Fatigue life prognosis of an oil well drill string using cascaded dynamic models

Abstract

Early estimation of the fatigue life of an oil well drill string reduces the risk associated with drill string fatigue failures. In this study, a low-order computationally efficient bond graph model of a vertical well drill string and a component-level higher-order finite element model of a drill pipe threaded connection are employed to predict the fatigue damage of a given drill pipe. The bond graph is a 3D lumped segment model developed using the Newton–Euler formulation and body fixed coordinates. It is parameterized using finite element modelling simulations. The stress history from the top-level model is applied to the component-level model that contains details such as threaded geometry. Then, a multi-axial, non-proportional, and variable amplitude (MNV) fatigue estimation is performed using an open-source finite element analysis code. The fatigue prognosis approach is then demonstrated in a drill string design case study that optimizes the placement of vibration stabilizers in the wellbore to avoid severe vibrations while minimizing fatigue damage. Optimal placement of stabilizers predicts a 200% increase in fatigue life of the most vulnerable component with reference to the worst-case scenario.