Why AI matters at this scale

CHS McPherson Refinery, Inc. is a substantial mid-continent petroleum refinery with a processing capacity of approximately 85,000 barrels per day. Operating since 1948, it transforms crude oil into essential products like gasoline, diesel, and jet fuel, serving a critical role in the regional energy supply chain. As a facility with over 1,000 employees, it represents a capital-intensive operation where efficiency, safety, and reliability are paramount. At this scale—large enough to have significant data generation but not necessarily the vast R&D budgets of super-majors—AI presents a powerful lever to gain a competitive edge, optimize complex processes, and manage escalating operational and regulatory pressures.

Concrete AI Opportunities with ROI Framing

1. Predictive Maintenance for Critical Rotating Equipment: Refineries rely on expensive, continuously operating assets like centrifugal pumps, compressors, and turbines. Unplanned failures cause massive production losses. An AI model analyzing real-time vibration, temperature, and performance data can predict failures weeks in advance. For a refinery of this size, reducing unplanned downtime by just 1-2% can translate to millions in annual saved revenue and maintenance cost avoidance, delivering a rapid ROI on the AI investment.

2. Real-Time Process Optimization: Refining is a complex, interconnected series of chemical processes. AI and machine learning can model these nonlinear relationships far better than traditional methods. By continuously analyzing data from thousands of sensors, AI can recommend optimal setpoints for distillation columns, reformers, and other units to maximize yield of high-value products (like gasoline) or minimize energy consumption (like fuel gas). A yield improvement of even 0.5% across the facility has a direct, substantial impact on the bottom line.

3. Enhanced Safety and Environmental Monitoring: Safety is the highest priority. AI-powered computer vision can monitor live video feeds to automatically detect safety hazards such as vapor leaks, flare smoke opacity, or personnel without proper PPE in restricted zones. Similarly, AI can analyze complex emissions data to predict exceedances and recommend adjustments. This reduces risk, ensures compliance, and avoids potential fines, protecting both people and profits.

Deployment Risks Specific to This Size Band

Companies in the 1,001-5,000 employee band face unique AI adoption challenges. They possess the operational scale to benefit greatly from AI but may lack the extensive in-house data science teams and IT infrastructure of larger corporations. A key risk is integration complexity—bridging the gap between legacy industrial control systems (like DCS and PLCs) and modern AI cloud platforms requires careful planning and often specialized partners. Data quality and governance is another hurdle; sensor data can be noisy, and siloed data historians must be connected. Finally, there is change management risk. Success depends on buy-in from veteran operators and engineers who may be skeptical of "black box" AI recommendations. A phased, pilot-based approach that demonstrates clear value on a single unit is crucial to building trust and scaling solutions across the refinery.