The steady state detection based on outliers identification for sodium aluminate solution evaporation process

In sodium aluminate solution evaporation process for alumina production, the measurement data are not accurate and contain outliers, which makes it difficult to identify the dynamic and the steady-state of the process. Therefore, an adaptive polynomial sliding filter multivariate steady-state detection method based on outliers identification is presented in this paper. Firstly, based on the analysis of the industrial process, suitable variables are carefully selected for steady-state detection. Before steady-state detection, outliers are detected and filled. Then, by adaptively determining the filter window size, an adaptive polynomial filtering method is proposed; the polynomial coefficient is then used to decide whether the measurement data is steady-state. Finally, single-points steady-state detection is fused to realize the multivariate steady-state detection. Simulation studies using the actual alumina evaporation process data show that the adaptive polynomial filtering steady-state detection method combined with outliers identification is effective, which is of great significance to the process modeling and optimization.