Monitoring the condition and stability of water retaining structures – dams, canals, embankments and levees – is crucial to the on-going security of water resources and hydropower supply, as well as the safety of communities living nearby. It is a multi-faceted task, involving the characterisation of hydraulic structures in terms of geometry, constituent materials, structural heterogeneities (such as voids and fractures), seismic risk, stability and, not the least, leakage.

In general, failures in earth-built hydraulic structures are related to internal erosion, an example being the Teton Dam failure in 1976 in the US state of Idaho. Internal erosion is defined as progressive fine particle motion under groundwater flow and is classed into a number of different physical processes, principally regressive erosion, dissolution and suffusion.

A statistical study of 11,192 earth hydraulic structures by Foster et al (1) in 2000 found defects in 136 of them; 6% due to sliding, 48% by overtopping and 46% caused by internal erosion.

Rapid and, ideally, non-invasive detection of erosion is far from straightforward, given the often significant scale and challenging setting of water power facilities. But a new geophysical tool, which is directly sensitive to groundwater flow, can offer this combination of fast, efficient investigation and accurate detection of leakage with minimum intrusion.

Commonly used
Geotechnical and geophysical methods are commonly used to assess the structural integrity of hydraulic installations and to detect structural heterogeneities. In these instances, cost effective geophysical techniques are applied first for broad characterisation with the aim of targeting subsequent geotechnical investigations to reveal a greater level of detail.

In the case of earth dams and levees, conventional geophysical methods such as electrical resistivity tomography (ERT) and electromagnetic (EM) surveys are often used to detect spatial heterogeneities and subsurface soil transitions (at depth). ERT has proved to be a useful tool for imaging earth dams and, in certain circumstances, for detecting leaks. Seismic tomography is typically used to determine compactness in masonry dams. Ground penetrating radar (GPR) can help determine the structure and condition of concrete and masonry structures, providing high resolution imaging to relatively shallow depth.

However, these techniques are not directly sensitive to internal erosion processes. While survey technology has progressed in many areas, periodic visual inspection is often the best method for detecting internal erosion, given that the observation of sand-boil in downstream flood areas and the presence of specific vegetation are the most pronounced indicators.

But visual inspections have their drawbacks. They are very often time consuming and, depending on the stage of internal erosion, may be ineffective in spotting all internal problems. For reasons of safety, sustainability and cost-efficiency, a geophysical method that locates preferential fluid flow would be a major step forward.

Self-potential method
The electrically passive geophysical method known as ‘self-potential’ (SP) offers one such solution. It has been proven in a number of commercial projects for the qualitative and quantitative determination of water flow paths in hydraulic structures.

Apart from temperature and hydro-acoustic measurement, the SP method is the only geophysical technique that is directly sensitive to ground water flow. It measures a specific electrokinetic effect in the electrical field generated by ground water flow that is detectable at the ground surface, in boreholes and under water.

The electrokinetic phenomenon is manifest in a source current density associated with the flow of pore water through a mineralised porous material. An excess charge coats the surface of the mineral grains in the porous matrix, and displacement of this excess charge by the water induces a net current density. This sets up a convective current (from preferential fluid flow) through a hydraulic structure which will have an upstream negative electric field signature of a few millivolts (in the range 0 to 100mV). The electric field magnitude depends on the fluid flow velocity and electrical properties of material within the hydraulic installation. Detection of a negative electric potential on the upstream side indicates a leakage area.

However, the SP technique is also sensitive to other passive electrical sources that may be present in a hydraulic structure including EM disturbances from power lines, for example, and oxido-reduction effects caused by corrosion of metal. Oxido-reduction generates a negative electrical potential in the range 0 to 1000mV. This magnitude of corrosion signal can hide a leakage zone, but may provide useful insight in terms of locating metallic structures or identifying corrosion related deterioration.

The SP method, therefore, has been mostly used to detect preferential fluid flow in earth embankments and levees where interpretation of results is not complicated by the effects of metal or EM interference. Specialists at Fugro, however, have had significant success in applying this approach to more challenging masonry and concrete structures.

Single channel SP acquisition
A typical site set-up for SP equipment consists of one set of roving and one set of reference non-polarizing electrodes connected to a high input impedance acquisition board.

The mobile electrode is dragged to the upstream toe of a dam or dyke using a boat while the reference electrode remains fixed at the ground surface (along the dam or dyke crest, for example). The electrical contact between the roving and the reference electrodes is provided by a reinforced cable – extending several hundred meters depending on the target structure. The SP signal is recorded at a frequency of 200Hz averaged every second. The spatial coordinates of the water SP acquisition is taken with an onboard global positioning system (GPS) acquisition device. Figure 1 shows an example of single channel SP results from a concrete dam.

Innovative multi-channel waterborne system – SCAN Leak
Where the scale of structure warrants it, the standard single channel system can be upgraded to multi-channel acquisition. In the assessment of canals, for example, a multi-channel system allows total coverage of the channel bottom in a single pass and, consequently, better visualisation of potential water leakage in both longitudinal and transverse view.

To harness these benefits, Fugro has devised a multi-channel waterborne system, SCAN Leak. Initial testing of this innovative approach was successfully carried out investigating a non-reinforced concrete-lined canal – 12km long and 60m wide – in the South East of France (Figure 2a).

Site performance and results indicate this method could lead the way for fast and accurate large-scale surveys of dams and other significant water containment assets.

The SCAN Leak canal set-up involved one boat and two floats on each side equipped with GPS antennae to record boat and float positions in real time during the survey (Figure 2b). A total of 10 non-polarising electrodes spaced at 1m intervals were dragged 30m behind the boat, five electrodes on each side of the boat in order to cover a survey width of 10m. Self-potential sensors were positioned at the water/canal bottom interface.

To make differential electric potential measurement possible, the onboard acquisition system was linked to a reference electrode positioned on the canal crest. The reference electrode stayed fixed for each leg of the 12km survey which consisted of 24 SP scans in 500m long sections.

In the example shown in Figure 3, the SCAN Leak anomaly is around 100m long and 10m wide.

These large dimensions can be explained by the fact that the electrical field is propagated in all electrical conductive domains, such as the canal water and soil forming the canal body. The electrical field reaches its maximum magnitude at the exact leak position and decreases with distance according to the anomaly strength and the electrical conductivity of the water.

The SCAN Leak system covered the entire 12km length of canal in five days. Its accuracy and effectiveness was demonstrated by its capacity to detect and locate the small passive electrical signals from corroded pipes underneath the canal. Other negative SP anomalies were interpreted as potential inflow areas, enabling highly targeted further investigation and repair.

The project demonstrates the potential for wider application of SCAN Leak for dam surveys. As well as surveying a large structure in a short time, the technique is entirely non-invasive and can be used with no disruption to operations and no environmental impact. Repeat surveys before and after leak remediation can verify the effectiveness of the work while periodic surveys can determine whether the performance of a structure is in a steady state or worsening with time. As the equipment is modular and portable, it is suitable for cost-effective mobilisation to remote locations worldwide.

Other techniques
The use of SP techniques for dam surveys needs to be considered alongside the growing range of intrusive and non-intrusive investigation methods. In most cases, application of the SP technique will provide the best return if used in the early phases of any evaluation. Anomalies pinpointed by the survey can then be analysed and interpreted with the benefit of complementary data.

These might include input from installed structural and geo-monitoring systems and from other surveys at locations pinpointed by the SP. Commonly used second stage survey techniques include hydro-acoustic measurements of passive signals associated with water pressure disturbance and targeted boreholes or cone penetration tests with associated measurement of parameters such as subsurface temperature gradients.

Standard self-potential and SCAN Leak measurements can provide important information regarding the presence of leakage areas in hydraulic installations.

With the potential for integration with other non-invasive geophysical techniques, the approach provides an opportunity to plan both time- and cost-efficient solutions to leak detection for every scale of structure and with minimised intrusive investigation.



Alex Boleve is a project manager at Fugro GeoConsulting, France.

(1) Foster M, Fell R and Spannagle M 2000. ‘The statistics of embankment dam failures and accidents.’ Canadian Geotechnical Journal, vol 37, pp.1000-1024.

Reproduction of geophysical results in Figure 3 is by kind permission of Electricité de France (EdF).




Other developments in dam investigation
Embankments (dams and levees) have important life-safety consequences if they do not perform as designed. Both aging and new embankment structures require state of the science evaluation for engineering design, structural stability evaluation, or mitigation planning.
Advances in technology, particularly remotely-sensed data acquisition, allow more robust characterisations especially when integrated with traditional geoscience approaches. This integrated approach that synthesises multiple technology-driven data sets with traditional subsurface exploration provides a powerful tool to evaluate geological hazards or foundation concerns for embankments.
In the past, interpretation of aerial imagery (e.g. oblique low sun angle photos) and photogrammatically collected topography data coupled with limited geophysical investigations (e.g. seismic refraction) constituted state of the practice. Over the past decade, geological and geotechnical evaluations for embankments studies have become more effective and affordable by integration of sophisticated geophysical survey and high-resolution landscape data with ‘traditional’ geotechnical data.
Techniques such as LiDAR (light detection and ranging) and, for more remote areas, INSAR (satellite radar) provide detailed elevation data for developing surface models that are used to identify hazards such as slope instabilities or fault scarps, among other potentially adverse near-surface features.
Geophysical tools, such as electrical resistivity imaging and seismic reflection imaging, enable the engineering geophysicist to interpret the type and distribution of subsurface conditions. This information can be used as:
1) a framework to develop targeted subsurface exploration programs as opposed to traditional layouts, and
2) continuous datasets for interpreting ground conditions between discrete exploration locations, thereby reducing uncertainties.
These recent technological advances have not replaced traditional geological or geotechnical approaches. But they provide an important complement, enhancing the ability to interpret site conditions and reliably characterise potential hazards, thus improving structural reliance, performance, and safety while minimising risk.