Determination of the system behaviour based on data analysis of a hard rock shield TBM / Analyse der Maschinenparameter zur Erfassung des Systemverhaltens beim Hartgesteins‐Schildvortrieb

Abstract

Both the technical boundary conditions and the usual advance rates achieved during hard rock TBM shield drives severely limit the ability of the on-site-personnel to document the geological conditions and the system behaviour at appropriate intervals. A systematic and continuous short-term investigation of the rock mass conditions is performed on the construction lot KAT2 of the Koralm Tunnel (obtained by impact drillings and geophysical methods). The difference in scales and the fact that no continuous inspection of the ground conditions can be performed lead to a gap between the prognosis and the observed behaviour and a direct comparison is seldom possible. Based on the daily comparison between the observed behaviour and the analysed machine data, a state-of-the-art interpretation method has been developed. This method allows reliable conclusions regarding the primary aspects of the system behaviour and thus enables deductions on the geological/geotechnical conditions which have been encountered. The identification of the relevant parameters and suitable analysis methods has been performed in a "trial and error" manner: machine parameters deemed relevant in theory have been continuously compared with the observations on site and depending on the results, found usable or discarded. Reliable knowledge about the face stability, qualitative degree of fracturing of the rock mass, occurrence of over excavation in the cutter head area, the state of the annular gap and the blockiness of the rock mass can be deduced by applying this method. The results are used on daily basis by the on-site geologist and geotechnical engineer in order to verify the prognosis and recommend additional/auxiliary construction measures.

Machine Learning in tunnelling – Capabilities and challenges

“…With the aim of deriving the rock mass behaviour from the TBM operational data of the main tubes, efforts are underway to correlate the data from the exploratory tunnel with the encountered geology [29]. The TBM data comprises different features such as the specific penetration or the torque ratio (after [30]) as well as a corresponding classification of the rock mass behaviour ( Figure 5).…”

Application of artificial neural networks for Underground construction – Chances and challenges – Insights from the BBT exploratory tunnel Ahrental Pfons

“…torque, advance force) on a ten seconds basis. The fea tures advance pressure [bar], torque ratio after [9] and spe cific penetration [mm/rot/MN] ( Figure 1) were chosen as input for the ANN as the highest accuracies were achieved with them. The output data consists of a sitespecific rock mass behaviour classification system -called Geological Indication (GI) -with four classes: GI1 (green) = good rock mass; GI2 (yellow) = good rock mass, with unfavourable discontinuity intersections; GI3 (orange) = squeezing rock mass, high degree of fracturing, weakened rock; GI4 (red) = geotechnically relevant fault zones [2].…”

From the geological prognosis to the TBM drive: Planning the main construction lots of the Brenner Base Tunnel

“…• Visual assessment of the system behaviour at the face and the extrados, • Evaluation of specific machine parameters to add more detail to the visual records, according to [9] and [19] (Figure 3), • Evaluation of the pressure development in the crown shield to estimate the rock mass pressures, • Performance of daily laser scans in the A1 area (about 15 to 20 m scanned section) and in the A3 area behind the backup during longer tunnelling stoppages in order to record the displacements in all construction phases (Figure 4), • Refraction seismology to locate fault zones ahead of the machine, • Spot sample taking with the appropriate laboratory tests according to the state of the art.…”