Elsevier

Engineering Geology

Volume 254, 2 May 2019, Pages 89-101
Engineering Geology

Induced polarization tomography applied to the detection and the monitoring of leaks in embankments

https://doi.org/10.1016/j.enggeo.2019.04.001Get rights and content

Highlights

  • There is a growing need for the detection of leaks in earth embankments and dikes.

  • An experiment is performed at a well-controlled test site.

  • The ability of induced polarization to detect and monitor leaks is demonstrated.

Abstract

During an induced polarization survey, both electrical conductivity and chargeability can be imaged. Recent petrophysical models have been developed to provide a consistent picture of these two parameters in terms of water and clay contents of soils. We test the ability of this method at a test site in which a controlled artificial leakage can be generated in an embankment surrounding an experimental basin. 3D tomography of the conductivity and normalized chargeability are performed during such a controlled leakage. Conductivity and induced polarization measurements were also performed on a core sample from the site. The sample was also characterized in terms of porosity and cation exchange capacity. Combining the 3D survey and these laboratory measurements, a 3D tomogram of the relative variation in water content (before leakage and during leakage) was estimated. It clearly exhibits the ground water flow path through the embankment from the outlet of the tube used to generate the leak to the bottom of the embankment. In addition, a self-potential survey was performed over the zone of leakage. This survey evidences also the projection of the ground water flow path over the ground surface. Both methods are found to provide a consistent picture. A 2.5D time lapse tomography of the electrical conductivity and normalized chargeability was also performed and evidences the position of the preferential flow paths below the profile. These results confirm the ability and efficiency of induced polarization to provide reliable information pertaining to the detection of leakages in dams and embankments.

Introduction

Embankment dams and dikes can be weakened by internal erosion and suffusion phenomena due to preferential flow paths and this on-going weakening can cause their failure (e.g., Foster et al., 2000; Peyras et al., 2008). Early warning associated with the concentration of flow paths in an embankment is therefore an important task to prevent its failure. On one hand, traditional geotechnical technics (e.g., cone penetration test and/or standard penetration tests) have limitations related to cost-effectiveness and the perturbation on the structures themselves (e.g., Fauchard and Mériaux, 2007; Cardarelli et al., 2014). Temperature measurements using fiber optics can be used to evaluate leakages (Beck et al., 2010; Khan et al., 2014), but such approaches are unfortunately intrusive. On the other hand, geophysical methods can provide fast and non-intrusive 3D and 4D tomograms of leakages using a variety of techniques (e.g., Rittgers et al., 2013; Ikard et al., 2015).

Different geophysical methods can be used to detect preferential flow paths in embankment dams and dikes including ground penetrating radar, passive and active seismic methods (Himi et al., 2018), and geoelectrical (galvanometric) methods including the self-potential method (Al-Saigh et al., 1994) and electrical conductivity and induced polarization methods (e.g., Martínez-Moreno et al., 2018). Ground Penetrating Radar (GPR) constitutes a fast and reliable method for shallow investigations. However, the so-called skin depth of GPR (i.e., the depth of penetration of the electromagnetic waves at a given frequency) can be very small in conductive media, such as, for instance, in presence of clays (Di Prinzio et al., 2010). Geoelectrical (galvanometric) methods (electrical conductivity, induced polarization, and self-potential) do not suffer such a limitation (Mendonça, 2008). A detailed description of the electrical conductivity and induced polarization method can be found for instance in Binley and Kemna (2005) and Revil et al. (2012). In the case of dikes and dams, electrical conductivity tomography is known to provide important information for the assessment of preferential flow paths (Perri et al., 2014; Cardarelli et al., 2014; Fargier et al., 2014). However electrical conductivity tomography is related to the water content and not to the flow of the ground water. In addition, electrical conductivity can be hardly used as a stand-alone technique. Indeed, two contributions control the electrical conductivity of porous soils. One associated with the conduction in the bulk pore space and one associated with conduction in the electrical double layer coating the surface of the grains. This second contribution is called surface conductivity and is especially strong in clay-rich materials and/or at low pore water salinity even for clean sands (Revil et al., 2014). Electrical conductivity tomography cannot be used to separate the bulk conductivity from the surface conductivity. This point is crucial as discussed below since bulk and surface conductivities have very different dependencies with the pore water content. Induced polarization can be used to separate the two contributions of electrical conductivity. Before discussing induced polarization in more details, it is worth mentioning another technique called magneto-resistivity, which can be used to track preferential flow paths as discussed in details in Jessop et al. (2018). Another method, directly related to the flow of the ground water, is the self-potential technique (Lapenna et al., 2000; Bolève et al., 2009; Revil et al., 2005). This method is a passive geoelectrical method in which the flow of water generates its own source current distribution, the streaming current, which in turn generates an electrical potential anomaly at the ground surface. This electrical potential anomaly can be sampled at the ground surface with a pair of non-polarizing electrodes (one used as a reference) and high input-impedance voltmeter. This method has been broadly used for the detection of seepages in embankments (e.g., Nzumotcha-Tchoumkam et al., 2010). However other sources of current exist in the subsurface (e.g., associated with the corrosion of metallic bars and ores, Mendonça, 2008) and can make the interpretation of self-potential signals more difficult (Revil et al., 2012).

We are interested to use the induced polarization method to detect leaks. Induced polarization measures the ability of rocks and soils to store reversibly electrical charges under the influence of an external electrical field and the relaxation time required by these charges to come back to equilibrium once the applied electrical field is suppressed. Induced polarization has a very long history in geophysics with early development done for ore body prospection (Vinegar and Waxman, 1984; Titov et al., 2010). Great progresses have been recently done regarding the underlying petrophysics of induced polarization and a model called the dynamic Stern layer polarization model seems to explain laboratory data for a broad range of porous media and environmental conditions (see details in Revil and Florsch, 2010; Revil, 2013). In parallel, Soueid Ahmed et al. (2018) recently developed an inversion code (ECT-3D) to invert electrical conductivity tomography with complex topography. This work was completed by Soueid Ahmed and Revil (2018) who developed a 3D joint inversion package for electrical conductivity and induced polarization tomography.

In this paper, we develop a field experiment over an experimental basin using induced polarization tomography. The aim of this study is to provide a way to analyze the advantage of using induced polarization to detect leaks in such infrastructures. The key questions we address are the following (1) Can we use a time-lapse induced polarization survey to determine soil moisture changes associated with an ongoing leakage? (2) Can we combine induced polarization tomography and the dynamic Stern layer model to image the change of the water content in an embankment?

Section snippets

Principle of the measurements

Induced polarization investigates the ability of porous materials to store reversibly electric charges under the action of an external (primary) electrical field (Vinegar and Waxman, 1984). Induced polarization measurements can be performed in time-domain or frequency domain (the so-called spectral induced polarization method). In time-domain induced polarization, we measure the secondary voltage decay after the primary current (and primary electrical field) is shut down (Fig. 1). In

Site description

Our study was conducted over an experimental basin (Fig. 2) developed by IRSTEA (National Research Institute of Science and Technology for Environment and Agriculture) and located in the vicinity of Aix-en-Provence, in the Southern part of France. The dimension of the basin is 22 m of length and 10 m of width. The inner side of the basin filled with water is covered by a geomembrane. The geomembrane is protected by a geotextile and pavement on the inner sides. The embankment consists of

Conductivity and induced polarization survey

A total of 7 profiles parallel to each other were acquired in both conductivity and chargeability using an ABEM Terrameter SAS-4000 (ABEM Lund Imaging System) resistivity meter. The position of the profiles (labeled P1 to P7,) is shown in Fig. 2b. Each profile contains therefore 2 sets of 32 electrodes (one set to inject the current and one set to measure the potential, as shown in Fig. 1b). These two lines are separated by a distance of ~20 cm. Along each profile, the spacing between the

3D tomography

The two 3D tomograms of the electrical conductivity and chargeability before and during the leak are shown in Fig. 8, Fig. 9. From these two tomograms, we can compute the normalized chargeability tomogram by multiplying, cell by cell, the conductivity by the chargeability. The normalized chargeability tomogram is shown in Fig. 10.

The three tomograms show that the structure of the embankment is quite homogenous in its center. The electrical conductivity tomogram shows that the embankment is

Conclusion

Electrical conductivity of porous rocks and soils has two contributions. A bulk contribution associated with current flow through the pore network of the medium plus a surface conductivity associated with conduction in the electrical double layer coating the surface of the grains. Electrical resistivity tomography alone cannot allow separating these two contributions, which have different dependencies with the water content of the porous material. Induced polarization can be used to map another

Acknowledgements

This work was supported by the project RESBA ALCOTRA funded by the European Community. The postdoc of Abdellahi Soueid Ahmed is funded by EDF through a contract with the CNRS. We thank the Editor and one anonymous referee for their useful comments.

References (46)

  • A. Soueid Ahmed et al.

    3D electrical conductivity tomography of volcanoes

    J. Volcanol. Geotherm. Res.

    (2018)
  • H.N. Al-Saigh et al.

    Detection of water leakage from dams by self-potential method

    Eng. Geol.

    (1994)
  • G.E. Archie

    The Electrical Resistivity Log as an Aid in Determining some Reservoir Characteristics

    (1942)
  • Y.L. Beck et al.

    Thermal monitoring of embankment dams by fiber optics

  • A. Binley et al.

    DC resistivity and induced polarization methods

  • A. Bolève et al.

    Preferential fluid flow pathways in embankment dams imaged by self-potential tomography

    Near Surf. Geophys.

    (2009)
  • H. Ciesielski et al.

    Determination of cation exchange capacity and exchangeable cations in soils by means of cobalt hexamine trichloride. Effects of experimental conditions

    Agronomie

    (1997)
  • T. Dahlin et al.

    Improvement in time-domain induced polarization data quality with multi-electrode systems by separating current and potential cables

    Near Surf. Geophys.

    (2012)
  • C. Fauchard et al.

    Geophysical and geotechnical methods for diagnosing flood protection dikes: guide for implementation and interpretation

  • M. Foster et al.

    The statistics of embankment dam failures and accidents

    Can. Geotech. J.

    (2000)
  • S.J. Ikard et al.

    Geophysical investigation of seepage Beneath an Earthen Dam

    Groundwater

    (2015)
  • M. Jessop et al.

    Magnetometric resistivity: a new approach and its application to the detection of preferential flow paths in mine waste rock dumps

    Geophys. J. Int.

    (2018)
  • M. Karaoulis et al.

    Time-lapse 3D inversion of complex conductivity data using an active time constrained (ATC) approach

    Geophys. J. Int.

    (2011)
  • Cited by (0)

    View full text