Hanan Samadi - MSc., School of Geology, College of Science, University of Tehran, Tehran, Iran,
Jafar Hassanpour - Associate Professor, School of Geology, College of Science, University of Tehran, Tehran, Iran,

The need to predict geotechnical parameters in soft ground is highly essential in evaluating the performance parameters of earth pressure balance machine (EPB-TBM) and ensuring the safety and efficiency of the tunnel boring machine during tunnel construction. In this study, several EPB operating parameters, including cutterhead torque, thrust force, chamber pressure, and the cutterhead speed revolution (RPM) were taken into account to estimate the geotechnical parameters such as friction angle (φ) and shear strength (τ) for engineering geological units (ET۱ to ET۵) in the tunnel route which are indicator units for the whole path. A soft computing technique called Support Vector Machine (Ls-SVM) was trained with EPB operating parameters and geotechnical specifications obtained from Tehran metro line ۶ -southern extension sector (TML-SE۶) and the East-west lot of line ۷, Tehran metro project (TML-EW۷). A comprehensive dataset consists of borehole logging results along the tunnel path was collected and ۸۵% of the data for training were randomly selected, while the remaining were considered for model testing. For the purpose of assessing the performance of the applied method and for evaluating the accuracy and precision, several loss functions, including mean absolute deviation (MAD), mean square error (MSE), mean absolute percentage error (MAPE), and relative absolute error (RAE) were brought on stream. The results of the proposed models indicate an acceptable and reliable accuracy of the approaches. Comparison of machine learning and multiple variable regression (MVR) method with measured data, it can be seen that the deviation intervals of the predicted values are small and the prediction values resulted from support vector machine (SVM) and MVR is in good agreement with the measured values of geotechnical parameters. In other words, the mentioned models, especially SVM, can predict the geotechnical parameters at the tunnel face. Hence, it can be applied for the accurate prediction of geotechnical parameters based on EPB operating data.

LS-SVM, GEOTECHNICAL PARAMETERS, BACK ANALYSIS, EPB-TBM, URBAN TUNNELLING