Development of Performance Prediction Model for TBM Using Rock Mass Properties

Tunnelling projects carried out by Tunnel Boring Machine (TBM) present several problems, the most significant of which is predicting machine performance. The improper prediction may result in the rescheduling of the entire project and also lead to high cost overrun. An established rock mass classification system like rock mass rating (RMR) is commonly employed for the development of empirical equations to predict the performance of TBM. However, such TBM performance prediction models have a limited scope of success due to the weights assigned to the input parameters. This issue can be overcome by performing simple, multiple linear, and multiple non-linear regression analyses for determining the most influential input parameters or modifying the weights of the RMR system such that their role in the performance of the TBM is better represented. In this research, an attempt is made to define the influence of different rock mass properties on the performance of the TBM, and a novel model is generated using simple, multiple linear, and multiple non-linear regression analyses. This model uses intact rock mass properties i.e., uniaxial compressive strength (UCS), and fractured rock mass properties like joint spacing (Js), rock quality designation (RQD), and alpha (α) for performance prediction of TBM. This research comprises 119 datasets from three different TBM tunnels sites in Iran for statistical regression analysis. The results showed that the utilized rock mass properties have a great influence on TBM performance and concluded that the TBM performance can be estimated using the newly developed equation.