(276g) Impact of Various Feed Configurations in Continuous Mixing of Segregating Powder Mixtures

Powder mixing is employed in many industries to produce pharmaceutical tablets and capsules, food products, metal powders, cosmetics... The mixing operation has long been carried out in batch but more and more companies are moving towards continuous mixing to avoid stocks and waiting times between processing steps, as well as to ensure an easier change in production volumes and to control quality in real time. However, the homogeneity of the mixture, and therefore its conformity, cannot be guaranteed during transient states, especially during start-up.

A phenomenon of segregation between two powders of different grain sizes has been observed in the GCM500 industrial continuous mixer by Ammarcha [1]. This author found out that at mixer’s start, the first particles that come out of the mixer are mainly coarse particles. Then, the composition of the mixture at the mixer’s outlet gradually reaches the composition of the feed. This means that during the start-up phase, fine particles accumulate in the mixer. Indeed, the fine particles percolate to the bottom of the mixer while the coarse particles roll over the powder bed and reach the outlet quicker. It was suggested that a change in feed configuration could be pointed out as a major improvement to the process.

The objective is to study the impact of feed configuration on the mixing process for different rotational speeds and mixture compositions. The different feed configurations consist in shifting one or both feeds in space and/or in time. This intensification of the process aims at reducing the start-up time. The mixing of two powders of different grain sizes is performed in a lab-scale modular continuous mixer, based on the GCM500’s geometry. The chosen powders are semolina and couscous because they are free flowing powders with a trend to segregate and are of the same chemical composition. Couscous, whose grains are larger, is colored in black. In order to measure the homogeneity of the mixture in-line, this one is poured on a conveyor belt at the outlet of the mixer and is analyzed by an image analysis system. The video is decomposed into a succession of grayscale images, which makes it possible to calculate the fraction in white (semolina) pixels for each image. The variance calculated on ten successive values of fraction in white pixels represents the intensity of segregation. Thanks to the sliding variance, it is possible to characterize the homogeneity of the mixture over time. Thus, the image analysis allows us to determine the time, denoted by t₁, for the first grains to exit the mixer and the time of the transient state, noted t₂, that refers to the time for which the fraction in white pixels becomes constant. We can then calculate Δt = t₂ - t₁ which refers to the time during which the mixture is not homogeneous. Those three time-parameters, t₁, t₂ et Δt, will allow to compare the influence of feed configuration on start-up for different operating conditions.

First, the influence of feed composition and rotational speed on t₁, t₂, and Δt are studied with a conventional feeding. The results are in agreement with Ammarcha's observations that, under certain conditions, semolina accumulates at the bottom of the mixer’s inlet during startup. Then, shifted feeding is studied by shifting one or both feeders in the middle or at the outlet of the mixer. The results show that it would be interesting to also shift the feeding in time by starting one feeder later than the other. Mixing at start-up could also be improved by staging the couscous feeding along the mixer so that the couscous grains are incorporated into the mixture progressively in order to reduce the start-up time. For this, a staging device has been developed to divide the powder flow coming out of a feeder along the length of the mixer and will soon be tested.

[1] Ammarcha C., Gatumel C., Dirion J.L., Cabassud M., Berthiaux H., Continuous powder mixing of segregating mixtures under steady and unsteady state regimes: Homogeneity assessment by real-time on-line image analysis, Powder Technology, 315 (2017), pp 39-52