Modeling and optimization of safety and availability for subsea all-electric Xmas Trees

Abstract

This dissertation is outlining the technological progress in subsea oil and gas production. The technological change, which is taking part in the subsea production environment, is creating new challenges and problems. This work is focusing on the functional safety aspect of distributed mechatronic systems in the subsea environment. The extreme conditions, which the systems have to endure without any planned maintenance, are leading to system designs with multiple redundancies for safety and for availability. The reliability calculations are state of the art executed for the first safety relevant failure which is occurring in the system. Afterwards a redundant system has to take over the control or a emergency shutdown has to be initiated. This emergency shutdown is causing production losses until the system is restored. This dissertation is analyzing existing methods for reliability estimation and combines different approaches to realize an embedded model for a realtime estimation of the remaining reliability of the mechatronic system. The Xmas Tree Controller, which is developed during this work, is capable of evaluating the reliability of the production system and is autonomously deciding, which action needs to be initiated to achieve a safe operating envelope. Different system architectures and operation strategies where analyzed and compared. The method of multi-redundant systems, which are sharing information about their State of Health and their capabilities is introduced. This novel operation strategy allows to increase the availability of the overall production system. The concept of cannibalization for availability optimization is introduced and applied to an already optimized system architecture. The result of the optimizations is a downtime reduction of 96 \%. The model is realized and verified for embedded usage and will be executable on the Xmas Tree Controller. The result will be a mechatronic deep-sea production system, which is determining its own safety based on the actual system condition, the environmental conditions and its statistical probability to fail. The validity of the model is proven based on classical reliability theory and compared to the state of the art IEC 61508 estimations, which is applied for this safety critical applications.