Business Objective:
To Increase Profit of a Hydrocracker at a Major Refiner in Saudi Arabia
A major refiner in Saudi Arabia approached Lummus Digital with the aim of increasing profit from its hydrocracker. The business objective included:
- To run the plant at the most optimum profit
- To respond faster to the changing market conditions by adjusting the unit to produce the optimal product mix i.e. increase the yield of the most valuable product
- To get more value from the plant, i.e., increasing yields, reducing energy / utilities consumption
Key Solution:
- Build a Digital Twin and have the Hybrid plant model mimic the physical plant behaviour with accuracy of ~98%
- Employ Supervised machine learning, which utilizes 1 year of historical operating process and lab data
- The Kinetics model for various conversion levels is instituted, which may be unknown to AI model in the event if the same conversions don’t exist in the data
- The AI optimizer is fed with configurable objective functions and required constraints of parameters to perform equipment duties like increasing diesel yield or increasing unit margin
Business Impacts & Outcomes:
- Increased profit due to increase in diesel yield : Avg. $6 Million per year
- Savings in energy/utilities consumption: Avg. $1 Million per year
- Solution enabled the hydrocracker to respond promptly to the dynamic market changes and determine the best optimal product mix.
- Derived more value from the plant, i.e., increased yields, reduced energy / utilities consumption
- Improved life of equipment as well as reduced labor and spare parts opex
