A Soft Sensor Modeling of Cement Rotary Kiln Temperature Field Based on Model-Driven and Data-Driven Methods

Abstract

In the cement manufacturing process, the temperature field in the rotary kiln has a significant influence on the quality of clinker, pollution emissions, and energy consumption. There do not exist methods that can directly measure the temperature field distribution in the rotary kiln. Therefore developing a soft sensor method to predict the temperature field has attracted wide attention. In this paper, we use model-driven and data-driven methods to establish a soft sensor, which combines computational fluid dynamics and multilayer perceptrons to predict the temperature field of the rotary kiln. The soft sensor takes axial air speed, swirling air speed, coal mass flow, material mass flow, secondary air temperature, and x, y, z coordinates as the inputs while the temperature at a certain position inside the kiln as the output. The proposed method has been tested on real industrial data. Compared to CFD simulation, the soft sensor reduces the computation time by three orders of magnitude from 3002.60s to 0.55s. Moreover, the predicted temperature at the kiln tail has a mean absolute percentage error of 6.10%.