An Implementation of Abrupt Changes Detection and Principal Component Analysis for Monitoring the Thermal Processes in a Cement Rotary Kiln

Abstract

Rotary kilns for the manufacture of cement, lime and other inorganic binders are widely used in the industry due to their technological advantages: high efficiency, increased reliability, and superior quality of the final product. Modern rotary kilns are equipped with infrared cameras for monitoring the processes inside the kiln, surface infrared scanners for monitoring the temperature at the outer shell of the kiln as well as control systems of all components of the kiln -burner, electric drive, materials supply. The complex chemical, physical, transport and thermal processes that occur inside the cement rotary kilns are stationary processes but space-distributed by sectors along the kiln's length. The main idea in this article is estimating the length of the sectors where these processes occur. We propose estimating the sectors length by means of the Abrupt Change Detection Algorithm implemented onto the temperature spectra provided by the surface infrared scanner. The results provided by the implementation of this algorithm are the estimated position of the abrupt changes onto the space series of the temperature at the outer shell of the kiln which is correlated with the processes inside the kiln. The analysis of the statistical correlations between the estimations of the locus and length of the processes' sectors on one side, and the average and the variance of the temperature observations on the other side, allow increasing the accuracy and demonstrating the repeatability of the proposed method. For this purpose, in this article we propose implementing the Principal Component Analysis on the observations of the kiln's outer shell temperature.