Terminert. Online adaptive decision support systems for risk-centred condition based maintenance in offshore drilling

The primary objective of this project is to develop next generation risk-centered decision support for maintenance planning in offshore drilling rig systems. The developed methodology and technology will lead to reduced operational risk and cost for offshore drilling operations. Large and increasing amounts of data are being logged from the actual drilling processes and from vessel power and propulsion systems. In addition, extensive data is also available from work processes, spare-part history, failure rates and other experience-related data. Taken together, all these data form a rich and complex multi-dimensional dataset that can provide valuable input to understanding risk and planning maintenance and operations. This value is not fully realized in current maintenance systems. The major innovation in this project is the merger of interactive visual data analysis, data mining and adaptive models integrated in an RCM (Reliability Centered Maintenance) context. Developed methodology will be implemented in prototypes that will be tested in operative situations. Several workshops have been arranged, including meetings with relevant condition monitoring vendors and service companies. The relevant maintenance systems and practices at Odfjell Drilling have been reviewed. We have furthermore defined criteria for selection of equipment and sub-system to be addressed in this project. Relevant equipment and subsystems have been analyzed and compared to these criteria, and two sub-systems have been selected: thrusters for floating rigs and top drives. We have performed a survey on related work to get an overview of the state-of-the-art in condition-based maintenance (CBM) and related concepts. OD has installed and upgraded condition monitoring systems so that required data, both historical and real time, can be available for the project. In 2014 we have worked on collecting data that can be analyzed to identify relevant methods that shall be implemented in the first prototype in this project. Vibration data have been collected for the thrusters on the floater Deepsea Atlantic and for the topdrive on the Grane platform. We have analyzed the collected data to identify properties and correlations. The findings can be used for selecting relevant methods for trending and classification of the condition of the monitored equipment. As a part of the analysis and method development, we have implemented software components for analysis and visualization that will be part of the first prototype.

Current maintenance regimes on offshore drilling rigs is often built around Reliability Centered Maintenance (RCM), and is based on offline analysis and risk-considerations based on standards and/or operational experience. Large and increasing amounts of data are however being logged from the actual drilling process, from infrastructure monitoring, vessel power and propulsion systems. In addition, extensive data is also available from work processes, spare-part history, failure rates and other experience- related data. Taken together, all these data form a rich and complex multi-dimensional dataset that can provide valuable input to understanding risk and planning maintenance and operations. This project will research and develop methodology for harvesting the potential that lies in these data for optimized maintenance strategies both in terms of risk and cost. In the project, we will develop systems that analyze historical and online technical condition data and cost information from critical components and subsystems using state-of-the-art visualization and data-mining techniques. Based on analysis results, the system will employ adaptive models for online condition-data based predictive decision support for maintenance strategy and planning. The deve loped methodology and technology will lead to reduced operational risk and cost for offshore drilling operations, addressing the topic "Cost effective drilling and intervention" of PETROMAKS.