The Effect of a Simplified Geotechnical Model for Predicting Surface Settlement Incorporating Bayesian Back Analysis

The intent of this study was to question if a simple geotechnical model could be used to undertake consolidation analysis to an acceptable level of reliability incorporating Bayesian back analysis. Simplification of the ground model is common in geotechnical engineering practice however the use of measured results to verify the model parameters is seldom. Use of a simplified geotechnical model can reduce the amount of computational time required, particularly for Bayesian updating. Oversimplification on the other hand will not capture the appropriate conditions for reliable settlement prediction. Bayesian back analysis can provide a way to update the prior belief of the model and adopted soil parameters using monitoring data. Therefore, a focus on adopting Bayesian updating through back analysis by treating the key parameters of compression ratio, recompression ratio, creep strain and coefficient of consolidation as random variables was considered. A three-layered simplified model was adopted and showed the surface settlement was well predicted using 117 days of observed data. The settlement data was used to update the parameters through Bayesian back analysis to fit the entire time-settlement history. The results of the predictions are discussed in this paper.

Keywords: back analysis, Bayesian updating, consolidation, embankment, settlement prediction