Engineers in refineries spend hours each week mass balancing different units to diagnose material losses, correct meter readings, troubleshoot the unit, improve feedstock intake and optimize production. Often the task is tedious and takes up hours, if not days of the engineer's time.
This paper discusses how a new approach to Industrial AI was used in the refinery to bring the mass balance around the Crude Unit within compliance in a scalable and repeatable way. The Industrial AI solution is used to identify stream flow readings that caused mass balance to go out of compliance, predict the correct values that bring it within the compliance window, and validate the predictions through simple visualizations. By using AI to analyze the entire crude unit, including both the atmospheric column, vacuum column, and the auxiliary Merox unit, the engineers were able to tighten their compliance criteria, improve production accounting and complete the mass balance analyses in half the time with significantly less effort. The solution can be easily applied to other units and used by engineers of all experience levels to scale the benefits.
The Mass Balance solution saves up to 40% of engineers' time, improving their efficiency and shortening ramp-up time for future engineers. It also provides better energy utilization, reduced emissions, and improved yield, resulting in an average refinery in the USA generating $1M/Year in additional profits with only 0.1% improvement in feed intake through corrected mass balance measurements. Get started with applying this Solution today!