Chaos, Solitons and Fractals 150 (2021) 111169

Contents lists available at ScienceDirect

Chaos, Solitons and Fractals

Nonlinear Science, and Nonequilibrium and Complex Phenomena
journal homepage: www.elsevier.com/locate/chaos

Frontiers

Design and simulation of mechanical ventilators

Abdellah El-Hadj a, Mohamed Kezrane a, Hijaz Ahmad b,∗, Houari Ameur c, S. Zamree Bin
Abd Rahim d,e, Abdelhakime Younsi a, Hanaa Abu-Zinadah f
a
Laboratory of Mechanics, Physics, Mathematical modeling (LMP2M), University of Medea, Medea, Algeria
b
Department of Basic Sciences,University of Engineering and Technology, Peshawar, Pakistan
c
Department of Technology, University Centre of Naama – Ahmed Salhi, P.O. Box 66, Naama 45000, Algeria
d
School of Manufacturing Engineering, Universiti Malaysia Perlis, Main Campus Pauh Putra, 02600 Arau, Perlis, Malaysia
e
Green Design and Manufacture Research Group, Center of Excellence Geopolymer and Green Technology (CEGeoGTech), Universiti Malaysia Perlis, 01000
Kangar, Perlis, Malaysia
f
University of Jeddah, College of Science, Department of Statistics, Jeddah, Saudi Arabia

a r t i c l e i n f o a b s t r a c t

Article history: During this period of COVID-19 pandemic, the lack of medical equipment (like ventilators) leads to com-
Received 29 December 2020 plications arising in the medical field. A low-cost ventilator seems to be an alternative substitute to fill
Revised 14 April 2021
the lacking. This paper presents a numerical analysis for predicting the delivered parameters of a low-
Accepted 7 June 2021
cost mechanical ventilator. Based on several manufactured mechanical ventilators, two proposed designs
Available online 25 June 2021
are investigated in this study. Fluid-structure interaction (FSI) analysis is used for solving any problems
Keywords: with the first design, and computational fluid dynamic (CFD) analysis with moving boundary is used for
Mechanical ventilator solving any issues with the second design. For this purpose, ANSYS Workbench platform is used to solve
Fluid-structure interaction the set of equations. The results showed that the Ambu-bag-based mechanical ventilator exhibited dif-
CFD, COVID-19 ficulties in controlling ventilation variables, which certainly will cause serious health problems such as
barotrauma. The mechanical ventilator based on piston-cylinder is more satisfactory with regards to de-
livered parameters to the patient. The ways to obtain pressure control mode (PCM) and volume control
mode (VCM) are identified. Finally, the ventilator output is highly affected by inlet flow, length of the
cylinder, and piston diameter.
© 2021 Elsevier Ltd. All rights reserved.

1. Introduction ries air and oxygen to the patient [7]. So, non-invasive ventilation
could be an excellent choice to avoid the multiplication of lung
COVID-19 outbreak has become a global issue as this new pan- damage induced by mechanical ventilation [8].
demic has strongly affected the world [1-2]. This virus causes a Faced with the large influx of patients, hospitals do not have
substantial global health problem with very significant economic enough respirators to meet their needs. The medical equipment
and social impacts [3-6]. Studies have shown that most people af- industry offers various highly sophisticated devices in the market,
fected by COVID-19 have mild symptoms, but other people, like but that available number still cannot meet the demand of hos-
adults over 65 and those with chronic illnesses, develop severe pitals. Thus, several enthusiastic research teams have used their
symptoms in the pulmonary system. This is reflected by the fact competence to develop a simple design and low-cost respirators
that the potentiality of these groups of people to exchange carbon manufactured in a large number within a short period [8]. Among
dioxide and oxygen between blood and lungs begins to fall once many low-cost ventilators, a published MIT student project using
they are attacked by the virus [7]. At this stage, patients are hos- Ambu-bag, which is available freely online, is referred [9]. How-
pitalized to support their lung functions with mechanical ventila- ever, a respirator is not just a pump that forces air and oxygen
tion, as it is the best medical treatment currently recommended into the patient’s lung. One of the critical problems encountered
by doctors. It has been reported that the early use of endotracheal by doctors during mechanical ventilation is Barotrauma [10-12]. In-
intubation (invasive intubation) can cause secondary lung infection deed, when there is difficulty adapting between the pressure de-
because of the risk of germs entering through the tube, which car- livered by the ventilator and that inside the lungs, an alveolar
overdistention can occur, further causing barotrauma [13, 14].
In the respiratory system (Fig. 1), breathing is controlled by the
∗
Corresponding author. diaphragm, which is a muscle separating the abdomen from the
E-mail address: hijaz555@gmail.com (H. Ahmad).

https://doi.org/10.1016/j.chaos.2021.111169
0960-0779/© 2021 Elsevier Ltd. All rights reserved.
A. El-Hadj, M. Kezrane, H. Ahmad et al. Chaos, Solitons and Fractals 150 (2021) 111169

Based on this review, there are no numerical simulations on

flow produced by mechanical ventilators. This type of simulation
is essential to evaluate the performance of the design before pro-
ceeding to the fabrication stage. The study of the mechanical ven-
tilation process requires the use of fluid-structure interaction (FSI).
The FSI problems play prominent roles in many scientific and en-
gineering fields, yet a comprehensive study of such problems re-
mains a challenge due to their strong nonlinearity and multidis-
ciplinary nature [19-22]. FSI method couples computational fluid
dynamics used for fluid flow with finite element analysis used for
the solid domain. It enables the investigation of the fluid behavior,
structural behavior, and how they interact and affect each other. In
a One-Way FSI study, the results obtained from the solution fluid
or structural domain are used as a boundary condition when solv-
ing the different domain. In a Two-Way FSI study, at each sub-step,
the fluid and structural domain solutions are solved in parallel. The
Fig. 1. The physiology of the breathing system [15]. solution must be converged before moving to the next step [23].
Thus, this present study is performed to contribute to improv-
ing these types of ventilators. ANSYS multiphysics software is used
to simulate the gas flow and volume produced by a low-cost ven-
Nomenclature tilator designed. Two ventilator designs are proposed and inves-
tigated. A two-way FSI analysis using a finite element method is
t Time Step, s used for the first design, while Computational fluid dynamics (CFD)
Lc Compressed length (m) analysis with a moving mesh is employed for the second design.
O2 Flow Flow rate of Oxygen (L/min) In order to control the ventilation process, the effect of several pa-
Volume Delivered volume (L) rameters on the flow is analyzed.
VT Tidal Volume (L) The paper is organized as follows: after achieving the literature
Upx Piston displacement (m) review and highlighting the purpose and originality of the work,
UAZ Vertical displacement of Ambu-bag (m) the physical problem is described in Section 2. In this section, the
Prel Relative pressure (Pa) mechanical ventilation process is well detailed. Then (in Section 3),
ρ Density (kg/m3 ) the required details on design and modeling of the mechanical
μ Dynamic viscosity (kg/m.s) ventilators are given. The geometry, boundary conditions, material
P Pressure (Pa) properties, and mesh adopted in this analysis are also provided in
PCM Pressure control mode this section. In Section 4, the computed results are presented and
VCM Volume control mode discussed.
FSI Fluid structure interaction
CFD Computational fluid dynamic 2. Mechanical ventilation process
PEEP Positive end expiratory pressure
The operating principle scheme of mechanical ventilation is il-
lustrated in Fig. 2. Frequently, the respiratory cycle consists of the
chest, and the intercostal muscles, which are located between the inspiratory time (TI) of 1 s and the expiratory time (TE) of 2 s
bones of the rib cage. During inspiration, the diaphragm contracts [24]. In this case, the respiratory frequency is often 20 cycles per
and moves toward the abdomen cavity, and the intercostal mus- minute. Once inspiration gets triggered, the insulator opens the in-
cles contract to lift the rib cage outward. As a result of these two spiratory valve VA (Fig. 2) and closes the expiration valve VB, caus-
muscle movements, the volume of the rib cage increases, and the ing the airway pressure to rise and then, compressed gas enters
pressure in the cavity where the lungs are housed, decreases. The the lung. During expiration, the mechanism is reversed. The inspi-
increase in volume and the decrease in pressure urge outside air ratory valve is closed and the expiratory valve opens, causing the
to fill the lungs, to balance the pressure in the lungs with external airway pressure to drop, and the gas leaves the lung (passive ex-
atmospheric pressure. During the expiration phase, the process is halation). These two valves always act in opposite ways. However,
reversed, where the pressure inside the lungs increases above at- this mechanism cannot generate a PEEP. The regulation of the ex-
mospheric pressure, thus resulted in a pressure difference which piratory valve causes the desired PEEP by taking into account the
draws air out of the lungs [15-17]. resistance of the breathing circuit [25].
Mechanical ventilation works entirely differently than the phys- There are two basic modes for mechanical ventilation accord-
iological breathing system. Instead of having depression during the ing to variables that are controlled during inspiration. These in-
inspiration, the mechanical ventilators push the gas by creating clude volume-controlled mode (VCM), which is commonly used,
high position pressure. The ventilatory parameters are carefully ad- and pressure-controlled method (PCM) ventilations [26]. In PCM,
justed by the doctor according to the ventilation mode that has the value of the maximum pressure is adjusted. However, the cur-
been chosen. It is possible to change the inspired oxygen fraction rent volume and flow delivered to the patient depended on the
between 21% and 100% while maintaining positive expiratory pres- importance of the inspiratory pressure and the set PEEP level [27].
sure (PEEP) in the circuit during expiration [8, 9]. In the VCM, the volume and the flow remain constant. Simulta-
A mechanical ventilator, inhalation circuit, exhalation circuit, neously, the pressures vary from one patient to another depending
and an artificial lung are required to simulate the complete breath- on the compliance and resistance of the respiratory system and the
ing process. Several pieces of research were conducted to experi- patient’s inspiratory contribution [28].
mentally and theoretically study the ventilation parameters [18]. An example of a typical curve of pressure, flow, and volume
It is important to note that the inspiration stage is related to the over time during VCM are illustrated in Fig. 3 (time is along the x-
ventilator. axis) [29]. The case of VCM is chosen because it is most frequently

2
A. El-Hadj, M. Kezrane, H. Ahmad et al. Chaos, Solitons and Fractals 150 (2021) 111169

Fig. 2. The operating principle scheme of mechanical ventilation [25].

Fig. 3. Typical curves of flow, pressure, and volume in the case of VCM [29].

used by doctors, as already mentioned. These curves are not de- 3. Design and modeling of mechanical ventilators
duced from a confirmed case of the ventilated patient but are
purely theoretical. From Fig. 3, it can be seen clearly that when in- As previously mentioned, a mechanical ventilator, inhalation
spiration starts (point A), the flow increases considerably to reach circuit, exhalation circuit, and an artificial lung are required to sim-
a plateau (point B) and stays at that target level for the duration ulate the complete breathing process. Therefore, this study is fo-
of inspiration before the expiration begins (point C). At this time, cused on the simulation of the flow inside the mechanical ven-
the pressure goes up gradually from PEEP and continues spreading tilator. Two different ventilators, namely Ambu-bag and piston-
until the preset tidal volume is delivered. The volume increases cylinder-based ventilators, are considered in Sections 3.1 and 3.2.
monotonically (linear rise) to reach the tidal volume (remember The flow rate and volume evolution as a function of time during
that volume = flow × time). During expiration, the flow curve goes the inspiration phase were calculated for each design (Circuit 1 in
below the time-axis and back to zero (point D). Simultaneously, Fig. 2).
the pressure drops to the set PEEP level, and the volume decreases In this section, mathematical formulations and solution proce-
and returns to zero. dures for the two proposed designs are presented. For the Ambu-

3
A. El-Hadj, M. Kezrane, H. Ahmad et al. Chaos, Solitons and Fractals 150 (2021) 111169

Fig. 4. Design of the Ambu-bag based on a low-cost mechanical ventilator [30]. .

Table 1
Solid material properties [34].

Material Properties

Silicon-rubber Density = 1120 Incompressibility parameter D1 = 1.1 × 10−9 (Pa−1) Tensile yield strength = 8.97 × 106 (Pa) Tensile ultimate
(kg/m3) strength = 8.97 × 106 (Pa)
Mooney-Rivlin 2
parameters
C10 = 2.59 × 105
(Pa)
C01 = 65 × 103
(Pa)
PVC plastic Density = 1330 Shear modulus = 10.983 MPa Tensile yield strength =1.95 MPa Tensile ultimate strength = 1.95 MPa
(kg/m3) Young’s
modu-
lus = 32.4 MPa
Poisson’s
ratio = 0.475
Aluminum Density = 2770
(kg/m3) Young’s
modulus = 71 GPa
Poisson’s
ratio = 0.33
Steel Density = 7850
(kg/m3) Young’s
modu-
lus = 200 GPa
Poisson’s
ratio = 0.3
ABS plastic Density = 1040 Shear modulus = 854 MPa Tensile yield strength = 41.4 MPa Tensile ultimate strength = 41.4 MPa
(kg/m3) Young’s
modu-
lus = 2.39 GPa
Poisson’s
ratio = 0.399

bag-based design (Section 3.1), the Fluid-structure interaction (FSI)

analysis methodology is illustrated with boundary conditions, ma-
terial properties, and mesh generated. However, for the piston-
cylinder-based design (Section 3.2), the CFD analysis is more ap-
propriate to evaluate the designed performance. The geometry,
boundary conditions, material properties, and mesh adopted in this
analysis are also presented. Therefore, flow rate and delivered vol-
ume by the ventilator are calculated. The pressure at the outlet is
not calculated due to the calculation of pressure at the outlet of
the mechanical ventilator requires the entire system to be simu-
lated. In this case, the gas delivered flows to the atmospheric en-
vironment.

3.1. Ambu-bag-based design

A modified version of the low-cost mechanical ventilator

(Fig. 4) proposed in this study was designed using the Solid works
software [30]. A one-step motor is used to deform the Ambu-bag
Fig. 5. Stress-strain curves for the silicon rubber material.
by the displacement of the top hand. FSI analysis was conducted

4
A. El-Hadj, M. Kezrane, H. Ahmad et al. Chaos, Solitons and Fractals 150 (2021) 111169

Fig. 6. Flowchart of the FSI solution procedure [36].

∂ ∂   ∂ P ∂ τi j
( ρ ui ) + ρ ui u j = − + (2)
∂t ∂xj ∂ xi ∂ x j
∂u ∂u ∂u
where, τi j = μ( ∂ x i + ∂ x j + 23 δi j ∂ x i ). with xi are the coordinates’
j i j
components, ui is the velocity components, ρ is the density,
P is the pressure, and μ is the dynamic viscosity. The den-
sity and dynamic viscosity are equal to 1.2999 kg/m3 and 1.919
919 × 10−5 kg/m•s, respectively.
The finite volume method (FVM) in the situation of moving
mesh is set as following [32]:
  
d →
∫ ρφ dV + ρ n u − −
ub φ ds − ρ φ n
∇ φ ds = ∫ S dV
φ (3)
dt V V
S S

where V is an arbitrary moving control volume, S is a closed sur-

face bounded V, n  is the outward unit normal vector on the bound-
 is the fluid velocity vector, −
ary surface, u
→
ub is the boundary mesh
velocity,  φ is the diffusion coefficient, and Sφ is the volume
Fig. 7. Boundary conditions and generated mesh for the structural domain. source/sink of φ .
To include the effect of the piston surface motion on the mesh
for this design. FSI occurs when Ambu-bag is bent, causing pertur- evolution during time variation, the mesh diffusion equation must
bation of gas flow, and then the gas generates a force on the solid. be solved to get the grid displacements as [33]:
 
3.1.1. Fluid flow formulation ∇ · disp · ∇δ = 0 (4)
The governing equations for unsteady laminar flow of a Newto- where δ is the displacement relative to the previous mesh loca-
nian fluid are given by [31]: tions. The mesh stiffness ( disp) determines the degree to which
∂ρ ∂ regions of nodes move together. The option of increase stiffness
+ ( ρ ui ) = 0. (1) near small volumes is used in this study to preserve the mesh
∂ t ∂ xi
5
A. El-Hadj, M. Kezrane, H. Ahmad et al. Chaos, Solitons and Fractals 150 (2021) 111169

Fig. 8. Boundary conditions and generated mesh for fluid domain.

stiffness matrix is updated at each time to take into account the

nonlinearities due to geometry, material, and contact.
The properties of materials used in this analysis are given in
Table 1. The hands are supposed in aluminum material. The Ambu-
bag valve is made of Acrylonitrile butadiene styrene (ABS) plastic
material. The Ambu-bag is made from Polyvinyl chloride (PVC) or
Silicon-rubber material, and all other parts are proposed to be fab-
ricated using steel material.
It is challenging to model the behavior of the Ambu-bag made
of hyperplastic material (Silicon-rubber) due to its nonlinearity,
large deformation, and material instability [35]. In this study, the
Mooney-Rivlin model was used for the Ambu-bag. The most com-
monly used model is the 2-term Mooney-Rivlin model. The two
Fig. 9. The second mechanical ventilator design. . parameters for the Mooney-Rivlin form of the strain energy poten-
tial is written as [34]:
1
W = C10 (I1 − 3 ) + C01 (I2 − 3 ) + (J − 1 )2 (7)
D
where the principal defines the strain invariants (I1, I2, J) stretches
at any deformation state (λ1, λ2, λ3) as:
I1 = λ21 + λ22 + λ23
I2 = λ21 λ22 + λ22 λ23 + λ23 λ21 (8)
J = λ1 λ2 λ3
The stretch ratio λ (or simply ‘stretch’) is defined as L
L0 =
L0 +u
L0 = 1 + εE , where εE is engineering strain. The parameters
Fig. 10. The mesh and boundary conditions of the fluid domain of the piston- of this model are calculated using the data presented in Fig. 5.
cylinder-based design. Through the strain energy function, the stresses can be calculated
as Sij = ddW
E
, where Eij are components of the strain tensor. For all
ij
of the preceding Mooney-Rivlin forms, the initial shear and initial
bulk moduli are defined as μ0 = 2(c10 + c01 ) and κ0 = d2 .
quality and limit mesh distortion. In this case,  disp, is obtained
by this equation [33]: 3.1.3. Formulation of fluid-structure analysis
V Cstiff In FSI, the fields are solved separately, which are coupled by an
disp = ref
(5) iterative procedure in the ANSYS workbench (Fig. 6) [36-38]. For
V
the two-way coupling procedure, the fluid pressure is transferred
where Vref is a mean control volume of the numerical domain. In to the structural solver. Using this pressure as an external load, the
this study, we use the recommended value of the stiffness model displacement obtained from the structure solver is transferred to
exponent Cstiff, which is equal to 2. the CFD solver. This procedure is repeated until convergence is ob-
tained [39-41].
3.1.2. Formulation of structural dynamic Fig. 7 illustrates the boundary conditions and the mesh of the
Transient structural analysis was used to calculate the response solid domain. In this case, the top hand is supposed to move ac-
of deformable solids, where the response at each time is the solu- cording to the vertical axis (UAZ) with a uniform velocity. During
tion of the general equation of motion [34]: ventilation process, the top hand is moved by the screw in two
    steps. During the first step (1 s), it moves to compress the Ambu-
[M] d̈ + [C] d˙ + [K]{d} = {F(t )} (6) bag for 9 cm. After that, the top hand returns to its initial position
during the second step (2 s), as given by:
where d is the nodal displacement, d˙ is the nodal velocity, and d¨

is the nodal acceleration. [M], [C], [K] and F(t) are mass matrix, −0.09 · Time, for 0s ≤ Time ≤ 1s
damping matrix, stiffness matrix, and force vector, respectively. The
UAZ =
0.09 · (Time − 1 ) − 0.09, for 1. < Time ≤ 3s
(m ) (9)

6
A. El-Hadj, M. Kezrane, H. Ahmad et al. Chaos, Solitons and Fractals 150 (2021) 111169

Fig. 11. Total displacement of the ventilator with silicon-rubber Ambu-bag, at 1 s. (PVC material, 100% O2, inlet flow = 12 l/min, UAZ = 0.09 m).

Fig. 12. Mesh history of fluid domain during one cycle of breathing. (PVC material, 100% O2, inlet flow = 12 l/min, UAZ = 0.09 m).

The design contains two fixed supports, as shown in Fig. 7. The while the fluid domain contains 35,376 tetrahedral elements
boundary conditions for the fluid domain and the mesh generated (Fig. 8).
are illustrated in Fig. 8. It is important to note that the ventilator To ensure the stability of the two-way transient FSI analysis,
flows to the atmospheric environment, and the flow of the oxygen the optimum range of 1 to 10 was set for external coupling iter-
(O2) is set as: ations and mesh displacement control. For CFD problem, a high-

resolution advection scheme was used with a second-order back-
L
12 min 0s ≤ Time ≤ 1s ward Euler discretization algorithm. The convergence criterion was
O2 Flow = (10)
L
0. min 1s < Time ≤ 3s set to 10−4 for the maximum residuals, which could be obtained
in all simulation runs. For the structural problem, the Newton-
In the FSI problem, the challenge is to couple the Lagrangian Raphson method is used for solving nonlinear equations.
structural mesh with the Eulerian fluid mesh. Once the new solid
position is computed from the structural solver at every timestep,
the fluid mesh has to be deformed to conform with the latest solid
place. ANSYS CFX utilizes the Displacement Diffusion as mesh de- 3.2. Piston-cylinder based design
formation method. The critical parameter in FSI analysis is the time
step. The time step is taken as a fixed value of Dt = 0.01 s. The second design proposed in this study is shown in Fig. 9. In
For the mesh of the two domains, tetrahedral elements which this design, a piston is moving in a cylinder actuated by a motor.
can fit better complex geometry are used. This type of mesh is The gas enters the cylinder from the inlet, compressed, and evacu-
suitable for problems involving mesh motion. In this case study, ates through the outlet.
the solid domain contains 10,756 tetrahedral elements (Fig. 7), The piston is displaced according to the two cases as follows:

7
A. El-Hadj, M. Kezrane, H. Ahmad et al. Chaos, Solitons and Fractals 150 (2021) 111169

Fig. 13. Surface streamlines history on the vertical plane for the first design. (PVC material, 100% O2, inlet flow = 12 l/min, UAZ = 0.09 m).

Fig. 14. Displacement uz of point1 and point 2 (PVC material, 100% O2, inlet Fig. 15. Outlet flow rate history (100% O2, inlet flow = 12 l/min, UAZ = 0.09 cm).
flow = 12 l/min, UAZ = 0.09 cm).

Fig. 10 displays the mesh and boundary conditions of the fluid

Case A: Uniform rectilinear motion domain, which contain 14,917 tetrahedral elements. The mesh is

−0.1 · Time, for 0 < time ≤ 1 changed at each time step using the mesh deformation technique.
Upx =
0.05 · (Time − 1 ) − 0.1, for 1 < time ≤ 3
(m ) (11) For the moving mesh motion, the stiffness model exponent Cstiff
is equal to 2.
Case B: Uniform decelerated rectilinear motion

1
(0.2 ) · Time2 − 0.2 · Time, for 0s < time ≤ 1s 4. Results and discussion
Upx = 2
2 (m )
2(
1
−0.05 ) · (Time − 1 ) + 0.1 · (Time − 1 ) − 0.1, for 1s < time ≤ 3s
As mentioned before, this study focuses on the inspiration pe-
(12) riod. The two proposed mechanical ventilators are studied for one
In the first case, the piston is moved according to a uniform cycle of breathing. Therefore, two stages are considered: compres-
rectilinear motion. On the other hand, the second case follows a sion (1 s) and decompression (2 s). The analysis results of the
uniform decelerated rectilinear movement. Here, the piston veloc- first design using the FSI analysis are presented in Section 4.1. In
ity varies linearly from the initial velocity to the stagnation point Section 4.2, the results of the fluid flow with a moving piston are
(zero rates). exposed. Finally, the effects of some essential parameters are in-
The gas flow, in this case, is transient laminar. The same equa- vestigated.
tions for the first design are used here. Thus, the equations of con-
tinuity, mesh displacement, and momentum are solved. The time 4.1. Results of AMBU-BAG based design
step in CFD analysis is estimated by the Courant Number (CN)
Uref · t
as Element Size
, where Uref is the reference velocity. This parameter Fig. 11 shows the results of the total displacement of the solid
ensures that the fluid passes through a number of elements in one domains for different instances. The maximum removal is obtained
timestep. Generally, CN is taken to be a value between 2 and 10 for a time equal to 1 s at the end of the top hand. The fluid do-
[26]. The time step (Dt) is set to a fixed value equal to 0.01 s. main’s mesh history is presented in Fig. 12, where each time step

8
A. El-Hadj, M. Kezrane, H. Ahmad et al. Chaos, Solitons and Fractals 150 (2021) 111169

Fig. 16. Outlet flow rate history for different compression displacements (UAZ) (Sil-
icon rubber material, 100% O2, inlet flow = 12 l/min).

corresponds to a new mesh. The value of the orthogonality an-

gle controls the quality of the mesh. During calculation over time,
the solver stops when the minimum value of the orthogonality an-
gle is below 0.2618 rad (15°) (written in Python language as min-
Val(Orthogonality Angle Minimum)@Default Domain < 15 [deg]).
Then, the re-meshing starts, and a new mesh is transferred to CFX.
The run continues until the end of the cycle.
Fig. 13 presents the streamlines on the vertical plane for dif-
ferent instances. It can be seen that some vortices are present
in the fluid domain either for the compression stage (1 s) or for
the decompression stage (2 s). The minimum pressure current in
the inside vortices cores will certainly cause perturbation in the
flow, mainly in the inspiration period. The fluid flows from high-
pressure zones to low-pressure zones. This is confirmed by the
zoom view of Fig. 11, in which it can be observed that the Ambu-
bag does not follow the motion of the top hand, and their displace-
ments are different from each other.
Also, Fig. 14 shows the vertical displacement of points 1 and 2. Fig. 18. Flow rate at the outlet and the inlet of Piston-cylinder ventilator for cases
Point 1 belongs to the Ambu-bag, and point 2 belongs to the top A and B. (100% O2, inlet flow = 12 l/min, Lc = 150 × 10−3 m, Dp = 37 × 10−3 m).
hand, as indicated by Fig. 11. It can be confirmed that the response
of Ambu-bag is more significant than the load applied.

Fig. 17. Mesh displacement at different moments of the piston-cylinder system. (100% O2, inlet flow = 12 l/min, uniformly rectilinear, Lc = 150 × 10−3 m,
Dp = 37 × 10−3 m).

9
A. El-Hadj, M. Kezrane, H. Ahmad et al. Chaos, Solitons and Fractals 150 (2021) 111169

Fig. 20. PCM and VCM modes of mechanical ventilation [42].

Fig. 19. Volume history of Piston-cylinder ventilator for cases A and B. (100% O2,
inlet flow = 12 l/min, Lc = 150 × 10−3 m, Dp = 37 × 10−3 m).

in Fig. 22. The depression created by vortices allows the reverse

flow from outside to inside the ventilator.
The flow delivered by this mechanical ventilator is presented in Finally, a parametric study is conducted to investigate the effect
Figs. 15 and 16 for different materials and compression distances, of different parameters on the outlet flow of the mechanical ven-
respectively. It can be seen from these figures that the system is tilator for VCM mode (Case A). Fig. 23a shows the evolution of the
not stable. The outlet pressure cannot be controlled because the outlet flows with varying inlet flows. The outlet flow increases as
flow contours are highly perturbed. The results obtained from the the inlet flow increases. Besides, the diameter of the piston (Dp)
first system (Ambu-bag-based design) are not satisfactory. Conse- and length of the cylinder (LC) have an essential effect on the out-
quently, another design is proposed. The following section repre- put data, as shown in Figs. 23b and c.
sents the results obtained from the second mechanical ventilator
(piston-cylinder-based design).
5. Conclusions
4.2. Results of piston-cylinder-based design
In this study, flow simulation produced by two proposed de-
The displacement of the piston wall is modeled using dynamic signs of low-cost mechanical ventilators was investigated. A finite
and deformable meshes, in which the mesh nodes are adjusted to element method was used to solve the set of coupled equations us-
new locations. Fig. 17 illustrates the mesh displacement as the pis- ing ANSYS software. A transient solution during one cycle of inspi-
ton compressing the gas for two instances (one at the beginning ration was considered. A fluid-structure analysis was used to ob-
and the other at the end of the piston motion). The higher values tain the fluid response resulting from the Ambu-bag deformation.
of deformation are felt near the piston wall. It was shown that the gas flow obtained is unstable. This first de-
The history of flow during 1 s of the inspiration period at the sign (Ambu-bag-based design) is uncontrollable due to the suscep-
inlet and the outlet of the piston-cylinder system are shown in tible hyper-elastic material of the Ambu-bag when force is applied.
Fig. 18. It is important to note that the difference between the inlet Utilization of this type of mechanical ventilators with this kind of
and outlet flow of the ventilator is due to the flow added by gas material could lead to health problems like barotrauma.
compression. On the other hand, a CFD study with a moving boundary was
The volume at any instant is estimated from the flow curve as performed for the piston-cylinder-based ventilator, and the results
t obtained were more satisfactory than the first design (Ambu-bag-
integral of flow over time as ∫ F low(t ) · dt. Fig. 19 illustrates the
0 based design). In this situation, two different ways of displacement
volume history during inspiration for the two cases of VCM and of the piston were used. Two modes of ventilation (VCM and PCM
PCM modes. The tidal volume (VT) delivered by the ventilator dur- modes) were evaluated. VCM was realized by a uniform rectilin-
ing the inspiration period is 668.31 × 10−3 (L) for cases A (VCM ear motion and PCM is realized by a uniform decelerated recti-
mode) and B (PCM mode). linear motion. In addition, the effects of inflow, cylinder length,
It can be seen clearly that the curve patterns in Fig. 19 are sim- and cylinder diameter on outflow were investigated. It was found
ilar to the inspiration stage presented in Fig. 20, where the two that the pattern of the outflow and the volume inspired could be
modes (PCM and VCM) are identified. controlled by manipulating these parameters. Finally, these results
Consequently, the parameters of case A can be used to produce could be further investigated in the future by including other pa-
VCM mode, in which the flow is fixed. However, PCM can be pro- rameters. This requires the addition of all respiratory circuits and
duced by using the parameters of case B, in which the pressure is an artificial lung to simulate the complete breathing process.
fixed.
Figs. 21 and 22 display the surface streamlines with isopressure
contours in vertical and horizontal planes. It can be seen from the Declaration of competing interest
figures that the flow is regular, and no vortices are present during
the inspiration stage, unlike the Ambu-bag-based design. However, The authors declare that they have no known competing finan-
at the second stage (during the return of the piston to its initial cial interests or personal relationships that could have appeared to
position), two large vortices were observed at time > 1 s, as shown influence the work reported in this paper.

10
A. El-Hadj, M. Kezrane, H. Ahmad et al. Chaos, Solitons and Fractals 150 (2021) 111169

Fig. 21. Surface streamlines on the vertical plane of the piston-cylinder ventilator. (100% O2, inlet flow = 12 l/min, uniformly rectilinear, Lc = 150 × 10−3 m,
Dp = 37 × 10−3 m).

Fig. 22. Surface streamlines on the horizontal plane of the piston-cylinder ventilator. (100% O2, inlet flow = 12 l/min, uniformly rectilinear, Lc = 150 × 10−3 m,
Dp = 37 × 10−3 m).

CRediT authorship contribution statement tion. Abdelhakime Younsi: Writing – original draft. Hanaa Abu-
Zinadah: resources.
Abdellah El-Hadj: Conceptualization. Mohamed Kezrane: Data
curation. Hijaz Ahmad: Writing – review & editing. Houari
Ameur: Formal analysis. S. Zamree Bin Abd Rahim: Valida-

11
A. El-Hadj, M. Kezrane, H. Ahmad et al. Chaos, Solitons and Fractals 150 (2021) 111169

References

[1] Jaimes A, André NM, Chappie JS, Millet JK, Whittaker GR. Phylogenetic analysis
and structural modeling of SARSCoV-2 spike protein reveals an evolutionary
distinct and proteolytically sensitive activation loop. J Mol Biol 2020. doi:10.
1016/j.jmb.2020.04.009.
[2] Tian S, Hu W, Niu L, Liu H, Xu H, Xiao SY. Pulmonary pathology of early-phase
2019 novel coronavirus (COVID-19) pneumonia in two patients with lung can-
cer. J Thorac Oncol 2020. doi:10.1016/J.JTHO.2020.02.010.
[3] Ayittey FK, Ayittey MK, Chiwero NB, Kamasah JS, Dzuvor C. Economic impacts
of Wuhan 2019-nCoV on China and the world. J Med Virol 2020. doi:10.1002/
jmv.25706.
[4] Pfefferbaum B, North CS. Mental health and the Covid-19 pandemic. New Engl
J Med 2020;383:510–12. doi:10.1056/NEJMp2008017.
[5] Velavan TP, Meyer CG. The COVID-19 epidemic. Trop Med Int Health
2020;25:278. doi:10.1111/tmi.13383.
[6] Yuki K, Fujiogi M, Koutsogiannaki S. COVID-19 pathophysiology: a review. Clin
Immunol 2020:108427. doi:10.1016/j.clim.2020.108427.
[7] Jiang F, Deng L, Zhang L, et al. Review of the clinical characteristics of
coronavirus disease 2019 (COVID-19). J Gen Intern Med 2020. doi:10.1007/
s11606- 020- 05762- w.
[8] A guide to designing low-cost ventilators for COVID-19; 2020 https://www.
youtube.com/watch?v=7vLPefHYWpY.
[9] Al Husseini AM, Lee HJ, Negrete J, Powelson S, Tepper Servi A, Slocum AH.
Design and prototyping of a low-cost portable mechanical ventilator. J. Med.
Devices 2020. doi:10.1115/1.3442790.
[10] Machntyre NR. Design features of modern mechanical ventilators. Clin. Chest
Med. 2016. doi:10.1016/j.ccm.2016.07.002.
[11] Boussarsar M, Thierry G, Jaber S, Roudot-Thoraval F, Lemaire F, Brochard L.
Relationship between ventilatory settings and barotrauma in the acute
respiratory distress syndrome. J. Intensive Care Med. 2002. doi:10.1007/
s0 0134-0 01- 1178- 1.
[12] Weg JG, Anzueto A, Balk RA, Wiedemann HP, Pattishall EN, Schork AS,
et al. The relation of pneumothorax and other air leaks to mortality in
the acute respiratory distress syndrome. N Engl J Med 1998. doi:10.1056/
NEJM199802053380601.
[13] Chan EY, Lynch JE, Loran DB, Zwischenberger JB. Chapter 20 - Barotrauma. Total
burn care. Herndon DN, editor. 3rd Edition. W.B. Saunders; 2007. doi:10.1016/
B978- 1- 4160- 3274- 8.50023- 4.
[14] Hess DR, Kacmarek RM. Essentials of mechanical ventilation. 4th Edition. Mc-
Graw-Hill Education; 2018.
[15] Lewith H, Pandit JJ. Lung ventilation and the physiology of breathing. Surgery
2020. doi:10.1016/j.mpsur.2020.03.005.
[16] Gould T, de Beer JMA. Principles of artificial ventilation. Anaesth Intensive
Care. Med. 2007. doi:10.1016/j.mpaic.2006.12.016.
[17] Feher J. The mechanics of breathing. Quantitative human physiology. Feher J,
editor. 2nd Edition. Academic Press; 2017. doi:10.1016/B978- 0- 12- 800883- 6.
0 0 060-4.
[18] Hao L, Shi Y, Cai M, Ren S, Wang Y, Hao Z, Yu Q. Dynamic characteristics of a
mechanical ventilation system with spontaneous breathing. IEEE Access; 2019.
doi:101109/ACCESS20192955075.
[19] Bazilevs Y, Takizawa K, Tezduyar TE. Biomedical fluid mechanics and
fluid-structure interaction. Comput Mech 2014;54:893. doi:10.1007/
s00466- 014- 1056- 7.
[20] Kwon YW. FSI study of structures containing fluid. Flu-
id-structure interaction of composite structures. Springer
tracts in mechanical engineering, Cham: Springer; 2020.
https://doi-org-443.webvpn.jnu.edu.cn/10.1007/978-3-030-57638-7_6.
[21] Bi Z. Chapter 11 - Applications-multiphysics systems, finite element anal-
ysis applications. Academic Press; 2018. ISBN 9780128099520. doi:10.1016/
B978- 0- 12- 809952- 0.0 0 011-X.
[22] Bodnar T, Galdi G, Necasova S. Fluid-structure interaction and biomedical ap-
plications. Birkhaeuser: Springerl; 2014. doi:101007/978- 3- 0348- 0822- 4.
[23] Hirschhorn M, Tchantchaleishvili V, Stevens R, Rossano J. Throckmorton: fluid-
structure interaction modeling in cardiovascular medicine - a systematic re-
view 2017-2019. Med Eng Phys 2020. doi:10.1016/j.medengphy.2020.01.008.
[24] Zavala E, Ferrer M, Polese G, Ramon Masclans J, Planas M, Milic-Emili J,
et al. Effect of inverse I: e ratio ventilation on pulmonary gas exchange in
acute respiratory distress syndrome. Anesthesiology 1998;88(1):35–42.
[25] Banner MJ, Lampotang S, Boysen PG, Hurd TE, Desautels DA. Flow resistance of
expiratory positive-pressure valve systems. Chest 1986. doi:10.1378/chest.90.2.
212.
Fig. 23. Effects of different parameters on the outlet flow. (100% O2, uniformly rec- [26] Ligori T. Modes of ventilation and ventilator strategies. Handbook of ICU ther-
tilinear Upx). apy. Fuller J, Granton J, McConachie I, editors. Cambridge University Press;
2014. doi:10.1017/CBO9781107323919.014.
[27] Rittayamai N, Katsios CM, Deloncle F, Friedrich JO, Mancebo J, Brochard L.
Pressure-controlled vs volume-controlled ventilation in acute respiratory fail-
ure. Chest 2015. doi:10.1378/chest.14-3169.
[28] Rabec C, Rosenstein D, Leger P, Rouault S, Perrin C, Gonzalez-Berme J. Ventila-
tor modes and settings during non-invasive ventilation: effects on respiratory
Acknowledgment events and implications for their identification. Thorax 2011. doi:10.1136/thx.
2010.142661.
[29] Magdy Y, Laurent B. The control of breathing during mechanical ventilation.
The authors wish to thank the DGRSDT/MESRS, Algeria, and
Clin Chest Med 2019. doi:10.1016/j.ccm.2019.02.009.
the Center of Excellence Geopolymer and Green Technology (CE- [30] Niazi K. AmbuBag Ventilator (COVID-19 crisis); 2020.
GeoGTech), Malaysia for their financial support of this study. https://grabcad.com/library/ambubag-ventilator-covid-19-crisis-1.

12
A. El-Hadj, M. Kezrane, H. Ahmad et al. Chaos, Solitons and Fractals 150 (2021) 111169

[31] ANSYS CFX CFX-Solver theory, ver. 12. ANSYS Inc; 2012. [38] Kesti J, Olsson S. Fluid structure interaction analysis on the aerodynamic per-
[32] Nakayama H. Application of dynamic mesh method in CFD to engineering de- formance of underbody panels, Sweden: Chalmers University of Technology
signs of needle-free liquid jet injector and diaphragm-less shock tube. Concor- Goteborg; 2014. M. S. Thesis.
dia University; 2013. M. S. Thesis„ Montreal, Quebec, Canada. [39] Ameur H. Pressure drop and vortex size of power law fluids flow in branching
[33] Mortazavinia Z. Two-way fluid-structure coupling methodology for modeling channels with sudden expansion. J Appl Fluid Mech 2018;11:1739–49. doi:10.
3D Flexible hydrofoils in viscous flow. Ecole Polytechnique de Montreal; 2018. 29252/jafm.11.06.28831.
Ph.D. Thesis„ Canada. [40] Arab MI, Bouzit M, Ameur H, Kamla Y. Numerical study on thoracic aortic
[34] ANSYS Mechanical ANSYS mechanical theory guide, ver. 12. ANSYS Inc; 2012. aneurysms: the aneurysm aggravation effects on the secondary flow motion.
[35] Kim NH. Introduction to nonlinear finite element analysis. New York: Springer; Mechanika 2020;26:407–15. doi:10.5755/j01.mech.26.5.23254.
2018. doi:101007/978- 1- 4419- 1746- 1. [41] Ameur H, Bouzit M. 3D hydrodynamics and shear rates’ variability in
[36] Barlas AK, Akay B. Optimization of morphing flaps based on fluid structure the United States Pharmacopeia paddle dissolution apparatus. Int J Pharm
interaction modeling. Wind Energy Symp 2018. doi:10.2514/6.2018-0998. 2013;452:42–51 2013. doi:10.1016/j.ijpharm.2013.04.049.
[37] Menter F, Sharkey P, Yakubov S, Kuntz M. Overview of fluid-structure coupling [42] Katira B, Yoshida T, Kavanagh BP. 34- Principles of mechanical ventilation,
in ANSYS-CFX, Terry Jones Pipeline Technology; Ocean Space Utilization. In: Kendig’s disorders of the respiratory tract in children. 9th Edition; 2019.
Proceedings of the 25th international conference on offshore mechanics and doi:10.1016/B978- 0- 323- 44887- 1.0 0 034-1.
arctic engineering; 2006.

13