MyTechDailyNews General: How Forge Plant Cognition Brings Industrial AI Into Refining Operations
Industrial AI is starting to be judged less by dashboards and more by whether it can help a plant make better operational choices. Honeywell and its Honeywell Technologies business have introduced Forge Plant Cognition for that purpose. The platform is assembled with NVIDIA AI and Dell Technologies infrastructure for AI workloads, while using NVIDIA accelerated computing capabilities. It is aimed at refining, petrochemical and LNG operators that want to create more value from existing assets while changing plant operations.
The first public deployment is at Dangote Petroleum Refinery in Lekki, Nigeria, a single-train site described in the announcement as one of the world's largest refineries. The objective stated for Lekki Refinery is to improve throughput while building up local engineering and operations teams. Using preliminary data, Honeywell Technologies said that Forge Plant Cognition's AI-enabled optimization could be worth more than 29 cents a barrel at Lekki Refinery. If the site runs at full-capacity throughput, that level of benefit could mean more than $60 million in gross annual benefit for the facility.
It is built on Forge Cognition, a Honeywell Technologies industrial AI foundation. The goal is to join process technology expertise, industrial automation and AI so plant teams can see performance limits earlier, predict problems sooner and carry useful insight into operational action. The framing is not merely to monitor; it also pairs real-time operating data, engineering knowledge, trusted calculations and physics-based models, with deep process expertise from Honeywell Technologies UOP.
Three operating goals, explained for a tech audience
The platform is organized around three plant-level outcomes rather than a single dashboard metric:
- Recover hidden margin: bottlenecks in linked processes can reveal chances to improve throughput, yield and product value.
- Protect production: emerging equipment and process problems can be detected, with corrective actions recommended to reduce unplanned downtime.
- Scale expertise: hard-to-replace process know-how is placed into AI workflows so operators and engineers can make faster and more consistent decisions.
Refineries are places where small variations in operating conditions can affect efficiency and throughput. Honeywell Technologies UOP is said to bring more than a century of experience in designing and operating refineries, and the platform uses integrated intelligence based on that operating knowledge to identify constraints and recommend actions.
Technology stack in plain terms
Dell AI Factory with NVIDIA is named as the platform infrastructure, and Dell Services is listed as support. The technical pieces named in the announcement are:
| Component | Stated role |
|---|---|
| NVIDIA Nemotron | Open models used for agents that understand and interpret operational data. |
| NVIDIA Omniverse | Libraries used to build a digital twin of the plant and its processes for simulation-based training and planning. |
| NVIDIA Cosmos | Open models used with the NVIDIA Metropolis VSS Blueprint for vision AI agents that support real-time alerts and detection of anomalies. |
| NVIDIA PhysicsNemo | Framework for building, training and fine-tuning efficient physics AI models at scale to support digital-twin optimization workflows. |
| NVIDIA RTX Pro 6000 and NVIDIA RTX Pro 4000 | Named computing components in the announcement; these applications run on NVIDIA RTX Pro 6000 and NVIDIA RTX Pro 4000. |
Why the Dangote deployment matters
Dangote Petroleum Refinery in Lekki, Nigeria, is set to combine operating data with process expertise from Honeywell Technologies and AI-powered optimization. The stated use is to identify constraints, accelerate troubleshooting and support more informed decisions. Beyond throughput, the project is tied to a workforce goal: by placing specialized process knowledge inside AI tools, engineers and operators can access expert guidance, solve problems faster and build capability as Dangote works toward a 70% Nigerian workforce and reduced reliance on external expertise over time.
Capacity growth and the limits of the estimate
Nearly a decade after the two companies began collaborating, they announced that Honeywell Technologies is helping the refinery raise capacity from 650,000 to 1.4 million barrels per day within three years. The stated route is optimization of existing assets and faster market delivery.
Honeywell Technologies also said the more than 29 cents a barrel and more than $60 million figures are not guaranteed. They are specific to this facility, gross of implementation costs, and actual results depend on configuration, operating conditions, crude slate, utilization, baseline performance, data quality, product margins and customer implementation.
FAQ: What general readers should ask
Is this only another AI monitoring tool?
The stated ambition is broader. It is designed to connect real-time operating data, engineering knowledge, process expertise, trusted calculations and physics-based models, then turn that into operational actions.
Why does the first deployment matter?
Lekki Refinery is a large single-train refining site, so the project gives Forge Plant Cognition a visible test bed for throughput, local capability building and measurable operational benefit.
Are the dollar figures a guarantee?
No. The estimate of more than $60 million is tied to this facility and depends on implementation and operating variables.
Common misconceptions
- A plant-level AI platform is not just a dashboard with alerts.
- Industrial AI value cannot be separated from process expertise and data quality.
- One facility's estimate should not be treated as a universal industry benchmark.
- Workforce targets are part of the story, not a footnote; the local-skills angle is tied to the Dangote deployment.
Action advice for readers watching industrial AI
- Check whether the vendor brings operational process knowledge, such as Honeywell Technologies UOP, and not only AI models.
- Look at the infrastructure named in the announcement, such as Dell AI Factory and NVIDIA tools, because plant AI depends on data, simulation and vision workloads.
- Treat early benefit numbers as conditional until they are observed on a comparable asset.
- Watch local workforce claims, such as the 70% Nigerian goal, because industrial deployments often combine technology transfer with operations.
For general-interest readers, the important takeaway is that industrial AI is being tested in a setting where small operating decisions can have large financial consequences. Forge Plant Cognition shows how vendors are trying to combine automation, process expertise and AI infrastructure to support throughput, stability and skills transfer in refining, petrochemical and LNG operations.