Saturday, August 9, 2008

3> First, there is the track

First there is the track. This is the most important part, and is half the reason the world needs to “open source” PRT design. I have not seen a single PRT system that I think will ever find any success in the U.S. and here is why. PRT can not compete with buses or light rail in situations where people can aggregate at a point A in order to make a straight line journey down a main drag to points B, C, etc. It is simply inefficient to load and transport many small vehicles as opposed to a few large ones.
As we all know, however, mass transit has real problems competing with cars because it can’t effectively reach less trafficked areas. Yet many people are coming from a less trafficked area, and are destined for a less trafficked area, and are effectively forced to commute through highway and intersection bottlenecks. Mass transit though these bottlenecks is unworkable because there is no infrastructure on either side of that bottleneck that would enable a commuter to complete the journey.
The answer is cheap, go-anywhere, go-over-anything, rails, and go-exactly-where-you-want cars to compliment them. We don’t need a couple of loops. We need a network. This is a critical mass situation. The system needs to have more destinations than is practical with the alternatives or else the alternatives should get the funding.
One problem with the current development model is that the companies hoping to build a PRT system covet the idea of the never-ending construction job that would be the track. It is no wonder that they tend to envision more grandiose guide-ways than the needs would seem to indicate. The same is true of the cars. Just like GM liked the business model of producing SUVs, instead of compact cars, so does the potential sole supplier of PRT “pods” covet the endless production of larger, more expensive vehicles.
Here is a picture of the Raytheon’s idea of a PRT vehicle.

PRT 2000

Is it any wonder that the cities that Raytheon approached eventually rejected it? Note the 72” x 72” track size. Clearly minimal cost was not the object here. The use of a 36” horizontal tubular support beam is a particularly odd choice, as ordinary “I” beams are cheaper, stiffer, less resonate, easier to handle, easier to bolt to, and more available.

5 comments:

cmf-seattle said...

http://www.seattlemet.com/issues/archives/articles/wikispeeds-100-mpg-car-january-2011/

I think we need to design something like Shweeb, but buildable by anyone in their garage/backyard. If it's simple enough for the DIY-er, I think it'd stand a chance of widespread adoption.

Dan said...

Does Shweeb have a way to switch? If so, how do you prevent a crash in mid-air?

Eric Dunn said...

I realize this is reviving an old post, but I had a couple questions/challenges:

First, the premise is wrong. "First there is the track". Uh. Nope.
First there is the need.
Then there is the vehicle that can meet that need.
Then there is the track that can support that vehicle.

The faulty premise that the track comes first leads to the false conclusion "Clearly minimal cost was not the object here."

It certainly seems to be the mythos of the Raytheon system that it wasfor some reason deliberately built too big and too expensive. I've even read that the design was based on left over pipe that Raytheon wanted to get rid of. But what possible incentive could Raytheon have to not build to cost estimates or to inflate those estimates?

The design decision that needs to be questioned and validated or rejected is the choice of vehicle size, capacity, and performance. THAT is what will then drive the specifications for the track/guideway.

The design/build disaster at Morgantown was two fold. First, in terms of PRT 'purity' the design decisions that led to bigger vehicles instead of smaller. Second, in terms of engineering and cost the quideway was overbuilt in the extreme as the vehicle design was not settled and getting bigger when the guideway was being designed. Waiting for the vehicle design to solidify would have resulted in a smaller guideway.

"as ordinary “I” beams are cheaper, stiffer"
Is there anyone invloved in PRT that has a civil engineering background that can validate or reject that statement? Or a beam design document that compares the advantages and disadvantages of different beam selection?

I'm not a subscriber to conspiracy theories. So I find hard to believe that either Raytheon in the past or Vectus more recently would either hire incompetent structure engineers or deliberately run up the cost of their test tracks.

Dan said...

Dan the Blogger Responds - You make great points and pose challenging questions, Eric! It took a lot of thought and a two part answer, since they limit you to 4096 charactors:

I stand by my statement. Clearly general purpose comes first, I’ll grant you that. But I think going from there to vehicles first and then to track second invites failure in the case of PRT. This post largely reflects my total agreement upon reading Ed Anderson’s anguished account of how Raytheon enlarged (and ruined) his well-researched and carefully prioritized design. I agree that they fundamentally blew it, partially by taking the very steps you suggest: If the vehicle is given design primacy, it will always end up being “family sized”. I believe that such a system cannot succeed without government funding because PRT needs such a large network to be competitive. (GRT and shuttles don’t, but they need much larger vehicles to be cost effective.) The math for expanding a first phase PRT system into profitability barely adds up as it is, and any public pushback, any restrictions on routing or station flexibility, even the cost of added steel makes PRT even less attractive as a business proposition than it already is. Also the track is permanent and cannot be easily modified while vehicles can be improved or replaced and will always pay their way because they will go to where the fares are and can be added or subtracted as needed. Suffice it so say that I believe even Anderson’s (now SkyWeb Express) track is as big as will work, and even this size is pushing it. I think a better way to design is to go from need to overall system viability, and then to the components, with track being the most challenging because of expected low ridership until a full network is established.

Dan said...

Raytheon does a lot of government work and transportation generally gets a lot of public (government) money and so, I think, they fundamentally misjudged the business model by comparing it to other spectacularly expensive public transportation solutions. Raytheon, let’s remember, was not making it to profit from collecting fares. As a system contractor, their main incentives were to make sure that it would be durable, robust, and yet simple and profitable to build. Their design reflected their priorities.

Both Raytheon and Vectus use a spine and rib design which is common in rollercoasters, and probably the simplest way to create banked turns. Raytheon’s enormous 36” pipe is perplexing, unless perhaps they secretly believed that the track would end up carrying more conventional, heavy vehicles. Or perhaps it was to be fitted with hatches so one could crawl through it for servicing or emergency evacuation. Even Vectus, (owned by a steel company) initially designed in super expensive in-the-track linear motors, before realizing that only subsidized systems could bear the cost. Now they let the customer choose the propulsion. In any case, both are as concerned with making profit by making track as whether PRT proves profitable to run.

As for I-beams vs round, I would point out the general truism that a beam that is taller is stiffer in spanning a long distance. A pipe has equal amounts of horizontal and vertical steel, and if that horizontal steel is repurposed into a vertical orientation, it can make a beam that is much taller. True, curved shapes in general have surprising stiffness that I do not have the engineering chops to quantify. I don’t think there is really an “expert” opinion on the matter, however, because of how each fails under load: Pipe deforms into a flattened oval and then folds, while I-beams twist before folding. I-beams are, therefore, generally braced from the sides to prevent this initial twisting. If the ribs of the pipe design can keep that pipe round, that stiffens the pipe greatly. These strategies make any apples to apples comparison impossible. And what criteria would be used? Weight of steel? Height/diameter?

Finally, I have to note that freeway overpasses are increasing using box beams with an isosceles trapezoid profile and internal bracing. This general V shape looks far stronger than either (per unit of steel) to me!