Condition-Based Maintenance Models: Application to Subsea Safety Systems

Abstract

This PhD thesis clarifies concepts and challenges in maintenance modelling for subsea systems. The main objective is to develop suitable condition-based maintenance (CBM) models for some typical safety systems in subsea. This objective is detailed in five tasks – degradation modelling, maintenance modelling, formalisms evaluation, numerical experiment and case study – in three corresponding parts at two levels: CBM models for single-unit system and CBM models for multi-unit systems.

·in the first part, we develop CBM models for single-unit systems with problems specific in subsea. Weextend and study delayed maintenance model of a choke valve where the degradation is characterised by a stochastic gamma process. We show for the choke valve the interest of using intermediate condition to trigger proactive actions and their impacts on system average availability.

·in the second part, we take a top-down approach to develop CBM models for multi-unit systems. We investigate Piecewise Deterministic Markov Process to analytically formulate a model of a control system. We give numerical solutions, and verify the model independently through Monte Carlo simulation. We illustrate the behaviour of this kind of formalism when maintenance modelling moves to multi-unit systems.

·in the third part, we take a bottom-up approach to develop CBM models for multi-unit systems. We use a modelling language AltaRica 3.0 and its underlying modelling formalism Guarded Transitions Systems (GTS) to develop CBM models based on probabilistic modelling and Monte Carlo simulation. We show the advantages of the pairs by a use case of High-integrity Pressure Protection System. The modelling formalism makes it possible to explore alternatives quickly and extend the model to industrial size.

The scope of the thesis is delimited to engineering assessment of subsea safety systems for maintenance decision making based on condition of the systems. It provides practitioners with guidance for maintenance modelling and performance assessment.