Evaluation of TBM performance prediction models and sensitivity analysis of input parameters

In recent years, anumber of models and formulas have been proposed to estimate the penetration rate of hard rock tunnel boring machines (TBMs). This study will focus on sensitivity analysis and examination of the effects of variation in the input parameters on the results of the commonly used performance prediction models for rock TBMs. The results of modeling were compared with the recorded machine performance to evaluate the validity and accuracy of these models. The sensitivity analysis of the four most common models indicates that uniaxial compressive strength followed by rock mass fracturing is among the most influential parameters used in these models. In addition, the result of our analysis shows that increasing the number of independent input variables does not increase the accuracy of prediction. The effect of discontinuity orientation in some models has also been considered. The results indicate that substantial changes in the angle between the discontinuity orientation and the tunnel axis have minor effects on penetration rate. The results of models were compared to recorded TBM penetration rate during excavation of Golab Tunnel, and the results indicate that the Gong model has the lowest error in comparison with other models. Grouping the rock types and using pertinent equations in the Farrokh model offered overall higher accuracy than other models in some structures of the Golab project. A brief set of recommendations are offered about the models and their applicability to various ground conditions.