NSGA-II Based Multi-objective Optimization of NOx Emission and Reheat Steam Temperature

摘要

It is challenging and imperative to reduce NOx emission with increasing reheat steam temperature in particular at lower load condition. To solve this problem, this paper proposes to optimize the NOx emission and reheat steam temperature simultaneously via a modified version of Non-Dominated Sorting in Genetic Algorithm (NSGA-II), in which the predictions of NOx and reheat steam temperature are achieved by a new version of multiple outputs nonlinear partial least squares model (MO-NPLS). To perform the proposed method well, some experiments are conducted on a real 1000MW ultra-supercritical unit. By comparison, the results suggest that the Pareto optimal solutions obtained by NSGA-II performs well with appropriate accuracy and outperform experimental value.