A Temperature Filtering Algorithm for Cement Burning Belt Based on Local Weighted Regression

Abstract

For the problem that burning zone temperature fluctuation of cement rotary kiln is large and the stability control is difficult to implement, a temperature filtering algorithm for cement burning belt based on local weighted regression is proposed in this paper to smooth the filtering of the burning zone temperature signal in the rotary kiln so that the temperature of the burning zone can be better applied to the control of the rotary kiln firing system. First, burning zone temperature can be obtained by visual inspection in a cement enterprise and is processed by using LOWESS. Different window widths are selected, then the temperature data is smoothly processed under the different window widths. By comparing and analyzing the impact of different window widths on the temperature smoothing degree of the burning zone, the appropriate window width is determined. Finally, the result is compared with the traditional moving average filter and the comparison result shows the proposed method has better data smoothing ability, and the smooth burning zone temperature can be better applied to the cement rotary kiln firing system.