No abstract

Abstract

Reliability assessment is one of the necessary and critical parts in structural design under uncertainties. The methods for structural reliability assessment aim at evaluating the probability of limit state by considering the fluctuation of acting loads, variation of structural component or system, and complexity of operating environment. Latin Hypercube sampling (LHS) method as advanced Monte Carlo simulation (MCS) has higher efficiency in sampling. It will be chosen and applied in this paper in order to obtain an effective database for building Kriging surrogate models. In this paper, we propose an effective method to have reliability assessment by Latin Hypercube sampling based Kriging surrogate models. This method keeps the certain level of accuracy in prediction of the response of a structural finite element model or other explicit mathematical functions.

A finite element model is established using MIDAS GTS NX 2018 software, in order to simulate the behavior of an instrumented large diameter bored pile, installed in multi layered soil and tested under three different loading and unloading cycles at Damietta Port Grain Silos project site. Modified Mohr-Coulomb constitutive model has been used to define the drained condition for sandy soil layers and undrained condition for clayey soil layers. Necessary soil parameters were determined from extensive laboratory and in-situ soil tests. Numerical results are compared with field loading test measurements and very good agreement is obtained. The effect of dilatancy angle on pile load transfer mechanism was investigated, and results of the study showed important effect for the dilatancy angle on the pile settlement values and the load distribution along the pile shaft. Results obtained also showed that the plastic zone below the base of the pile at failure extended laterally to about seven times the pile diameter and vertically to about 5 times the pile diameter.

Reliability based optimization (RBO) is one of the most appropriate methods for structural design under uncertainties. It searches for the best compromise between cost and safety while considering system uncertainties by incorporating reliability measures within the optimization. Despite the advantages of RBO, its application to practical engineering problem is still quite challenging. In this paper, we propose an effective method to decouple the loops of reliability assessment analysis and optimization by creating surrogate models. The Latin Hypercube sampling approach is applied to a structural finite element model to obtain an effective database for building surrogate models. In order to avoid premature convergence of the optimization process, the RBO problem is solved with metaheuristic methods such as genetic algorithm and simulated annealing. The relative efficiency of surrogate models and their relationship with metaheuristic search engine are discussed in the article.