Backmixing studies in tubular reactors—I: The middle injection technique

Abstract

A new technique of tracer injection to evaluate mixing inside tubular reactors with appreciable backmixing has been developed. Three models were tested; the dispersed plug flow model, the finite-stages with backflow model, and a multi-parameter mixed model. The relative response to a middle stimulus is found to be time-independent, independent of type of stimulus but dependent upon reactor geometry and flow conditions. Mixing parameters can be determined from the relative response to a middle stimulus with a minimum of experimental and computational efforts. The decay of the relative response with reactor length is compared to turbulence-decay behind a grid to develop the limits of the use of the middle injection technique. The middle-injection technique shows superiority over moment analysis, or curve fitting in the time or the Laplace domains. This method is also easier than the analysis of steady-state tracer profiles. The technique is useful in predicting the relationship between mixing parameters of different models.