Extension of operating lifetime of aging subsea pipelines is of great interest to oil and gas sectors. Corrosion and fatigue are the main causations of the condition degradation of subsea pipelines. A dynamic probabilistic condition assessment model based on continuous dynamic Bayesian network (DBN) is developed to support the life extension decision-making of aging subsea pipeline subjected to the corrosion-fatigue degradation. The methodology is built based on equivalent initial flaw size (EIFS) concept and a time-dependent prediction model implemented using DBN. The complex corrosion-fatigue degradation process is simplified by EIFS concept, and the crack propagation due to corrosion-fatigue is modelled using fracture mechanics model and DBN. A limit state function (LSF) is used to express the failure condition of subsea pipeline due to crack propagation. The dynamic reliability of subsea pipeline is estimated using Monte Carlo (MC) method with the probability distributions of the predicted crack sizes at different time slices. The estimated reliability is compared with the acceptable threshold to decide whether any measures are required to extend the life of subsea pipeline. The methodology is tested by a case study, and it is observed that it can be a useful tool to support life extension decision-making of aging subsea pipelines.

Spillways are essential parts of dams, in which the main task of these structures is to allow the passing of excess water and floods from the upstream to the downstream. In this regards, the main goal of this paper is proposing a novel framework for the probabilistic design of labyrinth spillway structures using a new developed reliability-based design optimization (RBDO) approach. In this RBDO approach, the total volume of the spillway is considered as the objective function of the optimization problem under uncertainties, while the labyrinth spillway parameters are considered as the design variables. Hereafter, to solve the formulated RBDO problem of the labyrinth spillway design, a new proposed model that consist of coupling the Monte Carlo Simulation (MCS) with a hybrid Artificial Neural Network (ANN) based Whale Optimization Algorithm (WOA) model is developed. The hybrid ANN-WOA is utilized to approximate the labyrinth spillway response in order to reduce the computational cost during the RBDO analysis. The proposed MCS-ANN-WOA model was implemented on the Ute dam labyrinth spillway at Logan, New Mexico (USA). The obtained results showed that the proposed RBDO model performance is more accurate and robust compared to the deterministic optimization (DO) approaches for an optimal design of the labyrinth spillway shape with the consideration of the safety levels.