An Experimental Validation of the Computer Fluid Dynamics of Normal Human Nasal Airflow Using a Precise 3-D Printing Model

Computer Fluid Dynamics (CFD) is a popular method for studying airflow of nasal cavities. However, the data of CFD studies has rarely been validated through experiments. To test the accuracy of CFD computation, we studied the consistency of the air pressure of nasal cavities in the CFD and the experiment. A proportional resin model of a normal human subject’s nasal cavities was created by a 3-d printer with a precision value of 0.1mm. The pressure of 63 check points in the nasal cavities in different breathing states was measured. The experimental data was compared with the data obtained by CFD simulation. At the flow rates of 180 ml s-1 and 560 ml s-1, the pressure in all check points remained highly consistent with the CFD data. At 1100 ml s-1 flow rate, there was a significant deviation in the posterior segment of the nasal cavity during exhalation. The method used in this study to measure the pressure in the nasal cavities can be used in experimental validation of CFD data. The computational methods and the boundary conditions used in this study resulted in a high agreement between the results of the CFD simulation and the experiment. Author Summary In the contemporary era, Computer fluid dynamics (CFD) is the mainstream method for studying air flow. Due to the complex anatomical structure of the nasal cavity, the CFD results of the nasal flow have rarely been experimentally verified. This study provides a method to verify the methods and results of nasal CFD. We printed an accurate model of a normal person’s nasal cavity with a high-precision 3D printer. In this nasal cavity model, we set 63 small holes to detect the air pressure of the places we concerned. Three different nasal flow quantity are used to represent different breathing conditions: high (1100 ml s-1), medium (560 ml s-1), and low (180 ml s-1). In medium and low nasal flow quantities, our CFD results are in good agreement with the experimental pressure values. On this basis, we analyzed the characteristics of nasal airflow in normal people. The method used in this study to measure the pressure in the nasal cavities can be used in experimental measurements of the partial resistance of the nasal cavity. With proper modification, it can be applied to the clinical practice for nasal resistance, giving more help for the design of the operation plan.


51
As the pathway of the respiratory system, the nasal cavity has important 52 physiological functions of passing air, filtering air, regulating temperature and humidity 53 of air, sensing smell, and providing immune defense. The change in the anatomic 54 structure of the nasal cavity leads to the change in the airflow of the nasal cavity. This 55 can cause the change of the physiological functions of the paranasal sinuses or even 56 illnesses. The study of the internal airflow in the nasal cavity is very important in terms 57 of guiding surgeries [1]. As far as the current clinical practice is concerned, only limited 58 measurements of the nasal flow parameters (such as rhinomanometry, acoustic 59 rhinometry) can be taken, and correlation between these measurements and clinical 60 practice is questionable [2]. 61 Due to the complex internal structure of the nasal cavity, it is difficulty to study its Resin model was one of the models employed in previous studies [6]. With the 67 development of computer technology, more researchers were using finite element 68 methods to create computer models and the CFD method to simulate the airflow in the 69 nasal cavity [6][7][8]. Not limited by the anatomical inaccessibility of the nasal cavity, CFD 70 enables the calculation of various parameters of the intranasal flow field with a high 71 temporal and spatial resolution [2]. This means computer modeling can provide detailed 72 and objective characteristics of airflow for all locations and periods of an individual nose. 73 Furthermore, the ability to remodel computer simulations provides a potential predictive 74 tool for planning nasal surgeries [9]. As the methods used by researchers are not 75 entirely the same, under what circumstance a method is more suitable for nasal cavity 76 studies needs further investigations [2,[10][11][12][13][14][15]. Although there have been some studies 77 comparing CFD simulation and experimental data [4,6,16,17], the models used are not 78 accurate enough. Therefore, they cannot fully explain the pros and cons among the 79 existing computation methods. 80 An adult's inspiratory nasal flow rate in restful breathing state is within the range 81 of 5-12L min -1 [18]. To find out whether the breathing equations for high (1100 ml s-1), 82 medium (560 ml s -1 ), and low (180 ml s -1 ) nasal flow rates are accurate, we made an 83 experimental measurement of a normal person's nasal cavity model, which was 3-d 84 printed with 1 : 1 precision ratio, and compared the results with the data obtained We first built a model of nasal cavity from a healthy female Chinese (35 years 89 old). The model in IGES format is shown in Fig 1A. The three blue lines were connected 90 by the air pressure check points which were located in superior, middle, inferior nasal 91 meatuses. The details of the grid for the nasal cavity surface are shown in Fig 1B. The 92 coordinate system of the model is showed in Fig 1C. The 1:1 3D printing model, with 93 holes inside for air pressure check, are shown in Fig 1D(a). The experimental device is 94 shown in Fig 1D(b).

108
Experimental and CFD data of static pressure in the nasal cavity during 109 exhalation and inhalation at different flow rates (180 ml s-1, 560 ml s-1 and 1100 ml s-Experimental Validation of the CFD of Normal Nasal Flow 6 110 1).As shown in Fig 2, the simulation data was close to the experimental results. When 111 the flow rate was 180 ml s -1 , the error was the smallest. With the increase of the flow 112 rate, the error increased gradually. With the increase of the flow rate, the pressure 113 fluctuated greatly. When exhaling, the area with the highest pressure was located in the 114 range of X / L = 0.6 ~ 0.8 of the middle nasal meatus. When inhaling, the area with the 115 lowest pressure was in the range of X / L = 0.6 ~ 0.7 of the middle meatus.

116
There was a slight difference between the left and right nasal meatuses. In the 117 superior nasal meatus, the pressure on the left side was slightly lower than that on the 118 right side when exhaling. When inhaling, the pressure difference on the left and right 119 sides was small at the flow rate of 180 ml s -1 or 560 ml s -1 . However, at the flow rate of 120 1100 ml s -1 , the pressure fluctuation in the right nasal meatus was much larger. In the 121 middle nasal meatus, the pressure of the left side was slightly higher than that on the 122 right side when exhaling, and the pressure difference on the left and right sides was 123 smaller when inhaling. In the inferior nasal meatus, the pressure on the left side was 124 slightly higher than that on the right side when exhaling, whereas the pressure 125 difference between the left and right sides was small at the flow rate of 180 ml s -1 or 560 126 ml s -1 when inhaling. However, at the flow rate of 1100 ml s -1 , the pressure fluctuation of 127 the left inferior nasal meatus was larger.

156
In the velocity nephogram of inhalation (Fig 4A), as the flow rate increased, the 157 flow velocity also continuously increased. But the maximum velocity of inhalation 158 occured in the front of inferior nasal meatus at the three flow rates. It is obvious that the 159 flow velocity in the right nasal cavity is larger than that in the left.

160
The velocity nephogram of exhalation ( Fig 4B) indicates that the maximum 161 velocity of exhalation also occured in the front of inferior middle nasal meatus. From the 162 Y-axis direction, it shows clearly that the flow velocity in the right nasal cavity was 163 greater than that in the left nasal cavity under the flow rate of 180 ml s -1 when exhaling, 164 which was consistent with inhalation. However, as the flow rate growing, the flow 165 velocity in left middle nasal meatus inreased much more rapid than that the right.   The air distribution between the left and right nasal cavities is shown in Fig 5B. pressure increase) also increased (Fig 2 and 3). This can be analyzed using Bernoulli  There was an obvious difference of flow velocity in space, inspiration, and expiration.

289
In terms of flow rates inside the nasal cavities, the nasal valve area had a higher

301
The flow velocity of inspiration near the middle turbinate and inferior turbinate was 302 significantly larger than that of expiration ( Fig 4C). This phenomenon, however, seldom increased as the flow volume increased (Fig 6). We also found significant inspiratory the olfactory mucosal uptake of smell created by the mixed turbulence at high flow rates 374 could be negligible as sniff duration was more important than sniff strength [26].

375
Therefore, the relationship between turbulence and olfactory detection needs to be 376 further studied.

377
The limitations of the study are: First of all, since our model was 3-D printed 378 using resin as its material, it was harder than the real nasal mucosa in terms of texture. The CT scan file in DCOM format was imported into the MIMICS software (19.0).

409
The segmentation threshold value was set at -1024~-475 HU. The file was manually Pointwise was used to generate a grid to calculate the internal airflow of the 415 nasal cavities. As the model was for the complete nasal cavities, which have a complex 416 structure, a tetrahedral unstructured grid was used as the grid type to ensure its 417 efficiency. Finely generated, the mesh well reflected the internal features of the nasal 418 cavities.

419
Computer simulation of the model of the nasal cavities and paranasal sinuses 420 The grid was imported into Ansys Fluent to find the solutions to the airflow of the 421 nasal cavities. The flow inside the nasal cavities was calculated in its steady state. K- 422 model was used for the simulation of turbulent flow. The CFD methods were used for 423 studying the airflow of the nasal cavities at three flow rates of 180 ml s -1 , 560 ml s -1 ,and 424 1100 ml s -1 , corresponding to restful breathing, medium sniffing, and strong sniffing [15]. Experimental measurement of the pressure in the nasal cavities