A hybrid model integrating improved flower pollination algorithm-based feature selection and improved random forest for NOX emission estimation of coal-fired power plants

Combustion optimization is an effective technique for energy conservation in power plants. An accurate NOX emission estimation model is crucial for combustion optimization. A novel NOX emission estimation model is proposed that integrates an improved random forest (RF) and a wrapper feature selection based on an improved binary flower pollination algorithm (FPA). Three improvements to the basic FPA are proposed, an elite-selection strategy, a mutation operator, and a dynamic switch probability. Moreover, the optimization precision of the proposed FPA is compared to those of five other optimization algorithms. A feature selection method based on the proposed FPA is employed to identify the optimal inputs for the NOX emission model. In addition, a fine-tuning mechanism to update the population (based on mutual information) is designed and embedded in the feature selection process to speed up convergence. To improve the basic RF, an ensemble pruning algorithm based on a back propagation neural network that can evaluate the importance of each tree and prune "detrimental" trees is developed. With the optimal inputs, the NOX emission model is constructed using the improved RF. The prediction results demonstrate that it is of higher accuracy and better robustness compared to the basic RF and the three other models. The designed feature selection could also decrease the computational complexity and improve the prediction accuracy.