Saturday, September 5, 2009

49> In Search of Gridlock and Opportunity

I have, for the last couple of posts, been trying to figure out how various PRT designs would fare with limited track and stations. My quest deals with the recognition of the fact that no investors in their right minds would sponsor a full network as a trial. (I am not abandoning the possibility of private money just yet) Like it or not, we have to look for the most cash flow from the least network. This exploration has led me to recognize the fact that a few dispersed low capacity stations may not introduce passengers to the system fast enough. Paying passengers must come from somewhere. If they do not come from dozens of smaller stations, then they must come from a few larger ones.

I understand how PRT is supposed to work. It is not a hub-and-spoke, mega-station type of technology. Nobody has described PRT vehicle storage because it is not supposed to be required. But if there is any way some track can be put down, and some vehicles deployed, expansion could start from there. The “foot in the door” it seems, would need to be some situation where people are so inconvenienced that they would be happy to pay a premium to take a short ride.

So here is a building block. This is a bare-bones seven-berth station that can have a 60 second turnaround time. (In one minute all berths have been loaded and moved out, and a fresh set of vehicles has taken their place.)

This illustration shows a high-capacity station comprised of six of the seven-berth stations arranged in parallel, with three boarding scenarios. The arrows indicate foot traffic. The first (A) shows how the station could be adapted to accept a majority of passengers who are boarding. Green lights or programmable signs would direct passengers to berths. Red lights would indicate that people are disembarking from that area. The middle figure, (B) shows a balanced load between incoming an outgoing vehicles. The last figure (C) shows the station accepting large numbers of incoming passengers with few heading out. Green lights mark the three berths accepting passengers. Such a station could, with the 60-second turnaround time, process 42 cars per minute or 2,520 cars per hour. (Note; This uses the gondola design: bottom track designs would require a means to get passengers to the boarding areas without crossing tracks.)

The illustration above shows a “hub and spoke” system with the main station connected to six satellite stations. In this scenario, the time required to completely empty the main station is one minute, the same as the satellite stations. For the sake of simplicity, let us assume that the satellite stations are a half-mile from the main station and that the travel speed is 30 mph. (I originally designed this to take people from a stadium to satellite parking, so that’s why the distances are so close) The number of stored vehicles is 42, so that with the 42 in berths and the other 42 from the satellite stations, the stations would never run out of vehicles. (126 cars total)

These examples point out a few key concepts. First, it is a closed system. PRT advocates tend to see PRT as an open-ended network, with an endless supply of vehicles, which will always self-balance. (i.e. a BIG network) In this example, all stations would run out of cars in 60 seconds, were it not for the stored ones. It is only the close proximity of the satellite stations that allows replacements to get there in time. For every additional half-mile, another 42 cars are needed.

More can be learned from the example. With only 3 miles of track, it is extremely limited, yet it would still probably cost over 30 million dollars. Could it pay for itself? One thing worth considering is that the full capacity would only be utilized for, perhaps, 3 hours per day, 5 days a week. If the cost of money were, say, 10%, the daily interest alone would be $8,219. What price point and usage pattern could bring in, perhaps, $13,000. to put the project solidly into the black?

Again, I designed this for a stadium, where the time to just get the whole family into and out of the parking lot would justify some very high fares. (not to mention that 90% of the track would be on private land) After checking the schedule of a local stadium, however, I realized that these big events are so infrequent that the system would never be viable, even though it could generate upwards of $40,000. per event. I now put it forth as something to ponder only. If the main station was broken into two or three, and the system stretched a bit, it might cut driving time by 10 minutes each way for some particularly congested downtown areas. My logic was to think up a scenario where passengers would gladly pour into a super cheap station to hardly go anywhere for top dollar. Investors, after all, are not known for altruism. We can all imagine a fully-grown, distributed PRT network. Can we come up with a viable PRT seed? For, say, 30 million?

18 comments:

Anonymous said...

Focus on connecting airports to CBDs and existing transit hubs. People pay lots of money for airport taxis. Airports also have a much steadier flow of passengers than stadiums, since planes arrive and depart 24/7, and air travelers are highly transit-dependent.

The AirTrain connecting JFK terminals to the LIRR and NYC subway system cost $230 million a mile. It carries 15,000 paying passengers a day at $5/passenger for a 1.5-3 mile ride--or $1.70-$3.30 per passenger mile. Even at that steep fare it only recoups less than 2% of its construction cost each year. And just look at it. When people complain about PRT cost, capacity and intrusiveness, show them this.

cmfseattle said...

you'd have to make the guideways stronger, but "coupling" vehicles might make sense.

Extending PRT Capabilities, by Ingmar Andreasson

the seattle monorail moves an average of 5,000psgrs/day between 2 food courts, and turns a profit. the capital costs were paid back by being able to handle 5,400pphpd during the 1962 world's fair.

Prof. Jerry Schneider's Bellevue Downtown PRT Planning Study used a $16M/mile figure and a limit of 12 guideway miles.

Anonymous said...

I don't think physical coupling is necessary, and it introduces some problems. But it might be possible for vehicles to be able to get arbitrarily close to each other and even touch if a shock-absorbing bumper system is introduced--say, one on each end extending a few feet in front of the rest of the vehicle. This, along with some creative fairing, can allow vehicles to draft each other while still maintaining a degree of safety should a sudden deceleration be necessary. That said, the high capacity that vehicle coupling or drafting can provide is not likely to be necessary in the initial systems. If the Seattle monorail's 5,000 passengers all rode within, say, four hours of the day, that would still only be 1,250 passengers an hour, well within even the most rudimentary PRT system's potential capacity range. And of course the fact that you can build multiple PRT lines for the cost of one conventional transit line means that single line capacity isn't really that great a limiting factor.

Ollie Mikosza said...

Dan,

It all depends if someone believes that PRT can be made to have a subsecond density vehicles separation. If not, then it ends the discussion, because then, as the PRT foes say, it will never be able to have high capacity, therefore profitability.

But I know, that my 1st 7 mile MISTER system (seed configuration, costing $50 mil) WILL be able to deliver over 5,000 ppdph, with some 35,000 passenger-miles per hour. This is more than any LRT or bus system of the same length and only 50% less than a subway (which would cost 30 times more). Of course, with the growth of the PRT system, it becomes better and more profitable.

And I also know, that a 7 mile LRT systems are designed to carry 50,000 passengers per day, even in such countries like Poland. With such parameters, and an average ticket price of $1,50 for an average 3 mile trip, such a system would entirely pay for itself in e.g. San Jose (or any other US city) witin 1-2 years.

But it seems that we can discuss it till we are blue, because nobody, even you, want to look at these numbers as something realistic. This is why you are searching for esoteric schemas of some large station, multi-level stops, remote and low cost stops etc. There is no need for such ideas, because simple, small yet numerous stops in densly built up city areas or around sports stadiums are more than feasible and more than enough.

Looks like you didn't read my Capability Analysis (http://www.mist-er.com/mister-description/capability-anaysis.html) or if you did, then you didn't tell me if and wrong you've found with it.

Anyway, once again all the best to your continued search for the best PRT system but it seems that I have nothing more to offer, other than what I did so far.

Cheers to all of you - Ollie

cmfseattle said...

"But it might be possible for vehicles to be able to get arbitrarily close to each other and even touch if a shock-absorbing bumper system is introduced--say, one on each end extending a few feet in front of the rest of the vehicle."

please read up on the french Aramis system. i'm not suggesting vehicles couple and uncouple at speed. i'm suggesting they be coupled just before rush hours in order to increase the main line capacity without headways below 3 seconds.

http://www.advancedtransit.net/atrawiki/index.php?title=Aramis

http://faculty.washington.edu/jbs/itrans/pulse.htm

once again, the problem is: how to pay for a starter PRT system? you'd have to charge passengers $3/trip if you can only handle 5,000/day. and 1,250/hour is near the upper limit at 3secs headway, 1.2psgrs/vehicle average.

ULTra are keeping costs down by limiting the amount of elevated segments and keeping it simple.

cmfseattle said...

my thoughts of late:
what are the top 3 things you'd change about transportation if you could?

1) cost -- both in money and time. there's insurance, maintenance and fuel. currently, 2nd biggest purchase, for most.

2) convenience -- cars can serve more functions, but need parking (add to cost, above). unless you live rural, there's a limit to how many can fit during rush hours. transit can often take twice as long and, like bikes/scooter/motorcycles, can mean dealing with the weather.

3) culture -- from rude drivers/passengers to international politics (social justice, energy, materials, etc.), it's better to lead by example.

no one (almost) suggests curbing population as a solution to congestion, right? so, the goal is to move more people and their stuff from home to work, school, play, shopping, recreation and back, in the most efficient manner possible.

maybe that means starting with fenced rights-of-way and low speeds. once you prove that it's safe and cost-effective, you can continue toward intercity maglev.

Asko K. said...

I think you're on the wrong foot in pricing individual trips on PRT usage. That's at least not how public traffic works here (Finland). We're trying to get away from paying for individual trips, and for people to use season cards instead. It's way more simple for everyone.

This offers another business model: selling the PRT system as a transport solution, guaranteeing certain bandwidth between stations and maintaining the service at any time.

In this case, the customer is responsible of charging the end users. It also means, it is the customer's concern to get much usage on the system. The PRT operator will get fixed price agreed up-front, not based on the number of passangers actually using the system. This takes a huge risk away from the PRT operator's shoulders, and more precisely focuses the roles of the players.

Whether there is any way such a business plan would work among the Americans, I don't know. But it would work in many places, and it especially makes sense where other traffic modes (buses, local trains) are being paid for with the same travel cards (s.a. in Helsinki region). We are looking for efficiency of transport (not multiple parallel services covering the same need) and still maintaining competition within that transport network. Maybe Helsinki and HKL is a world first in making this everyday reality already now.

(What I mean is that i.e. introduction of PRT in a certain area could be used to completely remove local buses from there. In a non-managed environment, the buses would remain and try to compete with PRT. In the transition there would be worst of both worlds - PRT track pillars and buses running half-empty. Not ideal. If both modes are part of the same overall traffic planning, this can be avoided. Also Heathrow has the same case, ULTra replacing airport bus lines.)

Dan said...

Dan The Blogger responds[

First of all, it is my style to break complex issues into small, digestible pieces. It is how I learn, design, and teach. This blog is ideally suited to this methodical approach. I hope to see a full set of design specifications come out of this effort one day, but that will take the talents of a much bigger audience than I have so far. In the meantime, I am exploring the issues one at a time. There are lots of myths, assertions, and assumptions that fly around the topic of PRT and any good system needs to grounded in reality.

The underlying inquiry here is two fold; what is the smallest, least expensive PRT configuration that makes economic sense, and secondly, can PRT stations keep up? As per the first question, I would like to note that this is important from a PR point of view. It is psychologically easy to take a workable concept and scale it up. Not so easy in reverse. As for the second question, this exercise indicates that, for one thing, the advocates of “track on the bottom” have a real design challenge in a high volume station, because of the “crossing the tracks” problem. (Andreasson, in the above link envisions a cumbersome scheme up to 65 PRTs in a row to handle transfers from a train)

Anonymous, Thanks for the perspective view. I wish there wasn’t often a commuter dead zone between airports and CBDs. In many cities the airport is way out of town. It would be nice to get part way to some long-term parking, though. Of course for that matter it would be nice to park the car and PRT into town.

Cmfseattle brought up the “linking cars together” discussion. Sorry, that’s a post I haven’t written yet, and won’t get into here. (Although you folks can, if you choose,)
I would like to point out a simple fact, however, than in regard to the example above, the following may be stated. Total track=3 mils=15,840 ft. Number of vehicles in the system = 126. (This is in the stations and storage as well as on route) 15,840 divided by 126 equals the minimum distributed spacing. (on center) That number is 126 ft. This assumes ALL vehicles are on the track and none are in a station. If a third of the vehicles are boarding, that number is 188ft. That kind of spacing would hardly seem to cry out for coupling. As I was saying back a few posts ago, for “mini” system the problem isn’t track, it’s the stations. I’ll try to post on this subject soon.

Ollie, This really isn’t a good forum to discuss numeric details and detailed performance claims, unless they can be very easily and concisely verified. The whole PRT industry has a credibility problem which is baked in to its fabric, because there are so few impartial observers to audit claims, and because so many projections must be premised on assumptions. That is why vendors generally take such pains to authenticate their data and stick to a rigorously scientific approach. I do not believe nor disbelieve. I am an optimistic skeptic, and for the purpose set forth in this blog, I think that is how I should be. I try to make a few good points and get people thinking each week, while I build readership. In the broadest context, this post was about how station capability and system scale might effect vehicle and track design. It is also an exploration of the boundaries of system design. I will generally try not to endorse or reject anything in this blog until it is time to commit to a design choice.

Dan said...

Oops! 4096 characters max per comment.
Dan continues
Hi Akauppi,
The exact method of payment would seem irrelevant at this stage. For the purposes of analyzing cash flow in a hypothetical mini system it would seem that the more straightforward way would be best. I guess that most systems envision a payment card of some type, probably including ID involved to keep down vandalism. If a card is swiped, what’s the difference? Isn’t it all the same except in the computer that is doing the accounting? I suppose later on a season pass or express check-in might save some time, though.

I’m really not sure who the “customer” is in your example. The city? In the U.S. I think there is a pretty complex relationship between the city, the state, the federal government, and the transit authority. I do not pretend to understand even half of it. Revenues include things like city services, property taxes, city-sponsored bonds, federal and state grants. Sometimes politicians “bring home the pork” as it is called, directly to their local constituencies in the form of a government grant. However it’s done, PRT should be revenue positive, especially if it is to start out as a public private partnership, such as an airport or campus that extends onto city property. I think it’s one step at a time, unless, of course, this would have a bearing on the physical designs. I guess my concern is that some payment schemes my hide, to some extent, the financial realities, which might make it unclear whether a project is a success or failure, and by how much.

cmfseattle said...

Dan, at 30mph line speed, 188ft = 4.28sec headways and at 4psgrs/vehicle, you'll take an hour to move 4,500psgrs. coupling doubles that number, so you could move 20% of a WWE event (they hold the attendance record of 50K at seattle's baseball stadium). then again, that would be double the national average transit ridership.

however, if you need ~$13K a day to turn a profit, you're barely squeaking by without the capability to couple at least some of the vehicles (especially with so few stations).

i think your comments about remaining grounded in reality are spot on. financing does affect PRT design.

John Greenwood said...

Hi

Can I point your readers to some rather simplistic but instructive simulations of PRT stations I did recently.

Download the EXCEL spreadsheet:

http://distart041.ing.unibo.it/~john/Berths/images/Downloads/Sequental.xls

You should open a spreadsheet with some cells coloured green and red. Press F9 to iterate and they will move up.

Now load: http://distart041.ing.unibo.it/~john/Berths/images/Downloads/Iterate_macros.xls

This contains a set of macros which repeat the interation a number of times as described in cells S3:T7 so, for example, if you press ctrl-U it will do a cycle of 3 iterations which is one time unit.

If this dosn't work for you, PLEASE let me know at John@john-Greenwood.co.uk.

There are models for other types of berthing and discussion at: http://distart041.ing.unibo.it/~john/Berths/Models.html

In all these models you can change the configuration, for example in the Sequential model, changing the logic state of cell L52 to FALSE will disable the berth on that line

The passengers have a randomly variable boarding time with a Weibull distribution. Parameters for four different passenger mixes are available on all the sheets.

So for example, to change the mix from "typical" to one with fit people only, copy the cells N21:N23 (coloured blue) to H21

I would not claim that the results are particularly accurate but good for comparative purposes. In particular it shows how throughput tends to be limited by congestion of vehicles.

John Greenwood said...

Following my contribution above, here are the links as html:

Sequential Berthing spreadsheet model

Iterate macros

Page with links to other models and discussion

Also relevant is a chart I made to try to put PRT passenger volumes in a wider contex.

cmfseattle said...

John, thanks, although the chart at the bottom of the "capacity of PAT guideways" (PATcapacity.html) is at first a bit confusing; seems like the left column should be 10, 20, etc. Km/H, and the bottom row label vehicles/minute?

anyway, i think that uncoupling at the arrival end of a station (to facilitate a parallel config), and coupling 2 or more vehicles with the same destination at the departure end, would be feasible. automatic couplers aren't new and allow emergency pushing (as in the taxi2000 design).

John Greenwood said...

cmfseattle, Thanks for pointing out the muddled axes labels, now corrected.

Coupling individual vehicles will increase capacity particularly if the speed is greater. This would not affect the number of berths for a given passenger flow but would mean that station capacities could be even larger.

If mass transit traffic is carried on the same guideway as PRT, it will have to negotiate the same tightness bends.

Here is my idea for such a vehicle.

Which would be articulated as in this Animation.

At higher speeds this has attractions for low drag, low energy consumption and passenger numbers similar to a bus.

cmfseattle said...

Capacity in a PRT system is mostly a function of the number of pods, and short headways between them. Congestion is avoided by having a set number of pods, in contrast to the continual increase in new automobiles being put on the roads. Capacity is the number of trips each pod makes, times the number of seats per pod, times the number of pods in the system. Just as an example, in a fleet of 1,000 four-seat pods each making five trips per hour, the capacity is 20,000 passengers per hour. Therefore on-demand service is the chief difference between PRT and light rail -- light rail is good at moving large groups in trains many minutes apart, along corridors; PRT serves the same number in smaller groups, with pods sometimes separated only by seconds, around a grid-like network.

i figured out that our 3-mile example system should be able to handle well over 100K vehicles/hr ((15840ft/132ft/sec)*20new trips/min)*60mins = 144K trips/hr.

however, once again, that's only if all parts of the network are being utilized to an equal degree (balanced load). i'm suggesting coupling in order to increase the maximum psgrs/hr along the most congested segment(s), using standard-issue vehicle$.

John Greenwood said...

There will never be enough guideway in a successful PRT network. There will be strong opposition to every km of new track on grounds of cost and planning.

The answer to the cost is simple: get as much revenue out of each bit of the guideway as possible. This can be done by:

- Coupling vehicles
- Reducing headways
- Higher speeds
- Mixed traffic, freight when passenger demand is low and
- Mass transit vehicles as well

I suspect the planning issues are less of a problem in the USA than in Europe. Streets are narrower and more chaotic in layout and the citizens are much more protective towards scenic aspects of historic buildings and street scenes.

A PRT system will be more acceptable if it has:

- Low bulkyness
- Compact stations
- Stations at grade
- Can go underground

The ideal network would have many parallel routes to a particular destination but I think that in real systems there will be regions of guideway linked by a single route. (eg. either side of a river) It will be the capacity of a single guideway that matters in this situation.

Dan said...

Dan The Blogger Returns-

Thanks for the posts, John. I don’t currently have Excel,…(maybe that’s a blessing…)
I have been meaning to download “Excel Viewer” from the MS site to see if it would work for your simulation but haven’t found the time. I’m afraid that this blog is something that I have to fit in weekends and after-hours.

You envision a much bigger system than I do, and, as you say, there may be some cultural bias for this. Anything even remotely resembling a train would run into huge opposition in most US cities, and would run squarely up against the forces of traditional light rail, which no initial PRT system here is likely to survive. I think as light as possible is the key. In the U.S., we have the experience of the Raytheon model of PRT, which was shown to be too massive for the tastes of the cities they approached. I will readily admit though, that getting the economies that you mention with a smaller lighter system will be a challenge.

Hi cmf.. That link only worked once for me and now gives a 404 message. It is a great 2-part summery and history of PRT, and can be read by searching for “podcars” in the Examiner’s search box. There is little telling what others will experience, however, because the site tailors itself to the city of the reader. I must say, though there is some really great old grainy footage of the Aerospace Corp’s early PRT test model doing its thing. I’m going to put that in a “best links” post one of these days.

Unknown said...

LibreOffice will run on MS and read Excel sheets.
But on topic, I think you have to sort out the stops further. On a two-way system, passengers arrive at a stop that is attached to one direction. How do they go in the other direction?