Prediction of Respimat® Inhaler Spray with an Eulerian-Lagrangian Approach

Prediction of Respimat® Inhaler Spray with an Eulerian-Lagrangian Approach

Authors: Maziar Kakhi, Brent Craven, Steven Chopski, Geng Tian, Ross Walenga

Affiliation: Food and Drug Administration, Silver Spring MD 20993, USA

Background

Currently, orally inhaled drug products typically use either a metered dose inhaler (MDI) or a dry powder inhaler (DPI) for drug delivery to the lungs. The Respimat® device (Boehringer Ingelheim, Ingelheim am Rhein, Germany) is a relatively new inhaler system that is claimed to increase drug delivery to the lungs as compared with MDIs and DPIs. To better understand how differences in the spray characteristics of a potential generic product may affect drug delivery, the U.S. Food and Drug Administration (FDA) is conducting in vitro and in silico research. For this project, an Euler-Lagrange computational fluid dynamics (CFD) model is being developed using OpenFOAM (OpenCFD LTD, Bracknell, United Kingdom) and predictions will be compared with particle image velocimetry (PIV) data.

Methods

Geometry creation of the experimental setup housing the inhaler was performed with SpaceClaim 2019 R2 (ANSYS, Inc., Canonsburg, PS, USA). A predominantly hexahedral mesh for the system was generated using CF-MESH+ (Creative Fields d.o.o., Zagreb, Croatia). A velocity outlet boundary condition was specified equivalent to an applied co-flow rate of 30 L/min. A pressure inlet boundary condition was specified far away from the inhaler’s air vents. Given the spectrum of Reynolds numbers in the various regions of the geometry, initial simulations of the (single-phase) Eulerian flow field were performed to compare the behavior of various momentum transport models. These included laminar, implicit Large-Eddy Simulation (ILES), the wall-adapting local eddy-viscosity variant of large eddy simulation (LES-WALE), and the Langtry-Menter k-ω shear stress transport (k-ω-SST-LM) models. Simulations including the discrete phase were performed using the reactingParcelFoam solver. A two-way coupled approach for simulating the droplets was investigated. For PIV experiments, the laser was placed in-line with the inhaler nozzle. Measurements with and without inhaler actuation are planned to quantitatively characterize the two- and single-phase flows, respectively.

Results

Initial Eulerian phase-only CFD simulations demonstrated unrealistic high velocities at or near the inhaler air vents, which was significantly remedied by adding a fillet to the inhaler’s air inlet vents and selecting solver parameters with limiters. Significant differences in the predicted velocity fields were found with the selected momentum transport models downstream of the inhaler mouthpiece where PIV measurements are being performed. Initial PIV results demonstrate that spray angle is sensitive to the amount of co-flow.

Conclusions Initial CFD simulations of the Respimat inhaler in a box that replicates experimental conditions have helped to identify a balance between the choice of mesh quality, solver settings, and the number of particles per parcel. Space- and time-resolved PIV measurements will further guide the appropriate choice of momentum transport and discrete phase models.

Disclaimer: Views expressed in this abstract do not necessarily reflect the official policies of the Food and Drug Administration; nor does any mention of trade names, commercial practices or organizations imply endorsement by the United States Government.