(148d) Applications of machine learning in production enhancement and field monitoring in Oil & Gas industry

Due to the fast advancement of machine learning technologies and availability of diverse types of data collection platforms, we have been seeing the growing machine learning applications in a wide spectrum of industries, spanning from emerging industries, e.g., autonomous driving, to traditional ones, such as Oil & Gas (O&G).

In this talk, we will discuss three different machine learning applications in the O&G industry. In specific, we will discuss a machine learning-based approach for understanding the multiphase flow system studied in the experimental fluid mechanics; which shows the potential of developing a complimentary methodology to assist the first principal physics; secondly, we will describe a deep learning-based object detection system that can be used to monitor the operation field or the natural environment, with the goal to improve operation safety and meet the regulation requirement ; Thirdly, we will introduce the content understanding of Process & Instrument Diagrams (P&IDs) using computer vision techniques.

Each of the above studies represents a unique usage scenario in the value chain of O&G industry. We would like to demonstrate the diverse types of machine learning applications and discuss their respective technical challenges that O&G industry presents.