Thanks for the information on Cybertran. I visited the inventor, John Dearien, and a couple of his colleagues in Idaho Falls several years ago and spent a whole afternoon with them, and indeed have a copy of a paper on the system on my desk right now. The first thing I asked when we were looking at the vehicle was: How does it switch? John mumbled something and tried to change the subject. Strange since he had been working on the design since about 1981.
Switching is fundamental when one is talking about off-line stations. His vehicle used steel wheels running on steel rails very similar to conventional railroads. The only way you switch such systems is to move the whole track. This is not a good idea for PRT or systems like it for reasons of safety and reliability at close headways. If an in-track switch hangs up in the middle you would experience a derailment.To minimize the chance of such a catastrophic failure, you have to run vehicles far enough apart so that after one vehicle passes the switch there is time to throw the switch and to stop at a reasonable braking rate if you cannot sense that the switch is locked in place. This has a marked effect on headway.
Consider steel wheels on steel rails in an elevated configuration. H-Bahn, a German GRT system, used steel wheels on steel rails. Their literature stated that the system was very quiet, yet upon watching a vehicle move along their elevated guideway it emitted an entirely unacceptable noise, which was due to small imperfections that made the guideway sing as the vehicle moved even at low speed. When steel rails are on the ground, such noise is damped, but in a minimum-weight elevated guideway, the noise is not damped. At the high speeds Cybertran plans to run, it is likely that the noise will be unacceptable in built areas or anything close to it.
Without saying so, Cybertran attempts to overcome the requirement for long headway by using larger vehicles than optimum for PRT. Their vehicles hold from six to 20 passengers and they claim that the vehicle size that minimizes system cost is about 12 to 15 seats. Our studies, which have been very comprehensive, have shown us that the smallest vehicle minimizes the system cost mainly because the guideway weight, and hence cost, is directly proportional to vehicle weight. Now, at higher speeds where air drag becomes more important, one can think that cost would be reduced by getting more people behind a given frontal area, but you have to think about how you get higher average load factors. I have looked at Cybertran data to try to understand how they came to a different conclusion than we did. It seems to be due to the design assumptions about various vehicles as the size decreases. In one of their illustrations, the smaller vehicle was just a larger one shortened without changing the crossection. I saw no attempt to optimize the design of a minimum-sized vehicle.
They talk about nonstop, origin-destination travel, bypassing intermediate stations, and on-demand service. Suppose they have just a line-haul system with no branching into a network. Then, if the travel is to be nonstop, at each station they need to wait to fill a vehicle with people all headed to the same destination. How do you do that? People, given their own choice, travel when they wish. To assume that you are going to get a random group of people going to the same destination takes time. If, for example you have ten stations of equal probability as a destination only every tenth person will be going to a specific station. When, in general, you calculate the mean time for a second person to arrive who wants to go to the same station you do, you find that that time grows as square of the number of stations, and quickly becomes unacceptable as the system grows. And to get six people at random going to the same station the wait time is proportionately longer. So, the only practical way to run Cybertran as a line-haul system is on schedule. Otherwise the average loads per vehicle will be in the range easily carried by the smallest vehicle. If you decide to dispatch vehicles on schedule, what should be the schedule frequency? Obviously it depends on demand. But now, not the total demand, but the total demand in people per hour at each station divided by the number of possible destinations. So you may have to wait quite a while to get six people going to the same destination, as a consequence of which you compromize and let each vehicle go with fewer people and then as the off-peak periods come you realize that to get the same load factor you have to decrease the schedule frequency quite a bit, but if the wait is too long too few people ride so you accept an even smaller average load factor and soon will find that you get only one group travelling together by choice. This is the inherent problem of scheduled service--the daily average load factor is very disapointing.
Now, suppose, as implied by Cybertran literature, that the system has line-to-line transfers, i.e., it can be built as a network. Then, when a vehicle enters a station with people in it, you can't board that vehicle unless it is destined on the same route that your destination station is on, and the number of possible routes increases very rapidly as the network grows. Moreover, extensive simulations performed in a Colorado RTD study of transit alternatives in 1974-5 of the operation of GRT stations showed that you can't tell which berth a vehicle will be stopped at until the vehicle is very close, which means that you have to stand somewhere near the middle of the station and be ready to dash to a specific loading berth in the last few seconds. All this is described in a paper by Johnson, Walter, and Wilde in PRT III, the proceedings of the third international PRT conference, held in Denver in September 1975. The bottom line was that only true single-party PRT or trains stopping at all stations are viable. The intermediate GRT systems work only in the simplest situations with very few stations.
Your correspondent says that the line-haul Cybertran makes more sense than a network PRT system. Ask him how people get from the few stations of such a system to their final destinations. You find that you are not really better off than in a conventional rail system. The advantage of Cybertran is that with smaller vehicles than conventional rail, the guideway is cheaper, but the service differs very little. If you drop people off at a few stops in an area like Silicon Valley, they generally have a long way to go to their destinations. A PRT network could provide that service. As our PRT matures the speed capability will increase. I estimate that with the same vehicles as used in an urban network, we will be able to go up to at least 80 mph, which will satisfy many of the line-haul needs of commuters. So, I see the market for Cybertran to move up to much higher speeds, they mention 135 mph.
|
