AI-based predictive control of waste-fired heat and power plants for improved energy and environmental performance

Purpose and goal

The project aims to develop and demonstrate a module-based, AI-based predictive control system for small-scale waste-to-energy CHP plants (2–10 MW). The new system will enable more efficient, flexible, and environmentally adapted operation than what´s possible today, and will focus on enabling higher electricity yield, lowering emissions and improving overall efficiency. The goal is to strengthen Swedish industrial competitiveness and contribute to Sweden´s energy and climate targets.

Expected effects and result

The solution features a modular AI system operating in parallel with the existing PLC-based control system. While the PLC system ensures safe, stable baseline operations, the AI system optimizes selected control parameters in real time. This enables optimization across the combustion process, flue gas cleaning, and auxiliary energy consumption, while dynamically planning electricity and heat production based on fuel availability, demand, and external variables such as market electricity prices.

Planned approach and implementation

The two-year project has six work packages: project management, data collection, CFD modeling, AI development, industrial validation, as well as utilization and dissemination of results. The objective is to advance the technology readiness level from TRL 2–4 to TRL 7 through demonstration in a real-world environment. The project is led by SiteConcept AB in collaboration with Luleå University of Technology and Kinnarps Holding AB, with support from Sigma Technology and SINTEF Energy Research.