Monitoring the vibration response of a tunnel boring machine: Application to real time boulder detection

This thesis is a continuation of Bryan Walter's dissertation work (2013), which analyzed ambient vibration of a tunnel boring machine (TBM) and proved this could be used to detect boulders. This work seeks to characterize vibration of a TBM consistent with boulder impacts. A network of accelerometers was placed on a TBM in Seattle to record the vibration response to tested impacts and machine operation (cutterhead rotation). This network was then secured in place for continuous monitoring during tunneling. The investigation was driven by the development of a boulder detection algorithm that would operate in real time. Other key goals included characterizing response to both known and unknown impacts and frequencies occurring during ambient vibration, which were intended to advance general knowledge of TBM vibration while improving the robustness of the boulder detection algorithm.;Several interesting results have been obtained from this study. Vibration is consistently best transferred in the longitudinal direction (along the axis of the TBM). Many high amplitude frequency components exist during ambient vibration, many of which were found proportional to cutterhead rotation rate by Walter (2013). A majority of these frequencies have been matched to mesh frequencies of the planetary gearboxes that drive cutterhead rotation. Knowledge of these high amplitude frequencies has allowed the development of a single boulder detection variable (BDV) which compiles the high amplitude frequencies present during an impact while avoiding the high amplitude frequencies present during ambient vibration.;Although this study is limited by lack of ground truth data (where cobbles and boulders are located), the presence of a cobble-heavy soil section in the path of the TBM allowed for calibration of the boulder detection algorithm and more reliable impact detection. No significant trends were found between the magnitude of BDV-detected impacts and the frequency response functions. However, BDV impacts were extremely well correlated to till-like deposits, where many cobbles were found exiting the TBM. No trends were found relating BDV magnitude to cutterhead rotation rate or pressure inside the TBM excavation chamber. Further investigation into ambient vibration is suggested to better characterize the relationships between vibration, operating parameters, and ground conditions, and advanced modeling of TBM components is recommended to bolster general understanding of TBM vibration.