On fixing the LRT
Previously I questioned those who thought it’s a good idea to scrap the LRT system.
Of course, the key logical fallacy needing to be called out there was that a majority of what was pointed out then were implementation-specific issues other cities have successfully overcome, and not a fundamental flaw with the concept, which has been adopted in other cities to greater success than we have, as explored by RMTransit:
But sure, the central thesis that needed addressing then was to scrap the LRT. What if we look at it the other way, from the perspective that we’re trying to save the LRT? This part will examine what the various renewal projects are trying to achieve, and how well I think they will achieve them. I’ve split this out from the previous post because it was simply getting too long.
Short of ripping everything out
Let’s start with the BPLRT where a series of comprehensive upgrades are currently underway to address some of the technical issues. But at least we are lucky in that there was a ready technical solution — the Scarborough RT, mentioned previously, was forced to shut down simply because Bombardier refused to make new vehicles compatible with their system and refitting the guideways like what Changi Airport did being too cost and time prohibitive.
Fortunately, this isn’t the case here, even if they might have considered such a rebuild based on the wording of the LTA press releases, and the case of Beijing Capital Airport shows that the APM 300R vehicle being supplied to us can also be supplied to other systems wanting to replace ageing CX-100 systems without expensive infrastructural modifications. There is a market, and Alstom-Bombardier are wise to chase it.
This may be a bit late, though. I personally believe that Bukit Panjang may have the distinction of having the world’s last APM 100 vehicles — these being the new fleet delivered in 2014–15, which was around the time Bombardier was pushing their APM 300 as the product of choice for new systems; the APM 300 using a different guideway supposedly with lesser issues. However, this isn’t something to be happy about — it is highly likely that despite being “newer”, the APM 100 vehicles from 2014–15 share a lot of design basis with their CX-100 predecessors, making it fundamentally a very old design and thus not taking advantage of many technological improvements in rolling stock made in the past decades.
EDIT: Looks like I’m wrong — SFO Airport received three new vehicles more recently— but the point still stands.
For example, the APM 100 still uses CX-100 era DC motors, when most of the world has shifted to AC propulsion for transport applications. This age may also be a cause of low reliability of the vehicles in general, as even if modernized components can be used, the ageing design core does not lend itself well to high-reliability applications. Even, for example, misting windows not working, I’d argue it’s a logical consequence of shoehorning such a feature on a design core now reaching 50 years old.
This is a likely reason why the LTA had to hype up such things as regenerative braking, which we’ve taken for granted on the MRT system since 1987, on the new APM 300R vehicles, and not before. There is a lot to catch up on. Likewise, with a newer, more capable propulsion system able to hit higher speeds, even if we are unable to benefit from the higher speed due to short station spacing, improvements made to vehicle stability may also improve ride comfort especially at higher speeds. In other words, they shouldn’t shake too much — though reports coming out of the Wenhu line in Taipei and their APM256s, or the Pujiang line in Shanghai and their APM300s, don’t lend much cause for optimism.
The upgrade of the signalling system, apart from hopefully cutting down travel times marginally, should also help with ride comfort. Trains can be programmed to drive in a gentler fashion with CBTC, which tells trains their location so they know where they are; and thus where to slow down and stop in a single, gentler motion. I’d argue that the current fierce braking is a direct logical consequence of a fixed-block signalling system like Cityflo550, which is old enough to actually have started on the San Francisco BART in 1972, 50 years ago.
In such a fixed block system, the train doesn’t know where to stop in advance, and can only brake on receiving either the necessary speed codes from the track circuit, or detecting the program stop tapes that tell it how far it is from the platform. Perhaps the relatively slower acceleration and braking of steel-wheeled trains may have cushioned its impact on the NSEWL, but on the LRT its blunt impact is there for us all to see. With CBTC, as we see on the NSEWL, the train now knows where it is. It can adjust for when and where it begins braking so it can enter the station more smoothly in varying weather conditions, or avoid speeding up as it finds itself approaching an oncoming curve.
Furthermore, radio-based control and communications mean that the only contacts to be made with the rail assembly will be for power, with no mention of track circuits. This is already triple-redundant with the 600V AC three-phase power rail, which will also be strengthened as part of the job. On the LRT, station spacing may be short enough that backup track circuits are not needed for train detection, and that in event of disruption, the control centre can perhaps simply use CCTVs to identify the physical locations of trains in stations; thus allowing only 1 train to be between 2 stations at any time.
But most importantly, there needs to be service, and upgrading works need to show results. This need not even be done in one straight job. Once enough replacement LRT vehicles are in service that drastic measures to reduce mileage on old vehicles are no longer required, it could be considered to reinstate Service C (yes, Service C) to operate the outer loop service during off peak hours, allowing 976 to be withdrawn.
This new Service C could either terminate at the crossover near Bukit Panjang station, or Ten Mile Junction could even be reopened to serve mall patrons. Trains now parked at stations in the off-peak could be used to operate this new service, mitigating the lack of depot space. In time to come, either this could be made a permanent off-peak arrangement with no more than 4–5 minutes of waiting time anywhere on the line thus providing a meaningful improvement over buses and trams, or the full Service A/B operation could be restored like the good old days. Due to new MRT development, providing all residents with a direct train to Choa Chu Kang may not be so necessary now compared to in the past.
Integration with buses could also be improved. At Pending station, the large amount of bus routes there (and reroute of 972M) could indicate that the role of the Block 223/270 bus stops could potentially be transferred here, especially if 190 and regular 972 are de-emphasized in favour of the MRT. Here and at Jelapang, The traffic light could be shifted nearer to the LRT station entrance if possible, making it more convenient for people to switch from the LRT to buses.
Secondary exits can also be considered at more stations, such as what was done at Fajar. Segar station, for example, could receive a new exit or concourse reconfiguration to shorten the walk to the park connector along the canal. Residents of Segar Road can then walk down this park connector to access the LRT station. And at Teck Whye, maybe an upper concourse of some sort could be built, with a direct overhead bridge connection to ITE College West.
But they still won’t solve the issue of imbalance, where having odd numbers of each fleet means only 15 pairs can be formed from 32 vehicles, as the remaining two vehicles cannot be coupled with each other and which limits their use. Perhaps an additional APM 300R vehicle could be delivered as compensation for project delays if it can be agreed, allowing 16 pairs of trains to be formed with a spare vehicle. After all, these trains were supposed to be here in Q3 2020, 1.5 years ago, and I’m not sure how much excuses can be made for the current international situation and of course also considering how R151s and CCL C851Es, due in 2021 after the APM 300R, have begun arriving.
Alternatively, they could just replace all 13 of the APM 100s instead of upgrading them for CBTC compatibility. Yes, I know the vehicles are relatively new, and Cityflo650 has been installed on APM 100 vehicles such as those at San Francisco Airport. As mentioned, if they are still based on an elderly baseline design, it could still impact the reliability of the new vehicles down the road. Current delays to the overall project may also make this a more appealing option, especially with order sequencing meaning we may get around 10 years of useful life out of the old vehicles, which means the old concerns about letting them go to waste may not be such an issue.
Apart from resolving the issue of leftover vehicles, the operational benefits brought on by a single fleet may also improve the LRT operations, both through consistent vehicle performance potentially enabling faster journeys without faster trains being held behind slower ones, speed restrictions due to vehicle noise. A single pool of spares and maintainer knowledge, which will be broken during the upgrade (if not already) due to the technological step change, will also be restored. The bidirectional coupling features of the new vehicles may also simplify depot operations by reduce excess mileage caused from having to find ways to rotate the trains in the LRT depot, so that they’re facing the right way (these are probably what the arrows outside are for?)
Things get more complex on the SPLRT system, although it does have a relatively newer design approach (the Crystal Mover may have been initially developed from Japanese AGT systems specifically for the SPLRT after all). As described previously, the builders were able to learn from the mistakes of its predecessor, and of course one of the most important things is that they chose to build on the relatively more open Japanese AGT systems. That last point means the door is left open for us to potentially turn to other Japanese or even Korean makers to build a compatible vehicle if we so wanted.
They announced the purchase of 17 new 2-car trains to replace 25 single units. What are left are 32 vehicles, 16 old and 16 newer ones, that I’d expect will be formed semi-permanently into 16 2-car trains. This means that, instead of creating 41 trains from 57 vehicles (16 2-car trains and 25 single-car trains), we now have 33 2-car trains. This is necessary for full 2-car operation, which should ease some of the headaches with capacity and actually put some of the infrastructure to work.
The good thing about fixed 2-car sets is the removal of excess equipment lockers and extra emergency exit mechanisms in the space between the two cars. After all, with fixed 2-car sets, things like signalling computers and propulsion packages can potentially be shared across both cars instead of being confined to one car, freeing up space inside the car. Renders are awkward on whether these vehicles will also come with an intercar gangway, which will further improve the carrying capacity of 2-car trains, — but not by much if it’s as narrow as the ones on the Sentosa Express.
And why not on Bukit Panjang? Bluntly put, it appears that depot equipment such as lifting jacks is incompatible with 2-car trains there. This means that 2-car trains at Bukit Panjang must be broken into their constituent units for maintenance (seen in picture below). This is not an issue at Sengkang-Punggol since the new depot expansion should be able to service fixed 2-car trains from the outset.
Less is more
Of course, having 33 trains instead of 41 may seem like a downgrade in terms of frequency even if all are 2-car, especially after I’ve spoken so much on how LRT systems should have better frequency due to their grade separated nature. On the surface, lesser trains are available for each loop — around 7 per loop after accounting for spares and vehicles under maintenance. But what the SPLRT might need to actually benefit from that much trains is an improved signalling system. Frequent riders will tell stories about how trains were held outside the stations just before the Town Centre, waiting for its turn to enter the Town Centre station.
Regretfully, it doesn’t appear that Mitsubishi has been contracted to do anything about this — only to provide signalling equipment for the depot expansion. But if they do fix this as part of an overall system upgrade, it may allow for trains to run closer together in the core sections. 90 seconds in the core means 3 minutes on the loops, better if they can push it even further down. Among other possibilities, Mitsubishi even has a CBTC system which they use in Macau and on the Tokyo Metro.
That’s even before talking about running even longer trains, something possible in Sengkang-Punggol due to the platforms (mostly) being larger than necessary. Additional capacity with larger trains minimize the impact of the shared track through town, if one does not want to upgrade train control systems to allow trains to move very close to each other in the area. Or many other tweaks that can be further made to the system. That said, I don’t see how immediate benefits can’t be put in place once the new 2-car trains begin to arrive. With creative operations such as stabling trains on the mainline tracks, it might even be possible to accelerate deliveries of new trains without the need for the new depot to be ready first, and/or to eventually accommodate more vehicles than the depot can eventually support.
Furthermore, as part of enabling works for the depot expansion, departure boards should be installed at all stations to advertise the arrival times of next trains. At the central stations, these can perhaps be placed over the existing queue lines, as an upgrade to the signboxes that tell people what loop the train at the platform will operate on. Or even find a way to place outward-facing electronic signs on the trains themselves, such as is done on buses.
Urban access still remains an issue, and to pretend otherwise would be to bury your head in the sand. Say you were a resident of Block 273C Compassvale Link. Where would you walk to, Buangkok MRT which you can probably see from your house, or Ranggung LRT? Or how about a resident of Block 269B?
All may not be lost — to reduce reliance on bus services, Sengkang Town Council (in this example) would probably want to construct covered walkways from as many points within a 400m radius from rail stations, and to provide as direct a link as possible. After all, what’s the good of 8 in 10 households being within such a radius if it’s difficult to actually cover that short distance anyway?
I need to be clear that this is out of the LTA’s control. This is a goal which MND and its subsidiary agencies would need to get on board with as well; road pedestrianization and other shift-to-public-transport projects will only go so far if MND is unable or unwilling to twist the arms of its agencies to get them to accommodate such goals. It might have been easier had we still had a Coordinating Minister for Infrastructure, but what’s built is built and the town councils will have to take on the slack of fixing HDB’s mistakes.
Even if trams or busways were used instead, whilst you might save the few minutes of stair-climbing, the same urban access issues in getting to the stop may still depress public transport ridership — and that’s without the uncertainty of how on-street stops are designed. Look at the walking route from Canberra MRT to other parts of Canberra estate, or to get to a bus stop within the newly built Queenstown/Dawson BTOs, or the recently launched Tampines BTOs and what it will take to get to Tampines West MRT.
Overall, I think the LRT experiment has actually been a valuable lesson, the hard way, on how to build and operate successful medium-capacity transit systems. Most importantly, we learn from our mistakes, but the question is whether people can understand that or are willing to just keep going with the old tropes.