ALERT-ME system installation, August 2010


In August 2010, we began the system installation which included:

  • Initial site investigation and track levelling to assist in siting the ALERT sensor arrays
  • Installation electrical resistivity electrode arrays and a geotechnical sensor network to enable high resolution, dynamic 3D volumetric imaging of the embankment
  • Installation of low powered field instrumentation and wireless data transmission to remotely control the ALERT resistivity monitoring system

Site investigation and track levelling

Static cone penetration resistance test (SCPT) vehicle
Continuous surface wave (CSW) survey

Various phases of site investigation activities have been undertaken at the site since September 2005. The 2010 phase included static cone penetration testing (sCPT), continuous surface wave profiling (CSW) and track levelling.

The site investigations were used to gather information about the materials, their properties and their distribution within the embankment, and how this affects drainage and track performance.

The sCPT technique (Gunn et al. 2007) uses a cylindrical cone, pushed vertically from a rig (as shown in Figure 1) into the ground at a constant rate of penetration of 20 mm per second. During penetration, measurements are made of the cone resistance, the side friction against the cylindrical shaft and, in piezocone tests, the pore water pressure generated at penetration by the cone.

The CSW technique uses sinusoidal surface waves generated by an electromagnetic vertical vibrator (the white pod behind the vehicle in Figure 2) seated on the ground surface. In practice, this produces a series of finite duration pulses, each at a single frequency over a range of frequencies, for example from 5 Hz to 10  Hz in increments of 0.5 Hz or 1 Hz.

Field data acquisition at each frequency is synchronized with the control to the vibrator. Field dispersion curves are generated from the recorded signals at two or more receivers using a method based on the steady state Rayleigh method.

The CSW technique is particularly suited to railway sites where ambient noise levels are high, and the field data are interpreted to provide a stiffness-depth profile (Gunn et al. 2006).

Test locations

Embankment plan

The embankment plan (Figure 3) shows the combined SCPT and CSW test locations to extend 300 m along the embankment.

A longitudinal 2D section will be produced along the axis of the embankment from which fill materials and their distribution will be interpreted.

From our partial interpretations we have managed to locate the ALERT-ME electrodes arrays at the interface between materials with differing engineering characteristics and performance.


Penetration resistance

The section below (Figure 4) shows upper interval from the embankment surface to approximately 0.8 m depth comprises successive layers of ballast built up from the original cobbles used in the initial construction.

Penetration resistances were above 10 MPa and up to 40 MPa within this interval. Beneath the cobbles and ballast, within the interval from 0.8 m to approximately 4 m depth, SCPT values are generally below 4 MPa from stations 0 m to 40 m along the section, but increase up to about 10 MPa within this depth interval from stations 40 m to 60 m.

Generally, the SCPT values continue to be less than 4 MPa below 4 m depth, continuing at this value into the bedrock (Cropwell Bishop Formation of the Mercia Mudstone Group), which extends from approximately 5.5 m depth.

One exception to these observations is where values peak to around 20 MPa in the fill just above the top surface of the bedrock. These occur on the interpreted section as a thin band at around 5.5 m depth from stations 0 m to 30 m and a disrupted band at around 5 m depth from stations 40 m to 60 m.

This correlation of peaks is consistent with a laterally disrupted feature that could have formed as an exposed surface by trafficking and/or drying out, or as a natural accumulation of large intact blocks in the early phase of embankment construction.

Strain stiffness profile

Small strain stiffness

The section below (Figure 5) shows the small strain stiffness profile constructed from the CSW tests also from stations 0 m–60 m. From 0 m to 35 m, stiffness values range from 80 MPa to 180 MPa in the upper interval from ground level to 0.8 m depth, whereas from 35 m to 60 m stiffness values range from 50 MPa to 110 MPa from ground level to 0.8 m depth.

Over much of the section this upper interval (above 0.8 m) is generally stiffer than the interval from 0.8 m to 4 m. n this lower depth interval, stiffness values are generally below 60 MPa from stations 0 m to 30 m, between 60 — 80 MPa from stations 30 m to 50 m and rise to around 100 MPa (with peaks to about 180 MPa) towards the 60 m station.

A high stiffness lens extends from the 44 m station to the 60 m station from 0.5 m depth to approximately 2 m depth. The position of this lens coincides with a zone of high penetration resistances observed in the SCPT profile (above).

The stiffness gradually increases with depth across the whole site, in a manner consistent with increasing effective stress, with stiffness values reaching around 130+ MPa at 8 m depth in the Cropwell Bishop Formation of the Mercia Mudstone bedrock.

Friction ratios (FR), measured during the SCPT surveys, of below 0.5% occur across the site from stations 0 m — 60 m in the upper ballast and underlying gravel and cobble layers to a depth of about 0.8 m.

From stations 50 m to 60 m a zone of low friction forms a lens of increasing thickness to around 1.5 m deep at station 60 m Pitting near the 60 m station revealed glaciofluvial sand and gravel, and occasional cemented sandstone to a depth of at least 1 m and is likely to contribute towards the increased SCPT penetration resistance, increased stiffness and lower FR values.

Track geometry

Track levelling indicated very poor track geometry over this short section of embankment. As shown below (Figure 6) the track appears to sag by around 20–30 mm from stations 20 m to 45 m where it meets the interface of the sand and gravel lens. Here there is a localised, raised bump in the track, which then appears to have sagged by around 20 mm from the 45 m station to around the 70 m station.

We are not sure of the causes of this sagging along either of these two sections.

For example, it could have been affected by the sleeper replacement programme or it may be due to gradual deterioration of the ballast integrity.

However, the change in fill materials and history of poor track geometry makes this an excellent location to site the ALERT-ME electrical resistivity arrays. This will enable us to assess the impact of the fill materials and also local vegetation species which comprises ash, hawthorne and oak on the moisture movement in this zone of the embankment.

Track geometry

Electrical resistivity electrode arrays and geotechnical sensor network

Layout plan showing the locations of the 3D electrode array and the geotechnical probe network

The 3D test area covers an area of approximately 22 m along the axis of the embankment from stations 50 m to 72 m and 32 m across the transect of the embankment from the toe of the east to the toe of the west side.

(There is also a 100  long 2D line running along the crest of the embankment from stations -20 to 80 m).

3D array installation

Cables laid across the embankment transect: drawing cables up the west flank

The 3D array installed in August 2010 comprised cables that were laid across the transect of the embankment (Figure 8.)

Twelve cable-pairs were laid from the toe to mid-crest with the first half of a pair covering the east side and the second half covering the west.

Each cable-pair provides coverage across the embankment transect from the east to west toe of 64 electrodes spaced at 0.5 m.

The approximate spacing of the cable-pairs along the axis of the embankment was approximately of 2 m.

Remote programming

Array configuration

The ALERT system at East Leake can be remotely programmed to gather many thousands of four-electrode measurements with a range of geometries across the area of interest.

Typically, we use the dipole-dipole array configuration (Figure 9) to gather field data.

The dipole-dipole command sequences comprised both normal and a full set of reciprocal measurements. Reciprocal measurements provide a robust means of assessing ERT data quality and determining reliable and quantitative data editing criteria (Chambers et al. 2007, 2008).

All the gathered field data will be quality checked and processed using a true 3D inversion to produce volumetric images of the resistivity distribution within the embankment.

An early result is shown below (Figure 10), whereby the volumetric extent of the sand-gravel lens in the fill can be identified as a high resistivity feature. This will form a basis from which further 3D images will be compared using resistivity differencing.

The differential resistivity images will be calibrated with resistivity-moisture content relationships and used to develop time lapse images of the moisture movement within the embankment.

Example of a 3D volumetric image of the resistivity distribution within the pilot section of the embankment

Probe network

Geotechnical probes

A network of geotechnical probes was also distributed throughout the resistivity electrode array as shown in the above figures. The network included probes located on the flanks and the crest next to every third electrode cable-pair within the central area of the image.

Along the length of the electrode cable-pair, the probes on the lower east and west flanks were 1 m deep, on the upper east and west flanks were 3 m deep and on the crest were 3.5 m deep.

Probes on the lower flanks comprised a surface (200 mm deep) combined, electrical conductivity, moisture and temperature sensor and a temperature sensor at 1 m depth (Figure 11).

Probes on the upper flanks comprised a combined, electrical conductivity, moisture and temperature sensor and temperature sensors at 1 m and 3 m depths.

Probes on the crest comprised combined moisture content and temperature sensors at depths of 200 mm, 0.5 m, 1 m, 2 m and 3.5 m deep.

The flank probes were placed into hand-chiselled and augered holes with surface sensor backfilled with local material removed from the hole. The crest probes were pre-packed within a mixture of sand and kaolinite (Gunn et al. 2009) within a cardboard tube which was packed into a 3.75 m deep borehole.

These probes will provide time series measurements of the variation in field moisture and temperature changes at each specific sensor location distributed within the volumetric electrical resistivity image.

These time series data will provide information on how rapidly the embankment conditions will change and can also be compared to the property changes of voxels within the 3D image that coincide in location with each sensor. In this way these measurements will also aid the calibration of the 3D volumetric resistivity image.

Field enclosures and instrumentation

Strengthened steel field enclosure housing the field ALERT system
Solar panels

The whole ALERT concept is captured in the ALERT flyer and includes systems and control software enabling the transfer, storage and visualisation of remotely streamed data from the field onto a web-based portal that can be accessed from anywhere in the world.

Field installations include systems to support the power and control required to switch the current source and make the voltage measurements on the electrical resistivity electrode arrays.

The ALERT field instrumentation is a fully automated, time-lapse, microprocessor-controlled electrical resistivity tomography system which will be interrogated remotely over the duration of ALERT-ME (and possibly beyond).

'Real-time' data streams will be transmitted using wireless telecommunications (e.g. wireless 3G modem with SIM card for mobile phone telecommunications). These are all housed in a strengthened steel field enclosure (Figure 12).

Electrical power for the whole field system is provided by lead-acid batteries which are charged by solar energy provided by two standard solar panels at the site (Figure 13).

The whole field system is in place and has been gathering data from late October 2010.

We will be posting monthly images from the site on the Dynamic 3D Images page of this website. So, look out for the resistivity images over the next few months and the moisture content images that will follow.


Dr David Gunn
Team leader — Geotechnical & Geophysical
Properties & Processes
Email: Dr David Gunn

Dr Jonathan Chambers
Team leader — Geophysical Tomography
Email: Dr Jonathan Chambers