EN | عربي

Optimizing Chemical Reactors

A framework for consistent performance in chemical manufacturing
Play Video

Project Description


Chemical Manufacturing

Principal Investigator

Mani Sarathy


Chemical manufacturing plant performance is typically controlled and optimized by process models. But those models often ineffectively account for things that can drive variations in a performance like:

  • Equipment aging
  • Feedstock variability
  • Catalyst deactivation

These factors can combine for lower efficiency and profitability.

The chemical and chemical product manufacturing sector in Saudi Arabia is expected to reach a value of $54.2 billion by 2024. Approaches that increase the efficiency of chemical reactors will help the industry grow more profitably.
KAUST has developed a novel framework that integrates reinforcement learning with economic model predictive control. This approach is agile enough to variable factors that would normally diminish performance. By using this framework, chemical reactors can be operated at, or near, optimal efficiency.
The KAUST approach also allows for control, optimization and model correction to be performed online and continuously. This makes autonomous reactor operation attainable, improving performance, chemical process yield, and profitability.