DEM investigation of mixing indices in a ribbon mixer

Mixing index is an important parameter to understand and assess the mixing state in various mixers including ribbon mixers, the typical food processing devices. Many mixing indices based on either sample variance methods or non-sample variance methods have been proposed and used in the past, however, they were not well compared in the literature to evaluate their accuracy of assessing the final mixing state. In this study, discrete element method (DEM) modelling is used to investigate and compare the accuracy of these mixing indices for mixing of uniform particles in a horizontal cylindrical ribbon mixer. The sample variance methods for mixing indices are first compared both at particle- and macro-scale levels. In addition, non-sample variance methods, namely entropy and non-sampling indices are compared against the results from the sample variance methods. The simulation results indicate that, among the indices considered in this study, Lacey index shows the most accurate results. The Lacey index is regarded to be the most suitable mixing index to evaluate the steady-state mixing state of the ribbon mixer in the real-time (or without stopping the impeller) at both the particle- and macro-scale levels. The study is useful for the selection of a proper mixing index for a specific mixture in a given mixer.