Skip to main content

Capacitive Imaging for NDE


This is a research project, funded by the EPSRC-funded Research Centre for NDE (RCNDE). It is a targetted project, jointly funded by four Industrial Partners: Airbus, BP, Shell and National Nuclear Labs.

The general approach of this technique is to use a co-planar capacitive electrode to detect changes in the local electrical characteristics within a sample.A schematic diagram of the capacitive imaging approach is shown in Figure 1. The co-planar probe, which contains two ormore metal electrodes, generates an electric field distribution within the test material when an AC voltage is applied between the positive and negative electrodes. The presence of the sample under test will affect the resultant electric field pattern.


Figure 1: Schematic diagram of the capacitive imaging approach



Typical capacitive imaging probes are shown in Figure 2, fabricated by etching a printed circuit board (PCB) substrate.


(a) (b)

Figure 2: (a) Photograph of a pair of triangular electrodes, mounted in a shielded metallic container. The triangular electrodes in this example were each 20 mm wide in the horizontal direction. (b) A concentric electrode design


The probe was part of an instrumentation system that could be used for imaging, and a schematic diagram of the basic instrumentation is shown in Figure 3. To measure the signal at any particular location, a single frequency AC signal was applied as the driving voltage to one of the electrodes. The frequency of operation could be adjusted from 10kHz to 1MHz, noting that all the images below were obtained at a frequency of 1MHz. An increasedsensitivity was obtained by using a lock-in amplifier, which converts the AC voltage signal into a DC voltage proportional to the amplitude of the received AC signal.


Figure 3: System block diagram of Capacitive imaging system


Figure 4 (left) gives the results of a scan of a 21 mm thick Perspex plate containing 4 2mm square holes of different depth (15 mm, 11 mm, 7 mm and 3 mm), and also one (right) for a metal plate. The images below the photographs show the presence of the holes as the lighter areas, where there is greater contrast for the deeper holes.



Fig 4a fig 4b

Figure 4: Capacitive scans of the samples shown. The images below are from scans within the yellow dashed areas of each sample.


Finite Element (FE) modeling

It is important to be able to predict how various probe geometries and excitation frequencies would interact with samples of different conductivity. Thus, finite element modeling has been used to investigate this. The FE model using the COMSOL package can be used to study the effect of defects on the electric field distribution. As an example, Figure 5 shows the field predicted by FE models for a circular defect in a non-conducting medium. It can be seen that the present of the air-filled defect causes a distortion of the field lines, and this would lead to a detectable change in signal.

fig 5a
fig 5b


Figure 5: Field lines inside an insulating specimen with (a) no defect and (b) containing a defect.



1. Corrosion under insulators


Figure 6: (a) Photograph of slightly corroded steel sample. (b) Capacitive image of the same sample, with corroded areas indicated.


2. Composite samples containing impact damage

Carbon fibre and glass fibre samples with impact damage were kindly provided by Airbus for use in this project. The sample below is a carbon fibre sample containing impact damage. This was scanned using a triangular electrode pair across the highlighted area to form a capacitive imaging scan. In addition, the sample was further quantified using an ultrasonic scan over a similar area. Note that water immersion would have changed the electrical characteristics, and hence it was decided to use an air-coupled ultrasonic scan. The results of both experiments are shown side-by-side in Figure 7. It can be seen that the air-coupled ultrasound scan, performed in through-transmission, gives a well-defined area of damage. The damaged are predicted by capacitive imaging is similar in extent, although some details differ. It is thought that this is due to the fact that through-transmission ultrasound provided that accumulated damage through the whole thickness, whereas capacitive imaging in this electrically-conducting sample will produce more detail of surface damage in comparison to that deeper within the sample.

fig 7 capacitive image scan Fig 7 air coupled ultrasound
(a) Capacitive imaging scan (b) Air-coupled ultrasound scan of the same area

Figure7: Top: photographs of both sides of the Airbus carbon fibre sample containing impact damage. The area highlighted in yellow was scanned experimentally. Bottom: results of both (a) capacitive imaging and (b) air-coupled ultrasound scans on the impact-damaged carbon fibre composite panel. .


3. Inspection of concrete


A sample containing multiple rebars is shown in Figure 8. It was of 1000mm (length) x 1000 mm (width) x 150 mm (depth), with two sets of rebars of 10 mm diameter passing through the concrete slab and parallel to the top and bottom surfaces. The first set of rebars (two parallel rebars with a 300mm spacing between their centres) was buried at a depth of 100mm from the top surface to their centres, and the second set (also two parallel rebars with a 300mm spacing between the centres), were oriented perpendicular to the first set were buried at a depth of 120 mm from the top surface to their centres. This sample provided a complex scanning environment, such as that likely to be met in practice, with sets of rebars at different depths and orientation. The rebars can be seen clearly as the darker areas in the image. Note that the two horizontal features are darker than the two vertical ones; this is because the former represent the two parallel rebars located closest to the scan surface.

image033.jpg image035.gif
(a) (b) (c)

Figure 8: Concrete sample with rebars (a) Photograph of sample, and (b) schematic diagram, (c) capacitive image.




G.G. Diamond, D.A. Hutchins, T.H. Gan, P. Purnell and K.K. Leong, “Single-sided capacitive imaging for NDT”, Insight 48, 724-730 (2006).