Intelligent Modeling Method for a Combined Radiation-Convection Grain Dryer: A Support Vector Regression Algorithm Based on an Improved Particle Swarm Optimization Algorithm

Abstract

This paper aims to investigate an intelligent model for a combined infrared radiation-convection (IRC) dryer by using a support vector regression (SVR) algorithm based on an improved particle swarm optimization algorithm (IPSO). The IPSO algorithm was designed to optimize the parameters of the SVR algorithm, which has improved the optimization ability of the IPSO [linear decreasing inertia weight (LDIW)] algorithm and the standard PSO algorithm by introducing a relative fitness deviation (FD) to the LDIW equation and by combining a concept of mutation. Based on the data collected from a practical drying experiment, the best IPSO-SVR (LDIW-FD) model for the IRC dryer was successfully constructed, and the prediction performance comparisons of different modeling methods were also made. The resulting IPSO-SVR (LDIW-FD) model has achieved a remarkable predictive accuracy compared with the other models, demonstrating the effectiveness of the proposed model. In addition, a model of concurrent-counter flow drying was also successfully established by using the same method, depicting the proposed method can be readily used to precisely predict different drying processes. This modeling method can give relatively good predictive output information of the nonlinear system, and it may provide an accurate model for the prediction control of grain drying.