Effect of urban tunnelling on surface structures: an AI approach
Loading...
Date
2023
Journal Title
Journal ISSN
Volume Title
Publisher
Ecole Nationale Supérieure de Technologie et d'Ingénierie. Annaba (Ex ENSMM)
Abstract
Tunneling causes deformations on the surface and damages surrounding structures. To reduce the vulnerability of neighboring structures, it is important to improve predictions of the deformations. For this aim, this thesis presents, explores and discusses the application of AI as tool. The database is built using technical information of the Algiers metro -Rabia Taher section. Following from that and basing on literature exploration, the necessary are defined .After that, the data extraction process is done and includes data collection and sorting in basic inputs of structural, technical and geotechnical characteristics, and field measurements of surface settlement with structures distortion.
The exploratory analysis of the data aims to identify outliers and potential links between the various parameters. Machine learning algorithms are used after to predict the settlement induced in structures during excavation. Three algorithms are tested and compared: Linear Regression, SVM and ANNs. The results of these predictions are evaluated with actual settlement measurements. The obtained results developed an operational tool capable of forecasting effectively, The database created has served as a source of data for machine learning algorithms and can be used for future applications,.