(2019050) DATA-DRIVEN PROGNOSTIC METHOD FOR EQUIPMENT IN OIL AND GAS INDUSTRY

Presenters

Yisha Xiang, Mario Beruvides and Lloyd Heinze
Texas Tech University

Catastrophic accidents in offshore drilling operations have greatly endangered human lives, environment and capital assets. Although risks in offshore oil and gas operations cannot be completely eliminated, a substantial amount of risks can be minimized through preventive and mitigative measures. A key aspect of the offshore drilling risk is the reliability of drilling systems. According to the World Offshore Accident Dataset and many other investigations, the overwhelming majority of disastrous accidents in offshore drilling operations were caused by equipment failures and human errors. The capabilities to predict the lifetime and provide early and effective warnings in real time are crucial to ensure reliable and safe offshore operations. The objective of this research is to mitigate offshore drilling risks by developing a scientific framework for data-driven failure prognosis for complex drilling systems operating in heterogeneous and extremely harsh environments. A novel data-driven reliability model in conjunction with a systems and economic impact analysis is developed integrating multi-source (e.g., operations and maintenance records, in-situ monitoring data) and multi-modal (e.g., lifetime data, degradation profiles) data. Numerical cases studies will be presented to demonstrate the proposed method.