown in, as the current trains’ 30-year lifespan would fit roughly the timing of the Seletar line’s opening in the early 2040s.

[embed]Don’t build MRT lines just to check boxes I think the newly announced Seletar and Tengah Lines shouldn’t aim to do everything.medium.com

In the 2030s, the Circle Line can be integrated with the Downtown Line as part of its system renewal, creating one single system with common technical specifications and trains able to cross between both lines and three depots. Orders for new trains designed for the conditions of both lines can be done at the same time, maximizing economies of scale. This is important especially if they plan to replace the Alstom C830 trains, as new trains are also needed on the DTL in the mid-2030s for the Sungei Kadut extension.

[embed]The ideal medium-capacity train Futureproofing means throwing out everything we’re used to.medium.com

Even in the shorter term, projects planned by the task force can also take advantage of this. Plans to build new power intakes and substations for the NEL can happen at the Greater Southern Waterfront, supplying power not only to the NEL, but also to the CCL, maybe even the Seletar and Tengah lines too.

That also reduces one of the CCL power system’s weakness — that half the Circle shares a power intake with the busiest stretch of the EWL from Tanjong Pagar to Jurong East. Any incident here can deal a significant blow to rail resilience, as we’ve seen a few times, while also limiting capacity if they can’t run more trains, because doing so will stress out the power systems.

source LTA

Systems less critical to rail operations, such as lifts and escalators, can also be integrated at a network level, with availability and maintenance contracts managed centrally by the LTA, instead of getting rail operators to sign their own maintenance contracts with lift and escalator vendors.

The end of private-sector involvement?

To sidetrack a bit, a report not so long ago rapped Singapore’s two sovereign wealth funds for their relatively poor performance against other SWFs. The government, at that point, claimed that it’s up to each SWF to manage their own risks.

[embed]Singapore defends GIC, Temasek returns as reasonable Returns generated by Singapore sovereign wealth fund GIC (GIC.UL) and state investor Temasek (TEM.UL) are reasonable…www.reuters.com

But, on the other hand, PM Lawrence Wong has also said at the 2026 Budget Debate that Temasek, owner of SMRT, is not there to “do national service” — that every asset owned by Temasek must eventually be able to make monetary benefit for Temasek.

Going back to the Rail Reliability Taskforce, the next big group of recommendations are to increase the amount of cross-pollination and knowledge sharing, not only through increased staff cross-posting, but even to outright turn the Singapore Rail Academy into a joint-venture of sorts between the LTA and the rail operators. And that’s not to mention on how things like spares supply are coming increasingly under the control of the LTA, as previously recommended by the task force.

Putting these two together, does it then make sense for SMRT to continue to be owned by Temasek, and for privately held companies like ComfortDelgro/SBS Transit and RATP Dev to continue to have a stake in our rail network?

Perhaps it may be time for SBST to be negotiated out of our rail network, and SMRT to be properly fully nationalized; wholly owned by the Transport Ministry, and no longer a Temasek asset. After all, SMRT must still make money as an asset of Temasek. Wanting it to focus on the rail business, with all these new requirements making it a form of national service, may mean it can’t do the job of making money so well.

Of course, by the MKBF metrics, SBST is doing good work. And if there is going to be cross-sharing of information between transport operators and the LTA, it may make more sense for SBST’s staff to continue in their jobs, just paid by a properly-nationalized SMRT. Good practices by SBST can then be more easily shared across the entire rail network, not held as a secret sauce to win operating concessions over SMRT.

In fact, this model already exists — it’s how our hospitals are run, ostensibly private subsidiaries of MOH Holdings, a wholly owned company of the Health Ministry. And in Taipei too, where the Taipei Rapid Transit Corporation is wholly owned by the Taipei City Government, allowing the Department of Rapid Transit Systems to apply political leverage over TRTC as dictated by their political masters — Taipei Metro even offered Travel Smart Journeys-esque discounts along the Xindian line during a period of roadworks expected to impact traffic.

This may be similar from a proposal that the Workers’ Party came up with before, but with the key difference that it only applies to rail. Buses, where policies and procedures make more impact than technical wizardry, can continue to still be contracted, allowing for international best practices to continue to flow into Singapore — provided, of course, the industry and workforce are willing to accept such changes.

[embed]Public private peril Has Singapore achieved a middle ground?medium.com

Don’t double down

I’ve long been convinced that the weakest link is still the human, something the taskforce has also admitted; in the early periods of rail disruption, someone has to go on site and investigate, diagnosing issues and implementing fixes in a complex environment where multiple systems interplay with one another. The 6-day EWL disruption was first recorded as a power fault, after all.

[embed]Meet the people where they are They have worthy goals, but should be careful to not reinvent the wheel.medium.com

In this case, improving standard operating procedures alone doesn’t work. Rehearsals and tabletop exercises may be ineffective when they’re pre-planned and highly choreographed — everyone knows there’s a planned exercise, perhaps a minister might come in and shake hands, so on and so forth.

They lack the element of surprise that a real train fault brings, as well as the cognitive workload and fog-of-war it brings to all involved staff — and if staff are in situations where they’re more likely to make unchecked judgement errors because they have too many things to do, then things can (and have) get worse.

Can they plan exercises that account for these? It may be possible. After all, human factors are a big reason why the Miracle on the Hudson played out as it did.

Systems should be able to assist rail workers on the ground in troubleshooting faults, providing better communications, and more, but that may not be the case today. The task force has already recognized the need for better passenger communications. Apart from SOPs, systems that help ease the communication workload and reduce the burden of staff in already high-pressure situations, should be built too.

Plans to improve the position of controls such as brake valves and train protection reset switches must also consider how these work in a packed rush hour train. If a member of staff must walk down the train to reset breakers and open valves, and they can’t easily do so in a packed train, that adds time needed to recover from train faults.

The answer may be that trains must be less crowded to facilitate easier recovery, which means running more trains during peak hours where possible, and encouraging decongestion by shifting passengers to use newer lines. They must also build them faster, so that their decongestion ability can be delivered faster.

The definition of insanity is doing the same thing over and over again and expecting a different result. I could say the same thing to the task force’s suggestion to stage more buses and drivers around the network —that was exactly what Task Force 50 was meant to do. In a time of bus driver shortages, as many bus captains as possible must be out there to deliver as much connectivity as they can, especially when the KPI is still how many numbers there are on the signs. Maybe they could get PHV drivers to help out — but at that point, why not recruit the PHV drivers into the bus industry?

Now the hard part is doing the right things. The plan has been laid out; but getting it done calls for a bigger change than just doing the same old things again. And the plan must also clearly reject old paradigms that aren’t effective.

Like what you read? Join the Telegram Channel for updates, or follow me on Instagram, Threads, or Bluesky for quick takes!