Multi-objective optimization of boiler combustion system based on NSGA-III algorithm with big data technology

Abstract Boiler combustion system optimization is crucial for both environmental and financial reasons. The study suggests an enhanced non-dominated sorted genetic multi-objective optimization method based on big data technology with the aim of increasing boiler thermal efficiency and lowering emissions. By analyzing the historical operation data of thermal power units, the improved K-means method of density clustering is used to optimize the initial clustering center, and the Apriori algorithm is used to mine the economic association rules. In the multi-objective optimization stage, fuzzy C-mean clustering and gray relational analysis are combined to form a two-stage optimization solution framework. The performance test results indicated that the proposed model of the study was able to reach the Pareto optimal frontier in the first 20 iterations. Furthermore, the running time on both datasets was within 15 s, which was significantly better than the other three comparison models. In the simulation experiments, the highest thermal efficiency of the boiler of the proposed model was 96.45%, and the lowest NOx emission concentration was 189.64 mg/Nm 3 . Meanwhile, the spread value of the model was 0.896. The outcomes reveal that the proposed model can provide a more accurate and stable multi-objective solution set for boiler combustion optimization, which is suitable for the practical application scenarios of improving the operating efficiency and environmental performance of thermal power units.