High Performance Ground Prediction Ahead of TBMs - The NeTTUN System “TULIPS”

"STATE OF THE ART The current situation of ground prediction in TBM is of great concern for tunnelling contractors, and the subject has been quite widely addressed over recent years, for instance by the French Eupalinos 2000 programme, which tested a variety of technological approaches. A number of interesting projects were executed, such as the SOUND and OnSITE projects, part of the German Geotechnologien programme: these projects are focused on seismics, while other techniques such as radar are not considered. Little information and even less industrial applications have been released apart from lab simulations and numerical modelling. A number of field tests in different configurations have been carried out for compression waves, but the results were often not reliable or convincing, and therefore little has been reported in open literature. The advances seem somewhat promising as far as hard rock tunnels are concerned, but ignore the problem raised by concrete lined tunnels such as those bored by a TBM in soft ground. In such tunnels, sending and receiving seismic waves reliably through the lining (and the grouting behind) is quite illusory. A careful analysis of the seismic prediction research results show that the main problem is the source which is poorly controlled and poorly coupled (for instance in concrete lined tunnels, or in approaches that use the TBM cutter head vibrations as a source). Generating waves perpendicular to the tunnel axis is also a problem: aligning with the tunnel axis makes more sense. Radar (Ground Penetrating Radar - GPR) has been used in geotechnical surveys for decades, and also for tunnel inspection. Attempts have been made at using it on a TBM (for example in Japan by Taisei) but mainly using off-the-shelf GPRs, which are not adapted to work in the specific environment of a TBM (because of its huge mass of metal, to start with). Very few results have been published, probably due to the somewhat disappointing performance of such implementations. At present, no TBM has a GPR incorporated in its cutter head that is able to predict ground characteristics while drilling. Another characteristic and drawback of most current methods is that they require dedicated periods of time during which other tunnel construction activities must be stopped. Due to the duration of such operations this leads to performing ground prediction on a weekly basis only, and therefore to a requirement for an investigation range of 100 metres approximately (in order to cover the TBM advance over the coming week). This is not achievable with boreholes made from the TBM, even less if a reasonable accuracy is required, as is usually the case."