Applied real-time Bayesian analysis in forecasting tunnel geological conditions

Unforeseen ground conditions not only affects our schedule and imposes large extra costs but may also introduce additional hazards in the tunnel project. Crucial pre-construction phase decisions and construction phase decision are strongly influenced by expected ground condition. A geological prediction model to quantify the risk of tunneling and predict the ground condition for unexcavated part of the tunnel using real time Bayesian analysis is proposed. One of the real time Bayesian analysis simulation techniques, Particle Filter Algorithm (PF), is used to simulate the geological prediction profile for unexcavated tunnel parts. In this paper a tunnel drainage project at Zhonghe area, Taipei county, Taiwan is used as a case study of the proposed model and a validation purpose. Compared with Iterative Bayesian Updating Approach (IBUA) proposed by Ioannou, the model that we proposed using the simulation technique gives a better prediction result. The results prove that the geological prediction model proposed is useful for decision making.