Evaluation of a new coded GPR system for Non-Destructive Testing

This thesis focuses on Ground Penetrating Radar (GPR) as a tool for Non- Destructive Testing, which englobes the techniques that allow testing without permanently altering the conditions of structures. GPR main parameters are studied and it is compared the performance of two radars working with different transmission principles. The first one, MALÅ Professional Explorer, is based on pulsed signals while the second one, Ilmsens m:explore, is a new device based on M:sequences. The main purpose of GPR in the project is detecting metal rebars and cuts buried inside concrete. Therefore, the equipment mentioned will be tested for this use.
The main features that make a radar useful and informative for GPR are a high Signal to Noise Ratio (SNR) and good power along with a high bandwidth. The last assures a good resolution in time domain. For the first GPR equipment analysed, MALÅ, the power conditions are not optimum and the Peak to Average Power Ratio (PAPR) is not the best. This is caused by the fact that pulsed systems only transmit energy in the small fraction of time that the pulse is sent. For Ilmsens, the power criterion is met, since M-sequences consist of a series of rectangles with amplitude +1, -1, resulting in an average power equal to the peak one, and a resulting PAPR close to 1.
The experiments for this thesis are carried out in a specimen concrete block, reinforced with several layers of rebars in both directions. The end of these rebars is visible from the outside of the block along with a cut, enabling a comparison of the measured results with reality.
In the measurements done with MALÅ, raw scans do not show a clear view of the concrete specimen profile. Therefore, they are filtered with the software ReflexW. This highly improves the results, making visible the detection of at least the two higher layered of rebars. The results correspond to the real location of rebars in concrete.
For Ilmsens m:explore, since the device is an open system it needs to be characterised first, along with selecting the best antenna orientation. The tests made for this purpose show that a 90º shift between the two antennas used gives the highest SNR. Another interesting result is that the antennas used have an important impact on the system’s bandwidth, highly reducing it. This implies that the antennas used are not optimum.
Real experiments with Ilmsens in the specimen block are carried out with four different antenna orientations. For analysing the results, some manual filtering is applied and then the scans are plotted next to the real images from the specimen. The results show a good space localization of the reinforcements. However, the lowest layer is not visible and the first one is detected better in some orientations. Lastly, an upper view of the specimen is compared to the scans. This results in an exact match of all the rebars and the cut present in the block is also visible.
When comparing the results of both equipment, MALÅ ones have more depth content. However, this is caused by the processing applied with ReflexW. For Ilmsens there is no tool that allows to do the processing in such an easy manner and therefore a manual approach is needed. This means that a further work on filtering Ilmsens scans will improve them. Besides that, the comparison of both equipment is not actually fair. This is caused by the fact that they use different antennas, which transmit different powers and work at different frequencies, along with the previously mentioned difference in the processing.
As a conclusion for the thesis, it can be stated that Ilmsens m:explore equipment can improve the use of GPR for Non-Destructive Testing, but still needs further work for obtaining optimum results. Its main advantage is that it is an open system, which allows to control all of its features, from the antennas used to the filtering applied. Therefore, a further work in all these aspects will highly improve the results.