Development of a generic methodology for probabilistically predicting the reliability of offshore mechanical components at the design stage
The continuously developing requirements of the offshore oil and gas Operators are placing more stringent demands on the designers to produce optimised solutions, with reduced development schedules, and application of new technologies for extreme environmental and operational conditions. The compounding uncertainty of service conditions and in the design capabilities is causing the designer to over-design, conduct extensive pre-service testing and introduce design redundancy. As such, designers have been forced to turn to reliability techniques in order to quantify the life of their designs. This has extended as far as needing to integrate reliability concepts, tools and methodologies into the design process. In recent times, the industry has attempted to apply conventional reliability tools within the design process, in terms of failure identification and reliability quantification. However, the use of historical reliability data has shown to be a particular downfall. As such, there has been a drive to predict the reliability of mechanical components based on their underlying degradation and failure processes and mechanisms. Consequently, this research is initially concerned with the development of a methodology, including an assessment of existing methods, that could be used to more readily understand the underlying failure characteristics of a mechanical component in terms of material, geometrical, environmental and operational characteristics. A particular underlying mechanism has been chosen and mathematical models were developed that simulate its physical behaviour and its degradation characteristics. Additionally, due to the potential uncertainty in the models and limited understanding of the characteristics of the underlying mechanism, the model was simulated within a probabilistic framework, fundamentally by application of the stress strength interference modelling approach. Finally, the model and its parameters were assessed to determine how uncertain governing parameters could appear to lead to variations in the reliability of the actuator.