DATASET:
In Silico Active Anterior Rhinomanometry - High Fidelity LES CFD Simulations

##########################################################################
AUTHOR:
Sverre Gullikstad Johnsen
SINTEF AS
sverre.gullikstad.johnsen@sintef.no
September 2023

How to cite:
Johnsen, S. G., SINTEF AS (2023).In silico rhinomanometry - A high fidelity LES CFD simulation study [Data set]. Norstore. https://doi.org/10.11582/2023.00126

##########################################################################
GRANT:
Virtual Surgery in the Upper Airways – New Solutions to Obstructive Sleep Apnea Treatment (VIRTUOSA)
Research Council of Norway grant no: 303218
https://prosjektbanken.forskningsradet.no

##########################################################################
The data are provided free of charge under the
Creative Commons Attribution-NonCommercial 4.0 International Public License (CC BY-NC 4.0)
https://creativecommons.org/licenses/by-nc/4.0/

##########################################################################
SIZE OF DATASET: 
Number of files: ~900
Size: ~5TB

##########################################################################
DATA TYPE:
Transient Computational Fluid Dynamics Data and Case-files
ANSYS Fluent R2019 format

##########################################################################
OVERVIEW OF DATA FILES:
There are two time series referred to as Left Open-Right Closed (LO-RC) and Left Closed-Right Open (LC-RO).
For each time series the following file-types are included:
	* cas-files for k-omega SST, pseudo-steady LES, and transient LES simulations (3GB per file).
	* dat-files saved at the end of initialization simulations with k-omega SST and pseudo-steady LES (12 GB per file). 
	* dat-files saved at every 0.1s containing all simulation data from transient LES (16 GB per file).
	* cdat-files saved at every 0.1s for pressure- and velocity-fields (for ANSYS CFD-POST post processing) (1.4GB per file).
	* out-files containing, for each timestep, in tabular format (2GB per file):
		* time step number
		* flow-time (s)
		* time-step size (s)
		* number of iterations per time-step
		* at selected cross-sections and in/outflow boundaries:
			* total mass flow rate (kg/s) 
			* total volumetric flow rate (m3/s)
			* mean value of:
				* pressure (Pa)
				* pressure squared (Pa2)
				* velocity magnitude (m/s)
				* velocity magnitude squared (m2/s2)
				* velocity components (ux, uy, uz) (m/s)
				* velocity components squared (ux2, uy2, uz2) (m2/s2)
		* pressure difference between right and left nostril (Pa)
		* total volumetric flow-rate at the open nostril (ml/s) 

	!!!!!!!!!!!!!!!!
	!!NB! The first 2 seconds of transient LO-RC dat-files are missing.
	!!!!!!!!!!!!!!!!

##########################################################################
DESCRIPTION OF DATA:
The data set contains raw numerical simulation data obtained from Computational Fluid Dynamics (CFD)simulations in ANSYS Fluent R2019, 
using the Large Eddy Simulation (LES) turbulence model.
The simulations are performed on a 3-dimensional geometry based on medical pre-operative CT images of the nasal cavity of an Obstructive Sleep Apnea (OSA) patient.
The geometry model is truncated at the nostrils and at the nasopharynx. 
The pharyngeal boundary was extruded approximately six hydraulic diameters to distance the flow boundary from the region of interest, 
to minimize the boundary condition (BC) influence on the  flow field.
In silico active anterior rhinomanomatry simulations were performed on both sides of the nose by:
	1. keeping one nostril closed and one nostril open (Left Open-Right Closed = "LO-RC", Left Closed-Right Open = "LC-RO")
	2. specifying total pressure BC at the open nostril and wall BC at the closed nostril.
	3. specifying sinusoidal, homogeneous velocity through the pharyngeal flow boundary (Qmax=600ml/s, T=5s).
Nasal cavity walls were assumed smooth, rigid, and no-slip.
The computational mesh was tetrahedral dominated with uniform size function 0.2mm and 5 layers of prismatic cells at the wall boundaries. 
This resulted in 44.7 million grid cells.
Time steps were 10 micro-seconds in the transient simulations.
Before transient simulations were performed, initialisation of the flow fields were performed by:
	1. steady-state simulation with k-omega SST turbulence model.
	2. pseudo-steady, transient simulation with LES turbulence model but constant pharyngeal velocity corresponding to the maximal inhalatory flow rate, for 1 second (flow-time).
The transient simulations were run for 15 seconds (flow-time), corresponding to three full breathing cycles (1.5 million time-steps).