Experimental Study and Large Eddy Simulation of Mixing Process in Improved INTER-MIG Impeller Classification Tank

Literature Overview

This paper by Lu Yuan and colleagues from Nanjing Tech University, published in the Chinese Journal of Process Engineering in 2015, presents a combined experimental and numerical investigation of the mixing process in a classification tank equipped with an improved double-layer INTER-MIG impeller. The research employs both dye-visualization experiments using high-speed photography and large eddy simulation (LES) with a dynamic Smagorinsky-Lilly subgrid-scale model to characterize turbulent flow and tracer transport within the tank.

Research Methodology

The study adopts a dual approach combining physical experimentation with computational fluid dynamics (CFD) modeling. The experimental component utilizes the dye-visualization technique, where a colored tracer is introduced into the tank and its dispersion is captured by high-speed cameras. The numerical component employs LES turbulence modeling, which resolves the large-scale turbulent structures while modeling the subgrid-scale motions through the dynamic Smagorinsky-Lilly approach.

Key Findings and Analysis

The research identifies several important characteristics of the mixing process in the improved INTER-MIG impeller classification tank:

1. When feed is introduced near the liquid surface, the LES-predicted mixing process shows good agreement with experimental observations, with the tracer dispersing in a spiral pattern. However, the predicted mixing time is larger than the experimental value, indicating that the numerical model may overestimate the diffusion resistance.
2. The optimal feed location is identified as the region near the lower impeller blades, where the mixing efficiency is 17.16% higher compared to feeding near the liquid surface. This finding has direct practical implications for process optimization in industrial classification tanks.
3. The minimum mixing time is observed in the mid-section of the tank body, with mixing time increasing progressively toward both the tank top and tank bottom. This reveals the existence of a mixing gradient along the vertical axis.
4. The bottom region of the improved INTER-MIG impeller classification tank is identified as a mixing-difficult zone, which represents a potential bottleneck for classification efficiency.

Flow Pattern Characterization

The spiral diffusion pattern observed experimentally and predicted numerically is characteristic of the INTER-MIG impeller design, which combines radial and axial flow components. The improved double-layer configuration enhances the interaction between the upper and lower flow fields, promoting better overall mixing. The LES simulation captures the large coherent vortical structures responsible for the spiral tracer movement, validating the model's ability to resolve the dominant flow features.

The 17.16% improvement in mixing efficiency when feeding near the lower impeller blades can be attributed to the stronger turbulent energy dissipation in this region. The lower impeller generates more intense local turbulence due to its proximity to the tank bottom and the interaction with the impeller wall clearance. This finding suggests that feed location optimization can significantly improve process performance without requiring additional energy input.

Engineering Practice Implications

For engineers designing or operating classification tanks in mineral processing applications, this study provides valuable guidance on feed location optimization. The identification of the bottom region as a mixing-difficult zone highlights the need for supplementary agitation or modified impeller design in this area. The quantitative improvement of 17.16% in mixing efficiency from optimal feed positioning represents a substantial process enhancement that can translate into improved classification efficiency and reduced energy consumption.

The validation of LES modeling against experimental data establishes confidence in using CFD as a design and optimization tool for mixing tank applications. However, the noted discrepancy in predicted mixing time suggests that further refinement of the subgrid-scale model or mesh resolution may be necessary for quantitative accuracy in industrial design applications.

Study Insights and Outlook

This research exemplifies the power of combining experimental observation with advanced numerical simulation in understanding complex multiphase mixing processes. The identification of specific mixing-difficult regions and the quantification of feed location effects provide actionable insights for process engineers. Future work could explore the effects of solid particle concentration on mixing patterns, as the current study focuses on tracer dispersion in a single-phase liquid system.