James P Hyslip PhD PE
Vice-President Engineering, Optram Inc
THE CONDITION of track substructure has a profound influence on track performance. Ensuring that it is in good condition and keeping it that way is essential on high tonnage lines, although determining precisely what that condition is may not be straightforward. The advent of tools such as ground-penetrating radar will make this process easier, and pioneering work with GPR has been carried out in North America.
Before exploring the development of GPR, it may be helpful to set out the background to the project. Track substructure is the term used to describe the different layers of rock and soil under the sleepers, including the ballast, sub-ballast - or formation protection layer - and the subgrade soil. Poorly performing substructure not only leads to high rates of track geometry degradation but also promotes higher rates of wear of the rail, sleepers, fastenings and other track components.
Substructure problems are typically associated with poor drainage, fouled ballast, subgrade failure or deformation, and longitudinal variation of conditions. Correction of chronic problems requires the root causes to be determined - typically one or more from this list. It is important to note that resources can be wasted on implementing an incorrect or incomplete solution.
A significant part of the track maintenance budget on US Class I railroads is allocated to correcting poor track quality caused by movements in the substructure under repeated train loading. Accurate knowledge of the substructure condition is important in effectively assessing the potential for serious degradation that would interrupt traffic and require speed restrictions.
The ballast and soil layers in the track system are subjected to unparalleled loading conditions, in that trains apply a very heavy load at exactly the same location over millions and millions of cycles. Consequently, the ballast and soil under the track often deform and fail in ways that are unique to the railway loading environment.
Substructure performance becomes increasingly critical as axleloads, and the number of repetitions of these loads, increase. Historically stable track can begin to deteriorate rapidly when heavy axleload traffic is introduced.
North American railroads are placing unprecedented demands on the track substructure, essentially because of higher axleloads, rising freight traffic and rationalisation of lines. Additionally, some existing main lines are being considered for mixed use with new high speed passenger services where the challenges of different dynamic loads and vibration will need to be addressed.
Recognising the problem
It is for these reasons that the North American railroad industry is working to develop more effective Substructure Maintenance Management (SuMM) techniques. For example, Burlington Northern Santa Fe, with the support of the Federal Railroad Administration, has recognised the need for improved substructure management and has made efforts to develop technology that will help to manage the track substructure more effectively.
Over the past 2 1/2 years, BNSF and FRA have sponsored GPR development by a team consisting of Ernest Selig and Jim Hyslip of Optram Inc, based in Hadley, Massachusetts, Gary Olhoeft from the Colorado School of Mines and Stan Smith of GeoRecovery Systems Inc. This has led to the development of a prototype GPR system mounted on a road-rail vehicle and the completion of GPR surveys on more than 300km of track in North America, including main lines used by heavy freight traffic and third rail electrified commuter lines.
The results of this work to date have shown that GPR can be used to assess the condition of track substructure and produce quantitative indices for use in the management of track maintenance.
How it works
The GPR method transmits pulses of radio energy into the subsurface, and then receives returning pulses that have reflected off layer boundaries below the track surface. GPR antenna pairs, consisting of transmitter and receiver pairs, are moved along the track with a continuous series of radar pulses, giving a profile of the subsurface. Reflections of the GPR pulse occur at boundaries in the subsurface where there is a change in material properties. Only a portion of the pulsed signal is reflected at a layer boundary, and the remaining part of the pulse travels across the interface to be reflected again back to the receiver from another interface boundary. The time the pulse takes to travel through the layer and back is controlled by the thickness and properties of the material.
The GPR equipment currently in use is mounted on a road-rail vehicle and includes multiple sets of 1 GHz air-launched horn antennae suspended above the track. The air-launched antennae permit fast survey speeds and high-resolution measurement. The GPR vehicle moves continuously at speeds ranging from under 3 km/h to over 40 km/h. However, higher speeds are possible, and the desired speed is a function of the longitudinal resolution required for the application. At 16 km/h the GPR system achieves resolution of a few centimetres horizontally and a few millimetres vertically to depths of more than 2 m.
Antenna configuration and surveying procedures are designed to reduce the influence of sleepers and rail. Antennae are located at both ends of the sleepers as well as in the track centreline, so variation of condition can be seen laterally across the track. This arrangement permits three continuous longitudinal profiles to be collected simultaneously as the pairs of antennae are moved along the track. In this way the subsurface condition variations are obtained across the track as well as along the track.
An accurate DGPS location system is supplemented by a vehicle distance-measuring encoder. There is also a digital video with integrated DGPS for identifying track features, clutter and obstacles along the track and to augment accurate positioning. The GPR software has automatic processing and modelling capabilities to provide substructure condition indices for use in maintenance management programmes.
The prototype can provide quantified information on substructure layers, such as thickness, lateral/longitudinal extent, and changes in layers from repeat surveys over time. It can detect the presence of trapped water, and further enhancement using proven radar modelling techniques will soon give it the ability to quantify accurately the thickness and water content of the substructure layers as well as the material composition of these layers. Methods are also being explored to determine the fouling condition of the ballast by analysing the textures of the GPR images.
The application of GPR in the railway industry can be separated into three categories: site-specific investigations, maintenance planning and monitoring.
In the first area, GPR provides a means to obtain a visual image of the substructure along a section of track with particular chronic problems. Along with railway geotechnology, this can be used to determine the root cause of the problem, the precise sections needing repair and cost-effective treatment.
GPR data is being used to determine if chronic track problems are due to poor drainage, fouled ballast, subgrade failure or deformation, or longitudinal variation of conditions.
In the second area of application, GPR provides the ability to survey many kilometres of track and provide a relative measure of maintenance priority using performance indices. Finally, GPR can be used to monitor chronic sites regularly and to observe the effectiveness of remedial treatment.
A common substructure problem where heavy axleloads are the norm is poor drainage caused by subgrade failure or deformation. This condition is illustrated in Fig 1, which shows the typical 'bathtub' condition that develops from subgrade deformation. The current GPR system has antennae suspended above the ends and middle of the sleepers, which enables the equipment to detect lateral variations of the subgrade surface.
Fig 2 provides an example of results obtained by GPR along a section of track approximately 150m long. A single-channel GPR image is shown, together with photos of the cross-trenches (test pit excavations) that were excavated for calibration to show the layers and composition of the substructure.
Fig 3 shows the actual longitudinal scans for approximately 250 m of track with the digitised subgrade layer superimposed on the images for emphasis. The white arrows indicate the distance from the datum to the top of the subgrade surface. Locations A and B are where the subgrade soil has deformed upwards, creating a situation similar to that shown in Fig 1. The results shown in Fig 3 indicate that the root cause of the problem is associated with the bathtub effect of deforming subgrade, and now can be used to prescribe the exact locations for remedial treatment.
The ability of the GPR system to travel at relatively high speeds and capture nearly continuous information on subsurface condition makes it an invaluable tool for use in track maintenance management. In fact, GPR data is already being used to derive quantitative indices of track substructure condition precisely for this purpose. These indices are based on measurements such as layer contours, moisture content in the different substructure layers and the amount of fouling material in the ballast. General indices of track substructure condition can be based simply on the longitudinal and lateral variation of layers, water and composition, as these variations often result in track stiffness variations, which translate into rough track.
From GPR data to information
In developing wide-area indices for robust application, GPR information needs to be calibrated to the substructure characteristics and must be correlated with other measurements such as track geometry, right-of-way features, track stiffness and maintenance records. The Optram Right-of-Way Infrastructure Management (ORIM) database and viewer (www.optram.com) is being used to help correlate the substructure characteristics derived from GPR with these other condition indicators. ORIM allows the relationship of GPR data to track condition and features to be seen, and helps the user to visualise the substructure effect on geometry trends and maintenance effort. It also analyses the integrated data and generates user-defined indices for prioritisation of inspection and maintenance work.
An example of a condition index based on GPR information is depicted in the ORIM screen grab in Fig 4. This shows two parallel longitudinal GPR profiles along a 2·4 km section of main line with heavy freight traffic. The images, shown with digitised layer boundaries, indicate ballast pockets that have developed in an embankment under the impact of heavy axleloads.
Ballast pockets are load-induced depressions in the subgrade surface directly under the sleeper. A Ballast Pocket Index was produced to indicate where the ballast pocket condition has developed. Fig 4 shows the BPI matched with the track layout and track geometry data, and careful examination of the illustration reveals that many of the geometry rough spots are being driven by the ballast pocket problem.
A common remedy to minimise the continued development of a ballast pocket is to excavate a cross-drain - essentially a ballast-filled trench - perpendicular to the track. GPR can delineate the bottom of the pockets to ensure that lateral drainage is provided in the most effective location, generally the lowest point of the ballast pocket.
The potential for GPR to improve substructure maintenance management on North American railways is quickly being realised. Specifications are now being developed for next-generation equipment to be installed on purpose-built recording cars able to survey substructure condition at up to 100 km/h.
These vehicles are still planned to be road-railers, but the ultimate goal of railroads like BNSF and the FRA is to fit GPR on rail-based track geometry vehicles, so that routine substructure measurements can be made as frequently as the track geometry is assessed.
Construction of the next-generation vehicles may begin as soon as early summer 2005 and, subject to certification by the Federal Communication Commission, the vehicles could be in use on US railways by the end of 2005.
Prototype road-rail vehicle with an array of three GPR units allowing examination of the substructure beneath both sides and the centreline of the track
Fig 1. Poor drainage is a common problem on heavy-haul lines with high axleloads. The typical bathtub shape of the ballast develops from subgrade deformation. L, C and R represent the position of the GPR antennae over the left, centre and right of the track
Fig 2. The results of a GPR test over a 150 m section of track with a single-channel image. The pictures to the left and right show cross-trenches excavated for calibration purposes which reveal the layers and composition of the substructure
Fig 3. Longitudinal scans for a 250 m section of track with the digitised subgrade layer superimposed on the images. The arrows indicate the distance from the datum to the surface of the subgrade
Fig 4. Example of a condition index from GPR data shown in a screen grab from the Optram Right-of-Way Maintenance Management database. It relates track geometry to subgrade condition along a 2·4 km section of a heavy-tonnage route
Demands on subgrade drive GPR development
Higher axleloads, rising traffic and route rationalisation are placing unprecedented demands on the track substructure of North American railways. In response, the industry is developing more effective substructure maintenance management techniques that include the use of ground-penetrating radar (GPR). Tests on more than 200 km of track have demonstrated that GPR assessment can be used to produce quantitative indices for use in managing track maintenance. A prototype system consisting of three sets of 1 GHz air-launched horn antennae mounted on a road-rail vehicle is already surveying below the track at both sleeper ends and along the track centreline, while specifications are being drawn-up for a new generation of GPR to be fitted on purpose-built recording cars operating at up to 100 km/h.
Les sollicitations des plateformes conduisent au développement du GPR
Sur les réseaux nord-américains, des charges par essieu plus élevées, un trafic en hausse et la rationalisation placent les sollicitations des plateformes de voies à un niveau jamais atteint. Pour y répondre, l'industrie met au point des consignes techniques plus efficaces de maintenance des structures sous-voies, qui incluent l'utilisation du radar à pénétration dans le sol (GPR). Des essais sur plus de 200 km de voies ont démontré que les évaluations données par le GPR peuvent être utilisées afin de produire des indices destinés au management de l'entretien des voies. Un système prototype, à base de trois ensembles d'antennes en forme de corne, de 1 GHz, fonctionnant avec lame d'air entre antenne et sol, montés sur un véhicule rail-route, analyse déjà les dessous de la voie, à la fois au niveau des extrémités des traverses et tout le long de l'axe de la voie, tandis que des spécifications sont élaborées pour une nouvelle génération de GPR à installer sur des voitures travaillant à 100 km/h.
Anforderungen an den Unterbau treiben GPR-Entwicklung voran
H”here Achslasten, höheres Verkehrsaufkommen und Rationalisierungen beim Streckennetz bringen unvorhergesehene Anforderungen auf den Gleis-Unterbau bei nordamerikanischen Bahnen. Als Antwort darauf entwickelt die Industrie leistungsfähigere Unterbau-Management-Techniken, welche unter anderem auch den Einsatz von Erdreich-penetrierendem Radar (GPR) umfassen. Versuche auf mehr als 200 km Streckengleis haben gezeigt, dass eine Vermessung mit GPR zur Erzeugung quantitativer Richtgrössen für den Gleisunterhalt herbeigezogen werden kann. Ein Prototypsystem, welches aus drei auf einem Hybridfahrzeug montierten Horn-Antennen im 1 GHz-Bereich besteht, wird bereits zur Untersuchung des Bereiches unterhalb des Gleises an den Enden der Schwellen und in der Gleismitte eingesetzt. Zudem werden die Spezifikationen für eine neue Generation von GPR erarbeitet, welche auf bis zu 100 km/h schnellen Messwagen montiert sein werden.
El uso de la plataforma impulsa el desarrollo de GPR
Las cargas por eje cada vez mayores, el aumento del tráfico y la racionalización de las rutas están generando exigencias sin precedentes en la plataforma de los ferrocarriles norteamericanos. Para darles respuesta, la industria está desarrollando técnicas más efectivas para la gestión del mantenimiento de la plataforma que incluyen la utilización del georadar (GPR). Las pruebas realizadas en más de 200 km de vías han demostrado que las evaluaciones con GPR sirven para producir índices cuantitativos que luego se pueden utilizar en la gestión del mantenimiento de la vía. Un prototipo de sistema, consistente en tres grupos de antenas de bocina aéreas de 1 GHz montados en un vehículo vía-carretera, examina ya el terreno por debajo de la vía en los dos extremos de las traviesas y a lo largo del eje de la vía, al tiempo que se están redactando las especificaciones de una nueva generación de GPR que se adaptará a coches de auscultación que operarán hasta a 100 km/h.