Emerson has announced the successful deployment of an AI-driven optimisation solution for Aramco. The collaboration commenced with the integration of Emerson’s Aspen Hybrid Models into Aramco’s existing refinery planning framework, resulting in the creation of one of the world’s largest multi-site, multi-period optimisation models.
By combining first-principles models, deep domain expertise, and purpose-built industrial AI, Aspen Hybrid Models™ capture nonlinear relationships in yield and quality responses, significantly enhancing the accuracy of refinery planning models. The deployment has already achieved yield and quality prediction accuracy of up to 98.5% in key refinery units.
These hybrid AI models have been implemented in Continuous Catalyst Regeneration (CCR) and Platformer Units, where they are enabling more precise feedstock blending, minimising gaps between planning and execution, and improving the accuracy of margin forecasting across Aramco’s global refining network.
Current efforts are focused on expanding the hybrid modelling approach to hydrocracker units across Aramco’s assets. This expansion is expected to further enhance model accuracy and demonstrate the scalability and robustness of this AI-driven optimisation strategy across the enterprise.
“This deployment represents a significant milestone in Aramco’s AI strategy and our long-standing relationship with Emerson,” said Ahmad Alkudmani, director of the global optimiser department at Aramco. “We are committed to leveraging innovative technologies for smarter, more efficient refining optimisation. With improved model accuracy, we are enhancing planning decisions, reducing the manual adjustments required from our engineers, and uncovering new value across our global assets.”
Key benefits Aramco aims to achieve with Aspen Hybrid Models include:
- 98.5% yield and quality prediction accuracy – Substantially increasing yield volume and enhancing stream quality across diverse feedstocks, operating conditions, and throughputs
- Efficiency through optimised feedstock blending – Diversifying feedstock selection and blending recipes to enable more profitable and sustainable operations
- Reduced planning-execution gaps – Minimising discrepancies between plans and actual plant performance, reducing the need for manual adjustments
- Enhanced model accuracy – Capturing complex non-linearities in critical unit operations such as reactors
- Improved operational efficiency – Automating model updates and reducing manual tuning requirements
- Scalable global solution – Maintaining model applicability across a wide range of refinery operations worldwide
Aramco is using Aspen Hybrid Models built and deployed in Emerson’s AspenTech Performance Engineering and Manufacturing and Supply Chain product suites. As a result, Aramco was able to create highly accurate nonlinear optimisations using thousands of converged simulation cases built upon rigorous first-principles models calibrated with actual plant data. The approach provides Aramco with a scalable, robust tool for global refinery planning.
“Aramco continues to set the standard for operational excellence through digital innovation,” said Claudio Fayad, chief technology officer of Emerson’s Aspen Technology business. “This deployment of AI-driven Aspen Hybrid Models to optimise complex, multi-site, multi-period planning workflows demonstrates the tangible value of combining deep domain expertise with advanced AI. We’re excited to expand our strategic relationship with Aramco as they advance their digital transformation goals.”
Discussion about this post