Original Research Paper Mixing experiments in 3D-printed silos; the role of wall friction and flow correcting inserts L.A. Fullard a,⇑ , A.J.R. Godfrey a , M.F. Manaf b , C.E. Davies b , A. Cliff b , M. Fukuoka b a School of Fundamental Sciences, Massey University, New Zealand b School of Food and Advanced Technology, Massey University, New Zealand article info Article history: Received 2 May 2019 Received in revised form 19 December 2019 Accepted 20 February 2020 Available online xxxx Keywords: Silo mixing Hopper inserts Mass flow Funnel flow Particle processing abstract The mixing of particles during silo discharge is of high industrial significance, yet quantitative studies of this process in literature are few. Here we present experimental measurement of particle mixing in 3D- printed silos during complete discharge, accounting for hopper half angle, the wall condition, and the addition of flow correcting inserts. Mono-sized mustard seed are dyed one of two colours (yellow or blue) to enable them to be distinguished during a step-change experiment. The silo is partially filled with yel- low particles, followed by an equal mass of blue, then discharged onto a rotating table apparatus which automatically collects particle samples for proportion analysis. The experiments are repeated for six hop- per half angles, and for three different wall conditions of different smoothness. Finally, three flow correct- ing inserts are added to a silo to quantify their influence. It is found that, while the wall condition has a large influence on the mixing and the size of the stagnant ‘‘pocket” region (the non-flowing region in a funnel-flow silo), the addition of flow correcting inserts has the greatest effect in reducing mixing and pocket region size. Our method of completely discharging the silo has the advantage that the pocket region mass can be easily quantified, in contrast to classical continuous flow residence-time experiments. Ó 2020 The Society of Powder Technology Japan. Published by Elsevier B.V. and The Society of Powder Technology Japan. All rights reserved. 1. Introduction The storage of granular materials in silo structures is common in the food and manufacturing, mining, pharmaceutical, and agri- cultural industries [1,2]. Often, to retrieve the stored material, the silo is emptied from an opening at the bottom, the flow itself driven by gravity. There is a vast literature describing aspects of this gravity driven silo discharge, including mapping flow regimes (mass, funnel, core flow) [3–6], silo and hopper design [7,8], influ- ence of particle and wall properties such as friction and cohesion [1,9], and the effect of flow correcting inserts on the dynamics [10–14]. One aspect of silo discharge for which there remains a need for further investigation is mixing. The mixing of dense gran- ular materials during flow can be extremely complicated, yet con- trolling mixing offers the potential of large cost saving and quality improvement for the granular processing industry [15,16]. For example, granular materials are known to ‘‘segregate” where there is dispersity in particle size, density, or other physical properties, and this segregation is enhanced with increasing shear rate on the material [17,18]. Yet, even in the mono-sized case (and mono-density, etc), where segregation is not present, ‘‘advective” mixing of granular materials due to velocity gradients in a silo is not well studied or understood [19,20]. In previous studies, the topic of advective mixing was described using step-change resi- dence time distribution (RTD) numerical experiments [21–23]. The idea of the step RTD experiment is that at some point in the silo there is a step change (say, from 0 to 1) in some measurable quantity (tracer). As the silo is discharged, the exiting material is sampled and the concentration of the tracer is measured as a func- tion of time (or discharge mass/volume). The shape of the resulting tracer concentration curve is then used as a proxy to describe the mixing properties of the silo. The RTD has also been studied for a cone-in-cone blender system to predict product quality [24]. Often, these RTD experiments are expensive and challenging to under- take, hence, most silo RTD studies have been numerical models, rather than physical experiments. One disadvantage of RTD exper- iments is that they are performed in a steady continuous flow regime. Therefore, any stagnant zones present in the silo are never discharged and are unable to be quantified. Mixing is also greatly affected by the silo flow regime. In a mass flow silo, all of the mate- rial contained within is in motion (although at different velocities) and, although advective mixing occurs, it is much less than in the funnel flow regime. In this funnel flow regime, there are stagnant https://doi.org/10.1016/j.apt.2020.02.024 0921-8831/Ó 2020 The Society of Powder Technology Japan. Published by Elsevier B.V. and The Society of Powder Technology Japan. All rights reserved. ⇑ Corresponding author. E-mail address: L.Fullard@Massey.ac.nz (L.A. Fullard). Advanced Powder Technology xxx (xxxx) xxx Contents lists available at ScienceDirect Advanced Powder Technology journal homepage: www.elsevier.com/locate/apt Please cite this article as: L. A. Fullard, A. J. R. Godfrey, M. F. Manaf et al., Mixing experiments in 3D-printed silos; the role of wall friction and flow cor- recting inserts, Advanced Powder Technology, https://doi.org/10.1016/j.apt.2020.02.024