The software’s reinforcement learning allows it to adjust in real time
The process industry now has its first closed-loop AI optimisation solution powered by reinforcement learning thanks to American software developer Imubit.
The Houston company’s Optimizing Brain Solution is a cloud-based industrial AI platform that takes a data-first approach, addressing the size, scope and quality of data before implementing any changes to avoid potential future delays and inaccuracies.
“ChatGPT and other foundation models are reshaping industries with their broad applicability. At Imubit, we’ve harnessed this approach for process manufacturing with our Optimizing Brain Solution,” said CEO and Imubit co-founder Gil Cohen. “It lets customers use data-driven models across operations, planning and engineering, breaking down plant silos and enabling all technical staff to thrive in an AI-driven future.”
At the heart of the Optimizing Brain Solution is Imubit’s Foundation Process Model paired with reinforcement learning, allowing the system to learn in real time through trial and error.
Because of this, Imubit’s solution can optimise industrial processing plants, improving margins by £.23 to £.38 per barrel. It also reduces natural gas consumption by up to 30% in rotary kiln operations.
“We see Imubit’s use of reinforcement learning for real-time process optimisation as unique and a strong differentiator in the market,” said Peter Reynolds, an industry analyst at ARC Advisory Group. “This approach prompted us at ARC to rethink how we categorise AI in process optimisation and control. Imubit falls into a new category, distinct from traditional APC and optimisation solutions, which we call closed-loop AI optimisation.”
The Optimizing Brain Solution was launched at Imubit’s Transcent AI Summit and Training event in Houston earlier this year. Refiners such as Citgo, Oxbow Calcining, HF Sinclair and Marathon Metroleum Corporation participated in a panel discussion to explore the ways in which the solution impacted day-to-day operations.
Some companies, for example, found that employees became more curious, asking questions and introducing more out-of-the-box solutions.
“Once your initial business case is solved, and you start seeing a lot of the excitement in the room around what's possible. That’s when you can really democratise AI,” said Kyle Kuchta, the senior director of business transformation at American refiner Delek. That’s when your operations team, P&E team, engineering team and IT team can explore all these other problems that you may not have even thought of and they’ll begin finding new ways to maximise value for the company using Imubit’s platform.”
Imubit’s Optimizing Brain Solution is the first reinforcement learning-powered AI process optimisation technology recognised by research firm ARC Advisory Group.
“The industrial sector is rapidly evolving,” said Imubit’s Cohen. “If you're not using AI to optimise your process operations, your competitors are.”
