Application of CFD and Artificial Intelligence for Prediction of Biomass Particle Burnout and Residence Time in the Reactor

In planning the development of the energy sector, increasing attention is paid to renewable energy sources, such as biomass. The process of (co)combustion of biomass in boiler furnaces is extremely complex with many coupled parameters. Because of that, the development and application of computational fluid mechanics and artificial intelligence are approached, as efficient tools for the analysis of physical and chemical processes that take place during combustion. The paper presents the applied CFD code and the methodology of application of adaptive neuro-fuzzy systems (ANFIS) in the field of machine learning for predicting the biomass particle burnout and residence time in a 150 kW reactor. Test cases for combustion of three types of pulverized biomass with different diameters and shape factors were considered. A database with the values of mass burnout and residence time of particles was obtained by means of numerical simulations using the in-house developed computer code. The results of ANFIS application on the formed base indicate the possibility of a reliable assessment of mass burnout and residence time of particles, based on knowledge of the type, diameter and shape factors of the fuel introduced into the furnace. The presented models represent a good basis for the implementation and application of CFD and ANFIS models at various thermal energy plants, in order to assess the efficiency of fuel combustion in the furnace.