Reliability Assessment of a Fixed Jacket Platform by Monte Carlo Simulation Using Neural Network

Fixed offshore structures are considered as an important structure in shallow water. Iranian oil and gas offshore structures which have been located in Persian Gulf are mostly fixed jacket. So reliability assessment of these kind of marine structures seems to be very important and is an essential part of offshore structure design. Mont Carlo is a powerful method which is used broadly for prediction of structure failure. The most advantage of this method is simplicity of implementation, the main limitation of this method is about the computational time due to the huge number of structural analyses. Incorporating the artificial neural network for the reduction of the sample size is used to eliminate of MCS‘s time bottleneck. An MCS based method is introduced to take advantage of precision in optimization part. To solve the scaling problem of a large reliability analysis, an artificial neural network is employed. In this paper, an almost new constructed fixed jacket platform in the South Pars is selected and modeled using SACS software. In this regard, the nonlinear static pushover analysis is performed by application of nonlinear soil-pile interaction. Analytical results show that the simultaneous use of these two techniques lead to more accurate and also faster reliability assessment. In MCS method probability of failure calculated using divide the number of failed sample by total samples which is concluded to the value of 9.4e-05 for the test case structure in the current study, and the reliability index is resulted to 3.73.