Session: 09-13-03: Design Analysis and Optimization of Energy Conversion Systems III

Paper Number: 173895

Multi-Objective Cfd Surrogate Optimization of Impeller Blade Design in Iron Electrowinning Processes 

Electrowinning presents a promising and sustainable approach for primary iron production by offering a low-temperature, water-based alternative to traditional high-emission blast furnace methods. This electrochemical process eliminates the need for fossil fuel combustion and enables iron extraction from ore using only electricity, thereby significantly reducing greenhouse gas emissions and aligning with decarbonization goals. A critical challenge in implementing electrowinning at industrial scale is the requirement for high reaction kinetics, which necessitates conducting the process within a narrow inter-electrode channel. While this geometric constraint enhances electrochemical performance by reducing ion diffusion distances and promoting faster charge transfer, it also severely restricts fluid movement, creating mass transport limitations that hinder the efficient supply of reactive species and removal of evolved gas bubbles.

To overcome these limitations, this study explores the use of a rotary impeller to actively enhance convective transport within the narrow channel. The impeller promotes bulk mixing, disrupts stagnant zones, and refreshes the concentration boundary layers adjacent to the electrodes, all of which are essential for maintaining uniform ore particle distribution and minimizing gas-induced blockage. The impeller’s geometric design and operational parameters are carefully tuned to meet the unique constraints of the narrow channel while minimizing energy consumption.

A high-fidelity computational fluid dynamics (CFD) model is developed using ANSYS Fluent. The simulation framework integrates the Volume of Fluid (VOF) method to capture multiphase flow dynamics involving gas evolution and electrolyte motion, and the Multiple Reference Frame (MRF) approach to model impeller rotation. Impeller geometry is defined by multiple parameters—blade number, length, thickness, tip shape, and rotational speed—allowing for extensive design space exploration. The Discrete Phase Model (DPM) is used to track inert particles, representing ore or tracer materials, and evaluate mixing performance using time-resolved Mean Squared Displacement (MSD), Mean Squared Radial Displacement (MSRD), and Dead Zone Ratio (DZR). Mechanical power input is assessed through torque extraction from the rotating region.

To accelerate the iterative design process, a surrogate model is constructed using Gaussian Process Regression (GPR), trained on data generated from CFD simulations. This surrogate model enables rapid evaluation of mixing and power metrics across the multi-dimensional design space and supports multi-objective optimization. Pareto front analysis is employed to identify optimal trade-offs between energy efficiency and mixing effectiveness under narrow-channel operational constraints.

While final optimization and experimental validation are ongoing, this work establishes a comprehensive simulation and data-driven optimization framework for impeller-assisted mass transport enhancement in electrochemical systems. The methodology is broadly applicable to electrochemical reactors operating in confined geometries, such as those used in metal refining, electroplating, and water treatment, where high reaction rates and uniform species transport are critical to process efficiency.

Presenting Author: Ben Xu University of Houston

Presenting Author Biography: Dr. Ben Xu is an Assistant Professor in the Department of Mechanical and Aerospace Engineering at the University of Houston. His research expertise lies in advanced manufacturing and computational modeling of multiphase fluid dynamics. His research has been supported by prominent funding agencies, including NSF, DOE, NASA, and DoD. Dr. Xu is actively involved in promoting STEM education and workforce development initiatives, particularly in clean energy and advanced manufacturing technologies.

Authors:

Shuqi Zhou University of Houston
Santosh Rauniyar University of Houston
Saurav Maraseni University of Houston
Munaiah Yeddala University of Nevada, Las Vegas
Jeremy Cho University of Nevada, Las Vegas
Ben Xu University of Houston