Optimizing condition monitoring for dynamic health and risk management

Abstract

This Ph.D. thesis explains the concepts of condition monitoring and associated challenges in maintenance modeling for subsea facilities. Thesis’s main objective is to develop systematic frameworks to assess the performance of the subsea system considering its degradation phenomenon. The primary focus of this thesis is on modeling the degradation behavior of the subsea safety systems. We also extended the degradation modeling concepts to study the subsea production systems with components experiencing stochastic deterioration. We have addressed four research questions explicitly: 1. In the first research question, we developed a framework to assess the reliability of a safety instrumented system that is subjected to destructive periodic tests. We utilized a multi-phase Markov process (MMP) to model the degradation process of the SIS. The selection of a multi-phase Markov process is motivated by mainly two factors: (i) It allows us to have intermediate performance levels between perfectly working and uniquely failed levels. (ii)The impact of destructive testing is modeled by altering the transition rate of the degradation process. We developed a dynamic failure rate model that depends on the current degradation level and the number of tests experienced. We also performed the case study on Down-hole safety valves (DHSV) to determine the optimum number of periodic tests that maximize the average availability of DHSV in given mission time. A high frequency of tests will reduce the probability for DHSV to be in an undetected failed state and not to act on demand. On the other side, the cumulative stress experienced due to tests may degrade the performance to failure. 2. In the second research question, we extended degradation modeling techniques in the qualification of novel subsea technology. All-electric systems are the novel subsea technology that is considered an upgrade of widely deployed electro-hydraulic systems. This novel technology promised more reliable equipment and a safer environment. The technology qualification requires the reliability assessment of new systems to provide sufficient evidence that the new technology is fit for the purpose without high risk. The current reliability assessment of such systems assumes perfect restoration during proof tests and no impact of degradation due to demands. Failure mode and effect analysis of all-electric actuation systems show that it is prone to degradation in performance due to power supply interruptions. These interruptions appear as random demands to the safety valves of the system. These valves may degrade its performance due to such demands. We utilized the MMP to model this situation. The impact of demand is modeled either by changing the initial condition of the MMP or by increasing the transition rates between two degraded states. We developed analytical formulae for realistic dynamic reliability assessment. 3. In the third research question, we studied the testing and maintenance strategies for a redundant SIS with imperfect detection of degraded state during proof tests. We studied the performance of redundant SIS under the combinations of staggered testing and simultaneous testing with preventive maintenance, corrective maintenance, and opportunistic maintenance. We developed analytical formulae for time-dependent unavailability and associated life cycle cost for finite mission time. This study’s main purpose is to incorporate and balance system availability and life cycle costs. We performed a case study on the subsea high-integrity pressure protection system. 4. In the fourth research question, we extended the degradation modeling techniques in the domain of subsea production systems. Subsea production systems are operated very aggressively to extract hydrocarbon quickly and as much as possible. This causes premature wearing of the systems, which increases maintenance and repair costs. There exists a trade-off between high maintenance and repair costs versus high production profits. In this study, we addressed this trade-off by developing a method that integrates the deterministic control laws to the stochastically deteriorating components. We utilized non-homogeneous MMP on the component level to describe the system dynamics. We considered that transition rates of MMP are proportional to the operational loads. This assumption ensured that high operational loads would lead to faster deterioration of components. At the same time, we also assumed the productivity in each performance level of MMP is also propositional to the operational loads. This assumption will ensure that higher operational loads will lead to higher production. The resultant optimization problem becomes a non-linear problem. We solved it numerically with the help of off-the-shelf non-linear problem solvers. The solver provides optimum values of operational load schedule, maintenance schedule, and maintenance efficiency, which maximizes the production. We performed a case study on the subsea compressor station to apply the developed method. The frameworks developed in this thesis are meant to provide support to operators/ engineers in informed decision-making. These decisions are based on the realistic performance assessment of the subsea system. It is assumed that the readers have familiarity with the general concepts in the domain of reliability theory.