Today, tunnel boring machines (TBMs) are widely used in construction projects and urban tunnels. These machines are divided into different types based on the type of working face maintenance system, excavated material transport systems and working modes. One of the most widely used TBMs in soft ground is Earth Pressure Balance (EPB) TBMs. Due to the increase in the price and the high costs of drilling, the eva‎luation of the performance of these machines is of particular importance. In this study, the advance rate parameter has been used to investigate and eva‎luate the performance of the excavator. Also, to check and eva‎luate the parameters needed to check the performance of the EPB-TBM in this research, the database, including the excavation information of the Isfahan metro line two project, which were drilled using the EPB-TBM, was used. Based on this database and the results of the collected studies, the most important influencing parameters during the drilling process on the advance rate include the Cutter-head rotation speed (CR), Cutter-head torque (CT), Thrust force (TF), Earth pressure balance average (EPBavg), Screw rotation speed (SR), Foam injection rate (FIR), which were used to build and develop intelligent models. Therefore, in order to predict the advance rate of the models, relying on data mining techniques including classification and regression tree algorithm (CART), support vector machine (SVM), artificial neural network (ANN) and random forest (RF) have been developed. It should be mentioned that the database used in this research contains 250 data items which were randomly divided into two training data sets and test data sets, which training data set (includes 90% of the total data i.e. 225 items) is used to build and develop the models, and the test data set (including 10% of the total data, i.e. 25 cases) is only used to eva‎luate the accuracy of the models. Therefore, relying on data-mining techniques using the training dataset, 4 intelligent models were developed to predict the advance rate of the EPB excavator. Then, in order to more closely examine and better eva‎luate the presented models to predict the progress rate of the statistical indicators, the coefficient of determination (R2), the normalized root mean square error (NRMSE) and variance account for (VAF) for each the model was calculated using the test data set. Then comparing the results of R2, VAF and NRMSE indicators showed that the model developed to predict the advance rate based on LS-SVM algorithm has more accuracy and less error than other models, that the calculated values of R2, VAF and NRMSE respectively It is equal to 0.898, 89.028 and 0.063. Finally, the comparison between the predicted and measured values of the advance rate showed that the presented models could predict the advance rate of the EPB-TBM with an error of less than 10%. It is worth mentioning that due to the high accuracy of advance rate prediction, the developed models can play a significant role as applied models in the mechanized drilling industry.

ابراهيم قاسمي ورنوسفادراني

خشايار باقرنيا

سعيد مهدوي , حسن طباطبائي