Road tunnel early cost estimates using multiple regression analysis

Among construction projects, the risk in the production of road tunnels is particularly high due to the inherent uncertainties of the underground conditions. Equally, tunnel costs are considerably higher as compared to those of normal surface roads. As published literature on early cost estimating tools for road tunnels is scarce, projects sponsors, decision makers and engineers lack reliable tools, especially during the important early conceptual and preliminary design stages, to establish the project budget. The problem is addressed in this paper aiming at developing early parametric estimating models for road tunnels based upon the application of multiple regression analysis on real-world constructed projects. Thirty-three road tunnels totaling forty-six Km in single bore length constructed for the Egnatia motorway in northern Greece were analyzed for this purpose. To do so, data related to the encountered geotechnical parameters of rock masses and the corresponding quantities of primary and permanent support were gathered and put into a database. This database coupled with statistical techniques was subsequently used to establish the correlation between geotechnical and construction parameters. Following this stage, tests were performed to validate the correlations produced. The analysis highlighted the most important parameters affecting construction cost.