Multivariate adaptive regression splines model for reliability assessment of serviceability limit state of twin caverns

• Main Authors: Zhang, Wengang, Goh, Anthony Teck Chee
• Other Authors: School of Civil and Environmental Engineering
• Format: Article
• Language: English
• Published: 2015
• Subjects: DRNTU::Engineering::Civil engineering::Geotechnical
• Online Access: https://hdl.handle.net/10356/107308; http://hdl.handle.net/10220/25427; http://dx.doi.org/10.12989/gae.2014.7.4.431
• Institution: Nanyang Technological University

Construction of a new cavern close to an existing cavern will result in a modification of the state of stresses in a zone around the existing cavern as interaction between the twin caverns takes place. Extensive plane strain finite difference analyses were carried out to examine the deformations induced by excavation of underground twin caverns. From the numerical results, a fairly simple nonparametric regression algorithm known as multivariate adaptive regression splines (MARS) has been used to relate the maximum key point displacement and the percent strain to various parameters including the rock quality, the cavern geometry and the in situ stress. Probabilistic assessments on the serviceability limit state of twin caverns can be performed using the First-order reliability spreadsheet method (FORM) based on the built MARS model. Parametric studies indicate that the probability of failure increases as the coefficient of variation of Q increases, and decreases with the widening of the pillar.