Optimize the NOx emission concentration of Circulation Fluidized Bed Boiler based on on-line learning neural network and modified TLBO algorithm

Abstract

The reduction of NOx emissions concentration of Circulation Fluidized Bed Boiler (CFBB) is an optimization problem, which can regarded as multi-inputs and single output problem. The first priority is to build NOx model. However, the combustion process of CFBB is complicated with strong coupling and nonlinear, etc. So a kind of on-line learning neural network is proposed to build online dynamic model of NOx emission concentration. Based on the established model, a modified teaching learning based optimization algorithm is used to optimize the boiler's parameters. Those parameters impact the NOx emissions concentration seriously. Experiment result shows that the NOx emissions concentration can be reduced by the two methods when the boiler runs with the optimizing operation data.