Sunday, December 14, 2008

10> A Couple of Issues

I need to address a couple of issues. First the matter of getting exposure. As I Google “PRT” I find that this acronym stands for more than Personal Rapid Transit, or Personal Rail Transit. I find this blog is nowhere to be found. But then again, did I even say “Personal Rapid Transit” in previous musings? If I say “Personal Rapid Transit” again and again, will it come up in a search? PERSONAL RAPID TRANSIT! PERSONAL RAPID TRANSIT!

Now on to other business. So far I have rejected the notion that “pods” sit on top of a track… Well my opinions go way further than that. So here’s a little something to chew on..

  1. Cars shouldn’t be just one size. They should be sized and in quantities based on need.
  2. Cars shouldn’t just carry passengers. They should carry parcels as well.
  3. Cars shouldn’t just go one way…(on each track) Why not go both ways? It would be handy when backing out of private property, good for third lane options where the third lane flows tward the city in the morning, and away from the city at quitting time.
  4. I believe that vehicles can go both ways on larger streets. (The “one way” assumption seems to be based on problems making tight turns, spacing of cars, making elevated stations, and other issues connected with cars that ride above the track. I believe that PRT could share stops with buses, on both sides of the street.
  5. Unlike previous models, I believe cars should be spaced dynamically, based on weight and speed. I do not think this is too complicated for present technology.
I think cars should be able be able to ascend and descend very steeply, even if the means is external. (roller coasters are easily and reliably grappled and lifted, so why not PRT vehicles?) They should also be able to turn (right, left, up, down) very tightly. This is just common sense, because the system should be as versatile as possible, and that should include track that can be constrained to placement over existing easements.

4 comments:

DSmith said...

There are a lot of debates over these issues. My take on them.

1. I think 4 is a good size. You don't want to deal with trying the get the right sized car to a station for groups a particular size. Larger groups can split up.

2. Agree completely.

3. You have an interesting idea with changing track direction for rush hour. A track on a west-bound street could be switched to east-bound if loads become too high in other west-bound streets.

4. I can't picture narrow city streets with more than one track. How would you do intersections. Would you have split grade with a full 8 interchanges, or a single grade traffic-circle approach?

5. Agreed.

timote said...

1. I think there is benefit to standardization here, you just hit your keycard and get a car. Now, if you want to offer a larger size for big groups (subject to additional wait time at the time of calling for the "car") or a freight-specific unit (additional space for freight say), that's fine. But let's not go down the personalized car route.

2. Agree.

3. Novel idea, I think it would complicate the design, perhaps too much (over-engineering possibly), but I'm open to the concept.

4. As per previous commenter, crossings are a big problem. Much easier with the "loop" concept usually associated with PRT. So long as you don't have to stop for red lights or stop signs, going around an extra block isn't a big deal.

5. Agree, I'd go further and say that speed is a very interesting point to me. Most PRT seems to be based on low speed (25-35 mph), which is fine and dandy through neighborhoods (I wish people would only drive that down my street at 2 am!). This would allow for a much more livable environment if the major transportation mode is quiet and doesn't shine headlights everywhere. However, there will need to be through corridors - like highways currently, where the speeds pick up to something like 50 mph or so... How does this concept intersect with your thoughts?

Anonymous said...

I really like your thinking on this. I have evolved to many of the same preferences: suspended operation, tight turning radius, steep climbing, bidirectional capability, etc. - these all increase the flexibility of application. Mister comes closest in the current set of designs. But I still like the more traditional designs like ULTra and Vectus, simply because they are close to cracking the market (the pragmatist in me ;-))

With respect to point 5: variable spacing and speed, I have put a lot of thought into this. I believe current designs have a top speed that is based on safety considerations in the most dense configuration, i.e. the entire guideway filled with vehicles.

The issue is this: maximum safe guideway capacity for a given system is a function of speed, and if you graph safe capacity vs speed, you will get a curve with a maxima around 15-30 mph. At zero speed, the capacity graph is obviously zero, and capacity also approaches zero as speed approaches infinity (due to safe stopping distance being proportional to speed squared). So there is a bump in the middle surrounded by zero at both extremes.

That bump seems to be the speed at which most PRT systems are designed. But my feeling is that there are times when you don't need raw maximum guideway capacity, and therefore can increase speeds without sacrificing anything but unused capacity.

Example: if you can run at 1 second headway at 25mph, but your headway goes up to 2 seconds at 40mph, obviously you have double the raw capacity at 25mph (3600 vph) than at 40mph (1800). However, if the traffic density is below 1800, you might as well travel at the higher speed and headway, which will clear the guideways faster than operating at the slower speed. Now, if you are standing at one point in the guideway and watching vehicles go by once every 2 seconds, this may seem to be a net loss because at 25mph they pass every 1 second. But from the system perspective, the vehicles have reached their destinations earlier and thus free up the system for other traffic. And you have lost nothing because the maximum capacity would not be reached anyways.

For illustrative purposes, consider the extreme case: infinite speed. Imagine the vehicles travelling as fast as pulses of light. Now, the theoretical guideway capacity in this case will be abysmal, because the speed of light squared would limit headways - probably days or years headway and a theoretical capacity of zero. However, assuming the guideway is a few miles long, the guideway will only have a vehicle on it for a few microseconds at a time, so there is no problem with safe following distance, so you could send out "vehicles" literally hundreds of times a second and *system* capacity is enormous.

This illustrates where point capacity is not the whole story. During busy times, 25mph may be optimal because you need that 3600 vph capacity, but at other times, it's better from a system perspective to increase speed at the expense of theoretical maximum capacity, thereby clearing the guideways faster for later traffic.

One other point about variable speeds/headways: at the user interface, there could be an option for a "smoother ride" for those with physical problems. This would decrease acceleration/jerk and increase headways (both ahead and behind) to ensure a smooth ride for those who need it, while providing maximum system benefit for those who can deal with a little more jerkiness (still comfortable for a healthy individual). Furthermore, the opposite can be done for freight traffic: a jerkier ride with shorter headways and faster accel/jerk, activated by a special command available only to freight-loading stations. This idea could also be applied to empty vehicles, as long as a foolproof way of detecting "emptyness" could be devised.

There are a few caveats to all this: guideways would have to be designed for higher speeds, e.g. larger turning radii and higher stress requirements. But variable speeds also means you can have guideway "speed limits" based on the guideway physical limits.

Of course, this all complicates the control system, but that's just a matter of getting the software right. Software is my thing, so my goal is to produce a control system that implements this kind of variable speed, variable headway, customizable vehicle movement, while preserving all safety constraints.

Dan said...

Dan the Blogger responds.

Thanks to all for your input. Dave, reverse is also useful in high density stations as shown in the MISTER simulations. Point 4, I really don't know, thanks for posing the question, it will have to be dealt with sooner or later.

Timote, I guess around the block is OK in very narrow streets. That's mostly an "Old" city problem. (Boston comes to mind) P.S. thanks for your thoughtful and intellegent input on so many posts. It has really helped get things going. If I don't respond, it's probobly because I agree completely.

A Transportation Enthusiast.. (Man what a long name.. how about I call you "A" for short)? That is some good thinking there. The problem is that I am going to address software and control in an up-coming post, so I don't want to deal with what you said just yet. Obviously it will be a long and involved discussion. I will say, however, I agree with basic premises, and will probably refer to it in a broad discussion about the structure of the information flow.