Squeezing more life from rails: TTCI, aided by AAR, studies ways to extend rail life and finds that North American railroads are doing well - Ttci R&D

Railway Track and Structures, April, 2003 by Kevin Sawley

Wear is the main cause of rail removal on freight railroads in North America (Figure 1). That is why engineers at Transportation Technology Center, Inc., are looking for ways to "squeeze more life from rails." Increasing the life of rails in track decreases railroad capital and maintenance costs.

Reductions in wear can be gained by the use of harder rail steels, by improved lubrication and by the adoption of trucks with improved steering performance. Most railroads are already pursuing one or more of these avenues. TTCI's project, funded by the Association of American Railroads, is focused on determining worldwide wear limit practices on freight railroads and how these practices influence live stresses in the rail, which will allow railroads to quickly and easily benchmark themselves against worldwide practices.

Railroads are highly-capital-intensive and one of the largest capital costs they bear is the cost of rail. As an example, North American freight railroads have more than 40 million tons of rail in-stalled in track and purchase upwards of 600,000 tons annually.

How much rail wear that is allowed can be highly railroad-specific, depending among other things on tie spacing, track condition, age, section and grade of rail, speeds and axle loads. Non-technical factors include the regulatory environment, railroad policy on rail defects and relative capital/labor costs. In track with significant numbers of joints, allowable rail wear can depend on joint-bar design and whether or not wheel flanges strike the bars.

Railroad survey

A six-page survey was sent to more than 80 railroads, known to carry only or mainly freight traffic, asking for information on rail-wear limits by track class, methods used to measure wear, rail sections and grades, tie types and spacings, rail defects and traffic information. Replies, most giving appreciable information, were received from 27 railroads, including North and South America, Western and Eastern Europe, Australia and Central Asia.

The survey provided useful general information, indicating that railroads around the world have common track problems and appear to be converging on common practices. For example:

* More than 85 percent of railroads use head-hardened rail as well as standard rail.

* 136-pound rail is the most common weight used, followed by 115-pound.

* Although one very-heavy-axle-load railroad uses head-hardened rail in all track, most railroads use the rail only in curves above a given limit. The most common limit is three-degree curves. As a rule of thumb, conventional three-piece trucks should start to contact the high-rail gauge face when the curve reaches approximately three degrees.

* 86 percent use vertical head loss or a combination of head and side loss to quantify wear, whereas 14 percent use loss of head cross-section.

* 78 percent of railroads use automatic rail-wear measurements.

* The criterion for when to re-rail because of rail defects varied only between two and four defects per mile per year.

In the survey, railroads were asked to rank their main rail-defect types. Figure 2 summarizes the replies. After transverse defects, the most common defects are vertical split heads, field welds (alumino-thermic), plant welds (flash butt), horizontal split heads and bolt-hole defects. (Figures 3a & b show some defects found in rail.)

Stresses allowed in rails

There are many types of stress in rails, including bending, contact, residual and thermal stresses. Of these, bending stresses are probably the most significant in terms of defective and broken rails. The other types of stress change very little with wear and are also difficult to determine by the survey. Contact stresses depend on wheel and rail profile shapes. Residual stresses depend mainly on rail manufacturing technique and thermal stress varies with ambient temperature through the year. Hence, survey results were analyzed to determine the level of bending stress allowed in the rails by the different wear limits.

Nominal rail-bending stresses, as used in this study, can be calculated using a simple model, called the "Winkler" model, which assumes the rail is supported on a continuous elastic foundation. The modulus of foundation (load per unit length of rail needed to produce a unit foundation deflection, k) depends, among other things, on ballast condition and tie type and spacing. Smaller tie spacings increase k, while larger ones decrease it. Using this model, the maximum bending moment ([M.sub.max]) and the maximum nominal bending stress ([[sigma].sub.B]) are related to the wheel load (P) by:

[[sigma].sub.B] = [M.sub.max]y/I = Py/4I 4[square root of (4EI/k)]

E is rail elastic modulus, y is distance from the neutral axis, and I is moment of inertia. (A single wheel load gives the highest rail bending stress. Because of the rail uplift effect, multiple wheels, for example from adjacent trucks, lead to a slight decrease from this maximum.) Using results from the surveys, the maximum bending stresses allowed in the rail head and base, for new rails and rails worn to their maximum limits, were found as follows: