Keith Hawthorne is Vice-President, Technology, at Transportation Technology Center Inc. R B Wiley is Principal Investigator I, and Tom Guins is Senior Engineering Economist
THERE ARE today 1·6million freight wagons operating in interchange service among the railroads of the United States, Canada and Mexico. Keeping track of their maintenance and operational condition is no small task, but advanced IT is opening up new ways to reduce the ongoing risk of accidents and derailments caused by defective vehicles.
Until the early 1980s the vast majority of wagons were owned by the railroads, with the exception of most tank cars and a few other specialised vehicles. This situation has now changed dramatically, and today shippers and other private owners hold 55% of the North American wagon fleet.
When most wagons were railroad-owned, interchange rules governing wagon inspection and maintenance were based on agreements between similar organisations. Today, owners range from large private companies with fleets of 180000 wagons and more down to investors owning far fewer vehicles and even single wagons. Wagon leasing arrangements also cover a broad spectrum, from turnkey leases where the entity hiring the wagon is responsible for maintenance to full-service agreements where it is the responsibility of the vehicle owner.
Thus the ability of railroads to ensure that all wagons operating in North America are maintained to the requisite safety standards has become a much more complex and demanding task. Through its Advanced Technology Safety Initiative, the Association of American Railroads is now providing wagon owners with the means to share the maintenance burden. Exchanging information on the condition of components as captured by wayside detectors enables owners to perform their own repairs when and where it is most efficient and economical to do so.
Tracking wagons on the web
Key to the rapid implementation of ATSI have been tools developed by AAR's research subsidiary TTCI and its IT counterpart Railinc. Prior to the start of the initiative, TTCI had made substantial progress in the development of its Integrated Railway Remote Information Service, InteRRIS¨, an internet-enabled system that gathers detector data from across North America. Railinc was able to provide a suite of databases containing detailed information on both railroad-owned and private wagons.
Work began in 2002 on the first ATSI implementation, involving the use of wheel impact load detectors to identify out-of-round wheels. Since 1996 AAR interchange rules have allowed the removal of any wheelset with a wheel impacting the rail at or above 41 tonnes, but until the advent of ATSI the burden of removing such wheelsets had fallen on the railroads.
ATSI went live on October 1 2004, with the Equipment Health Management System ready on schedule to support implementation. EHMS combines InteRRIS with Railinc's suite of information management tools. This first ATSI application also includes a new category of notification of the impending need for repair. This provides condition information to wagon owners in advance of a component becoming condemnable under the interchange rules.
Other strategies under review by the ATSI Task Force include two new repair categories, 'opportunistic repair' and 'final alert'. Opportunistic repair would allow for the replacement of high-impact wheels which are still below the AAR threshold when another defect is being repaired on the vehicle. Final alert would be substantially above the AAR limit, and would provide a mechanism for repairing railroads to recoup any additional costs arising from unscheduled repairs or train stoppage.
The idea of bringing together data from wayside detectors on all North American railroads was conceived at TTCI in late 1998. The concept has been driven by the interchange system whereby wagons typically operate over several railroads, as more than 50% of freight traffic is interchanged between two or more. The most complete picture of vehicle performance requires the consolidation of data from several railroads with their disparate communications, computing and data management systems. With at least 13 different detector technologies in operation and twice as many suppliers active in the market, the potential for integration was obvious.
TTCI began development of a prototype system in 1999, based on data provided by bogie performance detectors which measure vertical and lateral wheel/rail forces on a pair of reverse curves. A railroad working group established the fundamental requirements, which included the need to integrate data from multiple detector types and provide secure, internet-based access to authorised data for all wagon owners, including those other than railroads. The system would also be required to process complex alarm messages from multiple detector types, including time-based or trended alerts.
In the wider context, it was hoped that the system would provide industry decision-makers with useful information on vehicle condition and performance. This would require the development of a comprehensive vehicle inspection system, incorporating existing detector systems as well as new systems to fill gaps in current capability. TTCI would also assist railroads in the systematic deployment of detector systems to minimise costs and maximise industry-wide information quality, and work continuously to enhance and expand the usefulness of maintenance data accessible via the internet.
By mid-2000, development of a production system had begun in earnest and InteRRIS was born. The application interface was kept simple, yet powerful, with a focus on supporting actual workflows. A key to the latter was providing information requiring action in a reduced or summarised format. InteRRIS went live with data from bogie performance detectors by November 2000, with data provided by wheel impact load detectors added by the end of that year. An event-tracking module was implemented in 2001 when automated notifications were also introduced.
Subscribers to InteRRIS use the data in many different ways. Notifications are used directly to withdraw wagons for repairs, while some customers receive reports based on lower thresholds, allowing for preventive maintenance such as fitting tread-conditioning brake shoes to prevent wheel flats from worsening. Other subscribers have data forwarded to third-party agencies that provide fleet maintenance management on a contract basis for owners of small wagon fleets.
The data-driven design of InteRRIS, supporting core performance measurements, provides for rapid integration of new detector types. The complex event processor is independent of detector types, enabling new types to be added without modification. Boolean operators and compound measurement evaluations are stored as data items, to be assembled when new alert criteria are defined. Programming changes are therefore not required for event processing, only the addition of data tables for the new type of detector.
Specific notifications are sent to customers about vehicle performance based on their selected parameters, criteria and limits. To support preventive or predictive maintenance, an alert must be issued before component failure becomes imminent. Individual vehicle identification is necessary to compile performance history, with the most authoritative notification reports based on trending. The event-tracking processing engine is built in layers, using fuzzy logic principles, yet can provide the simple one-hit alert required by ATSI.
Acoustic bearing detector data was integrated in 2003 with wheel profile detectors targeted as the next data type to be added. Fig 1 shows the general system model of InteRRIS.
Leveraging existing systems
In late 2003, InteRRIS was selected to provide data warehousing and alerting for ATSI. Alert data was sent to Railinc's existing Early Warning System, and daily notifications were sent to parties responsible for maintenance, wagon owners or fleet managers.
To facilitate inter-system data flows, InteRRIS was migrated from a third-party hosting facility in Michigan to Railinc's own server centre in Pittsburgh, Pennsylvania. This transition began in August 2004 with the integrated systems going live. The first ATSI notices were processed on October 1 2004, on schedule. A high-level view of the ATSI system model is shown in Fig 2 as the Equipment Health Management System.
There are currently 107 detectors in North America reporting to InteRRIS, comprising 79 wheel impact load detectors, 21 bogie performance detectors and seven acoustic bearing detectors. The next major development effort will focus on providing a composite vehicle performance report, bringing together data from all supported detector types into a single display or report. This should highlight the correlation between different defect types or symptoms such as wheel impacts and bearing defects.
The cost of wheel flats
The effects of high-impact loads generated by out-of-round wheels are more rapid deterioration of track and vehicle components, additional fuel consumption and increased risk of derailment. The net economic benefit of removing high-impact wheels from service involves a trade-off between the cost of removing wheels before they normally wear out and the savings to be made on track and rolling stock maintenance, reduced fuel consumption and improved safety performance.
The loss of the remaining useful life of a wheelset removed because it is generating high-impact loads is an incremental cost to the wagon owner. The net present value of this incremental wheel replacement cost is equal to the difference between the NPV of the wheel replacement cost at a normal cycle and the NPV of early wheel replacement due to the high-impact condition. Using the average North American values of US$1000 per replacement wheelset, 400000 km for wheelset life, 40000 km for annual wagon usage and a capital cost of 10·5%, the incremental cost of removing a high-impact wheel at the time of detection is $619.
Damage to the track structure caused by high-impact wheel loads can lead to reduced rail weld life, accelerated growth in rail defects and damage to turnout frogs due to batter, metal flow and fatigue. Rail bearing plates are more likely to cut into the sleepers, and ballast condition deteriorates more rapidly.
TTCI has estimated the magnitude of each of these costs using track deterioration models to calculate the incremental annual cost of allowing high-impact wheels to remain in service. Fig 3 shows the NPV savings to be made by replacing wheels at different impact levels.
Mechanical damage to vehicle components associated with high-impact wheel loads is concentrated in the unsprung areas of the bogie. Roller bearings are subjected to high loading during impacts, as is brake equipment attached to the bogie side frames. Vibration of the wagon body resulting from wheel impact also reduces the life of brake control valves.
Using data collected by AAR member railroads, TTCI has estimated the damage to wagon components. The incidence of high-impact wheel loads shows high correlation with roller bearing failure and damage to brake beams and control valves. Based on analysis of the available data, TTCI was able to develop models to estimate the incremental cost savings resulting from the replacement of high-impact wheels.
Research has shown that an out-of-round wheel or flat resulting from wheel slip increases rolling resistance. This is because at every wheel revolution, the weight carried by the out-of-round wheel must be lifted somewhat to overcome the damaged portion of the wheel before the smooth portion contacts the rail. Out-of-round wheel condition can result from tread damage due to spalling or thermal shelling.
Savings in fuel costs are achieved by replacing high-impact load wheels with a smooth wheel, avoiding the incremental rolling resistance from the high impact wheel. TTCI has calculated the potential fuel savings using a model developed by Battelle-Columbus Laboratories. A typical North American combination of timber sleepers, 64 km/h operation and wagons with a gross weight of 120 tonnes at the rail was used to estimate the incremental fuel consumption associated with a variety of high-impact wheel shapes. Fig 4 shows the resulting increase in fuel consumption for impacts ranging from 23 tonnes to 68 tonnes.
Using a distribution of high-impact wheels based on in-service data from AAR member railroads, TTCI has estimated the economic benefits of replacing wheelsets under current interchange rules when the impact reaches 41 tonnes, producing a potential annual benefit of some $41m. As Table I shows, the break-even point is in fact 28·6 tonnes.
The ATSI Task Force is working to include more detector types in the initiative. For 2005, they have elected to investigate the inclusion of hunting (high speed instability) detectors. In this case, there are no current AAR interchange rules, but the benefits of eliminating unstable bogies come not only in the form of enhanced safety, but also in reduced damage to wagons and their payloads.
For example, over 70% of cars sold in North America are delivered by rail, and a large proportion of automotive components are delivered to assembly plants in this way. Markets such as these are highly sensitive to ride quality, making the elimination of bogie hunting highly desirable.
Table I. Cost:benefit summary
Total track maintenance benefits 19679
Rolling stock maintenance benefits1 11894
Total benefit 52146
Wheelset cost (11290)
Net benefit 40857
Break-even impact force 28·6 tonnes
1. Based on wagon damage sheet, including bearings2. Low based on CN/BNSF data
CAPTION: TOP: Eliminating high-impact wheel loads will generate benefits in terms of lower damage to track and rolling stock components as well as lower fuel consumption
RIGHT: Fig 1. TTCI's InteRRIS database gathers data from more than 100 defect detectors at key locations across the North American rail network
CAPTION: Fig 2. Launched on October 1 2004, the Equipment Health Management System integrates the InteRRIS database with a range of information management tools
CAPTION: Fig 3. Breakdown of NPVsavings by component at 40 225 km per car per year
CAPTION: Fig 4. Comparison of fuel consumption costs against wheel impact level
Wayside monitoring with ATSI aids wagon maintenance
The first phase of the Association of American Railroads' Advanced Technology Safety Initiative has enabled wagon owners to share data on wheels exceeding impact limits, thereby giving them the opportunity to undertake maintenance when and where it is most efficient and economical to do so. Internet-based data collection and analysis is used to flag up deteriorating trends, and analysis of the various costs has determined the condition levels at which alerts are issued
La surveillance au défilé avec l'ATSI aide à l'entretien des wagons
La première phase de l'Advanced Technology Safety Initiative (ATSI) entreprise par l'Association of American Railroads a permis aux propriétaires de wagons de partager les données relatives aux roues sur lesquelles les limites d'impact sont dépassées, leur donnant ainsi l'opportunité d'entreprendre l'entretien o?€” et quand cela est le plus efficace et le plus économique de le faire. La collecte des données est basée sur Internet et l'analyse est utilisée pour signaler les tendences à la détérioration; de plus, l'analyse des différents coûts a déterminé les niveaux de condition auxquels les alertes sont émises
Ortsfeste Überwachung mit ATSI hilft dem Wagen-Unterhalt
Die erste Phase der Advanced Technology Safety Initiative (ATSI) der Association of American Railroads erlaubt es den Wagen-Besitzern, Daten über Räder, welche Grenzwerte überschreiten, gemeinsam zu nutzen, was es dann erm?€?glicht, den Unterhalt wann und wo es am effizientesten und wirtschaftlichsten ist, durchzuführen. Internet-gestützte Datenerfassung und -Analyse wird zur Erkennung von Zerfalls- Trends benutzt, und die Analyse der verschiedenen Kostenfaktoren hat es erlaubt, die Zustandswerte zu bestimmen, bei welchen Warnungen ausgegeben werden
El control en línea con la ATSI contribuir? al mantenimiento de vagones
La primera fase de la Advanced Technology Safety Initiative (ATSI) que ha asumido la Association of American Railroads ha permitido a los propietarios de vagones compartir datos sobre ruedas que superen los límites de impacto, lo que les brinda la oportunidad de hacerse cargo del mantenimiento en el momento y lugar en que les resulte m? s eficaz y económico. La recopilación y an? lisis de datos basada en Internet permiten señalar las tendencias de deterioro, mientras que a través de un an? lisis de los diversos costes se han determinado los niveles de estado en que se emiten las alertas