Scale-up prediction of supercritical CO2 circulating fluidized bed boiler based on adaptive PSO-SVM

Abstract

Supercritical CO2 (S-CO2) circulating fluidized bed (CFB) boiler has broad prospects in the field of coal-fired power generation because of its high combustion efficiency, compact structure, and low pollution emission. Scale-up regularity research based on combustion characteristics of S-CO2 CFB boiler is the key to make it a wide range of industrial applications. In order to simplify the complicated workload and save the huge time cost of numerical simulations, it is of great significance to accurately make the scale-up prediction according to the operating performance of the S-CO2 CFB boiler. This study proposed a scale-up prediction model corresponding to the combustion characteristics of S-CO2 CFB boiler based on adaptive particle swarm optimization support vector machine (APSO-SVM). The parameters of the particle swarm algorithm were processed adaptively firstly combined with the boiler characteristics, and then the adaptive particle swarm algorithm was integrated with the support vector machine to solve the precocious convergence problem of particles being easy to gather in a certain position in the process of pattern recognition. The novel method effectively predicts the boiler in the scaling process from the aspect of boiler capacity, optimizes the scale-up regularity expression by numerical simulations, greatly saves time cost and applicability of enlarge design by altering complex numerical simulations, and lays the application foundation of S-CO2 CFB boiler in the industrial field with acceptable operation accuracy.