such as feed rate changes, cooling disruptions, and reflux limitations, directly into the existing AIO controllers, opera- tors could then run these ‘what-if’ scenarios in a controlled, offline environment, inputting how they would respond to specific upsets to get the unit back to a stable and optimal state while still meeting product specifications. • Results: As operators competed to achieve the best economic outcome, training sessions became interactive and engaging. This hands-on learning approach not only improved practical skills but also shifted the refinery’s cul- ture to embrace AI-supported optimisation. Big West has since extended this training methodology to other process units, further enhancing its workforce development strategy. Conclusions By moving beyond traditional linear controls and siloed strategies, the refiner consistently identified and relieved bottlenecks, such as the debutaniser constraint, while tightening key specifications, like diesel flash, to reclaim lost margins. Equally important, the refinery successfully integrated a robust change-management framework, ena- bling operators and engineers to trust, adopt, and fully lev- erage advanced AI-based applications. Moreover, a holistic approach to AI adoption extends well past the immediate closed-loop gains. By utilising the cloud-based platform for operator training and scenario planning, the refinery has cultivated an environment where data-driven decision-making is the norm. This investment in workforce development ensures that both current and future teams are prepared to navigate increasingly complex refining conditions. Looking ahead, refinery management aims to expand the scope of its AIO deployment, applying AI-enabled opti- misation solutions to additional refining units. As market demands shift, whether due to seasonal product specifica- tion changes, opportunistic feedstocks, or decarbonisation initiatives, refiners that harness next-generation technol- ogy solutions will be best positioned to adapt. In that sense, Big West Oil’s experience offers a roadmap for any refinery seeking to enhance margins, reduce downtime, and build an agile, tech-savvy workforce ready to tackle tomorrow’s challenges. Travis Legrande is Operations Technical Specialist at Big West Oil LLC. His current role builds on 14 years of expertise and a demonstrated history of working in the oil and energy industry. He is a seasoned pro- cess engineer with a chemical engineering degree from the University of Utah. Legrande has a proven track record of optimising unit eco- nomics, ensuring safety compliance, and successfully managing com- plex projects to enhance refinery efficiency, safety, and profitability. Mitchell McCloud is a Solution Manager Portfolio Lead at Imubit. With nearly a decade of experience in refinery process engineering, oper- ations, and industrial optimisation, he has led AI-driven optimisation projects and customer success initiatives, helping industrial clients maximise value through advanced technology. His work focuses on process optimisation, AI applications, and strategic efficiency improve- ments. McCloud holds a BS in chemical engineering from Mississippi State University and an MBA with business analytics focus from Auburn University.
limits. By shifting more C 5 material into cat gas, excess olefin production was also reduced, improving gasoline blending flexibility. This early success with the debutaniser established credibility for AI-driven optimisation and paved the way for subsequent successes. Case study 2: Diesel flash optimisation • Background: Crude distillate and light cycle oil (LCO) from the FCC are dewaxed and desulphurised at the facility through a series of hydrotreating steps. The hydrotreated distillate is then stripped of light material in the hydrodes- ulphurisation (HDS) stripper, where the net overhead is routed to the reforming unit. • Challenge: Typically, there are significant economic impli- cations of minimising HDS stripper naphtha to a ULSD flash specification limit. Historically, the site operated with a 5-10°F buffer above specification due to frequent tower flooding and concerns of measurement variability. This conservative operation limited diesel throughput, raising naphtha production. • Solution: Big West engineers worked with Imubit’s deliv- ery team to identify the two critical handles: reboiler return temperature and overhead temperature. These two handles together drive tower separation through manipulating the tower’s heat balance. As part of the application delivery, a robust ULSD flash inferential was built, a key enabler that provided accurate real-time flash predictions. • Results: Upon deployment of the HDS stripper AIO appli- cation, flash variability decreased significantly. These con- sistent, accurate predictions gave the operations team the confidence to run less conservatively, methodically lowering the flash target closer to the specification limit. The incre- mental closing of the gap between the flash target and spec limit ultimately reduced giveaway by approximately 4-5°F. Case study 3: Operator optimisation training • Challenge: Big West’s prior operator training programme involved on-the-job learning, shadowing others on shift. This led to a new operator’s development being restricted to the knowledge and experience of the person who came before them. There was awareness of the operator training simulator (OTS) market. However, these solutions tended to be costly and generic to a type of unit operation rather than a site’s specific operations. Large, integrated energy companies often maintain these systems through a central- ised team, which was not feasible for a single-site refinery like Big West. • An ‘aha’ moment: As the console operators gained famil- iarity with the FCC debutaniser AIO application, they began to recognise that a solution to tower optimisation could involve a combination of simultaneous, continuous adjust- ments, which outperformed the legacy approach of mak- ing step changes followed by resampling. This new way of approaching distillation optimisation began spreading to other operating consoles where AIO had not yet been rolled out. To further institutionalise this data-driven culture, Big West sought to incorporate its cloud-based AIO models for formal operator training in process optimisation. • Solution: By embedding common operational scenarios,
56
PTQ Q3 2025
www.digitalrefining.com
Powered by FlippingBook