To gather more insight into the performance of the transition zone, a monitoring programme has been designed that comprises both short-term and long–term measurements. The long-term measurements focus on capturing the track alignment (settlements), while the short-term measurements will be used to investigate the dynamic response, such as wheel–rail contact forces, sleeper–ballast forces, and sleeper displacements and accelerations.
To determine the vertical wheel–rail contact force between the two sleepers in a sleeper bay and the rail seat load on one sleeper, longitudinal strains at a given perpendicular distance (y) from the neutral axis (N.A.) of the rail at four sections in one sleeper bay will be measured using an FBG strain array comprising four gauges. The strain range is ±1500 μɛ, while the operating temperature range is -20°C to +60°C. By integrating the distributed force supporting the rail over the sleeper width s provides an estimate of the rail seat load. Further, by substituting x = v ∙ t (v is vehicle speed) into, it can be shown that the vertical acceleration a of the rail is determined by the curvature times the square of the speed. The acceleration inferred from the strain can then be integrated twice with respect to time to obtain the deflection w(x) of the rail.
FBG offers high sampling frequencies, a wide range of robust sensing options and immunity against electromagnetic interference (EMI). The optical fibre array is fixed to the bottom part of the rail web and can be interrogated using a single analyser. Each sensor cluster consists of an array with four strain gauges with temperature compensation, one displacement transducer, and one accelerometer (two accelerometers at clusters one and four). Each sensor cluster is connected to one channel of the interrogator. Further, to measure the rail seat load and sleeper displacement at four positions along the transition zone, four individual fibre arrays (cluster) at four sleeper bays (sleeper bays number 3, 5, 8, and 11 from the transition) will be deployed.
The vertical sleeper displacement is assessed by using an optical displacement transducer that is placed on the sleeper end. The displacement range is ±50 mm with a resolution of 30 μm, and the operating temperature range is -20°C to +60°C. Sleeper displacement is measured relative to a fixed reference anchor in the ground at four positions along the transition zone: 1.8 m, 3.0 m, 4.8 m, and 6.0 m from the 3MB slab. Further, absolute sleeper vertical acceleration will be measured using an optical accelerometer to assess track deflection at the same four sleepers and at two reference points. One of the reference accelerometers is placed on top of the 3MB slab, while the other is placed far away from the transition zone (sleeper number 30 from the transition).
Multiple levels of integrated wavelength referencing coupled with low-noise high-speed electronics allow for spectral feature tracking at a resolution of <20 fm at kHz–frequencies. The unusually high sample frequency of fs = 2 kHz enables the current combination of (slowly evolving) static and dynamic measurements. The interrogator used has no moving parts resulting in high reliability over a broad temperature range and forms an integral part of a rugged and reliable sensing system. Using optical sensing gives the possibility to have kilometres of fibre between the readout and sensors, adding sensors without compromising measuring speed and resistance to EMI. The acquisition system comprises a computer, an I4 interrogator, a heated enclosure (to keep the logger within operational temperatures), a junction box, and 120 m optical fibres to transfer data from the sensors to the interrogator.
The interrogator, which is connected to the rugged field computer in the heated enclosure, is continuously measuring all sensors at the pre-set fixed sampling frequency of 2 kHz and distributing those on a network socket. Subsequently, a custom script running on the field computer acquires the data from the interrogator, and formats and stores the data locally. The measurement script will be developed so that ‘dynamic’ snapshots, triggered by the far field accelerometer, are taken for each passing train and after post-processing (filtering, identification of the signal of interest) temporarily stored on the disk.
Subsequently, these data is synchronised with a server at Chalmers University of Technology. The expectation is that, given the amount of data obtained, size of the local storage on the field computer, quality of the telecommunication link (4g mobile data network), logarithmic nature of the degradation process, and total duration of the monitoring campaign, a trade off will be made on the measurement interval.
That is, the number of dynamic snapshots at the start and near the end of the monitoring period will vary. Here, the ability for remote configuration of the system will help adapt to the needs of the project. Furthermore, data reduction techniques that only keep track of certain evolving (static and dynamic) metrics (peak force, displacement amplitudes, key frequencies, permanent displacement) rather than storing the complete signal for each train passage will be investigated after some data becomes available.
Author: Kourush Nasrollahi, PhD student at Chalmers University of Technology