(409c) On-Line Measurement and Prediction of the Homogeneity of Binary Powder Mixtures in Continuous Mixing During Transitory Regimes

The mixing of solids is a crucial unit operation in many industries like the foodstuffs, cosmetics, detergents, powdered metals and drugs. This operation is commonly defined as the blending of different powder ingredients in order to obtain a uniform distribution of at least one key ingredient in the mixture. Unlike for the case of a liquid mixture, the homogeneity of a solid mixture cannot be realized spontaneously and the maximum uniformity of the final product must be measured and assessed. The need to control the blending process is therefore the greatest challenge faced today by the powder mixing technology. In these recent years, on-line and real-time methods for assessing the homogeneity of mixtures homogeneity, such as image analysis, and spectroscopic methods, are experiencing a flourishing development. It may also serve to a better understanding of how the mixing of different particulate flows occurs, and how the homogeneity changes with process variables, with essential aim to control the process.

There are numbers of factors influencing the quality of a final blend like physical characteristics the ingredients, chemical compositions, blending equipment, blending methods and finally process variables or conditions. In the present paper, we will concentrate on studying the effect of process variables variations in the homogeneity of binary mixtures in continuous mixing. In particular, we will focus on the effect of step-like variations in rotational speed of the stirrer, inducing a transitory regime in hold-up weight, flowrate and homogeneity. The apparatus used in this work is a pilot scale commercial mixer Gericke GCM500, for which steady-state operation, pharmaceutical mixing and Markov chain flow modelling has been reported earlier by our reserach team. The particles considered are two sieved fractions of semolina, one of them being coloured in black after iodine adsorption. A linear CCD camera is placed at the outlet of the process and measures the content in any of the two ingredients. The scale of scrutiny can be adjusted and mixture homogeneity can be calculated for this scale in real time. Changes of homogeneity, due to mixer's starting period, mixer's emptying or step-like changes in stirrer's rotational speed are recorded by image analysis. A non-homogeneous Markov chain model developped in a previous work allows the prediction of the homogeneity changes during these transitory regimes. The spatial distribution of the hold up weight inside the vessel for each component is also determined in the different operating conditions. It demonstrate the existence of a light particle segregation phenomena that takes place in the first part of the mixer. The model can also predict this effect and may serve as a good process control tool.