A convergence criterion for water conveyance tunnels

Abstract

One of the important methods used by tunnel construction engineers to examine the design, construction safety factor, and tunnel boring procedures is convergence measurement. Larger convergences may lead to safety problems such as rock destruction, boring machines jamming, delays in projects, or even geological disasters, in which case convergence measurement is considered the best method. The goal of this research is introducing a new convergence criterion for water conveyance tunnel based on the data from 31 convergence measurement stations of the Chehel Chay water conveyance tunnel. To achieve this objective, three different methods including statistical (SPSS), artificial neural network (ANN) and numerical modeling (3DEC) have been used. Observation points are defined on the tunnel wall in a realistic numerical modeling to compare the convergence results. The results revealed the success of the artificial neural network and indicated that the network managed to estimate the test data with a high correlation coefficient.

Application of Artificial Neural Network in Tunnel Engineering: A Systematic Review

“…It is concluded that the accuracy of the MARS method is lower than that of the MLP model. Zarei et al utilized SPSS and discrete element methods to introduce a new convergence criterion for water conveyance tunnel, and it comes out that the ANN is more suitable than the other two methods [69]. Note that the performance of different data mining methods and statistical methods varies depending on different data.…”

“…Nevertheless, the abovementioned conclusions can provide a reference in the tunnel engineering field. [21] WNN>ANN [22] GP>SVM>ANN [78] SVMFF>ANNFF [79] MLP>MRA [76] PSO-ANN>ANN [93] ANFIS>CFT>ANN [5] FLM>BPNN>MRA [20] ICA-ANN>ANN>LMRA [94] TL>RNN>SVM [49] PSO-SVR>PSO-BPNN>PSO-ELM [53] Stability BPNN>LRM [56] MLP>RBFNN [58] TBM performance BPNN>NMRA [59] ANN>SPSS [31] ANN>MRA [60] ANN>LMRA [30] ICA-ANN>PSO-ANN>ANN [61] SVM>ANN [39] PSO-ANN>ICA-ANN>ANN [24] Geological conditions ANN>XGBoost, CatBoost, RF, DT, SVR, KNN, BLR [18] Overbreak ANN>NMRA>LMRA [67] ANFIS>FLM>ANN>SVM>NMRA>L MRA [38] GA-ANN>ANN [68] ABC-ANN>ANN [40] ABC-ANN>ANN [11] Tunnel convergence MLP>RBFNN>MRA [36] SVM>ANN [35] MLP>MARS [69] MLP>SPSS [70] Rockburst and flying rocks PSO-ANN>ICA-ANN>GA-ANN Note: The '>' means the performance of the left model outperforms the right one.…”