PTQ Q3 2022 Issue

900 910 920 930 940 950 960 970 980 990

1

0.9

0.5

0.6

0.4

LCO

Slurry

0

PP

BB

CG

-0.4

-0.5

Without CLNN With CLNN

-0.9

-1

72

74

76

78

80

82

Feed rate bins (MBPD)

Figure 5 While CLNN is engaged, the average objective function over two months is consistently higher at fixed feed rate bins, with a more optimal shift in product yields

Optimising all operating strategies with CLNN Domain experts well know the complex reality of the refin - ery. While they are forced to break down and simplify the problem due to traditional technology limitations, plant groups attempt to reconnect their fragments to execute operating strategies. However, the true problem is never solved. Through a holistic, dynamic process model, an experienced-based con- troller and optimiser, and the ability of the entire organisation to adopt and understand such a technology, CLNN enable plant experts to optimise all refinery operating strategies with AI, adding millions of dollars to the plant’s bottom line. Geraldine Hwang is Product Manager at Imubit in Houston, Texas. Hwang earned a B.S. in chemical engineering from the University of Texas. Abishek Mukund is Head of Product Marketing and Strategy at Imubit, Houston, Texas. Mukund has a B.S. in chemical engineering from Georgia Tech University.

constraints continually changed due to accelerated fouling. CLNN were able to better model and control the chang- ing unit dynamics to respect operating limits effectively, building confidence and trust with operators to drive engagement. Over several months of engagement, the CLNN were able to capture previously unattainable value. To clearly see value generation, the feed rate was normalised to show that CLNN could effectively manipulate composition han- dles and conversion to drive the objective function higher than the refiner’s traditional solutions (see Figure 5 ). In addition, this value was attained by the CLNN under- standing when to use specific handles, faced with several tradeoff constraints, to make the most valuable products per the refiner’s defined price deck. CLNN control and opti - misation of the refiner’s FCC yield improvement operating strategy led to a $0.53/bbl increase in the feed incentive of the FCC.

Easy access to a growing database of technical articles, company literature, product brochures, webinars, videos, and news

89

PTQ Q3 2022

www.digitalrefining.com

Powered by