Mixed-mode Personal Rapid Transit (PRT) and Group Rapid Transit (GRT) - Some Reasons Why It Should Be Considered
by Nathan Koren
Vectus PRT is adding GRT vehicles for a very good reason. Although PRT can be designed to handle sharp fluxes in demand, it isn't necessarily economical to do so. In many cases, a mixed-mode PRT/GRT system is the best response, and I'm glad to see Vectus PRT advancing the industry this way. (See the recent Vectus PRT video for illustrations of their mixed-mode concept at: http://www.youtube.com/watch?v=S1rf_lOb3b0
Let me illustrate this with a somewhat contrived example. Imagine you have an application which operates 18 hours per day, and during most hours, has 1,000 passenger-trips heading in all directions. Let's say that each trip takes five minutes and carries 1.5 passengers per vehicle, and 30% of the vehicle-trips are empty (to accommodate repositioning vehicles). Do the math, and you'll find that this system requires 94 vehicles to serve that level of demand on a continuous basis. Not bad, to carry 18,000 passenger-trips per day with just 94 vehicles, right?
However there is also a rush hour in the morning, and another one in the evening, when 4,000 passengers trips per hour need to take place. How does a PRT-only system handle this problem?
1.) You can add more PRT vehicles and run them as before. Now you've just bought a fleet of 376 vehicles, 282 of which will go unused for almost 90% of each day. You also need to make the stations equivalently larger in order to accommodate this number of berthing vehicles. You can't amortise the cost of a PRT vehicle (or berth) by utilising it for just 2 hours per day, so that's going to significantly worsen your business case.
2.) You can create incentives for ride-sharing and ways of accommodating it within the ticketing and passenger-flow systems. In practice, it's possible to get average vehicle occupancies of up to three people in most circumstances, although this requires quite a bit of creativity to do successfully. Now you only need to buy 188 vehicles (and equivalently larger stations), but you're still not going to be able to amortise 94 of them.
3.) Because most of the commuting fluxes like this typically appear along predictable corridors, you can serve the rush-hour demand with GRT vehicles. Unlike a PRT vehicle, GRT vehicles can carry people until they are standing-room-only. So, in this hypothetical system, you can serve the excess 3,000 passenger-trips during each rush hour with 24 18-person GRT vehicles. You'll still have difficulty amortising these vehicles, but since they're only (at a guess) about twice the price of a PRT vehicle -- and the passenger-side space requirements for GRT vehicles are also relatively modest, reducing the cost burden on the stations as well -- you'll get a much better business case from this scenario than from the other two examples.
This might seem like a rather contrived scenario, but in practice, many of the business cases that we're seeing actually look quite a lot like this. This is mostly because interest in transport solutions is highest in areas where peak commuting fluxes are causing the severe problems. There are certainly a lot of urban transport problems that don't resemble this example, but that isn't where the market is at right now. Vectus is doing a very smart thing by catering to the demand pattern of the market, as well as making sure that the GRT infrastructure is fully PRT capable as well.
I'm sure that when PRT vehicles are drawn from a production line which produces millions such vehicles per year, vehicles will cost about $10k apiece and the financial calculus will be quite different. But in today's world (2012), PRT vehicles are still built in very small custom-production runs, and the cost per vehicle is considerably more than $10k each -- enough so that optimising the fleet size is a very real and necessary concern.
Also, it's not just about the number of vehicles; it's also about the size and cost of the stations. To give you an idea of how that math works:
* A single PRT berth is 5 meters long and can launch an occupied vehicle roughly every 30 seconds (this number can vary a lot, depending on system specs and station design, but this is a pretty optimistic figure I'm using.
* If each PRT vehicle carries 1.5 passengers, then that's 180 pax/ berth/hr, or 36 pax/hr per meter of platform length.
* In practice, to accommodate shorter-duration demand fluxes than the hourly rate would indicate, you need to leave a margin of berth space that is well above the minimum. Let's call it a 100% buffer for our purposes; so we get 18 pax/hr/m of linear station platform.
* To serve 1,000 pax-trips per hour, you thus need about 56 linear meters of platform space.
* To serve 4,000 pax-trips per hour in the same way, you'd need at least 222 linear meters of platform space -- and again, 75% of that would go unutilized 90% of the time.
* You'd also need to find a place to park the other 282 vehicles, which would add up to another 1.2 kms of off-guideway buffers somewhere. That would be very expensive.
On the other hand:
* A GRT system which can inter-operate with 2 PRT berths can also launch an occupied vehicle every 30 seconds.
* If each GRT vehicle carries 18 people, then that's 1080 pax/berth/ hr, or 216 pax/hr per meter of platform length.
* This doesn't really need to have an extra margin on it, because the GRT runs during rush hour on a regularly-scheduled interval, but let's give it one anyway
* So to take up the extra 3,000 pax/hr during rush hour, the GRT needs an extra 28 meters of of platform space.
* For storage outside of rush hour, the 24 GRT vehicles would need about 0.3 kms of off-guideway buffers somewhere.
So, to compare the two options:
* 376 PRT vehicles, vs. 93 PRT vehicles and 24 GRT vehicles
* 222 linear meters of station platform, vs. 84 linear meters of station platform
* 1.2 kms of extra off-guideway buffers, vs. .3 km of extra off-guideway buffers
And keep in mind that with this arrangement:
* During 90% of the day, the system is still PRT-only.
* Even during rush hours, trips that are not along a natural GRT-served demand corridor can still be taken by PRT
* Waiting time for the GRT vehicles, during rush hours, would not be significantly longer than for the PRT vehicles
* Capital costs are dramatically lower
* Space utilisation is dramatically more efficient
Hope this helps you to see why Vectus PRT and others will be evolving in this direction in the future.
Again, in scenarios where the demand curve has low peaks and high valleys, all-PRT systems make sense. But where the demand curve has high peaks and low valleys, what makes sense is some kind of mixed-mode system, with the complementary mode being one that is able to accommodate "crush" capacities. This can be rail or bus, but running GRT off the same infrastructure that you'd have anyway for PRT, often makes the most sense. And, mixed-mode PRT/GRT should be able to offer higher levels of performance with lower capital and operating costs than conventional rail or bus systems.
Other papers that discuss the ride-sharing option:
Extending PRT Capabilities, by Ingmar Andreasson - explores ways to increase the capacity and speed of PRT
Ride-sharing on PRT, by Ingmar Andreasson - includes detailed simulation results for various ride-sharing strategies
PRT as a Feeder/Distributor to Rail, by Ingmar Andreasson, October, 2011
Ridesharing Methodology for Increasing PRT Capacity, by Peter Muller (2012 TRB presentation)
MicroRail is currently developing a mixed-mode system in Texas
Cabintaxi was the first system to offer mixed-mode service. It was developed the 70's in Germany and is currently awaiting a revival effort.
Last modified: February 08, 2012