Yokogawa and Saudi Aramco Achieve Milestone in Industrial AI at Fadhili Gas Plant

Advancing Industrial Automation with AI

Yokogawa Electric has partnered with Saudi Aramco to deploy multiple autonomous AI systems at the Fadhili natural gas plant. This collaboration marks a historic milestone in applying industrial AI for process optimization. The initiative demonstrates how industrial automation can evolve toward autonomous operations, enhancing efficiency, sustainability, and safety.

Yokogawa and Saudi Aramco implement autonomous AI at Fadhili gas plant, optimizing AGR process with PLC, DCS, and industrial automation technologies.

Autonomous AI for Acid Gas Removal Optimization

The deployed AI solution uses multiple cooperative agents to autonomously control and optimize the Acid Gas Removal (AGR) process. These agents rely on Factorized Kernel Dynamic Policy Planning (FKDPP), a reinforcement learning algorithm developed by Yokogawa and NAIST. By dividing the AGR process into three sections, the AI optimizes each step, achieving fully autonomous control while maintaining operational reliability.

Safety and Integration with Existing Control Systems

To ensure safety, Yokogawa first created a plant simulator to train the AI agents and validate performance. The AI solution then integrated with the CENTUM VP production control system, leveraging existing safety mechanisms. This approach demonstrates how PLC and DCS control systems can coexist with autonomous AI, bridging traditional automation with next-generation intelligence.

Measurable Operational Benefits

Preliminary results at Fadhili show approximately 10–15% reduction in amine and steam usage and 5% lower electricity consumption. Process stability improved, and operator interventions decreased significantly. These gains highlight the potential of industrial AI solutions to enhance resource efficiency while maintaining process safety under dynamic conditions.

Industry Leadership and Strategic Vision

Khalid Y. Al-Qahtani, Senior Vice President of Aramco Engineering Services, stated: “Deploying autonomous AI is part of our strategic plan to improve efficiency and sustainability. Yokogawa’s solution exemplifies our commitment to advanced technologies in the energy sector, supporting operational excellence and shareholder value.”

Kuniaki Harano, CEO of Yokogawa Electric, added: “We are honored to introduce this innovative solution at a key facility. Our vision of transitioning from industrial automation to industrial autonomous operations (IA2IA) is realized, showcasing our leadership in safe and reliable autonomous control.”

About FKDPP and Autonomous AI

Yokogawa defines autonomous control AI as systems capable of deriving optimal control strategies and responding robustly even in unfamiliar situations. The FKDPP algorithm, jointly developed with NAIST, won the 2023 Japan Prime Minister’s Award for Industrial Technology, reflecting its technical excellence and real-world applicability.

About Saudi Aramco

Saudi Aramco is a global integrated energy and chemical company committed to sustainable, efficient, and reliable resource utilization. Its projects span traditional oil and gas operations to advanced energy technologies, supporting global economic growth and productivity.

About Yokogawa Electric

Founded in 1915 in Tokyo, Yokogawa Electric provides advanced measurement, control, and information solutions for energy, chemical, pharmaceutical, materials, and food industries. With 128 offices in 62 countries and over 17,000 employees, Yokogawa leverages digital technologies to help clients optimize production, assets, and supply chains, enabling the transition to autonomous operations.

Practical Application: Autonomous Process Control

The Fadhili AI deployment illustrates how industrial automation, PLC, DCS, and AI systems work together to improve energy efficiency, reduce resource consumption, and enhance operational safety. Companies can replicate such solutions in chemical, energy, and manufacturing plants to drive factory automation toward a fully autonomous future.