Multi-scale predictive modeling for achromic power in a Lithopone Calcination process based on EMD

In order to predict achromic power, one of the most important quality indices of the semi-product in a Lithopone rotary kiln, a predictive model based on time series modeling technique was focused on. According to the Lithopone calcination mechanism, the multi-scale feature of the process was introduced. Empirical Mode Decomposition, an adaptive data-driven method, was applied to settle the coupling of different time scales. Autoregressive model was used to model every Intrinsic Mode Function decomposed, and GM(1,1) model was used to model the trend series. By combining these models, a multi-scale predictive model for achromic power was obtained eventually. The simulation results show its promising performance. This approach serves as a motivation for similar processes such as cement kilns and ceramic kilns.