Selected Publications

Mixing of fluids in a coaxial jet is studied under four distinct viscosity ratios, m = 1, 10, 20 and 40, using highly resolved large-eddy simulations (LES), particle image velocimetry and planar laser-induced fluorescence. The accuracy of predictions is tested against data obtained by the simultaneous experimental measurements of velocity and concentration fields. For the highest and lowest viscosity ratios, standard RANS models with unclosed terms pertaining to viscosity variations are employed. We show that the standard Reynolds-averaged Navier-Stokes (RANS) approach with no explicit modelling for variable-viscosity terms is not applicable whereas dynamic LES models provide high-quality agreement with the measurements. To identify the underlying mixing physics and sources of discrepancy in RANS predictions, two distinct mixing modes are defined based on the viscosity ratio. Then, for each mode, the evolution of mixing structures, momentum budget analysis with emphasis on variable-viscosity terms, analysis of the turbulent activity and decay of turbulence are investigated using highly resolved LES data. The mixing dynamics is found to be quite distinct in each mixing model. Variable viscosity manifests multiple effects that are working against each other. Viscosity gradients induce additional instabilities while increasing overall viscosity decreases the effective Reynolds number leading to laminarization of the turbulent jet, explaining the lack of dispersion and turbulent diffusion. Momentum budget analysis reveals that variable-viscosity terms are significant to be neglected. The scaling of the energy spectrum cascade suggests that in the TLL mode the unsteady laminar shedding is responsible for the eddies observed.

Turbulent shear-driven mixing in a coaxial and co-flowing configuration is studied using experiments and computations to understand and model the effect of viscosity gradients in the flow field. Two liquids with a large disparity in dynamic viscosity are mixed, with a low viscosity, high momentum jet directed into a high viscosity, low momentum co-flow in a pipe. Simultaneous experimental measurements of the velocity and concentration fields are made using high-resolution PIV and PLIF to obtain their turbulent cross-correlation statistics for viscosity ratios of 1 and 40. LES simulations are also performed using dynamic mixing sub-grid model to investigate the three dimensional mixing structure of the flow for the two cases. The overall structure of turbulent mixing in the coaxial confined jet configuration is studied to identify the mixing regions of the flow and the effect of viscosity gradients on the dynamics of the same. Besides the effect of Reynolds number between the two cases that manifests as reduced mixing, it was noted that the transport of turbulent kinetic energy and scalar concentration variance shows significant asymmetries that arise from the viscosity gradients in the field. The scalar mixing structure between the two cases is studied in detail with relevance to complex mixing-limited reactions frequently encountered in such environments. It was found that turbulence production and associated scalar mixing is highly skewed in variable viscosity flows, where the low viscosity regions show enhanced turbulence activity. The implications of such turbulence skewness on the chemistry of reaction systems involving variable viscosity environments are discussed in further detail.

The performance of vacuum membrane distillation (VMD) modules can be optimized through careful selection of design parameters. The present study examines how the addition of cylindrical filaments in the feed channel increases momentum mixing and the overall performance of VMD modules under different operating inlet conditions. Three-dimensional transient Computational Fluid Dynamics (CFD) simulations are conducted using Wall-Adapting Local Eddy-Viscosity (WALE) subgrid-scale Large Eddy Simulation (LES) turbulence model. Local concentration, temperature, and flux are coupled at the membrane surface to predict the rate of water vapor diffused through the membrane by Knudsen and viscous diffusion mechanisms. The predicted and measured vapor flux agrees reasonably well; validating the employed model. The small-scale eddies induced by the presence of spacer filaments promote mixing in the module, thus the temperature and concentration polarization is alleviated and the water vapor flux is immensely improved. The insertions of filaments in the feed channel increase the water permeate rate by more than 50% at higher feed flow rates and inlet temperatures. The pressure drop by the spacer reduces the allowable module length by one order of magnitude, but the module length increases two folds at feed temperature 80°C. Even though the power consumption of the module containing the filaments is increased, the addition of filaments is strongly recommended since the required power for the process could be supplied from readily available low-grade heat source.

We develop a machine learning tool useful for predicting the instantaneous dynamical state of sub-monomer features within long linear polymer chains, as well as extracting the dominant macromolecular motions associated with sub-monomer behaviors of interest. We employ the tool to better understand and predict sub-monomer A2 domain unfolding dynamics occurring amidst the dominant large-scale macromolecular motions of the biopolymer von Willebrand Factor (vWF) immersed in flow. Results of coarse-grained Molecular Dynamics (MD) simulations of non-grafted vWF multimers subject to a shearing flow were used as input variables to a Random Forest Algorithm (RFA). Twenty unique features characterizing macromolecular conformation information of vWF multimers were used for training the RFA. The corresponding responses classify instantaneous A2 domain state as either folded or unfolded, and were directly taken from coarse-grained MD simulations. Three separate RFAs were trained using feature/response data of varying resolution, which provided deep insights into the highly correlated macromolecular dynamics occurring in concert with A2 domain unfolding events. The algorithm is used to analyze results of simulation, but has been developed for use with experimental data as well.

We show how a simple design adaptation creates desirable flow structures that dramatically enhance performance in hollow fiber membrane (HFM) water desalination modules using computational fluid dynamics (CFD) simulations. Conventional HFM desalination modules encase thousands of co-axially aligned HFMs in a mutually parallel flow of brackish water. HFMs are subject to a phenomenon called concentration polarization (CP), which leads to fouling and will eventually prevent clean water production. We found that the twisted HFM module mitigates CP effects and increases transmembrane permeate flux by 5–9% for three flow rates considered. Twisted HFM bundles induce swirling flow structures inside desalination modules that increase momentum mixing throughout. Frictional energy losses and increased pumping power associated with this subtle design alteration are small relative to projected gains in clean water production. We predict system performance increases about 70% for the twisted modules herein considered, and there are in principle no additional required components associated with this geometry adaptation. With our findings, we identify how the twisted module design induces desirable flow structures that increase membrane separation performance by mitigating CP effects and increasing HFM efficacy.

Hollow fiber membrane (HFM) modules are among the most common separation devices employed in membrane separation applications. Three-dimensional steady-state computational fluid dynamics (CFD) simulations are carried out to study flow past hollow fiber membrane banks (HFMB). The current study investigates the effects of flow behavior on membrane performance during binary mixture separations. Carbon dioxide (CO₂) removal from methane (CH₄) is examined for various arrangements of HFMs in staggered and inline configurations. The common HFM module arrangement is the axial flow configuration. However, this work focuses on the radial cross-flow configuration. The HFM surface is a functional boundary where the suction rate and concentration of each species are coupled and are functions of the local partial pressures, the permeability, and the selectivity of the HFM. CFD simulations employed the turbulent k–ω shear stress transport (SST) model to study HFM performance for Reynolds numbers, 200 ≤ Re ≤ 1000. The efficiency of the inline and staggered arrangements in the separation module is evaluated by the coefficient of performance and the rate of mass flow per unit area of CO₂ passing across the membrane surface. This work demonstrates that the module with staggered arrangement outperforms the module with the inline arrangement.

Computational fluid dynamics simulation was carried out for three-dimensional desalination modules containing triangular and square ribs attached to the membrane surface. The solution-diffusion membrane transport model was applied along the surface of the corrugated membrane. The membrane flux model which couples the water permeation rate with local salt concentration impose a selective removal of components in the feed channel. The local water flux, salt concentration, wall shear stress, and Sherwood number were monitored over the surface of membranes to determine effects of eddy promoter corrugations on mitigation of concentration polarization and the total water flux. Simulations are conducted using a laminar model for Reynolds number of 100 and k–ω SST turbulence model for Reynolds number of 400 and 1000. Mathematical model and numerical methods employed are validated by comparing predictions against measurements reported earlier. Predicted results agree quantitatively and qualitatively with previous experimental results for membranes with semi-circular cross-sectioned ribs. The results show that corrugated membranes especially triangular chevrons enhance membrane performance profoundly at all flow rates. Water permeation rate is increased, concentration polarization is alleviated, and the potential fouling in the module is reduced by introducing corrugated membranes.

Water desalination by reverse osmosis hollow fiber membrane has been widely used to produce fresh water. This work numerically characterizes flux performance of the membrane, concentration polarization and potential fouling sites in the reverse osmosis desalination module containing hollow fiber membranes arranged in an inline and staggered configuration. Steady k-ω SST turbulence model is utilized to study membrane performance. An accurate membrane flux model, the solution-diffusion model, is employed. Hollow fiber membrane surface is treated as a functional boundary where the rate of water permeation is coupled with local concentration along the membrane surface. The rate of water permeation increases and concentration polarization decreases as the feed flow rate is increased. Hollow fiber membranes in the staggered geometry perform better than those in the inline geometry. It is proven by the present study that desalination modules containing hollow fiber membranes should be designed and optimized by careful consideration of their configurations. It is demonstrated here that flows in the hollow fiber bank becomes strongly time dependent at high flow rates and that transient effects could profoundly influence hollow fiber membrane flux performance and characterization of concentration polarization.

Computational fluid dynamics simulations are conducted for binary fluid flows over banks of hollow fiber membranes. Separation of carbon dioxide (CO2) from methane (CH4) is studied using hollow fiber membranes structured in different arrangements. The membrane is considered as a functional surface where the mass flux and concentration of each species are coupled and determined as a function of the local partial pressures, the permeability, and the selectivity of the membrane. k–ω Shear Stress Transport (k–ω SST) turbulent model is employed to study steady flow over banks of hollow fiber membrane for values of the Reynolds number up to 1000. Lattice Boltzmann method is used to study transient flow pass an array of diamond shaped hollow fiber membranes. The flow structure around the hollow fiber membranes has strong influence on the separation performance. This study demonstrates that good mixing in the bank of hollow fiber membranes enhances separation. The results show that hollow fiber membrane module with staggered arrangements performs much better than that with inline arrangements. It is also demonstrated here that the transient nature of flows has significant influence on the membrane performance.