|NDT.net Sep 2005 Vol. 10 No.9|
Fast Inspection of Railway Ballast By Means of Impulse GPR Equipped with Horn AntennasDr. A. Kathage*, GSSI, North Salem, N.H., USA
J. Niessen, GBM Wiebe, Achim, Germany
G. White, N. Bell, Allied Associates Geophysical Ltd., Dunstable, U.K.
*Corresponding Author Contact:
Published in Proceedings of RAILWAY ENGINEERING-2005, The Eighth
ABSTRACTThe inspection of railways ballast by use of ground penetrating radar devices has been performed for several years now. A continuous and non destructive profiling of the ballast and subsoil offers obviously significant advantages compared with the traditional way of coring and sampling. The latter method is appropriate for layered structures which can be described by statistical methods but for example local structural defects can easily be overlooked because statistical drilling patterns can not be used for addressing this kind of problem.Last but not least it is well known that the application of non destructive inspection methods like GPR is reducing the inspection costs compared with the traditional approach. One serious handicap for the application of GPR on railways is survey speed. Exploiting the full economical potential of GPR would allow users to fit the surveys in between regular train schedules.This was not possible until newly developed equipment became available. The risk of hitting switches or similar obstacles close to the ground was the main reason for slow data collection speed with bow-tie antennas. Bow-ties need to be operated within a quarter of a GPR-signal wavelength.Practically this forced the GPR operators to mount their antennas not higher than 10 cm above the ballast. This low height allowed a maximum data collection velocity of only 30 km/hour.
The operation of horn antennas avoids this problem because they can be mounted about half a metre above the ballast. The development of a new 400 MHz horn antenna for railways ballast and subsoil inspection was additionally triggered by the availability of new GPR control units like the GSSI SIR-20. These units allow data collection rates of several hundred scans per second with a time resolution of 5 picoseconds for 512 or 1024 samples per scan. Using the 400 MHz horns with 50 nanoseconds time range offers survey velocities of more than 100 km/hour with 20 scans per metre. This scan separation has been identified to be an important parameter for good data quality. Less scans per meter would mean less information between the sleepers.
The use of the newly developed hardware and software for collecting high speed GPR data in combination with other sensors like RTK-PS,Doppler radar, video as well as the streamlined data processing and data interpretation routines will be presented in this paper. Examples of typical survey data and the final survey results will be shown for demonstrating the high performance of this new technology.
KEYWORDS: GPR, horn antennas, ballast inspection, computer analysis,
INTRODUCTIONThe traditional method for investigating the quality of the ballast uses sampling and analysing of drill cores.The sampling interval depends on local parameters.Usually drill cores are collected every 50, 100 or 200 meter. Local defects can be missed by this approach because they are not distributed along the railway route according to statistical laws. They can be located just far enough from a drill hole to be missed.
Nowadays big machines are employed for carrying out the required maintenance work. Unexpected standstill times caused by unforeseen obstacles in the ballast like muddy spots, bedrock, compact layers can lead to enormous rise of costs. A solid economical planning can not be performed based on these technical boundary conditions.
The application of ground penetrating radar (GPR) solves this information problem.Four GPR profiles are collected along the railways route. The planner obtains a non destructive section of the ballast structure from surface down to about 4 meters depth. Based on this continuous data the sampling of cores is guided exactly to the spot where defects have been detected by the GPR scan.
Standard bow-tie antennas have been used for this application in the past. These antennas need to be operated closely to the ground surface allowing a maximum survey speed of about 30 km per hour.Due to the increasing traffic on the railways network especially in east-- west direction this speed limit posed a serious handicap on the application of this method. Most of the GPR surveys were forced to be performed during night time when the traffic situation allowed the slow moving survey. That procedure turned out to be risky antennas could get damaged on near surface obstacles ((like switch boxes) and too expensive due to many long waiting periods (regular traffic had to pass by) during the data collection. Horn antennas needed to be used in this situation.This antenna type can be operated half a meter above the ground surface. Therefore the speed limitation would only be depending on the data collection rates of the GPR control unit. 1 GHz horns or even higher frequencies were commercially available. But the detection range of 4 meters required a lower frequency to be applied. Therefore customized 400 MHz horns were developed.
METHODOLOGYThe following table illustrates the typical key parameters of a railway line GPR survey project.
a) Data collection:
Two GSSI SIR-20 control units are used for collecting the GPR data (see figure 1). These systems allow scan rates of more than 800 scans per second (single channel, dual channel operation is multiplexed) which means that in combination with horn antennas a survey speed of 120 km/h can be achieved while collecting 20 scans per meter.A Differential GPS (see figure 2) in combination with a DMI (distance measuring device) and a Doppler-radar provide accurate distance-/position information. A PC is used for managing and saving of position as well as video data while an external Firewire hard-disk serves for backup of all survey data.
The antennas are mounted in front of the train. The mounting system allows hydraulic changing of the antenna height.
In practice, the following applies: One metre of measuring distance along the rails means: space between sleepers (approx.40 cm) +sleeper (approx.20 cm) +space between sleepers (approx.40 cm). In total this distance corresponds to approx. 1 metre. At twenty scans per metre, a total of 16 scans can be collected in the gaps between sleepers and four scans each through the sleepers. Four pulses per sleeper are sufficient for being able to image these sleepers clearly visible in the radargrams. In most cases the sleepers can then be counted individually! Sleepers are part of the building structure and therefore they need to be visibly imaged in the data. Ringing effects in the data due to sleeper reinforcement are caused by an inappropriate polarisation of the antennas. The figure 4 shows an example of the four GPR sections which are typically collected.
b) Data processing and interpretation:
Each profile can be described with standardized abbreviations. All detected features and signatures can be
saved in a special railways data base.The position accuracy of these data base elements is +/-1m.That
means also that the tracked layer interfaces are stored in that data base.
c) Data Examples:
CONCLUSIONSPR has proven to be a successful as well as beneficial method for the non destructive investigation of railways embankments and building grounds. Meanwhile thousands of kilometres of railways lines in several European countries like Germany, Austria, Switzerland, Hungary, Norway, etc. have been investigated with GPR for the benefit of safety as well as the optimized allocation of financial resources in times of tight budgets. The success of this method was based on the combination of dedicated hardware and software. Fast GPR control units in combination with air coupled horn antennas of 400 MHz and 1000 MHz allow high speed data collection,the specialised software allows fast data processing and interpretation. Due to the high quality of the survey results and not at least due to the time efficiency of this method PR already plays an important role for the geotechnical inspection of railways with high potential for the future.
REFERENCESPublications in two different German railways journals:
1) Der Eisenbahningenieur