Bundesanstalt für Materialforschung und -prüfung

International Symposium (NDT-CE 2003)

Non-Destructive Testing in Civil Engineering 2003
Start > Contributions >Lectures > Corrosion: Print

A multi-dipole mobile array for the non-destructive evaluation of pavement and concrete infrastructures: a feasability study.

Michel Chouteau, Ecole Polytechnique-CERCA, Montreal, Canada
Sébastien Vallières, Ecole Polytechnique, Montreal, Canada
Edgar Toe, Ecole Polytechnique, Montreal, Canada

Abstract

Concrete used for bridge decks, road pavements and other transportation facilities experience loss of integrity over time caused by poor initial quality, action of de-icing salts, temperature changes, fatigue and, above all, delamination caused by corrosion of reinforcement bars. Also poor support from the subgrade layer can cause significant problems to the concrete slab. Electrical resistivity is sensitive to those losses of integrity. Poor initial quality reduces compressive strength and causes low bulk electrical resistivity. Chlorides can also penetrate concrete and drastically reduce resistivity; in turn, lower resistivity accelerates corrosion of rebars. The method has been tested on a bridge deck in Montreal and compared with other geophysical and geotechnical data. The results were encouraging; however acquisition was very slow and time-consuming. Traffic must be closed to perform the survey. In order to fully make the method operational and flexible for NDE, we have investigated the development of a multi-dipole mobile array. A compromise between resolution, depth of investigation and dipole combination must be found.

Using 3D numerical modeling we first examine resistivity responses associated with the various problems mentioned above that affects concrete. In the case of rebars, differences occur if they are corroded or not and if cracking and/or delamination is taking place; resistivity anisotropy is also indicative of rebar distribution. The system is also very sensitive to fine cracks that develops over the concrete surface and to internal defects. The system can also acurately estimate asphalt thickness and characterize the condition of the subgrade layer supporting the concrete slab in case of paved roads. The system has been tested in the lab over concrete slabs simulating various conditions. Results suggest good agreement with numerical modelling data. Finally we suggest future development in data acquisition and interpretation with this NDT system.

Introduction

The use of the resistivity method for non-destructive testing (NDT) of concrete structures has been very limited until now. Cabasut et al (1997) have investigated loss of integrity of water pipes by conducting resistivity surveys inside. Usually the method is slow because of the physical contact of electrodes with the material under investigation. The contact may be difficult to achieve and when concrete is involved special techniques must be use to penetrate the material. Also engineers will rely more on mechanical properties (seismic, sonic methods) than electrical resistivity to diagnose problems. On another hand, non-seismic methods like GPR would be much faster to perform surveys and generally resolve the geometry better. However, preliminary studies using numerical and scale modeling (Chouteau et al, 2002) demonstrated the potential of the method. A resistivity tomographic survey carried out in 1999 over an aging overpass in Montreal was able to successfully map quality of concrete and determine the position of rebars (Chouteau et al, 2000?). It was soon realized that for investigating highway concrete infrastructures the method was very slow and was confronted to high contact impedances in case of carbonated concrete surfaces and asphalt cover. Resistivity measurements had to be done in a continuous mode and fast. Capacitively-coupled resistivity systems would be the solution. Such systems already exits for a few years (Corim, from IRIS; OhmMapper from Geometrics, MPR-2R developed by University of Paris). The concepts can be found in Grard & Tabbagh (1991), Tabbagh et al (1993) and Panissod et al (1998). However they are not used for NDT of concrete and not adapted in anyway to those applications. It was therefore decided to study the performance of a electrostatic system designed especially to map various factors affecting integrity of concrete structures. As a compromise has to be found between depth of investigation and resolution we came with the concept of a wheel train of dipoles (figure 1) that could be used either in in-line or equatorial modes. The first dipole transmits a AC current (frequency in the kHz's). The remaining dipoles are receiver potential dipoles. The distances between the poles are from 5 cm to 20 cm. The total length is between 1,0 to 2,5 m.

Fig 1: the train of dipoles used to continuously measure resistivity either in in-line or equatorial modes.

It has been shown (Tabbagh et al, 1993) that for the quasi-static case and low induction number, the resistance measured is the same as would measured the DC resistivity method. Therefore in the following presentation all results have been obtained from 3D DC resistivity modeling.

There are many factors that could affect integrity of concrete and reduce its compressive strength: initial composition, corrosion of rebars leading to cracking, scaling, delamination. The compressive strength is dependant on the water/cement ratio. The higher the ratio the lower is the strength. Electrical resistivity of concrete decreases almost linearly with increasing water/cement ratio for a given cement content (Neville, 1998) and therefore resistivity can be diagnostic of the compressive strength. Electric current is conducted through moist concrete by ions (electrolytic mechanism). Figure 2 shows the resistivity response with curing time for three concrete slabs with two differing water/cement ratio and compressive strength.

Fig 2: Resistivity as function of curing time for three concrete slabs; B30-05 and B30-10 have the same compressive strength (24.3 MPa) and water/cement ratio (0.46); B15-10 has a compressive strength of 9.2 MPa and a ratio of 0.80. The Wenner spread is used.

Progress of corrosion of steel reinforcement is controlled by the electrical resistivity of concrete and therefore it is better to maintain high resistivities within the concrete. Addition of chloride ions originating from de-icing salts for example will decrease electrical resistivity and accelerate the generation of corrosion. Studies have shown that for resistivities lower than 100 Wm risk of corrosion are high while it is negligible for values higher than 1000 Wm.

Experiments have shown that bulk electrical resistivity of concrete is dependant upon its integrity (Cabasut et al, 1997). Damages can vary from cracking, spalling and delamination induced by steel reinforcement corrosion to scaling caused by weathering (frost-thaw). Presence of water or not in the cracks will modify the bulk resistivity of the material.

Therefore, loss of integrity, chloride ions and compressive strength will affect resistivity of concrete and a method based on the mapping of the resistivity distribution will help diagnose defects in concrete structures.

Numerical modelling and imaging

In order to provide diagnostic information about the concrete structure the method must be sensitive to change in concrete integrity. Here we will show how the method responds to concrete slabs and reinforced bars, corrosion, de-icing salts, cracks, defects, asphalt thickness and also how well theybe imaged.

Concrete slab and rebars
Fig 3: response to a thin (20 cm) and thick concrete slab and a thin 20 cm slab with a level of rebars.
A 300 Wm concrete slab is 20 cm thick and may include a row of rebars at a depth of 7 cm. Figure 3 compares apparent resistivities ra measured with the in-line train of dipoles over the slab (air below) without and with rebars and over a very thick (>>20 cm) slab without rebars. The method is sensitive to the thickness and to the presence or not of rebars, these inclusions been displayed as a conductive layer between to more resistive zones.

Using 2D resistivity inversion for the modelled data, the resistivity of concrete and the depth to the rebars are well resolved Figure 4). The rebars show up as a conductive layer. Below, resistivity rises caused by the resistivity of concrete and then the air.

Severe rebar corrosion may cause delamination of concrete. Then an air filled resistive horizontal crack develops at the rebar level that shows up as a resistive layer in the interpreted model (figure 5) compared to a conductive layer when no delamination takes place (figure 4).

Fig 4: (top) pseudo section of measured ra, (middle) pseudo section of computed ra from the resistivity model (bottom). True resistivity of the concrete and depth to the rebars are well determined. Colour scale is in log10(ra).
Fig 5: in presence of delamination the resistivity section (bottom) shows a resistive zone at the depth of the rebars. Again the resistivity of concrete is well resolved.

Cracks
The method is very sensitive when measurements are done across lateral discontinuities Figure 6 shows the response to a thin vertical crack extending from the surface down to a given depth (1 to 6 cm). The crack is located at x = 0,5 m and the in-line train is used with the transmitter centered on x = 0. The change is resistivity at the location of the crack is large and increases with depth extension. The resistivity method is therefore an excellent crack detector; detecting small cracks at the surface is important since they represent conduits for water and salts to the level of rebars.

Fig 6: response to a thin crack of variable depth on top of the concrete. Slab.

Defects
In figure 7 the train of dipoles (in-line) moves over a small cavity 5 cm x 5 cm located at a depth of 5 cm within the concrete slab. A small positive bump located over the cavity moves on top of the "normal" response of the slab in air as the system moves. Inverting these data yields the image of the resistivity model at the bottom of figure 7. The cavity can be well detected and located.

Fig 7: (top) the train of dipoles moves across the crack on top of the concrete slab; (bottom) shows from top to bottom the pseudosection of measured ra, the pseudosection of computed ra and the interpreted resistivity model. The latter displays the cavity as resistive anomaly located at the right location and depth.

Heterogeneous subgrade layer
Slabs used for concrete paved highways are not reinforced but they are lying flat on top of a compacted subgrade layer. Water infiltrationand cycles of frost and thawn cause modifications to the subgrade layer (deformation, heterogeneities, etc). In turn those modifications lead to poor support of the concrete slab and may cause cracking. Modeling shows that a train of dipoles long enough (10 times the slab thickness) can detect those variations in the filling below the slab.

Asphalt thickness
Assessing asphalt thickness is one of the most important parameter required by departments of transportation. It is often estimated by coring the paved roads at a pre-established distance interval. This desructive technique suffers from undersampling and now GPR is increasingly used to map asphalt thickness. Figure 8 shows the reponse of the method to an asphalt layer of varying thickness (5 to 20 cm) overlying a concrete 20 cm slab. The train of dipoles is 2 m long. Asphalt is a highly resistive material (~ 5000 Wm). The method shows that it is very sensitive to asphalt thickness and that thickness and resistivity of concrete can be resolved from the modelled data (figure 9). Here asphalt is 15 cm thick.

Fig 8: ra response from a train of dipoles over an asphalt layer overlying a concrete slab.

Fig 9: resistivity images of the observed data, of the computed data and of the interpreted model from data collected over a 15 cm-thick aphalt overlying a 20 cm concrete slab.

Configurations
All modellings were done with two configurations, the train of in-line dipoles (TDL) and the train of equatorial dipoles (TDE). Results show that in general the TDL configuration is sensitive to lateral variations and therefore recommended for detecting cracks. The TDE configuration is weakly sensitive to lateral variations but highly sensitive to the vertical variations (material thickness, delamination, asphalt thickness). One or/and the other configuration may be used depending on the problem to be solved.

Testing on concrete slab models
The Jacques-Cartier Bridge is an old structure, vital to the traffic between the island of Montreal and the south shore. In order to reduce traffic annoyance, the concrete deck of the bridge is cut and replaced gradually by prefabricated reinforced concrete slabs. Those slabs are made by a company near the south end of the bridge. The geometry of the reinforcement framework, the geometry of the slab and the composition and mechanical properties of the concrete are controlled and well known. To test the method we have made measurements over one of the prefabricated slabs using various dipole-dipole combinations. Data were collected over three lines separated by 20 cm; electrode separation was 5 cm for 21 electrodes on each line. A 1.5 GHz GPR survey was also performed over the same investigated zone to correlate electrical properties with the resistivity tomography data and to precisely position the rebars with regard to the resistivity survey. Data inversions line by line or all together clearly delineate the first level of rebars and even each rod has a marked anomaly associated with it. The resistivity of the concrete is also well resolved and shows, as anticipated, values (~ 1000 Wm) for a good quality concrete.

Fig 10: (top) resistivity pseudo-section for Line 1 over a Jacques-Cartier concrete slab.; (bottom) interpreted resistivity model showing high concrete resistivity (~ 1000 W.m) for the first 4 cm and then low resistvivity (blue) caused by the presence of a rebar level.

Conclusion

Bulk electrical resistivity of concrete is dependant on (1) the intrinsic quality of concrete (composition, compressive strength), (2) the condition of surfacial/volumic deterioration caused by cracking, delamination and salts, and (3) the distribution of reinforcement bars within the concrete structure. Using numerical modeling and imaging, we have shown that the resistivity method can be used to map corrosion potential (resistivity of concrete), vertical cracks, delamination, defects and asphalt thickness. helping diagnose possible damage. Experiments on bridge decks and concrete slabs show that the method is applicable and could help diagnose the state of concrete structures; however a new generation of NDT equipment needs to be developed to expedite surveys over transportation structures.

References

  1. Cabasut, L., Bourget, M., Frappa, M., Al Heib, M., Viguier, C. and Guerin, J.F., 1997, Contribution of electrical and acoustic methods in the evaluation of the condition of accessible underground pipe work and environment, in McCann, Eddleston, Fenning and Reeves (eds), 1997, Modern geophysics in Engineering Geology, Geological Society Engineering Geology Special Publication No. 12, pp. 193-200.
  2. Chouteau, M. and Beaulieu, S., 2002: An investigation on application of the electrical resistivity tomography method to concrete structures; Geophysics 2002, Los Angeles, CA, April 15-19, 2002
  3. Chouteau, M., Beaulieu, S., Rucareanu, M., Leblanc, J.-M., 2000: développement et validation de la tomographie électrique comme méthode d'évaluation non-destructive de l'état du béton; application au viaduc Dickson; rapport final, GPR International, École Polytechnique, 11 novembre 2000, 64 p.
  4. Grard R., Tabbagh A., 1991, A mobile four-electrode array and its application to the electrical survey of planetary grounds at shallow depths, J.G.R , 96, B-3, 4117-4123.
  5. Handbook on Nondestructive Testing of Concrete; ed. V.M. Malhotra and N.J. Carino; CRC Press, 343 p.
  6. Loke, M.H., and Barker, R.D., 1996a, Rapid least-squares inversion of apparent resistivity pseudosections by a quasi-Newton method. Geophysical Prospecting, 44, 131-152.
  7. Loke, M.H., and Barker, R.D., 1996b, Practical techniques for 3D resistivity surveys and data inversion; Geophysical Prospecting, 44, 499-523.
  8. Neville, A.M., 1998, Properties of concrete, 4th edition, Addison Wesley Longman, 844 p.
  9. Panissod C., Dabas M., Hesse A., Jolivet A., Tabbagh J. & Tabbagh A., 1998, Recent developments in shallow electrical and electrostatic prospecting using mobile arrays. Geophysics, 63-5, 1542-1550.
  10. Tabbagh A., Hesse A. & Grard R., 1993, Determination of electrical properties of the ground shallow depth with an electrostatic quadrupole: field trials on archaeological sites. Geophys. Prospect. 41-4, 579-597.
STARTPublisher: DGfZPPrograming: NDT.net