Modern off shore oil production can produce 40.000 different possible alarms, and typically 500-1000 displays. As the control of off shore oil production has increased in complexity throughout the years, with increasing amount of information available for dynamic and nonlinear process system, decision making is highly complex. As reported by the US Dept. of Energy in 2009 20% of incidents off shore are related to equipment failure and 25 % to human errors. The human errors are the result of such system complexity. All disturbances, alarms and shutdowns leads to a loss in productivity.
In order to increase productivity and safety of process systems in off shore oil production, a decision support system is developed, to aid operators in faster and more reliable decision making. The system provides situation awareness, by diagnosing faults and providing this information to the operator.
The Decision Support System builds on Multilevel Flow Modeling (MFM), a functional modelling methodology, for building process system models. The models are used for generating fault trees which can be reasoned amongst in order to diagnose the origin of faults and consequences.
The PhD project aims to scientifically employ the methodology in real-time to diagnose faults and to use HAZOP and dynamic process simulations, to identify and analyse fault scenarios. The PhD additionally aims to investigate how to validate knowledge based models and causal models by using dynamic process simulations and HAZOP.