Research on SO2emission prediction model of desulfurization system based on GRU

Limestone-gypsum wet flue gas desulfurization system of thermal power plant is the primary manner to solve SO2 pollution from the flue gas of coal-fired units. However, the system has the characteristics of multi-input variables, high complexity of environmental impact, and time-varying nonlinearity. Therefore, the establishment of accurate system model is the basis of optimal control of wet flue gas desulfurization system. Too many influencing factors will bring interference to the modeling. In this paper, Pearson's correlation coefficient method was accustomed to evaluate the correlation between input vector set and output, so as to select the optimal input variables, and then the optimal variables are smooth-processed as the input quantities of GRU model. Data mining technology and case data model of wet desulfurization system are used. An example on account of the operation information of a power plant shows the model of this method has the advantages of low complexity, high precision and generalization ability.