Monday, 25 January 2021

Are Trackless Trams the solution to our problems?

The CRRC Autonomous Rail Rapid Transit system, aka Trackless Tram (via N509Z)

Plenty of lobby groups and government bodies have floated the idea of installing "trackless trams" in cities around the world, but at the time of writing, the original 6.5km test track in Zhuzhou has only been succeeded by one line in revenue service, a 17.7km line in Yibin. This presents a huge problem - with basically no real-world examples to look at, people can make huge claims about how great (and especially how cheap) this technology is, and it's hard to sort fact from fiction. 

But the Federal Minister for Cities and Urban Infrastructure, Alan Tudge, announced a while back that the government would fund a business case for a trackless tram in Perth; and it's something that various lobby groups like to float for places closer to home like Geelong and Ballarat from time to time as well. So I'm going to do my best to cut through the hype, and see how they really stack up. 

The trackless tram in Yibin (via 来斤小仓鼠吧)

So-called trackless trams are sort of pitched as being halfway between a bus and a tram, and able to capture the best of both worlds - the quality of a tram without the price tag. So it makes sense for this to be the comparison we make. However, we need to draw the distinction between average buses and good buses here, because we're considering what our buses could be if we were willing to invest big money (like, trackless tram kind of money) into the services. So on each claim, I'm going to rate the trackless trams against two alternatives - conventional trams and good buses. To simplify the environmental questions, I will factor in the procurement of electric buses for any "good bus" project. 

Vehicle price 

It is surprisingly hard to find a firsthand quote for how much a trackless tram costs; Committee for Geelong quotes $3-4 million, compared with $10 million plus for an E Class tram, but it's not clear where that figure comes from. This quite frankly doesn't pass the sniff test - there is no way they'd be a third of the cost of what is fundamentally a very similar product. The bogies could potentially be cheaper, but the body is basically the same; and if it's powered by batteries rather than overhead wires, you might save money on infrastructure but the vehicle costs will be higher since big heavy batteries will be way more expensive than pantographs. 

Conventional trams are expensive, but trackless trams likely will be too (via Liamdavies)

Right now there is only one manufacturer of trackless trams, and they haven't built up much in the way of economies of scale yet; also, the new tech is all proprietary, which prevents them from facing competition to drive prices down for the next several years. Whereas conventional trams have a whole industry, with a number of players competing for contracts around the world, driving prices down - indeed, an internal review by Transport for NSW raised the lack of supplier competition as a key concern. It's certainly possible that CRRC may be willing to let the first order go cheap just to get their foot in the door, but in the relatively short term, the competitive forces at play suggest that they'd be at least as expensive as a conventional tram, if not more expensive. 

On a per-vehicle basis, a bus is always going to be heaps cheaper than either of them. A new battery electric bus costs around $800,000, while a 2015 government order of 20 E-Class trams cost an average of $13.7 million each. (EDIT: a 2018 order came out to $8.3m each, and the 2015 order may have included things like power or depot upgrades as well as the trams themselves. So consider the $13.7m figure as me erring very much on the side of caution.)

These trams have more than double the passenger capacity of a standard bus, of course, so the per-passenger cost is a little bit closer; but when you can buy about seventeen buses for the cost of one tram, it's clearly no contest. 

Verdict: The claims the trackless tram would be substantially cheaper than a conventional tram are implausible; most likely they'd be about the same price. Buses smash them both. 

Electric buses are substantially cheaper than trams (via Ryanmirjanic)

Infrastructure costs 

Trackless trams are often sold on the idea that the infrastructure costs required to implement them are very low - the idea being that they can use the existing road network with just a bit of paint for the guidance system, and all you really need to build are the "stations", which can be something like an accessible tram platform stop (or something more elaborate if you so desire). This is contrasted with conventional trams, which require the same outlay for stops, but also require you to rip up the road surface and lay steel rails. If we assume that that's true, then buses are still the cheapest - you can run them on the existing roads too, and their ability to "kneel" down at kerbs means you really just need a concrete pad and a bus shelter to make for an accessible system, rather than platform stops. But, assuming that it is true, trackless trams would represent a significant cost saving compared to conventional trams. 

However, there are very good reasons why we can't assume that this is true. The technology has only been shown to work on brand-new, extremely high-quality and freshly-painted road surfaces, and the reality is that most of Australia's suburban roads do not currently meet this standard - they're old and fragmented and, in many cases, not able to carry significant loads of weight sustainably. In the report I mentioned earlier, Transport for NSW noted that existing bridges may not be engineered strong enough to carry them. 

It's impossible to be sure without any real-world experience on real-world legacy roads, but it seems likely that in most cases, roads and other infrastructure for trackless trams would need to be completely resurfaced and/or rebuilt before being repainted. This might be cheaper than installing steel rails, but it's still a cost that needs to be accounted for - it certainly isn't as simple as just slapping on a coat of paint. 

In Yibin, the trackless trams run on fresh, smooth tarmac (via 来斤小仓鼠吧)

The other big infrastructure cost that needs to be kept in mind is maintenance. These trackless trams simply haven't been running for long enough on a wide enough variety of roads for us to have a robust estimate for what the long-term costs for maintenance are likely to be, but we can make some broad assumptions based on first principles. The first thing to note is that heavier vehicles do exponentially more damage to road surfaces than light vehicles do - when you double the weight on a vehicle's axle it does sixteen times as much damage - and trackless trams are much, much heavier than buses. So a road that carries trackless trams not only needs to be re-engineered for heavier loads in the first place, it's also going to need more costly maintenance on an ongoing basis. Steel rails, by contrast, are much more resilient and capable of handling big loads like this, so the maintenance costs are lower. 

Another common claim is that trackless trams are punctual and don't get stuck in traffic. The idea for the trackless trams is that they would have their own dedicated lane, and this does of course make for fast and punctual public transport - but if you give conventional trams, or buses, their own dedicated lanes, they are equally punctual. The trackless tram does have the flexibility to go around obstacles which conventional trams generally don't - but conventional buses do have this same flexibility. (Also, the "flexibility" thing is only useful when there's space to go around, which isn't the case for the most common problem - car traffic). 

Any vehicle will get stuck behind cars if it has to run in mixed traffic (via Robert Merkel)

Verdict: The trackless tram's claims of low infrastructure costs are very dubious because they exclude highly relevant factors; in truth, they are probably the worst of both worlds on this criterion. If you want something you can run quickly and easily with minimal infrastructure costs, buses are the way to go. 

Operational costs

When it comes to the marginal cost of running a public transport service, the biggest factor is always the cost of the driver. There will of course be "fuel" costs (in this case electricity) and the trackless tram is again the worst of both worlds here - there's much more weight to lug around than a conventional bus, and there's considerably more friction with rubber tyres on tarmac than steel wheels on steel rails - but this doesn't matter much, because the fuel costs are dwarfed by the driver's wages. 

Despite the "autonomous" label, trackless trams do require drivers (via Mikeshouts)

Trackless trams are sometimes described as "autonomous", which might lead you to believe that you could save all those driver wages, but you manifestly cannot. It's true that there is quite a lot of automation involved in their normal running, but they still have a steering wheel and you still need to pay a human to sit behind it, and be ready to intervene if necessary. And it's not like it's an unskilled job - monitoring a mostly-autonomous system is arguably harder than just driving the damn thing yourself - and it's therefore likely that it will attract reasonably high wages. It would be extremely optimistic to expect them to be much lower than conventional tram driver wages. 

We obviously have a lot more precedent when it comes to tram and bus driver wages. The annual base rate for a bus driver in Victoria is around $62,000 while for a tram driver it's about $72,000, so it's cheaper to run one bus than one tram. However, this is where capacity and service levels come into it - buses are slightly under half the capacity of trams, so if you have enough demand that your trams (trackless or otherwise) would be full, you'd need to run at least two buses for every tram to cope with this - which means paying twice as many drivers. 

That said, being at capacity is not the only reason to run services more frequently - if you're planning to run a service every 10 minutes regardless of demand, so that you're providing a useful and attractive service, this requires exactly as many buses as trams. 

Verdict: Trackless and conventional trams are likely to require basically the same wages. Buses have lower wage costs per vehicle, but higher wage costs per passenger capacity. So which option is cheapest really depends on whether your service frequency is driven by the desire to providing a turn-up-and-go system (bus) or by high demand for capacity (trams). 


So which is the best option? It seems pretty clear that it will depend on how much capacity you need. 

If you've got a highly populated area like Melbourne, and you're trying to solve traffic congestion, then you probably need quite a bit of capacity. Even in a big city it's still worth doing your due diligence on a bus option, because you can move a hell of a lot of people with buses running frequently enough - but nonetheless this is the environment when a high-capacity tram will shine. So should you go for the conventional tram or the trackless tram? 

Despite the claims, a trackless tram is probably not going to be dramatically quicker or cheaper to set up in the first place. You might make some small savings initially, but these will be offset by higher maintenance costs pretty quickly. So it's likely that a conventional tram is the way to go here. 

Buses done well can be surprisingly successful (via Smiley.toerist)

In a lower-demand situation like Ballarat or Geelong, a good bus system will be the way to go. The "cool" factor of a tram can only bring so many people onto a system - it still needs to run frequently and for a long span of hours, or people will just stick to their cars. 

Either kind of tram would have substantial infrastructure costs upfront, and the vehicles themselves would be quite expensive too; whereas the infrastructure costs are basically zero for a good bus network, and you could buy a whole fleet of new electric buses for less than you'd spend on a handful of trams of either flavour. As I mentioned last time, brand-new electric buses can also gain a bit of the tram's cool factor, with bells and whistles like WiFi. 

In a relatively low-demand situation like this, your service frequency isn't generally determined by responding to passenger crowding, it's determined by fundamentals of providing a useful service - which means you'd be running the vehicles at the same frequency regardless of whether they're buses or trams. This eliminates the tram's advantage in operational costs, because you're not having to run two buses for every tram, it's just one-for-one - and bus drivers are cheaper than tram drivers. 

With the amount of money it would cost to implement a basic trackless tram network that'd have a much lower service level than the average conventional tram line in Melbourne, we could have an absolutely whiz-bang world-class bus network across the whole of Ballarat. So for me, that clinches it. 


  1. The high-speed stability of a (presumably) multi-articulated vehicle on tyres on a road, instead of on steel wheels on rails, will be an issue. And by high-speed, I mean over 10 km/h. There's a very good reason multi-articulated passenger vehicles don't exist yet - they're simply not sufficiently stable. Are these prototype units built with some kind of active coupling management system so that they don't work like a hauling unit and a couple of trailers? I bet they're not.

  2. A comparison of trackless tram and light rail, including a few extra areas like noise and the environment:

  3. As a concept, the idea of a bus with doors on both sides, with a driver's cabin at each end is beneficial, giving you a lot of flexibility with stop design and requiring less space to turn back. This can make it harder to share infrastructure with conventional buses though. When we examined them as an alternative to light rail for a government client, the cost were quite close. Land take for TTs was greater as stations needed passing lanes to benefit from the bus-like flexibility, and more of the TTs were needed to carry forecast demands. This meant headways would be close enough that the ability to provide effective signal priority was constrained.