By Bob Tuzik • July, 2007
You’ve heard it before, you’ll hear it again. Delegates to the 13th Annual Wheel/Rail Interaction Seminar presented by Interface Journaland Advanced Rail Management heard it time and again over the course of the seminar: The wheel/rail interface is a system. And when there are problems, as there often are, it takes a systems approach to solve them.
Sometimes the problems are related to performance and wear, and the costs associated with them. Sometimes they’re related to component failures that sometimes lead to derailments.
The Hatfield derailment in late 2000 caused a severe disruption and loss of passenger confidence in the UK. With the bad, however, came some good: “The derailment stimulated investigation into and understanding of some of the root causes of poor rail/track conditions, such as cracks in the rail and rolling contact fatigue, in particular,” said Kevin J Sawley, Principal Consultant, Interfleet Technology Ltd.
Field investigation and analysis indicated that there were three primary modes of crack initiation:
— Steady state curving, which is caused by high steering forces and occurs in curves of less than 2 degrees.
— Bi-stable contact, which occurs when small wheelset lateral shifts cause a sudden large change in rolling radius and, thereby, generate a longitudinal force. Bi-stable contact is symptomatic of conformal wheel/rail profiles, and occurs in curves of 1 – 2 degrees.
— Convergent motion, which is caused by short-wavelength lateral rail misalignment that the wheelset cannot follow. Convergent motion causes a sudden large change in rolling radius, along with the associated longitudinal force, and occurs in very shallow curves and nominally tangent track.
Remediation methods for the crack-initiation modes include rail grinding (primary) and lubrication (secondary) for steady state curving; grinding (primary) and track alignment (secondary) for bi-stable contact; and track alignment (primary) and rail grinding (secondary) for convergent motion.
“Grinding and friction modification are essential, but they are band aids,” Sawley said. “We need to find ways to reduce the amount of energy spent and wear.”
Through extensive research into wheel/rail interaction, methods have been found to predict wear and RCF. Software programs such as NUCARS® and VAMPIRE® can now apply “what if” queries that can be used to guide vehicle and track design and maintenance, and the business decisions associated with them.
Norman E. Hooper, Project Engineer, Hatch Mott MacDonald reported on measures that can be taken to reduce broken rail derailments in dark territory. During a three-year period ending in September 2006, broken rails accounted for 40% of all FRA-reported, track-related derailments on main track. Temperature differentials, residual stress, axle loading, accumulated tonnage, wear, track modulus, wheel impacts and other factors can contribute to the frequency of broken rails. While these factors affect all railways, they are particularly problematic in “dark territory” where the absence of a signal system that can alert a railway to a broken rail can go undetected until “found” by a train — a circumstance that often ends in derailment, Hooper said.
Hooper shared some of the risk-management strategies that helped reduce service failures and broken rail derailments by 60% on a Class II railway. Frequent ultrasonic rail testing is a must. But, depending on the size of the defect and the surface condition of the rail, testing equipment doesn’t always find a defect in the rail. Industry guidelines indicate that testing equipment should be 65% to 95% accurate, depending on the size of the defect. Accuracy of 65% leaves at a lot to be desired, Hooper said. “In cold weather, you don’t need a very big defect.”
If anything, the risk of a broken rail derailment is greater in dark territory, he said. “And if you are applying the methods you’re using in signaled territory directly onto dark territory without some consideration, you’re missing some opportunity.”
The chance of finding some of the smaller defects can be improved by cleaning up the rail through grinding or by turning down the lubricators prior to testing. Track inspection is also important. “It is much more important to review the quality of the track inspection than the pieces of paper (it’s reported on),” he said. “What is missed or omitted from those inspections is as important as what’s entered.”
Good data allows railway managers to identify trends, such as defect growth compared to tonnage, in order to determine the appropriate testing frequencies. “Test the hot spots more frequently,” Hooper said. A branch line with 85-pound rail may require more frequent testing than a main line with 136-pound rail. Late fall or winter testing is also beneficial. While no one wants to change out rail in December, that’s when defect-related problems are likely to appear.
Freight Car Performance
The increasing load and stress environment of primary railcar structures, as evidenced by top-chord and axle failures during extended 286,000-pound gross rail load (GRL) service, has become a growing concern on North American railroads. Design, inspection and maintenance practices may have to be revised to ensure that the vehicles can withstand 286,000-pound GRL operating conditions on a sustained basis.
The Transportation Technology Center, Inc., (TTCI) conducted a series of four over-the-road tests to measure the wheel forces, axle strain/stress, bolster loads, and top chord strain/stress under typical revenue service conditions. Instrumented wheelsets were used to measure vertical and lateral forces and torque. Dynamic wheel forces and axle strain were recorded during the first test. Transducers were added to measure top chord strain during the second test, which was conducted at higher average speeds than the first. Additional transducers were added to measure the vertical bolster loads during the third test and fourth tests, which were conducted over different routes emanating from the Powder River Basin in Wyoming.
Results from the four tests indicate that the primary variables affecting dynamic vertical wheel forces, bolster center bowl loads, and top chord compressive strains in hopper cars appear to be train speed and track surface deviation, said Kevin Koch, the TTCI’s Manager of Track Inspection and Development.
The most significant car response occurs with the following combination of parameters:
—Train speed in the range of 48 to 51 mph.
—Track surface vertical deviation greater than about 1.25 inches peak-to-peak.
—The period or distance of the deviation at about 40 to 60 feet.
The most obvious track geometry features associated with these events were vertical track perturbations, he said. High load events were correlated against track geometry data collected by the railways.
While ambient temperatures, which ranged from 15 degrees F to 80 degrees F during the tests, did not appear to affect car response, the use of a premium (or M976-compliant) truck did. The use of premium trucks provided a general reduction of 15% in top cord stresses and as much as an 18% to 20% reduction in vertical center bowl stress, Koch said.
Out of Round Wheels
Out-of-round railway wheels represent another source of problems in railway operations. Paul Mittermayr, Managing Director of Mittermayr Scientific Consulting GmbH, pointed out that out-of-round wheels increase wear, cause damage to the track structure, compromise comfort on passenger lines, generate noise and increase costs. As a result, there is a need to be able to reliably detect cars that do not meet the operational specifications. “With the liberalization of railway traffic in Europe, this will be important for the calculation of usage fees in the near future,” Mittermayr said.
Identifying out-of-round wheels is one thing, understanding the causes and their impact on wheel/rail interaction is another, though essential aspect of identifying and controlling maintenance and life-cycle costs. New simulation tools such as simOOR (Simulation of Out-of-Round Railway Wheels) model rolling motion under realistic operating conditions, using representative out-of-round wheels, rather than new wheel conditions, as is done on some existing models, to simulate vehicle/track interaction, Mittermayr said. The actual wheel/rail contact is characterized as non-elliptical contact, which provides deeper insight into the causes of out-of-roundness and the mechanisms associated with it.
Whenever railroaders talk about wheel/rail interaction, friction management is part of the discussion. Over the past several years, interest has grown in the use of top-of-rail friction modifier technology as a means to control friction at the wheel tread / railhead interface. Benefits associated with controlling friction at this interface include reductions in lateral curving forces, gauge widening, fastener fatigue, rail wear and rail rollover potential.
In tests on a full scale rail/wheel test rig, Kelsan Technologies and Voestalpine Schienen GmbH examined the relationships between rail strength, contact geometry and friction conditions to determine wear rates and the initiation of rolling contact fatigue (RCF). They found that in addition to reducing lateral loads in curves, the application of a TOR friction modifier every 50, 250 or 500 cycles per 100,000 cycles simultaneously reduced rail wear, surface and sub-surface plastic flow, and RCF crack generation, said Richard Stock, Voestalpine Schienen’s Manager of Research & Development. While cracks and some wear developed on “dry” rail (friction modifier was applied every 500 cycles), no cracks developed after 100,000 cycles with friction modifier applied every 50 or 250 cycles. And since the friction modifier is a dry film, though applied in a water-based solution, the risk of hydro-pressurization of cracks is eliminated.
Still, applying TOR friction modifier technology to heavy axle load traffic in heavy-grade, high-curvature areas presents challenges. NS installed 24 Portec wayside systems to apply Kelsan Technologies’ Keltrak®, TOR friction modifier at 14 locations on its Virginia Division Bluefield-Narrows coal route, on which trains apply air brakes through the application zone, much of which has sustained, greater than 1% descending grades. Curves range from 3 to 12 degrees, with very short spirals and very little tangent between the curves. NS makes extensive use of dynamic and air brakes and sanding to hold the speed on the descents on this route.
It’s a high stress environment with plenty of evidence of high lateral loadings, including wide gauge, negative cant, gauge corner cutting and low rail wear in curves, said NS Research Engineer Kevin Conn. “The lateral force not only transmits itself into the tie plate, but into the fasteners, resulting in a large number of broken spikes and broken screw spikes.”
NS wanted to determine whether TOR material would withstand the high heat generated by nearly continuous braking on descending grades, and under heavy sanding on ascending grades. NS also looked at the effect of superelevation on train handling.
Data from the NS track geometry car showed a reduction in loaded gauge in the test area on descending grades. Additional measurements conducted on back-to-back 6.3- and 6.8-degree reverse curves (in conjunction with the TTCI’s Eastern Heavy Axle Load “Megasite” monitoring program) initially indicated that reductions in lateral loads were not in the anticipated 30% range, however. After experimenting with the application rates, NS, along with Kelsan and Portec, found that an application rate of 140% of the normal rate enabled NS to achieve the desired 30% reduction in lateral loads.
NS also looked at the effects of multiple wayside TOR units on wheel conditioning. NS achieved a 15% reduction in lateral loads with the installation of the first TOR unit. With the activation of another TOR system in front of it lateral loads were reduced by another 12%, or a total of 27%. “We concluded that you need two top-of-rail systems to fully condition the wheels before they are fully saturated,” Conn said.
On an ascending grade, with five TOR sites established at distances up to 2.5 miles from the L/V measurement site in a 4.5-degree curve with heavy locomotive sanding, NS found that it was able to achieve similar reductions in lateral forces. It learned as a byproduct of the study, however, that 2 – 3 inches of excess superelevation can increase lateral forces by as much as 66% in a 4.5-degree curve.
“We found that while top-of-rail application is effective in reducing lateral forces under a range of conditions, issues such as speed, superelevation and many others must be considered simultaneously,” said co-presenter Kevin Oldknow, Kelsan Technologies’ Manager of Friction Control Technology.
“Trains are often configured and powered without consideration of the alignment of the track that they will run over,” Jude Igwemezie, President of Applied Rail Research Technologies, Inc., pointed out in a presentation on Controlling Energy Spent through Wheel/Rail Interaction. “Operating Departments’ provision of minimum power just to get over the ruling grade in the hope that the trains make it exacerbates this problem. As a result, different trains that ply the same route require superelevations that are different from what can be provided by the track. “And when something goes wrong,” Igwemezie said, “the blame game begins.”
In most cases, the governing standards of the Engineering Departments cannot meet the requirements of the trains. “The power distribution throughout the train can have a direct impact on track forces and the energy directed toward destruction of track assets,” he said. Distributed Power is one way to improve overall performance. It can make long trains behave like short trains, reduce fuel consumption (wasted energy), reduce braking distances and reduce track asset damage.
Louis T Klauder, Jr., of Track Shape & Use, LLC, outlined a form of improved spiral geometry that overcomes an inherent deficiency. Whereas the traditional geometry is conceived as a shape for the track, the definition of the improved geometry begins with a decision about how to rotate vehicles as they move from a fixed roll angle on one curve to a different fixed roll angle on a following curve or tangent, Klauder said. Simulations of vehicle responses show that improved spirals have noticeably better dynamic characteristics than their traditional counterparts.
In improved spirals, the axis about which track roll takes place is raised above the track in order to reduce the forces required to rotate vehicles and to reduce the lateral accelerations perceived by passengers. A test carried out by Amtrak showed that improved spirals, which can be implemented using conventional tamping procedures, can provide substantial improvement in ride quality.
Another important aspect of managing wheel/rail interaction is the implementation of an effective rail grinding program. After achieving unsatisfactory results, Norfolk Southern applied Six Sigma “quality” concepts to its grinding program in order to determine how efficiently and how effectively it was grinding rail. Efficiency deals with the cost of the program; effectiveness deals with the performance on the finished rail. “You can grind very efficiently by using only one pass, but you can also leave finished rail profiles that cause the rail to fatigue prematurely or that cause high lateral forces that contribute to fastener and tie degradation,” NS’s Kevin Conn said.
NS made two fundamental assumptions at the start of this project: 1) Longitudinal creep forces are the main culprit in rail fatigue damage; 2) Grinding is necessary to control fatigue and prevent defects from propagating. As with any six-sigma project, NS began by collecting data. Overall, it determined that field relief of the low rail accounted for nearly 75% of the reasons for performing multiple passes on a track segment. NS further determined that the reason for this lay in the use of templates that required a grinding depth of more than 80 thousandths of an inch on the field side of the low rail — a template designed to reduce the potential for false flange contact and low-rail rollover. NS worked with the TTCI and its WRTOL (Wheel/Rail Tolerance) software to develop an alternate profile that required half of the metal removal required by the previous template. The new template also improved wheelset steering in curves.
NS noted accelerated rail wear in the first two months after grinding. “We concluded that we were removing too much metal in a single grinding visit, and that, in fact, we were removing the work-hardened layer of the rail,” Conn said. NS began performing pre-grind inspections with a laser-based optical rail measurement system in order to determine the most appropriate rail profile before grinding. “We are now grinding efficiently but we still have work to do to ensure that the finished rail profiles provide the best operating conditions,” he said.
By using the Six Sigma process, NS found that there were a number of interactions between different variables, Conn said. “If you only look at one variable, you don’t really understand how other variables influence it.” The trick, he said, is to identify the weakest link that limits the performance of the wheel/rail interface.
Grinding in Europe
Rail grinding in Europe has migrated toward a “product-based” specification, which identifies a specific profile to be achieved for each section of track. In this approach, the rail grinding contractor is typically given a specification, and must be able to verify that the specification has been met. “While the railway may indicate where to grind, it’s the contractor’s duty to plan and execute the grinding plan,” said Wolfgang Schoech, Manager of External Affairs at Speno International SA. Certification that the finished work meets the specification is required for payment.
The standard European specification, which was developed by the European Committee for Standardization (CEN), is essentially an envelope of all of the individual railways’ standards, and deals with longitudinal and transverse profiles, and the condition of the rail surface. Requirements relating to the surface “finish” are also included in the European specifications. Facet width, for example, should be less than 10mm on the crown radius, 7mm on the shoulder, and 4mm on the gauge side. The variation of the facet width should not be more than 25%. Additionally, the rail may not exhibit coloring or bluing after grinding.
Roughness is also specified. The rail grinding contractor must measure six points on each rail after grinding; only one may be outside the specified value of 10 microns. While metal removal rates are not specified, the typical metal removal rates are 0.3mm for new rail, and 0.2mm for cyclical preventive grinding programs. Metal removal rates for corrugation removal typically are 0.1mm below the corrugation troughs.
Since payment is tied to performance, rail grinding contractors utilize various high-tech measuring tools and systems, such as laser-based optical rail measurement systems to measure and document their work.
You’ll hear the wheel/rail interaction “systems” theme picked up on and advanced at the 14th Annual Wheel/Rail Interaction Seminar in Chicago in May.