Residence Time Distribution (RTD) study of a screw reactor, a DEM approach

Screw reactors are widely used in industrial granular processes, yet their residence time distributions (RTDs) remain difficult to predict due to complex flow patterns and sometimes bimodality. Classical RTD models—axial dispersion, PF-CSTR, and Zusatz models—are usually applied but often fail to capture the asymmetry and bimodal behavior of screw-driven granular flows. Using DEM simulations, this work evaluates the accuracy of these literature models and introduces two new formulations: a PF-CSTR sigmoid model that incorporates diffusion effects, and a bimodal model based on radial partitioning of the powder bed. The PF-CSTR sigmoid model consistently outperforms the commonly used literature models with a clear physical interpretability of the fitting parameters, while the bimodal model further excels, achieving fitting errors below one percent. A parametric study on clearance gap, filling rate, and pitch length demonstrates how operating conditions control the transition between unimodal and bimodal RTDs and prove the consistency and the interpretability of the new models. These results provide practical tools for analyzing and optimizing granular transport and reaction in screw reactors.