1 of 17

Flow of air inside a�Nasal cavity using SU2 and OpenFOAM

Praveen kumar Krishnapur1 , Hrishikesh Kulkarni2,a) , Mithilesh Maurya2, Dilshad ahmad2

​

1Indian Institute of Technology Hyderabad, Telangana, India - 502285

2Tata Consultancy Services (TCS) Research, 54-B, Hadapsar, Pune, India - 411013

a)Corresponding author: hrishikesh.kulkarni2@tcs.com

​

SU2 Conference 2022

2 of 17

Outline

  • Need for CFD Simulation
  • Problem definition
  • Simulation with OpenFOAM
  • Simulation with SU2
  • Comparison of Results
  • Advantages
  • Conclusions
  • Acknowledgements
  • References

3 of 17

Nasal cavity air flow: Physiological problems and Surgical challenges

Fig 2. Paranasal sinuses

Objective: A steady state CFD simulation for calm breathing level (inlet airflow rates through each nostrils) and evaluate the pressure drop, velocity patterns  across the nasal passage.

Fig. 1 Nasal cavity cross section [1]

Fig 3. Healthy (left) and Unhealthy Sinuses

Fig 4. Sinus after surgery

Frontal Sinuses​

Ethmoid Sinuses​

Maxillary Sinuses​

Sphenoid Sinuses​

4 of 17

  • In recent times, study of airborne particles and their transmission has become increasingly important due to pollution, viruses and their hazardous effects.
  • A detailed study of these transmission routes into our body will not only help in developing ”filtration devices”(Masks, PPE) but also in treatment of the diseases that accompany them.

Need For Simulation (Contd..)

5 of 17

Simulation case: Geometry, Mesh and Physics

  • Anonymous nasal cavity [2] and meshing with SnappyHexMesh for OpenFOAM
  • Flow physics for calm person’s breathing : A single phase, steady, Incompressible, turbulent, viscous

MR/CT Image

3D Model

Nasal 3D Model(NC01)

CAD Geometry Cleaning

Mesh Generation, Checking and Independence

CFD Simulations

Postprocessing of Results

Medical Diagnostic and Decision

Geometry

Mesh

Total Volume: 6.4266E − 5 m3

Cells: 0.48 million

Bounding box dimension: Length: 0.04m, Width: 0.118m and Heigh: 0.105m

​

6 of 17

Simulation case: Geometry, Mesh and Physics

Fluid properties

Density, ρ = 1.2886 kg/m3

Kinematic Viscosity, ν = 1.45*E−5 m2/s

Fig 5. Nasal Cavity STL geometry

Fig 6. Unstructured mesh SnappyHexMesh

7 of 17

Simulation case: Geometry, Mesh and Physics

Reference state conditions of the flow

8 of 17

Simulation with OpenFOAM*

* Open-Source CFD Analysis of Nasal Flows by Balatinec, L., Uroić, T. and Jasak, H., 2021.

Boundary

U BC

Value

P BC

Value

Nasal Cavity

No Slip

0

Zero Gradient

-

Inlet 1

Fixed Value

Uniform (-1.77802336 1.57547665 2.33074331)

Zero Gradient

-

Inlet 2

Fixed Value

Uniform (0.954729319 1.30676162 3.50552511)

Zero Gradient

-

Outlet 1

Zero Gradient

-

Fixed Value

uniform 0

Outlet 2

Zero Gradient

-

Fixed Value

uniform 0

Boundary

K BC

Value

Omega BC

Value

Nasal Cavity

Wall Function

Uniform (0.10)

Wall Function

Uniform (63.56)

Inlet 1

Fixed Value

Uniform (0.10)

Fixed Value

Uniform (63.56)

Inlet 2

Fixed Value

Uniform (0.10)

Fixed Value

Uniform (63.56)

Outlet 1

Zero gradient

-

Zero gradient

-

Outlet 2

Zero gradient

-

Zero gradient

-

9 of 17

Simulation with OpenFOAM*

Term

Numerical Schemes

Gradient

Gauss linear

Divergence

Gauss upwind

Laplacian

Gauss linear limited corrected 0.33

Interpolation

Linear

Surface Normal Gradient

limited corrected 0.33

Variable

Solver

Preconditioner Smoother

Pressure

GAMG

Gauss Seidel

Velocity

Smooth Solver

Sym Gauss Seidel

K, Omega

Smooth Solver

Sym Gauss Seidel

* Open-Source CFD Analysis of Nasal Flows by Balatinec, L., Uroić, T. and Jasak, H., 2021.

10 of 17

Simulation with SU2

Boundary

U BC

Value

P BC

Value

Nasal Cavity

No Slip

0

Zero Gradient

-

Inlet 1

Fixed Value

Uniform (-1.77802336 1.57547665 2.33074331)

Zero Gradient

-

Inlet 2

Fixed Value

Uniform (0.954729319 1.30676162 3.50552511)

Zero Gradient

-

Outlet 2

Zero Gradient

-

Fixed Value

uniform 0

Outlet 2

Zero Gradient

-

Fixed Value

uniform 0

Multigrid parameters

Multi-Grid Levels

3

Multi-grid cycle

W_CYCLE

Damping factor

0.9

Linear solver for the implicit formulation

BCGSTAB

Numerical parameters

Convective numerical method

LAX-FRIEDRICH

Monotonic Upwind Scheme

MUSCL

Turbulence Convective numerical method

SCALAR_UPWIND

Turbulence Slope limiter

VENKATAKRISHNAN

Turbulence Boundary Conditions: FREESTREAM_TURBULENCEINTENSITY= 0.05 ,FREESTREAM_TURB2LAMVISCRATIO= 10

​

11 of 17

Hardware* and Time

Simulation using Open Foam

Simulation using SU2

CPU

Intel i5-10th Gen

CPU

Intel i5-4th Gen

RAM

16 GB

RAM

08 GB

Number of Iterations

815

Number of Iterations

61

Time taken

26.5 Minutes

Time Taken

12.2 Minutes

* Because of remote working locations different computers were used to perform simulations

12 of 17

Comparison of Results

a) Open Foam Results

b) SU2 Results

13 of 17

Advantages of SU2

Numerical Schemes

    • Automatic selection of best scheme
    • Better convergence rate due to higher order schemes

Simulation time

    • Faster simulation for the same mesh and physics
    • Convergence improves due to multigrid method

Learning curve

    • Less steep learning curve than OpenFOAM
    • Algorithm knowledge is not required

14 of 17

Conclusions

  • Since the nasal flow simulations will be mainly used by doctors or medical professionals, SU2 is a better choice for them for performing simulations.
  • The above statement is backed by the argument that SU2 doesn’t have a steep learning curve, is faster and one doesn’t have to be from a CFD background to use it.
  • It is to be noted one still must understand fundamentals of CFD in order to set up the configuration file, but in the aspect of the advantages that the nasal simulation provides, a little study can go a long way in saving a life.

15 of 17

Acknowledgements:

  • I would like to thank :
    • Office of Career Services , IIT Hyderabad for providing an opportunity to work with TCS.
    • TCS Research & Development for providing me with an Internship opportunity and funding the work.
    • Some people within TCS I would like to acknowledge:
      • Dr BP Gautham (Chief scientist at TCS research)

​

​

​

16 of 17

References:

  1. Oliver Jones. The Nasal Cavity. Sept. 2019. url: https://teachmeanatomy. info/head/organs/the-nose/nasal-cavity/.
  2. Balatinec, L., Uroić, T. and Jasak, H., 2021. Open-Source CFD Analysis of Nasal Flows. OpenFOAM® Journal, 1, pp.2-26.
  3. Inthavong, K. (2017). NC01 - Nasal Cavity 01 (Version 2). RMIT University. https://doi.org/10.6084/m9.figshare.5398384.v2
  4. Jasak, H., Jemcov, A. and Tukovic, Z., 2007, September. OpenFOAM: A C++ library for complex physics simulations. In International workshop on coupled methods in numerical dynamics (Vol. 1000, pp. 1-20). IUC Dubrovnik Croatia.
  5. Economon, T.D., Palacios, F., Copeland, S.R., Lukaczyk, T.W. and Alonso, J.J., 2016. SU2: An open-source suite for multiphysics simulation and design. Aiaa Journal, 54(3), pp.828-846.

​

​

17 of 17

THANK YOU