Swarm intelligence algorithm on combustion optimization of coal-fired boiler

In this paper, we study the problem of how to improve the boiler combustion efficiency and reduce the NO
_x
emissions. In order to solve the problem, we research on several swarm intelligence algorithms and propose two new algorithms. The first is the improvement of genetic algorithm and the second one is a algorithm which combined with the first algorithm and the traditional particle swarm optimization algorithm. When compared to the existing literatures, our contributions are at least three-folds. First, due to the complexity of the boiler combustion model, the Least Squares-Support Vector Machine (LS-SVM) is applied to the system to predict and get more accurate system outputs. Second, the Improved Genetic Algorithm (IGA) and Particle Swarm Optimization-Improved Genetic Algorithm (PSO-IGA) are proposed to applied to the combustion optimization. Third, we optimize the system both in single-objective optimization and multi-objective optimization. Finally, numerical simulations show that the LS-SVM algorithm has a accurate prediction and the prediction result can use to estimate the system output in the optimization process. The simulation also compare the new algorithms with the traditional algorithms, and shows that the PSO-IGA has the best optimization effect among all the four swarm intelligence algorithm. So, PSO-IGA is a efficient way to improve the combustion efficiency and reduce the emissions.