Study on Optimal Control of Desulfurization Process of Fluidized Bed Boiler Based on Neural Network

According to the desulfurization mechanism of the fluidized bed boiler, the factors affecting the desulfurization efficiency are determined, and the weight coefficients are determined according to the contribution of each factor to the desulfurization efficiency. A neural network-based model of the desulfurization system of the fluidized bed boiler is established. The structure, training and simulation capabilities of this model have been deeply and systematically studied. The simulation results show that the neural network intelligent integrated model can simulate and predict the desulfurization efficiency well. The reactor outlet temperature approaches adiabatic saturation temperature, calcium-sulfur ratio and cycle rate are the most significant parameters affecting the desulfurization efficiency. As the adiabatic saturation temperature increases, the desulfurization efficiency decreases, while the desulfurization efficiency increases with the calcium-sulfur ratio increases. Moreover, the use of recycled materials significantly improves the desulfurization efficiency. In addition, the control optimization method has high prediction accuracy and fast convergence speed, and is suitable for industrial processes with severe nonlinearity, especially for industrial processes with many input variables.