Issue link: https://www.e-digitaleditions.com/i/1475629
14 August 2022 Inhalation to be spherical, having relevant properties updated at every timestep with respect to position, velocity, size, temperature and chemical composition. Ide- ally, the particle moves no further than one cell per timestep, otherwise it would miss the opportunity to exchange mass, momentum and energy with all of the air through which it travels. Coupling equations Coupling equations are solved to calculate the exchange of mass, momentum and energy between the Lagrangian and Eulerian phases, e.g., causing fast particles to slow down, transferring momentum into the gas; cold droplets to heat up, cooling down the surrounding gas; and volatile droplets to evapo- rate and reduce in size while transferring their con- tents into the gas. e coupling can be "one-way" for certain simulations when the aerosol is dilute [10], modeling only the effect of the gas on particles. In Figure 2(a) and 2(b), results from a coupled aerosol spray simulation with the USP-IP throat geometry are shown, illustrating the prediction of the fate of different size particles, where the large particles (> 20 µm) have deposited on the throat walls before the 90° bend. In Figure 2(c), a one-way coupled solution is shown for the same problem: very similar particle behavior is observed, with a 20% reduction in com- putational time. Lagrangian parcels Dense pharmaceutical aerosols contain thousands of microparticles. It is computationally inefficient to model every droplet, particularly given the further reduced timesteps required for fast-moving droplets and particles. A reliable way of greatly reducing com- putational time is to group hundreds or thousands of simulated particles into parcels. ese parcels behave like one particle, with a defined position, size, velocity, temperature, etc., but contain the mass of multiple particles, co-located. In this way, a reduced number of parcels can be simulated, in a reasonable time, to represent the overall aerosol. "Engineering-level" simulation and workflow e "engineering-level" simulation seeks to simulate just enough of the physics of the fluid behavior to provide accurate prediction of trends and relatively accurate prediction of numerical quantities. Typically, engineering-level CFD includes a type of turbulence modeling called Reynolds-Averaged Navier-Stokes (RANS). Turbulent flow is important in several applications relevant for inhaler devices. RANS solv- ers produce a time-averaged prediction of flow, or an ensemble-averaged version for transient flow events. Although this means some turbulent detail will be missed, it is far more computationally efficient. By contrast, simulations in a research environment may use a more complex turbulence-modeling Figure 2 (a) USP-IP center line air velocity and particle size, simulated with a coupled aerosol and 30 L/min gas flow; (b) as image (a) with particles colored by parcel size; (c) as image (a) with a one-way coupled solution—shown by absence of gas jet at throat entry. (a) 3.4e-05 3e-5 2e-5 1e-5 1.3e-06 8.7e+00 6 4 2 3.5e-04 d U Magnitude (b) 3.0e+01 20 10 1.0e+00 8.7e+00 6 4 2 3.5e-04 n Particle U Magnitude (c) 3.4e-05 3e-5 2e-5 1e-5 2.2e-06 4.9e+00 4 3 2 1 3.5e-04 d U Magnitude