Search & filter

Search for a publication

Search & filter





New Artificial Neural Network Model to Predict Surface Settlements Caused by

Artificial Neural Network Model to Predict Surface Settlements Caused by


U. Ates / I.s. Binen

Utilization of TBMs in urban tunnelling is increasing due to the infrastructure requirements of modern-day cities. On the other hand, settlement related risks are still one of the most important considerations which require comprehensive preliminary studies and calculations with very sensitive excavation procedures. In this study, deformation measurements obtained from the deformation stations located on the surface, related with the excavation data of 3 different EPB TBMs and geological information, combined and analysed. A unique script has been prepared in MATLAB to evaluate the deformation data and calculate the corresponding settlement curves. With the help of the script, deformation measurements obtained from approximately 3000 different deformation measurement stations were separated, classified, and associated with the related excavation and excavation data. For each advance, 3 different circular impact zones were then determined according to the parameters such as tunnel depth, excavation diameter, and inflexion point of the settlement curve, making it possible to match the excavation data of a specific advance with the surface deformation data. In the second step of the study, artificial neural networks method has been adopted to predict settlements and understand the weight of different parameters.

More details

10,00 €