Probabilistic assessment of tunnel construction processes using statistical learning to include past experience

Olga Špačková, Jiří Šejnoha, Daniel Straub

Research output: Contribution to conferencePaperpeer-review

Abstract

A Dynamic Bayesian Network (DBN) model for probabilistic estimation of tunnel construction time (and costs) is presented. It considers two main types of uncertainties: those associated with ordinary variability in the construction performance and those associated with severe delays caused by extraordinary events such as tunnel collapses or major organizational problems. A method for learning the model parameters from analysis of data from past projects is shown. The proposed DBN model is applied to a case study of a two-lane road tunnel excavated with a conventionnal tunneling method. The model parameters are learnt using data from one reference tunnel.

Original languageEnglish
StatePublished - 2012
Externally publishedYes
EventISRM International Symposium - EUROCK 2012 - Stockholm, Sweden
Duration: 28 May 201230 May 2012

Conference

ConferenceISRM International Symposium - EUROCK 2012
Country/TerritorySweden
CityStockholm
Period28/05/1230/05/12

Keywords

  • Bayesian networks
  • Construction time estimation
  • Probabilistic modeling
  • Risk assessment

Fingerprint

Dive into the research topics of 'Probabilistic assessment of tunnel construction processes using statistical learning to include past experience'. Together they form a unique fingerprint.

Cite this