Motivated by the goal of generating numerical methods for the simulation-based steering of tunnel boring machines based on meta models, an artificial neural network (ANN) is presented in conjunction with particle swarm optimization (PSO), denoted as ANN-PSO to substitute computationally far more expensive three-dimensional numerical simulations of the tunneling process for the purpose of real-time predictions of surface settlements, parameter identification and process optimization. A comprehensive, process-oriented simulation model for mechanized tunneling is used to provide a rich data base for the training of the meta model. In order to systematically generate data, this numerical model is integrated in an automatic data generator, for setting up, running and postprocessing the numerical simulations for a prescribed range of parameters. This data generator provides files for training, testing and validation of the meta model. After an optimal substitute model is established and used for the prediction of settlements, PSO is applied for parameter identification, model correction and steering of tunneling-induced effects. The goal of this paper is to improve the accuracy of the numerical model and provide the foundation for computational steering of mechanized tunneling processes. Therefore, PSO is used as an identification tool for the inverse analysis to adapt the input parameters of the simulation model such that the results for the predicted settlements coincide with the measurements. Since the inverse analysis requires a large number of realizations, the meta model is used instead of the actual numerical simulation model. This is a considerable advantage in comparison with standard optimization procedures that require the execution of numerical simulations, which, in a realistic case, may require days or weeks for only one calculation. The second advantage is that a similar procedure can be further used for instantaneous optimization of process parameters for a real-time control of surface settlements.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £42 +P&P)