Train motor power

Can too many trailer cars be a bad thing?

This video was going around in my circles as sort of a joke. Not exactly a laughing matter, but it is what it is. In the video, a 3-car train hobbles out of Botanic Gardens station likely due to a fault with the train’s propulsion system.

Train fault aside, this illustrates how a key component of high frequency is the ability to attain, and maintain, high acceleration rates to allow trains to spend lesser time getting in and out of platforms. It’s why for the longest time people thought that rubber tyres were necessary for high frequency, automated operations — an impression no doubt reinforced by the success of the Paris Metro or VAL systems in the frequency domain.

I cannot say for sure without most datasheets in the public domain, but I can go off Wikipedia.

Technology marches on

But are rubber tyres necessary now? Ask different people at different times and you get different results.

In 1967 the Victoria Line was possibly one of the few automated systems on steel wheels, and that was almost completely underground apart from the depot. Contemporary systems like the Paris Metro and Kobe New Transit used rubber tyres and managed to remain aboveground. Then you have Vancouver, who went against the mold and used linear induction motors on steel-wheeled trains.

This might be the reason why, before the completion of CBTC upgrade works on the NSEWL, trains had to be driven manually in wet weather. The technology might have worked well for an underground metro line, but maybe not so well in the open air, and worse still in tropical monsoon conditions where there can be poor braking conditions caused by slippery rails in wet weather. With Malaysia and Indonesia (Sumatra?) being amongst the highest in thunderstorm occurrences, we are in good company.

Another explanation is that the better grip of rubber tyres allows for trains to climb steeper slopes. This isn’t wrong, though it may not be necessary in Singapore since we try to avoid anything greater than 3% on heavy rail. The MTR has 4% inclines such as on the South Island Line, and Taipei Metro even has 5.5% inclines on the steel-wheel Circular line — the latter of which were previously thought the exclusive domain of rubber tyred vehicles.

How do they do this? Simply put, they have more motors on the trains; and in the case of the Taipei Metro, albeit less powerful ones. Both the MTR SIL and Taipei Circular line trains are fully motored:

• 3-car SIL trains, approximately the same size as our Circle and Downtown line trains, have 12 powered axles each with 200kW motors, resulting in 2.4MW available.
• 4-car Circular Line trains, also about the same size as each individual car is smaller, have 16 powered axles each with 102kW motors — resulting in 1.6MW of power available.
• And 4-car SBK Line trains in KL, while having only 2 motor cars, have a total installed power of 1.84MW across 4 cars.

As for us, this is what we have:

• Our newer 6-car MRT trains, with 140kW motors on 16 powered axles, thus have 2.24MW (Those with PMSM motors, though, are only rated for 135kW — only 2.16MW across the train)
• According to Wikipedia, 6-car NEL trains have 150kW motors making up 2.4MW of installed power, and 3-car CCL trains have half that, at 1.2MW.

Dragging on

What is the difference? As mentioned, the Taipei Metro and MTR vehicles quoted have fully motored axles. On the contrary, the middle car of our 3-car trains are designated as trailer vehicles, which means they have no propulsion equipment. This does make them lighter.

But it also means that the entire train, being less powerful overall, may be unable to produce the tractive effort to maintain a given acceleration rate for long. Another concern is that should there be a fault of the traction package on one of the two motor cars, the other motor car has to move the entire train by itself, as seen in the video above. This means the train essentially has to hobble around and slow down the service; in the case of the DTL train in the video, it was likely withdrawn to the depot at Bukit Panjang station, but it’s easy to see how a jam would have built up behind had it been in a less convenient place.

An improvement made with 4-car TEL trains was to have the additional car be a motor car. Reliability should thus be improved — should power equipment on one car develop a fault, the remaining two motor cars can still propel the train possibly at a better rate than a train on the three-car lines with a faulty motor car. This matters especially with how far out train depots are; imagine if that DTL train had its motors failed at Telok Ayer or something and had to hobble all the way back, or at least to be withdrawn midway to a siding.

The MTR was also rather anxious to replace its old MLR trains on the East Rail Line for similar reasons. In a new weekend timetable, by solely replacing the old underpowered MLR trains with new Hyundai Rotem trains, they were able to complete the entire extended journey to Admiralty (6km longer) in 3 minutes less. Travelling along just the then-current line to Hung Hom, the Rotem trains were 8 minutes faster. Yes, Rotem trains may be shorter at 9 cars to the MLR’s 12, but they have 6 motor cars compared to the MLR’s 4. Faster trains means lesser are needed to maintain a given service level; you could conversely increase service frequencies for the same amount of trains. For the MTR, this was meant to offset the shortening of the trainsets.

Another benefit of having higher tractive effort potential through more motor cars, and perhaps more powerful individual motors, is the ability to leave platforms faster, as higher acceleration rates can be maintained for longer. And no, you don’t really need rubber wheels for that anymore. Spec sheets for the rubber-tyred Wenhu Line vehicles indicate an acceleration rate of 1m/s2, similar to that for the steel-wheeled London Underground S Stock.

As mentioned, a key component of high frequency operations is platform occupation time, which high acceleration rates can reduce. This may not matter so much on the 3-car lines, and to a lesser extent on the TEL, due to the relatively shorter platforms. But it would matter on the NEL especially if they want to break the 90–100 second headway barrier there, or if at any point in the future there’s interest in pushing the boundaries on the other lines.

Of course, this is assuming external factors fall in place too — it may be possible that the Circle Line trains’ unique acceleration patterns is to avoid overloading the power system with high startup and inrush currents; a power system upgrade may also be necessary to deal with this.

Over the hills and far away

One can’t help but wonder whether tractive power issues might have been a factor in switching away from using TEL trains to operate the RTS as well, apart from being able to build literally lighter structures. Earlier promised economies of scale may remain as CRRC Zhuzhou-Batu Gajah had been contracted to deliver the trains for the Klang Valley’s LRT3 project, and with that order being scaled down due to the cost-cutting efforts of the previous Pakatan administration, some part of the order may be redirected towards the RTS in order to retain promised economies of scale.

Petrol heads will be familiar with power to weight ratios, so I go with those since the concept is similar. Despite their three motor cars, 4-car TEL trains probably have the worst power to weight ratio amongst all our MRT passenger trains—compared to the above, Wikipedia indicates that a 4-car TEL train has 12 powered axles, each with a 120kW electric motor. Thus, the total power that can be generated is only 1.44MW for a 4-car train, to move 153t of vehicle and up to 1280 passengers. This compares poorly to other lines.

Based on information released by Malaysia’s MRT Corp during the public inspection for the RTS, there is also a very long up incline for the RTS to be able to climb up to the 25m high bridge crossing from the underground Woodlands North station. Information from the public inspection showed that this upslope is almost 1km long at a nearly 3% incline, the maximum that can be tolerated on the MRT system. Speed, supposedly the best part of the TEL train design, is not so necessary either considering the short length and two major 90 degree turns along the RTS alignment.

What this means is that, had we continued using TEL trains, they run the risk of this creating some operational issues. Apart from the poorer ability to climb the slopes with a full load, if there was a motor fault on one car, it might not be possible for that train to remain in service. Either the train might have to be withdrawn (which may be inconvenient at Woodlands North), or one car might have to be taped off, as was custom on the old KCR East Rail 12-car trains when similar issues occured.

It also means that lesser interventions can be taken to make the most out of space on the RTS trains. For example, seats cannot be removed or even flipped up to accommodate extra standing passengers — a useful thing to do considering how transient RTS passengers are — if the train does not even have the capability to haul those passengers in the first place. Deploying the flip seats on the T251 trains to different positions depending on whether they’re on RTS or TEL duties thus might not be able to happen. Similarly, this could also be a factor on why they aren’t removing seats on the Circle Line trains in order to make space for more standing passengers — the trains might not be able to carry the additional load.

Conversely, using smaller vehicles means there is a lighter load that needs to be hauled up the slope to defy gravity. Of course, this is assuming that the overall motor power on the trains to be provided by CRRC Zhuzhou are equivalent or greater to the 1.44MW provided on the TEL-spec T251 trains. This would result in a higher power to weight ratio since the physically smaller vehicles can also be expected to be lighter, in addition to carrying lesser passengers.

The need for speed, redux

It’s interesting to note that despite higher top speeds and a much straighter alignment, the NEL still has a similar average speed as the NSEWL mostly due to low station spacing within the Circle Line — or in other words, between HarbourFront and Serangoon.

There’s not much that could be done about this apart from some claims of design-time whisperings of putting NEL Chinatown station nearer to Hong Lim Park than the Garden Bridge and removing what is now Clarke Quay station, but what’s done is done. Sections of newer MRT lines also have stations similarly close by anyway, and closing stations wouldn’t save much time.

What is really more important is to speed up the existing MRT network. I know what I said before was less optimistic on the matter, but as the NEL shows, top speed alone doesn’t make for good average speeds. It’s the ability to reach those speeds faster and to maintain those top speeds for longer, that can also improve speeds, which adding more motor power facilitates.

For comparison, apparently Klang Valley MRT is faster than our Singapore MRT:

• SSP Line Phase 1: 12 stations, 16km, ~24min
• TEL1/2: 9 stations, 16km, ~23min (before noise related slowdowns)
• SBK Line (pre-SSP Line): 31 stations, 51km, ~78min (apparently slowed down too)
• EWL: 33 stations, 50–51km, ~85min

TEL1/2 compares poorly to the SSP Line especially when you consider that three additional stations mean they stop more than us but still manage to achieve the same average speeds. Now, of course, noise mitigation gets harder at higher speeds; but it’s useful to note that the TEL trains may have poorer noise mitigation to begin with. Improved motor power may be something for the next generation to deal with, by which time there may be better noise mitigation technologies too.

The CRL will probably never be as fast as 8km in 9 minutes, but with better train performance, it may yet be able to shorten travel times again despite the lack of express services and relatively shorter interstation distances in the suburban sections of the line. And perhaps this might enable something to be done about the large curves planned in the CCNR tunnels as well.