Session A02: Industrial Applications: General

Microfibre filtration in washing machines

Fibres from our clothes make up around 35% of all of the microplastics in our oceans, and it is estimated that each person produces on average 243g of microplastic fibres per year when washing their clothes in a standard washing machine. Thus, we need to find ways to reduce the number of microplastics, in the form of microfibres, that are drained with the wastewater during laundry. Working in collaboration with the home appliance brand, Beko plc, we are exploring simple mathematical models to capture microfibre shedding and filtration within a washing machine. In this talk, we focus on exploring the concept of removing these microfibres from dirty fibrous water using ricochet separation, a solid–fluid separation technique used by manta ray fish to separate plankton from water. Our aim is to understand the interplay between fluid mechanics and clothing materials, using homogenization techniques, so that we can maximise filtration efficiency within a washing machine.   

Investigation of the complex 3D flow structure within a front opening unified pod (FOUP) semiconductor wafer carrier

Minimizing moisture infiltration into a front opening unified pod (FOUP), which is a semiconductor wafer carrier, is crucial in suppressing surface defects. Moisture reacts with airborne molecular contaminants (AMCs) on the wafer surface, creating residual particles that cause defects. Wafers are transported between various processes by the FOUP and enter through the equipment front end module (EFEM). During this procedure when the door is open, the FOUP is inevitably exposed to moist air coming from the EFEM. Thus, nitrogen gas is used to purge the FOUP to reduce humidity and AMCs. To enhance purge performance, it is essential to understand the complex flow structures within the FOUP-EFEM system. This study analyzes the 3D flow structure within the FOUP-EFEM system using magnetic resonance velocimetry (MRV). Additionally, we conducted a parametric study for various flow conditions using computational fluid dynamic (CFD) simulations. As a result, this study provides a deeper understanding of the purge flow structure and identifies effective methods for defect prevention.   

Flow characteristics of a semiconductor fabrication plant using large eddy simulation

We investigate the flow characteristics of a semiconductor fabrication plant (FAB) using large eddy simulation with a dynamic-coefficient Vreman model. An immersed-boundary method is used for no-slip boundary condition at solid surfaces, and the momentum sink terms are included to model the pressure drop across the perforated floor of main FAB. The Reynolds number based on the velocity of fan-filter units (FFUs) and the mean height of machines is 67000. Inside the main FAB, the downward flows are dominantly induced by FFUs. However, upward flows are locally observed between FFUs, which are mainly due to the flow entrainment near jet flows induced by FFUs. Also, upward flows are found near the lateral ends of main FAB due to the presence of unperforated floor and wall. Below the main FAB (sub-FAB), the velocity magnitude is minimum at the middle region and maximum near the recirculation duct, because of the downward flow from the main FAB. Finally, we investigate the dispersion of air pollutant by varying source position, and the results will be provided during the presentation.   

Experimental investigation of turbulent flow characteristics in a cross-flow fan

Cross-flow fans are utilized in applications such as air-conditioner indoor units due to their advantage of generating uniform airflow across the entire length with a compact design. Compared to axial or centrifugal fans, cross-flow fans have a significantly longer spanwise length and a structural characteristic where the impeller consists of multiple blocks, leading to turbulent flow variations in the spanwise direction. However, cross-flow fans are a relatively recent innovation compared to conventional fans, resulting in a relative scarcity of research and data on their flow characteristics. In this study, we aim to investigate how turbulent flow characteristics change in the spanwise direction by experimentally measuring the flow around a cross-flow fan used in an actual air-conditioner indoor unit. The flow field of the cross-flow fan was measured in the region from the center to the edge of the impeller using Particle Image Velocimetry (PIV). The flow rate discharged from the joint disk (the boundary between impeller blocks) was lower than that from the center of the blocks. Additionally, the joint disk increased flow instability around the stabilizer compared to the block center. More details will be discussed in the presentation.   

Numerical investigation on the aerodynamic effects of underbody blockage ratio on high-speed train using full-housing bogie

Most of the air resistance in high-speed train (HST) is generated by the underbody, which includes the bogies [1]. With international interest in increasing the speed of HST, significant research is being conducted on improving bogie designs [2].

In this study, the shape of full-housing bogie (FHB) was suggested, and its effects on the underbody flow field of HST were numerically examined. FHB was designed based on the underbody blockage ratio (BR) with consideration of gauge limits of vehicle. 3 designs of FHB with BR = 0%, 50%, and 75% were selected to compare drag coefficient and near-train velocity magnitude with the conventional bogie.

KTX-Cheongryong, the next-generation HST of South Korea, was used as a baseline model with the speed of 400 km/h. Although the installation of FHB increased the underbody BR, the aerodynamic performance such as mitigating the wake turbulence, reducing shear stress due to the slower growth of the boundary layer, was improved compared to the conventional bogie. In particular, FHB with BR = 50% showed 13.3 % lower air resistance than the conventional bogie.   

Assessment of crosswind safety of a 400 km/h high-speed train through reduced-scale wind tunnel testing.

South Korea is conducting research to operate a high-speed train (HST) with a maximum speed of 400 km/h in 2034 [1]. As the speed of a HST increases, the possibility of overturning due to crosswind increases [2], so it is necessary to assess crosswind safety when introducing a new HST. 

In this study, the aerodynamic coefficient was measured to evaluate the crosswind safety of the HST through reduced-scale wind tunnel test. Wind tunnel testing was conducted in Low-Speed ​​Wind Tunnel (LSWT) of Korea Aerospace Research Institute, where the specifications were satisfied with the requirements in EN 14067-6. 4.3% reduced-scale model being designed with a maximum speed of 400 km/h was installed on a split plate equipped with turntable located 1 m above the floor of the test section to prevent boundary layer disturbance. The turntable was rotated from 0° to 90°, and the tests were performed at the maximum wind speed of LSWT. Forces and moments were measured by using a 6-axis load cell, and the rolling moment coefficient around the leeward rail (C_mx,lee) to evaluate the crosswind safety was calculated. As a result, C_mx,lee of the present HST model showed 8% lower than that of KTX-Cheongryong, currently in commercial operation with a maximum speed of 320 km/h, implying that the crosswind safety has been improved for the newly designed model.   

A Semi-Simulation Method Integrating CFD and Thermal Resistance Networks for Tubular Heat Exchangers

Tubular heat exchangers are essential in various industrial applications and often require costly CFD analysis due to the use of thousands of thin-walled tubes. This study proposes an efficient semi-simulation method that combines a thermal resistance network with CFD via outer wall surfaces, thereby reducing computational amount. In the semi-simulation method, the flow outside the tubes is simulated firstly with given temperature at the outer surfaces of the tubes. The calculated wall heat flux is in turn employed as boundary condition for the thermal resistance network. The thermal resistance network models heat transfer within the circular tube walls and the tube-side flow. The importance of radial, axial and circumferential heat transfer within the tube wall varies depending on the tube material and the flow type (laminar or turbulent) and must be considered properly. An empirical relation for the Nusselt number of the tube-side flow and the Newton's cooling law are employed to capture the heat transfer rate between the inner wall surface and the fluid inside the tube. The semi-simulation method is validated against a full-simulation of a 16-tube cross-flow heat exchanger. Under high accuracy requirements (1% error in the fluid bulk temperature), our method saves 11% of CPU time compared to the full simulation.   

Exploring Flow Dynamics: A Computational Investigation of Spiral Air Jet Mills

This work focuses on analyzing and optimizing air jet mills, which are widely used for fine-grinding powders to sizes under 10 microns. Air jet mills typically have an efficiency of 2% to 4%, so even small improvements can save significant energy. Spiral air jet mills are particularly notable for their lack of moving parts, low-temperature rise during grinding, better temperature control, minimal maintenance, and ability to classify ground material, setting them apart from other machines.

We studied gas flow in an air jet mill using CFD with OpenFOAM. The spin number, the tangential to radial velocity ratio, determines particle cut size by balancing inward drag and outward centrifugal forces. Simulations with a simpler geometry of air jet mill - wherein the entire curved surface of the outer cylinder is an inlet instead of jets. Studies on various design and operational parameters were carried out. Results showed the emergence of toroidal recirculation zones formed similar to the Taylor-Couette flow.

In the base case, simulations with the chamber's curved surface as a rotational inlet, airflow reached a steady state. Recirculation zones strengthened with higher spin numbers, concentrating maximum velocity near chamber walls and causing misclassification. The spin number showed a quadratic relationship with zone size. While laminar and turbulent streamlines differed slightly, recirculation zone enlargement was evident in both, with distinct velocity profiles. Flow patterns with jets were complex due to many dead zones inside the mill. The variation of the complex flow with the parametric variation is also studied. The study is useful in optimizing the dead zone size to balance the throughput and product size tread-off, needed for a particular application.    

Development of a multiphase PIC model for slurry flow modeling

MFIX-Exa is a recently released multiphase CFD code originally developed for the simulation of particle-laden gas-solid flows. Due to its high-performance computing capabilities, MFIX-Exa is an ideal candidate for scale-up studies of slurry reactors, specifically the coarse-grained particle-in-cell (PIC) model with its statistical treatment of the particle phase. Unfortunately, several physical models that were neglected during original development because they are not relevant for high-density ratio gas-solid flows are important in slurry flows where the particle-to-fluid density ratio is near unity. In this preliminary work we focus on the effective (suspension) viscosity. The models of Brinkman (1952), Krieger and Dougherty (1956), and Cheng and Law (2003) are considered. The impact of the effective viscosity model is studied on horizontal pipe flow. The experimental data of Gillies et al. (2002) is used to assess the pressure drop predictions.