American Steam Locomotives

I read American Steam Locomotives, Design and Development, 1880-1960, William L. Withuhn, 2019, Indiana University Press, ISBN 978-0-253-03933-0, 451 pages.

Withun wrote:

To me, that is the first responsibility of any historian on any subject: to reveal the subtlety and complexity of issues that confront the human beings who preceded us.

I believe he carried out that responsibility with panache.

This is the book I’ve been looking for off-and-on for several years. It explains how steam locomotives were designed, among many other things.

The Acknowledgments page, written by his wife, Gail Withun, says that he wrote chapters over many years. Withun’s health failed, and he discovered he had waited too long to put together this book. The book was published posthumously, Withuhn died in 2017. “Friends and admirers” finished the book.

Withun writes this in the first chapter:

As others have said, engineers are human, and the make decisions in a conflicting context of management objectives (designing things for a presumed marked, with management objectives well or badly defined), economics (ratios of effectiveness to cost, with imperfect knowledge of either effectiveness as a changing market may define it, or costs), professional goals (how one’s work may gain rewards from employers and standing with peers, in the context of engineering practices of the day), and numerous other direct and indirect pressures.

This is a book about engineering, but it doesn’t contain lot of equations, and not one single indicator diagram. The engineering is more about how human taste and lack of complete knowledge affect design, and even large scale systems like railroads.

Chapter 4 is called Big Wheels Turning: A History of Counterbalancing, or Science and Snakeoil. A steam locomotive’s “drive train” has a lot going on, dynamically. There’s a large number of moving masses, some reciprocating, some revolving. Pistons go back and forth, pushing on one end of the drive rod, which reciprocates with the pistons. The other end of the drive rod connects to a crank pin on which sticks out to the side of one wheel. That wheel revolves, so that end of the drive rod goes around in a circle. A side rod connects all the drive wheels. It goes around with the crank pins on the drive wheels. The mass of the crank pins was enough to cause some lumps when the wheels rotated. The crank pins on the main driving wheels are 90° apart to prevent the drive train from getting stuck at top or bottom dead center when the train comes to a stop. The lumps they caused during rotation were therefore asymmetrical.

Locomotive designers didn’t have the theory to work this all out, and they didn’t have the computational horsepower to deal with out-of-balance rotation and reciprocation numerically. They used educated guesses, and trial and error.

I was surprised to learn that a lot of locomotive design was trial and error, or using heuristics. Any modern style numerical engineering was very simple. Required factors of safety were very large. For example, boilers were said to have a (legally required!) factor of safety of 4. In my aerospace engineering days, we used factors of safety of 1 for limit load (based on yield strength), and 1.50 for ultimate loads.

Just to see this play out, the Union Pacific Big Boy locomotive had a cylindrical boiler of 107 inches outside diameter, a thickness of 1.375 inches, and a 300 psi operating pressure according to the Big Boy Wikipedia entry.

Famously, the hoop stress in a thin walled cylinder is σ = Pr/t

The hoop stress in a Big Boy boiler at 300 psi is:

σ = (300)(52.8)/1.375 = 11,500 psi

A safety factor of 4 would mean a design stress of 46,000 psi, somewhat higher than yield stress, but less than the 58,000 psi ultimate tensile stress given for mild steel today. I can see mild steel’s ultimate strength in 1941, when Big Boys were designed, being a lot less than it is today. Or maybe they used steel that wouldn’t qualify as mild steel today.

One big part of the empirical engineering was Pennsylvania Railroad’s Altoona test plant. They actually set up big rollers and ran steam locomotives inside a building. The locomotives were instrumented to actually measure performance. PRR actually published reports with data on their testing instead of keeping it proprietary and secret. That seems like something that simply wouldn’t happen today, in an age of “intellectual property” hoarding.

The number of wheels on a locomotive varied enough that there’s Whyte notation to describe wheel configurations. This always confused me. Why have so many arrangements? This book makes clear the engineering tradeoffs behind the variation, balancing the “tractive effort”, which derives from friction between wheels and rail, and the load bearing capability of track. Locomotive stability at speed, having to fit inside tunnels or under bridges, stay on track while negotiating curves, and the need to support the firebox end of the locomotive only complicate things. There isn’t one solution to these constraints, so people tried a lot of different wheel arrangements to see what worked best.

The very human side of locomotive engineering is embodied in a chapter on safety. Steam locomotives blew up every once in a while, but the introduction of mechanical stoking seems to have changed the jobs of fireman and engineman completely, in terms of safety.

The chapter on railroad employment also showed the human impact of converting to diesel-electric locomotives. Railroad employment goes way down with diesels. There’s just not the same need for all the support workers, the mechanics with huge wrenches, or the foundry workers that cast steel locomotive frames. The conversion to diesel had wide reaching economic effects all over the USA. Diesel locomotives didn’t have to refuel as often, or take on water. Coaling and water stations in small towns all over the US disappeared, and took the workers with them.

All told, a great book, a companion to How Steam Locomotives Really Work, and a worthwhile read.