Data-driven, Real-time and Integrated Modeling and Multi-objective Optimization of Cement Clinker Burning System

This paper proposes a data-driven framework for real-time integrated modeling and multi-objective optimization of cement clinker burning process for energy conservation and emission reduction. This framework leverages real-time data to adaptively capture operational conditions, and selects frontier data to model the Pareto trade-off relationships effectively. By integrating Radial Basis Function Networks (RBFNs) modeling with the Non-Dominated Sorting Genetic Algorithm II (NSGA-II) optimization, RBFNs capture the conflicting relationships between controlled variables and performance indicators, while NSGA-II is enhanced by initializing its population with the hidden layer centers of the RBFNs. This integrated strategy helps NSGA-II explore high-quality Pareto-optimal solutions by guiding the search toward promising regions in the solution space. Numerical experiments further validate the framework's effectiveness, and application to a real-world 10,000-ton cement production line dataset achieves a 1.67% reduction in average coal consumption and a 5.07% reduction in NOx emissions.