The nasal epithelium is an important target for drug delivery to the nose and secondary organs such as the brain via the olfactory bulb. For both topical and brain delivery, the targeting of specific nasal regions such as the olfactory epithelium (brain) is essential, yet challenging. In this study, a numerical model was developed to predict the regional dose as mass per surface area (for an inhaled mass of 2.5 mg), which is the biologically most relevant dose metric for drug delivery in the respiratory system. The role of aerosol diameter (particle diameter: 1 nm to 30 µm) and inhalation flow rate (4, 15 and 30 L/min) in optimal drug delivery to the vestibule, nasal valve, olfactory and nasopharynx is assessed. To obtain the highest doses in the olfactory region, we suggest aerosols with a diameter of 20 µm and a medium inlet air flow rate of 15 L/min. High deposition on the olfactory epithelium was also observed for nanoparticles below 1 nm, as was high residence time (slow flow rate of 4 L/min), but the very low mass of 1 nm nanoparticles is prohibitive for most therapeutic applications. Moreover, high flow rates (30 L/min) and larger micro-aerosols lead to highest doses in the vestibule and nasal valve regions. On the other hand, the highest drug doses in the nasopharynx are observed for nano-aerosol (1 nm) and fine microparticles (1–20 µm) with a relatively weak dependence on flow rate. Furthermore, using the 45 different inhalation scenarios generated by numerical models, different machine learning models with five-fold cross-validation are trained to predict the delivered dose and avoid partial differential equation solvers for future predictions. Random forest and gradient boosting models resulted in R² scores of 0.89 and 0.96, respectively. The aerosol diameter and region of interest are the most important features affecting delivered dose, with an approximate importance of 42% and 47%, respectively.

Particle-laden jet flows can be observed in many industrial applications. In this investigation, the horizontal motion of particle laden jets is simulated using the Eulerian–Lagrangian framework. The two-way coupling is applied to the model to simulate the interaction between discrete and continuum phase. In order to track the continuum phase, a passive scalar equation is added to the solver. Eddy Life Time (ELT) is employed as a dispersion model. The influences of different non-dimensional parameters, such as Stokes number, Jet Reynolds number and mass loading ratio on the flow characteristics, are studied. The results of the simulations are verified with the available experimental data. It is revealed that more gravitational force is exerted on the jet as a result of the increase in mass loading, which deflects it more. Moreover, with an increase in the Reynolds number, the speed of the jet rises, and consequently, the gravitational force becomes less capable of deviating the jet. In addition, it is observed that by increasing the Stokes number, the particles leave the jet at higher speed, which causes a lower deviation of the jet.

Super-resolution reconstruction of turbulent flows using deep learning has gained significant attention, yet challenges remain in accurately capturing physical small-scale structures. This study introduces the Conditional Enhanced Super-Resolution Generative Adversarial Network (CESRGAN) for reconstructing high-resolution turbulent velocity fields from low-resolution inputs. CESRGAN consists of a conditional discriminator and a conditional generator, the latter being called CoGEN. CoGEN incorporates subgrid-scale (SGS) turbulence kinetic energy as conditional information, improving the recovery of small-scale turbulent structures with the desired level of energy. By being aware of SGS turbulence kinetic energy, CoGEN is relatively insensitive to the degree of detail in the input. As shown in the paper, its advantages become more pronounced when the model is applied to heavily filtered input. We evaluate the model using direct numerical simulation (DNS) data of forced homogeneous isotropic turbulence. The analysis of Q-criterion isosurfaces, energy spectra, and probability density functions shows that the proposed CoGEN reconstructs fine-scale vortical structures more precisely and captures turbulent intermittency better compared to the traditional generator. Particle-pair dispersion simulations validate the physical fidelity of CoGEN-reconstructed fields, closely matching DNS results across various Stokes numbers and filtering levels. This paper demonstrates how incorporating available physical information enhances super-resolution models for turbulent flows.

This paper introduces a deep learning-based super-resolution framework specifically developed for accurately reconstructing high-resolution velocity fields in two-way coupled particle-laden turbulent flows. Leveraging conditional generative adversarial networks, the generator network architecture incorporates explicit conditioning on physical parameters, such as effective particle mass density and subgrid kinetic energy, while the discriminator network is conditioned on low-resolution data as well as high-frequency content of the input data. High-fidelity direct numerical simulation datasets, covering a range of particle Stokes numbers, particle mass loadings, and carrier gas turbulence regimes, including forced- and decaying-turbulence, serve as training and testing datasets. Extensive validation studies, including detailed analyses of energy spectra, probability density functions, vorticity distributions, and wavelet-based decomposition, demonstrate the model's accuracy and generalization capabilities across different particle parameters. The results show that the network utilizes particle data, mainly in the reconstruction of high-frequency details modulated by particles. Additionally, systematic assessment of the model's performance in capturing previously unseen flow regimes further validates its predictive capabilities.