Integrated modeling for roller kiln temperature prediction

Aiming at the complicated mechanism of the sintering process of the roller kiln and the large delay in the temperature change of the kiln, which brings difficulties to the precise temperature control. This paper presents the integrated model of roller kiln temperature prediction combined with the fusion mechanism and data model. Firstly, based on the analysis of the heat transfer mechanism in the kiln, a mechanism model of temperature output is established. Then, considering the dynamic characteristics of the atmosphere outlet temperature in the mechanism model, a local weighted least squares support vector machine model (KPCA-LW-LSSVM) based on kernel principal component analysis is used to dynamically simulate the dynamic characteristics of the atmosphere outlet temperature. And the mechanism model and the data model are effectively combined to establish a integrated model for roller kiln temperature prediction. Finally, the particle swarm optimization (PSO) method is used to identify the integrated model parameters. Combined with the industrial field data, it is proved that the model can effectively reflect the change trend of the temperature of the roller kiln and lay a foundation for optimal control of sintering temperature in the sintering process of roller kiln.