A Comparative Study of Artificial Intelligence Techniques to Estimate TBM Performance in Various Weathering Zones

Abstract

AbstractThis study aims to propose a practical intelligence way for the prediction of tunnel boring machine (TBM) performance in various weathering zones. To do this, after reviewing the available literature, the data collected from the tunnel site and doing laboratory investigations, five important parameters, i.e., rock mass rating, Brazilian tensile strength, weathering zone, cutter head thrust force, and revolution per minute, were set as model inputs to predict penetration rate (PR) of TBM. Then, two intelligence techniques, namely, group method of data handling (GMDH) and artificial neural network (ANN) were applied to the collected data (i.e., 202 data samples). In developing these intelligence techniques, a series of parametric studies were conducted on the most important parameters of these techniques. After developing GMDH and ANN models, some important performance indices were selected and calculated to select the best one among them. It was found that the GMDH model receives a higher accuracy level compared to the ANN model. It can be established that the GMDH is an applicable and powerful technique in the area of TBM and tunnelling technology.KeywordsGroup method of data handlingTunnel boring machineTunnelling technologyArtificial neural networkPredictive model