Machine Learning in Oil and Gas 2017_Agenda - Energy Conference Network

In early December 2016, an oil spill in Belfield, ND that measured more than 175,000 gallons was discovered by a local landowner. It is unknown how long the pipeline may have been leaking and while the pipeline had monitors and was regularly inspected, the leak was not detected. In late January 2017, an Iowa pipeline that crosses several mid-western states, spilled almost 50,000 gallons of diesel fuel caused by an excavator not checking on the presence of underground utilities. Monitors, sensors and rapid responses coupled with algorithms provide strong foundations for managing operations. However, all machine learning algorithms have flaws.