By selecting different types of solid waste samples for cement kiln co-pyrolysis experiments, the distribution patterns of sulfur and nitrogen in solid, liquid, and gas products were explored under the influence of different pyrolysis temperatures, residence times, carrier gas flow rate, and heating rate. Based on the principle of Back progatiom （BP) neural network and using Matlab neural network toolbox, a distribution model of pyrolysis product yield for different types of waste under different reaction conditions was established. The input conditions of the model are reaction conditions and sample characteristic parameters, and the output results are the proportion of sulfur and nitrogen in the three-state products. The predicted values of the model are in good agreement with the experimental values, and the accuracy of predicting the distribution of nitrogen and sulfur elements is high; Furthermore time, the composite algorithm was used to optimize the sample characteristic parameters of solid waste, and the sample composition and ratio were optimized. The results met the goal of sulfur and nitrogen control in cement kilns, indicating the feasibility and effectiveness of the model for simulating the pyrolysis process.

Practical Implementation of the Underlying Design for Digital Models in Cement Plants [J]. Cement Technology, 2025, 1(6): 1-8.