The problem they discussed is related to the use of thief hatches in oil tanks. A thief hatch is a closable aperture used for taking samples, determining the level of the tank, and protecting the tank from overpressure and excessive vacuum. However, leaving the hatch open for an extended period can result in safety hazards, such as fire and harm to workers, due to combustible gases. It may also result in environmental damage and regulatory fines or penalties. Additionally, the hatches may be opened due to process events, such as high operating pressure being pushed into the tanks. Therefore, it is important to close the thief hatch after finishing tasks to avoid such hazards.
Their BPX Challenge problem was to develop an algorithm to classify time-series data for determining the status of the thief hatch in oil tanks and the probability of it being open at any given time based on sensor pressure data.
For more information, please visit https://www.mines.edu/mcneil/innovationchallenges/