EXPERIMENTAL AND NUMERICAL INVESTIGATION OF MIXING MISCIBLE LIQUIDS WITH HIGH VISCOSITY CONTRASTS IN TURBULENTLY STIRRED VESSELS

Abstract

In the vast landscape of process industries, including food, consumer goods, pharmaceuticals, and paint, mixing is a pivotal process. Complications arise when complex fluids are involved in the process of mixing. In the consumer goods industry, focussing on the production of shampoos and conditioners etc., the highly viscous nature of the ingredients involved (e.g., surfactants and thickening agents), causes uncertainly in the prediction of mixing performance, particularly, the mixing time. Historically, industries have depended on trial-and-error experimentation to ensure the scalability of these processes for high-quality production. This reliance, however, leads to extended production times for new products, increased costs, thereby reduced competitiveness, and increase waste in material and energy. Over the course of this thesis, using an electrical resistance tomography-based technique, we set out to experimentally explore the limitations of mixing time correlations in turbulent regimes. Such correlations have conventionally been used to set a benchmark for mixing performance comparisons in different mixing equipment and configurations. By doing so, a new trend and a deviation from the conventional correlation were observed, resulting in developing a new correlation to compliment and extend the use of older mixing time formulations. The main benefit of having such correlations being that at any given initial mixing geometrical configurations and material properties, the process designer is able to predict the mixing time of a particular mixing process involving a Newtonian viscous liquid additive, in a bulk with lesser viscosity. These experimental measurements set the benchmark required for validating future predictive models. Further experiments were performed to also explore the effects of addition strategies and volume fraction of the liquids involved in a batch mixing process to find potential optimisation rules when it comes to blending highly viscous materials. This thread of work showed promising prospects and novel findings in this area. In addition, in light of the advances in digital manufacturing, computational fluid dynamics has provided powerful tools to potentially mitigate some of challenges within the mixing industry. This thesis leverages high-fidelity simulation models to enhance understanding of batch mixing processes within turbulently stirred vessels. With a particular focus on the simulation of blending two Newtonian miscible liquids with high viscosity contrasts in order to predict their mixing times. These predictions were validated against our measured data and the results showed good agreements and followed the trend established in the developed correlations.

Date of Award	1 Aug 2024
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Amir Keshmiri (Supervisor) & Claudio Pereira Da Fonte (Supervisor)