Emerson’s Nikolaos Pratikakis and Michael Painter presented Modernizing Power Plant Control with Aspen DMC3: AI-Enhanced MPC for Efficiency and Reliability at the 2025 Ovation Users’ Group Conference. Here is his session’s abstract.
As power generation assets face increasing pressure to improve efficiency, flexibility, and environmental performance, advanced control technologies have become increasingly essential. Aspen’s DMC3™, AspenTech’s next-generation adaptive model predictive control (MPC) platform, delivers powerful optimization and constraint-handling capabilities across diverse power plant configurations—including combined-cycle gas turbines, coal-fired boilers, and hybrid systems with renewable integration.
This session will explore how Aspen DMC3 applies AI-enabled adaptive modeling, real-time multivariable optimization, and automated tuning to enhance load response, fuel efficiency, emissions control, and equipment reliability. Attendees will learn how features like Smart Tune™, Aspen Maestro™, and the AVA™ virtual advisor simplify deployment, reduce manual engineering effort, and extend controller lifecycle performance. Real-world use cases will demonstrate DMC3’s potential to impact accelerating startup ramps, improving heat rate, reducing thermal cycling, and coordinating complex subsystems across GT, HRSG, and ST units. Whether optimizing a baseload unit or enabling flexible grid-following in a renewables-heavy environment, Aspen DMC3 provides a scalable, software-based solution to elevate operational excellence in modern power plants.
They began by describing the asset optimization digital reference architecture, which spans from forecasting and planning through design, operations, maintenance, and distribution. The Aspen DMC3 Model is used to define process relationships between Independent and Dependent Variables. Its characteristics include a model-based predictive controller, multivariate inputs and outputs, an embedded LP/QP (linear & quadratic programming) optimizer, and constraints are also modeled.
Aspen DMC3 is designed to automatically adapt and optimize for changing process conditions. Its embedded AI handles process nonlinearities and builds seed APC models from historical data combined with adaptive process control to sustain peak plant performance.
The workflow involves seeding the model and then entering the continuous improvement cycle, which includes calibrating the model, updating the model ID, adjusting the controller, reducing the calibration ratio, and then recalibrating the model. The seeding phase uses PID loop tuning, instrument review, controller design, and gathering historical data. The calibrate mode and model ID include multivariate step testing, model quality analysis, and user-triggered model updating.
You can safely manage model mismatches due to ever-changing plant conditions with background testing & model updates by adjusting the calibration ratio from complete testing to full optimization. The DMC3 Smart Tune online strategy view enables quick control adjustments to align with changing market or plant conditions.
The AI-based Aspen Virtual Advisor (AVA) provides guides to the operators on controller actions. It enhances operational decision-making by providing answers to common process control questions & offering actionable guidance.
Aspen Centralized Performance Monitoring monitors this advanced process control (APC) with secure remote access. This capability helps improve operating margins by providing greater visibility into APC performance.
Nico shared the results of a leading chemical company over a three-year period. These results included:
- Increased production capacity by 11%
- Unit variability reduced by 45%
- Eliminated off-spec production when the controller is running and reduced recycling by 20%
- Cut start-up time by 60% on average
- Emissions reduced by 80+%
- Feed rates increased by 8%
Visit the Aspen DMC3 section on Emerson’s AspenTech website for more on these capabilities to drive improved operational performance.
