Prediction of Maximum Settlement in EPB Mechanized Twin Tunneling by Using Supervised Combined Artificial Intelligence Model

In this paper, the effective parameters on settlement due to EPB twin tunnelling excavation including face sup-port pressure, back-fill grouting pressure, penetration rate, pitching angle, groundwater level, tunnel depth and soil characteristics (standard penetration test number, soil elastic modulus, dry density, internal friction and cohesion) belonging to a part of the Tabriz metro line 1 twin tunnels in a distance between Qunqa and Gazran stations was se-lected as the intelligent network input data's and designed to optimize artificial intelligence models using artificial neural network, fuzzy logic methods for predicting of maximum surface settlement. Comparing of obtained results with actual measured data's showed that despite of the ability of both artificial intelligence models on estimating the maximum settlements in mechanized excavation, it is still possible to precise the results by applying a combined arti-ficial intelligence model. Therefore, the output of two single models was used as the input of the Nero-fuzzy model and the obtained results (R2 = 0.77 and RMSE = 0.78) indicate a decrease of at least 27% RMSE compared to the in-dividual models.