Health-aware advanced control applied to a gas-lifted oil well network

Abstract

Currently, preventive maintenance is the common practice in offshore oil production processes. Maintenance stops are scheduled in regular intervals, even if the equipment is working well, sometimes decreasing significantly the plant profitability. In smart production, the goal is to change to a predictive maintenance paradigm, where process monitoring algorithms use process data to detect the system degradation state and predict when the equipment needs maintenance. Adopting this perspective, we propose a model predictive control approach that incorporates process monitoring. This allows us to steer plant degradation actively, preventing violation of health-critical constraints, while optimizing the economic production of the system. In this paper, we present a case study, a gas lift oil well network subject to sand erosion in the choke valves. The results indicate that we are able to maximize oil production while keeping the erosion of the choke, the critical process equipment, below a safe threshold. Additionally, we discuss some of the implementation issues of the proposed approach.