Distillation control configuration selection based on product variability prediction

A novel technique is presented to evaluate control configuration selection for distillation columns. The technique uses process disturbance data and applies signal-processing techniques to extract its amplitude and frequency information. This information is combined with the closed-loop Bode plot of the process to predict the closed-loop product variability for the column. The closed-loop Bode plot is developed using a linear dynamic tray-to-tray model of the process. The approach is demonstrated using two distinctly different distillation columns by identifying the best control configurations from among nine potential configurations for each column. The limits of this approach for control configuration selection are also analyzed.