Long Read: Mass Rapid Transit in Wellington

Posted by Kerry Wood

CRRC_Autonomous-rail_Rapid_Transit_train_at_Metro_Trans_2018.jpg

The latest version of ‘trackless tram’ (TT) has been developed by CRRC in China. A trial system has been running in Zhuzhou since 2017, and should be coming onto the market about now. It is of interest in Wellington because of potential cost-savings over light rail, but comes with corresponding problems and is barely commercial at this stage. TT is distinct from BRT but shares some important characteristics.

Key messsages

The TT feature of interest in Wellington is capacity. It is the highest-capacity BRT-like vehicle on the market, presumably with a much better ride than a bus, and may be able to meet Wellington needs on a two-lane route.

Any decision to adopt TT will require careful studies; Wellington has already run into costly problems created by a casual attitude to supposedly minor issues. In a more difficult situation, we must get it right this time:

  • BRT using conventional articulated buses is well-established but an unlikely option for Wellington. High-capacity BRT is generally used in cities having wide streets, unlike Wellington.
  • TT might be an alternative to BRT, if it can offer sufficient capacity, and when ‘the kinks have been ironed out.’
  • At a time of very rapid change, uncertainties are inevitable and require good management. In this case high-capacity would be a low-risk approach, favouring either light rail or four-lane BRT.
  • Decision-makers need to bear two things in mind:
    • First, light rail becomes cheaper than either BRT or buses at a relatively low ridership.
    • Second, BRT also benefits from a properly segregated route, to minimise congestion, and from diverted underground services to minimise delays.

Light rail may well be the lowest-risk option, or even the cheapest option.

An independent conclusion comes from Matt L at the Greater Auckland transport blog:

I do think that this [TT] technology is promising and definitely worth keeping an eye on, but I’m not convinced that Auckland should be so quick to jump on the bandwagon. Let’s at least wait till at least a handful of cities have successfully rolled this out and ironed out all the kinks… Let’s also wait till there are multiple suppliers with inter-operable systems. Unfortunately, even without the capacity/frequency issues that I think would be an issue for the city centre, I don’t think Auckland can afford to wait. We need to get on fixing transport in this city and so should get on with installing light rail as soon as possible.

↑ Contents

Route and capacity

The LGWM route has recently been challenged, with proposals for a Mt Victoria tunnel for buses, walkers and cyclists.

Residential density on the MRT route

A tunnel for walkers and cyclists seems sensible, but a new bus tunnel would be a backward step. The existing Bus Tunnel is adequate for serving Hataitai, and a much better MRT route is through Newtown, because of high residential density. Densities are too low for MRT in Hataitai and through to Miramar and the Airport. The Newtown route offers substantially greater residential density, on both sides of the route, as well as potential for future density. Adelaide Rd and Kilbirnie are designated WCC development areas. A Mt Victoria route was proposed in the 2013 Spine Study, apparently to save time, but the real time-savings come from good detail design on the chosen route.

Bypassing Wellington Hospital is itself a planning error for MRT: BRT in Brisbane went as far as a stop within the Hospital building.

It is not a criticism to recognise that LGWM’s modal demand estimates for 2036 contain serious errors. Ideas and assumptions in transport are changing very quickly, among professionals and through society as a whole. Engineering NZ’s latest Transport Group Conference had the theme ‘Change is in the air.’ Who could have imagined, twelve months ago, that school children would be going on strike to demand action on climate change? Will we really see a third of CBD commuters still travelling by car in 2036, as predicted by LGWM? We don’t know.

With so many uncertainties to manage, LGWM might be wise to plan for generous spare capacity on primary public transport routes: rail into Wellington and MRT further south. This might even extend to purchasing delivery options, or more vehicles than needed. If world-wide demand shoots up, small orders for a city like Wellington might take too long.

The combination of highly uncertain demand and high-capacity MRT suggests that mass-transit might usefully be over-provided, within reason. Under-providing seems likely to be the greater risk.

↑ Contents

Light Rail

At this stage, light rail seems to be the only option clearly suited to Wellington and the chosen route. It is also available from multiple suppliers; light rail is well-established and supply-competitive. BRT is also available from multiple suppliers, but TT is only available from CRRC.

The example vehicle chosen by FIT is seven-section, similar to the Gold Coast (G-link) vehicle in the photo. It is 63 m long with a capacity of nominally 470 passengers. Shorter vehicles might be best for the early years, reducing costs, but longer vehicles might be cheaper in the long term. The costly parts of a modern tram are the control system and cabs, and operating cost-differences are almost independent of vehicle length. If lack of capacity is a risk, then longer vehicles could usefully be introduced at once.

Leaving Broadwater Parklands on the Gold Coast Light Rail

The obvious drawback of light rail is the cost of track and diverting underground services. The usual arrangement is that services running along the light rail route are relocated beside it, and services crossing it are relaid in ducts, so that they can be replaced without disturbing light rail. Large drains are generally an exception because they can be repaired from the inside.

↑ Contents

BRT

A new route study can be based on the ITDP BRT Standard. In 2017 LGWM’s consultant WSP recommended design to the ITDP ‘Bronze Standard,’ and gave these assumptions:

  • Full separation from general traffic flows (dedicated lanes), except intersections.
  • High priority at traffic signals.
  • Requires integration with surrounding walking, cycling & traffic network.
  • Fully electric vehicles.
  • High frequency 2.0–2.5 min/direction/peak hour (“realistic/normal” operating frequency of BRT on Golden Mile).
  • Less transfers/interchanges for passengers.
  • Maximum Capacity 150+ passengers.
  • Medium potential to attract car users to PT.
  • Modern low floor articulated bus vehicles.
  • Flexible/less physical infrastructure.
  • Generally fixed route, some flexibility (if required).

BRT is likely to cost roughly the same as conventional buses.

In practice, BRT seems very unlikely to be satisfactory in Wellington, because lack of space in the CBD will require a two-lane route. This might be sufficient with good management, of bus lanes, but can never be enough at stations. BRT stations in Brisbane (scaled from an aerial photograph) are typically about 27 m wide, compared with a street-width of 15.1 m in Wellington’s Manners St, for all purposes. BRT stations need two lanes each way, for buses overtaking buses. Also needed are more bus-berths, dedicated berths for each route (so that passengers know where to wait), and substantial platform width to handle passenger numbers. Some principal CBD junctions may need flyovers, to allow adequate junction time for traffic crossing the busway.

WSP (bullet point 5 above) anticipate a reliable maximum time between buses of two or two and a half minutes between buses on the golden mile, only 24–30 bus/hr.

The only real alternatives to the golden mile are two lanes on the waterfront or two lanes on the ‘secondary spine’ proposed in the Spine Study, using Featherston and Wakefield Streets southbound, and returning on Jervois Quay. Neither is wide enough, with very poor passenger access and legibility.

↑ Contents

Trackless Trams

Chinese developer CRRC is now the world’s largest manufacturer of railway rolling-stock (Newman et al. (2019), p 33, The Trackless Tram: is it the transit and city shaping catalyst we have been waiting for?). CRRC’s Autonomous Rail Rapid Transit (‘trackless tram’ or TT) system is now being trialled in Zhuzhou.

TT might prove an attractive option, but there are surprising uncertainties here. Detailed information from CRRC is still scarce, and some sources seem very unreliable. Much of what is available is dated 2017, and an apparently official video is remarkably amateur. It is not even clear that CRRC have yet begun to market TT.

TT uses digital steering of all six axles to track a pair of painted lines, with supplementary data from GPS and LIDAR. CRRC have paid close attention to ride quality, using high-speed rail technology. The vehicles are battery-powered (in fact condensers), with an anticipated range of 50 km after a ten-minute charge, backed up by an overnight ‘deep recharge’ and a brief top-up at each station (Newman et al. (2019), p 38).

CRRC is offering, or planning to offer, vehicles 30 metres long, in three sections, with a five-section option planned. See the photos below.

CRRC's Autonomous Rail Rapid Transit Vehicle

CRRC now has the largest vehicles on offer, with probably the best ride and the most effective batteries and charging systems. Other manufacturers are also in the market, including Alstom, Van Hool and Irizar (Newman et al. (2019), p 34), offering shorter, bus-based vehicles.

The route capacity achievable using light rail is about 10,000 passengers an hour in Wellington, which seems a reasonable target for TT. A lower target would be more easily achieved but might risk running into capacity problems.

Three-section TT vehicles are 31.6 m long and 2.65 m wide (the standard light rail width). The claimed capacity is 250–300, which seems very high. A standard figure in Europe is a preferred maximum of 4 standing passengers per square metre. Using this figure, and comparing on a floor-area basis (after subtracting two metres at each end, for the drivers’ cabs), gives a TT vehicle capacity of about 220 passengers. A further correction is needed, because TT vehicles have wide wheel-boxes for six axles (like the front wheels of a bus), and the boxing is continued beneath side-facing seats: the seats are set forward from the windows (photo above right). The full vehicle width is only available to passengers around the doors. An estimated width-correction of 300 mm reduces the capacity to 200 passengers, or 330 on a five-section TT, about 50 m long. This is about 70% of the assumed light rail capacity of 470 (FIT example vehicle).

An animated video suggests that two TT vehicles can run in convoy only about a metre apart. If such an option becomes practical, TTs might be capable of running together without coupling, matching light rail capacity and eliminating the need for a four lane route. However, stop-length is another consideration. Finding space for platforms longer than about 50 m becomes progressively more difficult, and extremely difficult beyond about 70 m.

Two potential TT risks are:

  • A typical modern European tram (Siemens Avenio, 63 m long) weighs nearly three times as much as a full load of passengers, but TT vehicles weigh only about 15% more. The risk here is that long vehicles need adequate ‘buffing strength’ to protect passengers in the event of a crash. The whole vehicle needs to be strong enough to absorb the kinetic energy of the rear end with minimum risk to passengers.
  • TT in New Zealand will need careful checking for compliance with regulations, regardless of whether the system is treated as bus or light rail. In either case, new regulations will be needed, and may need legislation. Wellington would gain a dual advantage from choosing ‘the same as Auckland’: no regulatory costs, and cheaper vehicles and equipment because of repeat orders.

In Looking past the hype about trackless trams, Wong (2018) points out that TT is not really revolutionary, and alternatives to light rail have been available for years. However, Wong also challenges TT’s ride quality, which might be unfair, but his paper is still of interest.

A guide and manual with application to Trackless Trams, a paper by Peter Newman et al. (2018), develops a new method of assessing public transport, specifically with TT in mind:

Traditional transit planning does the transport engineering first and then adds the land use planning as a supplement after finding government funding; the approach being presented here starts with the land development planning and then does the transport engineering after achieving the funding/ financing from the land development potential. [p 6]

Four approaches to capital are used: broadly, all-public; mostly public; mostly private; and all-private.

While the paper seems very useful (and note the BCR below), explicitly applying it to TT seems doubtful:

By integrating higher value into land development within cities, rather than having further land development on the urban fringe, there are significant public and private benefits that vastly outweigh the costs. Some BCR calculations have seen a simple light rail project with a BCR of 1.5 increase to around 7 because of the increased land development. This not only saves public money in infrastructure costs (usually 1.5 times as much as redevelopment) but also provides transport time savings for those living in the [Transit-Oriented Development areas (such as WCC’s plans for Adelaide Rd)] (based on all transport usage). Thus, it is important to ensure land value increases are integrated into the full transit and land system upgrade process. [p 6]

Clearly, the model also works with light rail, but perhaps more worrying is this:

Towards the end we show that a Trackless Tram is likely to be the new ‘rail’ system for cities as it does all the things light rail does but costs one tenth of it. This low cost makes it possible for entrepreneurial developers to build such systems as it will unlock their developments. [p 14]

TT at a tenth of the cost of light rail is implausible. While the four-level model is interesting, other sources suggest that saving 90% of light rail costs is unrealistic. One of Newman’s errors has been picked up by Matt L:

The press for the trackless train claims the vehicle can hold 300 people. This seems highly unlikely given the vehicle is only about 30m long. As a comparison, AT say that a 66m light rail vehicle will hold up to 420 people. The interior of the vehicle doesn’t suggest a huge amount of standing space either and a capacity of 180–200 people seems more realistic. But even if it could hold 300 people, it’s not enough, which is why AT are going for higher capacity vehicles.

Newman himself notes (Newman et al. (2019), p 39) an Australian estimate of a third of the cost of light rail, which seems a reasonable starting-point; real-world costs must cover more than painting double white lines.

Trackless trams, like BRT, look tempting because they seem far more cost-effective than light rail. This has gone on for a long time, and Wong (2018) refers to a 1994 paper, by Henscher and Walters, titled Light rail and bus priority systems: Choice or blind commitment?

Perhaps the largest single risk when adopting alternatives to light rail is the simplest. Decision-makers have repeatedly demonstrated how easily they can convince themselves that anything without tracks must be better than light rail.

An example is that UCL, in Innovative technologies for light rail and tram: a European reference resource Briefing paper 1 Tyre innovation–rubber tyred trams (a 2015 review of earlier versions of trackless trams), commented:

All (BRT) systems installed to date have been more expensive than conventional tramways.

At least two of those systems were replaced by light rail.

A related blind-commitment temptation is assuming that only light rail needs to disturb underground services. The ignored risk is that underground services can disrupt TT, just as they have always disrupted present-day motor traffic:

  • TT/BRT proponents, including CRRC, claim the benefits of being able to avoid a crash by manually steering around the obstruction. This is as much a disadvantage as an advantage, because the converse is motor vehicles running on TT/BRT ‘tracks.’ Light rail experience in Britain is stoppages when parked cars obstruct the track, and TT/BRT must also address these risks. The light rail photo on page 3 shows a kerb outside the tracks (at right), with prominent ‘TRAM ONLY’ signs painted on the road, to discourage motor vehicles. Light rail has to maintain an exclusive corridor, and effective TT will need to do the same.
  • If TT/BRT is seen as not needing underground services diversion, decision-makers have unwittingly accepted the risk of delays or damage when underground services fail. Motor traffic is frequently delayed in this way, and drivers manage it by travelling at other times or taking an alternative route. Road signs warning of future disruptions are commonplace. Neither management option is available to either TT or BRT, and Wellington has recent experience of the effects. When the Hutt railway line was washed out in 2013, motor traffic also came to a standstill, for several days. Ignoring the need for services diversion for TT/BRT will tend to have the same effect, rarely over days, but even ten minutes can be very disruptive.

Wellington decision-makers need to face facts here. Two major studies, the 2011 Bus Review and the 2013 Spine Study, were wiped out by ill-considered cost-savings. Ten years after the problem was first identified, Greater Wellington still has a heavily overloaded bus route and no plans for improvement. This process, of unconsciously working towards a substandard outcome, is well-known; blind commitment is one term, but Wikipedia calls it BRT Creep:

BRT creep comprises several types of gradual erosions in service that sometimes affect a bus rapid transit (BRT) system, resulting in a service that is not up to the standards promised by BRT advocates. In its ideal form, BRT aims to combine the capacity and speed of a light rail system with the flexibility, cost and simplicity of a bus system. BRT creep occurs when a system that promises these features instead acts more like a standard, non-rapid bus system… The most extreme versions of BRT creep lead to systems that cannot even truly be recognised as “Bus Rapid Transit”.

This is what happens when the bus lobby sidles in and whispers, “we can do exactly the same for half the price.” They do, and they can’t.

↑ Contents

Costs

Costs for TT vehicles are roughly comparable with light rail; say about $80 million to run a five-minute service.

Other cost estimates vary wildly, but real-world costs must cover more than painting double white lines:

  • Road re-grading as needed; TT videos show well-levelled surfaces everywhere. TT vehicles use the same low floor-level as light rail, and will tend to need similar large-radius vertical curves.
  • Heavy-current, high-voltage power at all stops, termini, and especially the depot.
  • Stations, including platforms, shelter, passenger access; ticketing machines and connections at hubs.
  • A depot, with scope for expansion.
  • Motor traffic realignment to make room for TT.
  • Integration with traffic signals for TT priority.

Any TT cost-estimates for Wellington will need great care, using data from existing users.

Ensuring a dedicated and separated corridor would future-proof TT to support fully autonomous operation when the technology matures: light rail is future-proofed by design.

The first light rail line in Montpellier opened in 2001, and in 2008 was carrying 30 million passengers a year. A cost analysis from Marc le Tourneur (2011), Making the case for trams and regional trams, showed that buses and BRT both cost about 45% more than light rail:

light rail (actual figures)
Investment cost per passenger€ 0.93
Operating cost per passenger€ 0.53
Total€ 1.46
buses (actual figures)
Investment cost per passenger€ 0.49
Operating cost per passenger€ 1.61
Total€ 2.12
bus rapid transit (simulated using data from Nantes)
Investment cost per passenger€ 0.84
Operating cost per passenger€ 1.27
Total€ 2.11

Montpellier (populaton 290,000) now has four light rail lines, with a total length of 60 km.

Data from Transport for London gives equal costs for buses and light rail at about 3200 light rail passengers an hour; a little higher and light rail is cheaper than buses, and a lot cheaper when light rail is running at capacity. One reason is that savings on operations cost are sufficient to pay for greater capital costs. Roughly 70% of operating costs are driver’s wages, for either buses or light rail, but one light rail driver replaces some four to six bus drivers.

↑ Contents