Probabilistic Uplift Resistance of Pipe Buried in Spatially Random Cohesionless Soil

The lower bound finite element limit analysis technique in conjunction with the second-order conic optimization is used to determine the uplift resistance of pipeline buried in spatially random cohesionless soil. The Cholesky decomposition method is employed to produce the spatially random discretized soil domain. The Monte Carlo simulation technique is implemented to obtain the probabilistic responses. The mean uplift factor and the failure probability of the pipeline for a wide range of practical cases of soil spatial variability are provided as the design charts. It is found that for the smaller value of correlation distance, the deterministic uplift factor is always higher than the mean uplift factor; however, with the increase in the correlation distance, the difference between the deterministic uplift factor and the mean uplift factor reduces. The probability of failure of pipeline is found to be increasing with the reduction in the magnitude of correlation distance and factor of safety. The influence of the soil spatial variability on the failure mechanism is also studied in detail.