Research on SCR Precise Ammonia Injection Control Based on BP Neural Network

Nowadays, selective catalytic reduction (SCR) is usually used in denitration optimization strategy of thermal power units. It can remove nitrogen oxides from flue gas. This denitration optimization system is influenced by chemical reaction and flue arrangement, and it has the characteristics of great inertia and great delay. Nitrogen oxides are usually measured by CEMS analysis instruments. They produce pure delay in flue gas extraction and analysis. Therefore, it is difficult to have good control effect on denitration system only by using traditional PID controller feedback regulation. Based on operation parameters of denitration system, this paper uses BP neural network algorithm to predict the export nitrogen oxide concentration. It is used to calculate ammonia quantity and realize precise control of ammonia injection.