By Roy E. Smith • March, 2006
The introduction of new, low-floor vehicles into existing North American transit systems has provided an attractive means of meeting current operating requirements. By the same token, their introduction has created a number of vehicle/track compatibility issues—particularly when these modern cars are introduced into older systems. Paramount among these issues is that the Independently Rotating Wheels (IRWs) that are used on low-floor vehicles present a significantly higher derailment potential at lower L/V ratios than conventional, fixed-axle wheelsets.
Low-floor vehicles tend also to be multi-body cars, which are more sensitive to flaws in track and the vehicle’s leveling system. Even leveling systems themselves can contribute to the track-unworthiness of low-floor cars. All of these issues must be taken into account when introducing new, low-floor vehicles on older transit track, because unless the track was designed specifically for low-floor and multi-body cars, problems can arise.
North American transit track is, for the most part, built and maintained to safe standards. But the question is, safe for what kind of vehicle? What kind of vehicles did the engineers have in mind when they designed the track and the maintenance standards? For many years, this vehicle was the PCC car (named for the Electric Railway President’s Conference Committee), which was commonly used on street track in North America. PCC cars were rugged and reliable vehicles, capable of handling track irregularities with relative ease. On the other hand, the PCC car was operated at limited speeds (of approximately 30 mph), and the ride quality was not very good.
Despite improvements to vehicle design over the years, the modern, low-floor vehicles that have been introduced on old street track that was once PCC car territory have not been able to match the performance of the PCC car, in terms of speed and safety. The Massachusetts Bay Transportation Authority (MBTA), for instance, has had to restrict the speed of its new low-floor cars to 15 mph and 20 mph in many parts of its system where conventional articulated vehicles have historically operated at 40 mph and 50 mph. The reason for this disparity is poor compatibility. Something has to change—either the track or the cars—in order to bring the two into balance. The question is: Which to change? The answer is: Both.
Independently Rotating Wheels
One of the common characteristics of low-floor vehicles is their use of Independently Rotating Wheels (IRWs). Many of the challenges of operating these cars are also related to the use of IRWs. The use of IRWs has advantages. Most significant is that the floor structure of the vehicle can rest much lower on IRWs than on conventional solid-axle wheelsets. This makes it easier for cars to meet ADA (Americans with Disabilities Act) requirements.
But there are mechanical downsides to IRWs. Conventional wheelsets generate a self-steering effect because the taper of the profile allows for one wheel to run at a larger radius than the other. This self-centering effect brings the wheelset to an equilibrium on the track. IRWs do not produce this effect. Even if the wheels have tapered profiles, with no axle to connect them, there is no steering moment. Lacking this self-steering mechanism, IRWs tend to remain straight through a curve, resulting in heavy flange contact with the rail, as well as a higher L/V ratio, which produces a greater potential for wheel climb derailment. And while a conventional wheelset might move relative to the train or truck frame, the wheels remain parallel. With IRWs, it is possible for the wheels to move out of parallel alignment, which again raises the risk of derailment.
Poor suspension and improper leveling often contribute to the already inflated derailment potential for low-floor cars. Most vehicles use either a three-point or four-point leveling system. These systems typically consist of four air springs under a rigid car body—two at each end. The height of these springs is controlled by valves; the pressure must be continually adjusted in order to keep the car level. The valves detect where the carbody is, relative to the truck frame, and regulate the air spring pressure to keep the height constant as passengers enter and exit, and as the car passes through various track types and conditions.
In a three-point leveling system, two of the air springs at one end of the car are controlled independently; the other two are controlled by the same valve. In a four-point system, each valve is controlled independently. The major difference between the two is that in a three-point leveling system, the airbag pressures at one end will always exactly match. At the other end, the independent valves control the attitude of the car relative to the ground. If the car is sitting on unlevel track, such as a superelevated curve, the single-valve end keeps the height constant while the two independent valves adjust the attitude of the carbody. Airbag pressures will be different at that end—but only sufficient to overcome any unbalanced loads.
If a train with four-point leveling rests on unlevel track, the two valves at one end attempt to keep the carbody parallel to the truck frame at their end, while the valves at the other end attempt to do the same. In theory, this doesn’t present a problem on level track, but on unlevel track where, for example, one rail is tilted relative to the other, the dual valves at either end try to twist the carbody into position. Since the carbody is a rigid tube that will not twist, the valves go on pumping up diagonally opposite air springs without actually leveling the car. Once the car moves back onto level track, it takes a long time for the pressure to equalize. In practice, it almost never does. The resulting imbalance in weight distribution, which is frequently skewed to 70/30 from one side to the other can stress the vehicle’s frame. It can also overload the wheels on one side and unload those on the other.
Three-point leveling is a better choice—particularly for low-floor and multi-body cars, which generate enough stability concerns without having to take suspension unbalance issues into account.
The multi-body car is an important aspect of the modern transit line, and most low-floor vehicles are multi-body cars. The most common type of multi-body car is a two-carbody, three-truck, single articulated car, with the center truck under the articulation joint. Three- and four-carbody, double or triple articulated vehicles are available too, but their behavior is very similar to the single articulated car, as long as the intermediate trucks coincide with the articulation joints. The cars that behave differently are the double-articulated cars with one truck under the center carbody. In a configuration like this, the center truck typically has no swivel motion relative to the carbody; it’s just a pair of axles mounted to the carbody. As a result, controlling the pitch of the center carbody is difficult. There are two solutions to this problem: The center carbody can be fastened to the other two carbodies, or the center carbody can be secured to the axle in the direction of pitch so that the axles themselves prevent forward and backward tilt.
With a non-swiveling truck on the center carbody, there are operational difficulties to take into account. For example, should the lead carbody have a slightly stronger braking effort than the trailing carbody, the three-carbody system will buckle laterally. Without a swiveling truck this forces the center axles themselves to swivel, which forces the wheels hard against the rail. If the car in question is a low-floor vehicle with IRWs, there is a strong possibility it will climb the rail, since there is next to no self-steering force to correct the misalignment. The same buckling can occur if the rear body section is accelerating harder than the other two. This effect is illustrated in Figure 1.
This same sort of buckling can, and frequently does occur in the vertical direction. When a car is braking, the center of gravity is higher than the truck, so there is a moment that tends to push the back of the car up, and the front of the car down. The lead car body pulls up on the front of the center section and the trail one pushes down on the rear of it. If there is tilt control acting through the roof unit, there may not be a problem with vertical force. But if the tilt control acts at the point at which the carbody is fastened to the truck, the axle itself may be lifted, posing a derailment potential—particularly given that the IRWs on these cars are less forgiving of L/V ratios in general. The vertical buckling effect is illustrated in Figure 2.
A similar situation can occur in acceleration as well. The trailing axle of the center section will be lifted in that case. The two buckling effects do not occur in isolation, of course. The same action can cause both effects; they can, and do, occur simultaneously. The wheels in the center section, therefore, are lifted and forced sideways at the same time, even on tangent track. This, combined with any other design factors that might be generating high L/V ratios, can lead to a very high propensity for derailment—and one that is very difficult to predict analytically.
Derailments that occur due to these effects can be particularly difficult to identify after the fact and a false sense of security may well be generated if a minor track perturbation is identified at or near the scene and falsely identified as the cause. The problems described here can be exacerbated if a disabled vehicle is being pushed, especially through a curve.
Multi-body cars have been in service on various types systems for many years. The vehicle/track incompatibility issues associated with the introduction of single-articulated cars more than 30 years ago have been resolved. The introduction of multiple articulation cars with short center sections and non-swiveling trucks, and the introduction of low-floor vehicles with IRWs, in particular, have created a new set of vehicle/track interaction issues that require a new round of technical analysis to ensure compatibility and safe operation on existing North American track.
This article is based on “Considerations for Operation of Low-Floor Vehicles on North American Track Designs,” a presentation by Roy E. Smith, President of RESCO Engineering, at Interface Journal and Advanced Rail Management’s Rail Transit ’05 Wheel/Rail Interaction Seminar.
Roy Smith is President of RESCO Engineering, a railway dynamics consulting and design organization in Kingston, Ontario.