Application of response surface method to the prediction of TBM penetration rate

Performance prediction of the tunnel boring machine (TBM) is one of the crucial issues for estimating excavation costs and construction time of tunnel projects. TBM performance highly depends on an achieved penetration rate. The aim of this study is to develop TBM penetration rate prediction models using Response surface method (RSM) and then to compare the results obtained from various meta-heuristics optimization techniques including Differential Evolution (DE), Hybrid Harmony Search (HS-BFGS) and Grey Wolf Optimizer (GWO). To achieve this aim, the database uniaxial compressive strength (UCS), intact rock brittleness (BI), the angle between plane of weakness and TBM driven direction and distance between planes of weakness are assembled by collecting data from Queens water tunnel project. According to the results, it can be said that the proposed model is a useful and reliable means to predict TBM penetration rate provided that a suitable dataset exists. From the prediction results the squared correlation coefficient (R2) between the observed and predicted values of the proposed model was obtained 0.939, which shows a high conformity between predicted and actual penetration rate. The performance of different predictor models controlled by Mean Absolute Percentage Error (MAPE), Route Mean Square Error (RMSE), Variance Absolute Relative Error (VARE), Variance Account for (VAF) and Correlation Coefficient (CC). Response surface method based model with higher VAF and CC as well as lower MAPE, RMSE, VARE will show better performance.