Why reinvent the wheel?

On my blog post last week about the Wheel-Rail Interaction Conference, a commenter asked “Why reinvent the wheel?” and noted that the same basic wheel shape has been used since at least the 1870s. 

I don’t mean to single out this commenter. It’s a good question. 

My answer, and I welcome others in the comments, is in two parts:

1. Because everything above and below the wheel has also been reinvented

2. Because the process isn’t intended to change the function of the wheel, but to optimize its performance based on an ability to measure data which wasn’t possible before now

First, let’s put some numbers on how much things have changed. In 1870 a typical freight car was around 30 feet long and weighed perhaps 50,000 pounds distributed across eight wheels. Freight trains were typically 20-40 cars long, depending on the grades and the locomotives in use. 

Today a freight car is twice the length and weighs up to 286,000 pounds, yet still rides on the same number of wheels. The pressure experienced by the contact patch of each wheel on the rail can reach 70,000 psi, up from perhaps 13,000 in 1870.

(For those doing the math at home, my back-of-a-napkin figuring assumes a contact patch of 1/2 square inch today and 1/4 square inch in 1870. Mathematically, the intersection between a circle and a line is a single point, but wheels and rails distort to create an oval-shaped contact patch. The heavier the weight, the more distortion and the larger the contact patch. This is called Hertzian contact.)

This force is first transmitted into the rails, which today are often 136 pounds per yard, up from 50 pounds per yard in the 1870s. Rails and the rest of the track structure — tie plates, ties, ballast, and subgrade — distribute the load in order to reduce it to the 20-40 psi typically tolerated by natural ground. 

Vertical weight is the easiest measure of how much the wheel-rail interface has changed since the 1870s, and wheels and rails deforming does cause much of the wear, damage, and risk in the wheel rail interface, but it also comes from friction and forces in other directions besides down. These forces scale up with the weight of the train.

Solving problems created by rail and wheel wear isn’t just a modern quest. In the early 1900s, train wrecks caused by broken rails with undetectable internal cracks were steadily on the rise, with railroads powerless to stop them. The Interstate Commerce Commission was the federal agency which investigated wrecks, and its Chief Inspector of Safety Appliances wrote in a report about a 1911 Lehigh Valley wreck which killed 29 in Manchester, N.Y.:

“... it would thus appear that the danger zone in the use of steel rails as at present manufactured has been reached, and, since it is supposed that transverse fissures are the direct result of high wheel pressures acting on hard steel, a complete investigation should be made for the purpose of scientifically determining the matter and ascertaining a remedy. Until such an investigation has been made, the danger of similar accidents will exist.”

Today we know that transverse fissures are caused by rolling contact fatigue, namely the repetitive deforming of the rail by passing trains which drives the creation of tiny cracks which, if not ground off, can dive deep into the rail and cause a rail break. Railroads gained the ability to detect internal cracks a few decades after that report with the development of ultrasonic testing by Dr. Elmer A Sperry, and axle loads continued to increase. 

Now to my second point: new wheel profiles are designed to increase the life and performance of wheels.

A relatively new innovation in the wheel-rail world is the ability to easily measure the shape of things. It’s easy to measure distances or whether a shape matches a profile. Tape measures and templates have long been used to tell if wheels meet legal standards: is the flange wide enough? Is the tread too hollow? Measuring the exact shape of a worn wheel was too labor-intensive for routine maintenance, and not worth it. But now you can roll a sensor wheel across the wheel or even move a handheld laser in an arc across the wheel tread to measure its exact shape. It takes ten seconds to measure, then that data can be uploaded to the cloud where you can track trends in how a wheel wears over time, how it compares to other vehicles in your fleet, and what it means for any wheel profile experiments you’re running.

One of the most widespread improvements in wheel profiles since 1870 is the advent of worn wheel profiles. Think of it like breaking in a new pair of shoes. When you get them from the factory, they have some tight spots, maybe some hard edges, and the shape isn’t quite right. After a few months, they fit your feet comfortably and still work just fine. When you take care of the leather, do you try to reshape it to factory condition or maintain them in the comfortably broken-in state? 

Before 1990, a common wheel profile on North American freight railroads was the AAR 1:20. In the 1980s, the Transportation Technology Center (TTCI) in Pueblo, Colo., tested four new wheel profiles which were based on a worn AAR 1:20 wheel. The best of the four, AAR-1, rolled and curved better than the AAR 1:20 but couldn’t go as fast without hunting: 49 mph down from 70 mph. The revised AAR-1B could match the speed of the old wheel and became the standard wheel around 1990. In 2018, the industry began transitioning again to the new AAR-2A, designed in the 2000s and again based on worn wheels: 122 wheelsets plus 210 pairs of rail profiles. The new profile underwent three major tests: on a five-unit double-stack car passing a wayside hunting detector 109 times, ten coal cars rolling 25,000 miles, and 200 grain hoppers in revenue service on two railroads. After rolling 115,000 miles in the last test, the new profile had 40% less tread wear than the AAR-1B, plus a host of other improvements.

If you haven’t read my previous blog post, it includes another wheel profile success story on the Los Angeles Metro: You’ve got to spend money to save money.

In the example from my other post, the LA Metro originally used the AAR 1:20 profile and got a disappointing 19,000 miles out of each wheelset. Switching to the AAR-1B increased that to 25,000 miles, but it was only after switching to the custom RESCO profile that they achieved their current 600,000 miles per wheel.

The simple reason is that transit vehicles have to negotiate sharper curves than freight vehicles, and freight wheels usually can’t steer well enough. The angle of a wheel’s taper lands somewhere on a spectrum where one end is a highly tapered wheel with exceptional steering but a lower top speed; at the other a fully-cylindrical wheel with great high-speed stability on straight track but which flanges around even moderate curves. Most transit wheels use a wheel with more taper than a freight wheel because they need the curving ability and don’t run as fast as mainline trains. San Francisco’s BART subway was one of the few which went all the way to cylindrical wheels to help them run their trains at 80 mph, but they recently finished reprofiling their wheels to a more traditional tapered wheel to reduce wear. 

In summary:

• Wheel profiles need to be the right tool for the right job

• Modern measuring equipment has allowed railroads and research organizations to improve wheel profiles since the 1980s

This post, like my others in the past couple weeks, originated at the Wheel-Rail Interaction Conference, of which Trains is the presenting sponsor. Look for more stories from the conference in future issues of Trains Magazine.