Neural network based prediction of ground surface settlements due to tunnelling

Ground surface settlement due to tunnel excavation varies in magnitude and trend depending on several factors such as tunnel geometry, ground conditio ns, etc. Although there are several empirical and semi-empirical formulae a vailable for predicting ground surface settlement, most of these do not sim ultaneously take into consideration all the relevant factors, resulting in inaccurate predictions. In this study, an artificial neural network (ANN) i s incorporated with '113' of monitored field results to predict surface set tlement for a tunnel site with prescribed conditions. To achieve this, a st andard format (a protocol) for a database of monitored field data is first proposed and then used for sorting out a variety of monitored data sets ava ilable in KICT. Using the capabilities of pattern recognition and memorizat ion of the. ANN, an attempt is made to capture the rich physical characteri stics smeared in the database and at the same time filter inherent noise in the monitored data. Here, an optimal neural network model is suggested thr ough preliminary parametric studies. It is shown that preliminary studies f or generating an optimal ANN under given training data sets are necessary b ecause no analytical method for this purpose is available to date. In addit ion, this study introduces a concept of relative strength of effects (RSE) [Yang Y, Zhang Q. A heirarchical analysis for rock engineering using artifi cial neural networks. Rock Mechanics and Rock Engineering 1997; 30(4): 207- 22] in sensitivity analysis for various major factors affecting the surface settlement in tunnelling. It is seen in some examples that the RSE rationa lly enables us to recognize the most significant factors of all the contrib uting factors. Two verification examples are undertaken with the trained AN N using the database created in this study. It is shown from the examples t hat the ANN has adequately recognized the characteristics of the monitored data sets retaining a generality for further prediction. It is believed tha t an ANN based hierarchical prediction procedure shown in this paper can be further employed in many kinds of geotechnical engineering problems with i nherent uncertainties and imperfections. (C) 2001 Published by Elsevier Sci ence Ltd. All rights reserved.