Leveraging machine learning and data analytics for equipment reliability in oil and gas using predictive maintenance
Authors/Creators
- 1. Department of Renewable Energy Engineering, Henriott Watt University, Edinburgh, United Kingdom.
- 2. Department of Petroleum Engineering, University of Benin, Edo State, Nigeria.
- 3. Department of Banking and Finance, Ekiti State University, Ado Ekiti, Ekiti State, Nigeria.
- 4. Department of Chemical Engineering, Lagos State University of Science and Technology, Ikorodu North, Lagos State, Nigeria,
- 5. Department of Computer Science Adekunle Ajasin University, Akungba-Akoko, Ondo state, Nigeria.
- 6. Department of Biomedical Technology, Federal University of Technology, Akure, Ondo State, Nigeria,
- 7. Department of Electrical and Electronics Engineering, University of Ibadan, Ibadan, Oyo State, Nigeria
Description
The oil and gas industry operates under very extreme conditions, posing a huge challenge when it comes to equipment reliability. Predictive maintenance; which is now possible through machine learning and data analytics, has transformed the way one looks at equipment management by making real-time failure prediction possible, reducing unplanned downtime, and optimizing maintenance schedules. The review of the technological advances in predictive maintenance methodology focuses on supervised and unsupervised machine learning, deep learning models, and integration with IoT-big data analytics. The paper also summarizes a number of case studies from some of the leading IOCs such as Shell, BP, ExxonMobil, Chevron, and Total Energies. While emphasizing respective KPIs between traditional and predictive maintenance methods; advantages, challenges, and future opportunities in the use of predictive maintenance systems were analysed. This review will be very helpful to both academics and field professionals with research and professional interests in pursuing operational efficiency and sustainability for the oil and gas industry.