Development of hybrid intelligent models for predicting TBM penetration rate in hard rock condition

Abstract

The aim of this research is to develop new intelligent prediction models for estimating the tunnel boring machine performance (TBM) by means of the rate pf penetration (PR). To obtain this aim, the Pahang-Selangor Raw Water Transfer (PSRWT) tunnel in Malaysia was investigated and the data collected along the tunnel and generated in the laboratory via rock tests to be used for the proposed models. In order to develop relevant models, rock properties including uniaxial compressive strength (UCS), Brazilian tensile strength (BTS), rock quality designation (RQD), rock mass rating (RMR), weathering zone (WZ), and also machine parameters including thrust force (TF) and revolution per minute (RPM) were obtained and then, the dataset composed of both rock and machine parameters were established. After that, using the established database consisting of 1286 datasets, two hybrid intelligent systems namely particle swarm optimization (PSO)-artificial neural network (ANN) and imperialism competitive algorithm (ICA)-ANN and also simple ANN model were developed for predicting the TBM penetration rate. Further, developed models were compared and the best model was chosen among them. To compare the obtained results from the models, several performance indices i.e. coefficient of determination (R2), root mean square error (RMSE) and variance account for (VAF) were computed. It is found that the hybrid models including ICA-ANN and PSO-ANN having determination coefficients of 0.912 and 0.905 respectively for testing data as that of the simple ANN model are 0.666. More, the RMSE (0.034; 0.035) and VAF (90.338; 91.194) of hybrid models are also higher than these of simple ANN model (0.071; 66.148) respectively. Concluding remark is that the hybrid intelligent models are superior in comparison with simple ANN technique.