Evaluation on a combined model for low-rank coal pyrolysis

Highlights: • A combined model of low rank coal pyrolysis for process design is put forward. • The FG-DVC model with an adjusted van Krevelen diagram can well predict products. • Predication results from the NLP model fall within 95% confidence interval. -- Abstract: Pyrolysis is an initial step of the upgrading lignite that exhibits a structurally complex connection between physicochemical changes and unknown pyrolyzed compounds, which complicates process simulation for downstream processing. Combined the functional group-depolymerization vaporization cross-linking (FG-DVC) model with non-linear programming (NLP) theory would link between coal pyrolysis and process simulation. First, we adjust the range of the van Krevelen diagram and predict the char and volatiles yields from coal pyrolysis using the FG-DVC model. The tar ultimate analysis is then estimated based on mass/element conservation, and the tar group composition is calculated using the NLP model on the basis of the total tar yield and ultimate analysis. Upon completion of these steps, the process simulation and energy consumption distribution of coal pyrolysis is carried out using Aspen Plus. Results show that the FG-DVC model with the adjusted van Krevelen diagram can accurately predict coal pyrolysis products with better performance than that obtained using empirical correlations. Results show that the energy consumption of drying coal was the largest with 653.2 MJ when drying 1000 kg of coal, followed by pyrolysis with 482.2 MJ. The combined coal pyrolysis model, being independent on experiments, can be used for process design.