Case Studies in Thermal Engineering 63 (2024) 105225

Contents lists available at ScienceDirect

Case Studies in Thermal Engineering

journal homepage: www.elsevier.com/locate/csite

Dynamic characteristics of the hydrogen injector-ejector unit in a

PEM fuel cell system
Jiquan Han a , Yuhang Chen a , Zihui Pang a , Jianmei Feng a, * , Anna Diao b ,
Yanchen Yao b , Xueyuan Peng a, c
a
School of Energy and Power Engineering, Xi’an Jiaotong University, Xi’an, 710049, Shaanxi, China
b
Shanghai Marine Diesel Engine Research Institute, Shanghai, 201108, Shanghai, China
c
State Key Laboratory of Multiphase Flow in Power Engineering, Xi’an Jiaotong University, Xi’an, 710049, Shaanxi, China

A R T I C L E I N F O A B S T R A C T

Handling Editor: Huihe Qiu The ejector, as a fluid-dynamics controlled passive component, is a promising hydrogen recir­
Keywords:
culation device in proton exchange membrane (PEM) fuel cell systems. Nevertheless, its perfor­
Ejector mance is significantly affected by dynamic operating conditions and the behavior of the injector.
PEM fuel cell This study investigates the dynamic performance of an ejector integrated with a hydrogen injector
Hydrogen recirculation in a 100 kW PEM fuel cell system. A dynamic computational fluid dynamics (CFD) model of the
Dynamic performance integrated injector-ejector unit is developed using a dynamic mesh method and validated through
theoretical analysis and experimental data. A thorough analysis of the global performance and
local fluid flow characteristics of the injector-ejector unit is conducted. The results show that the
hydrogen injector, with a throat diameter of 2.00 mm, exhibits an adjustable linear flow range
from 0 to 1.62 g/s under an inlet pressure of 2.0 MPa. Additionally, the primary mass flow rate
can be linearly regulated by adjusting the valve gap, while the secondary mass flow rate exhibits
asynchronous behavior with the primary flow. The velocity and temperature distributions within
the injector-ejector unit undergo significant changes under dynamic conditions. These findings
contribute to optimizing the design and control strategies of hydrogen supply components in PEM
fuel cell systems.

1. Introduction
The proton exchange membrane (PEM) fuel cell is an efficient energy converter that harnesses the chemical reaction between
hydrogen and oxygen to produce electricity [1]. The overall efficiency and lifespan of a PEM fuel cell stack heavily depend on the
reliability of its key components, which encompass auxiliary subsystems for hydrogen and oxygen supply, as well as heat and water
management. Moreover, the widespread commercial application of fuel cell systems is hindered by challenges related to the cost and
durability of these auxiliary components [2,3].
Hydrogen recirculation devices, such as mechanical pumps, ejectors, and electrochemical pumps, are key components in hydrogen
supply subsystems [4]. Among these options, ejectors offer several advantages, including their simple structure, low cost, and absence
of parasitic power. Consequently, the ejector has been popular in refrigeration, desalination and fuel cell systems [5,6]. Nonetheless, as
a fluid-dynamics controlled component, the ejector performance is susceptible to changes in operating conditions [7].

* Corresponding author. School of Energy and Power Engineering, Xi’an Jiaotong University, 28 Xianning West Road, Xi’an, 710049, China.
E-mail address: [email protected] (J. Feng).

https://doi.org/10.1016/j.csite.2024.105225
Received 18 October 2023; Received in revised form 3 June 2024; Accepted 28 September 2024
Available online 29 September 2024
2214-157X/© 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC license
(http://creativecommons.org/licenses/by-nc/4.0/).
J. Han et al. Case Studies in Thermal Engineering 63 (2024) 105225

Previous studies have investigated the working characteristics of ejectors under various operating conditions in fuel cell systems [8,
9]. The exhaust gas from the fuel cell stack, specifically the humidity and temperature of the recirculation flow, can significantly
impact ejector performance [10]. Yin et al. [11] and Hailun et al. [12] observed that higher humidity and temperature of the recir­
culation flow resulted in a decrease in the hydrogen entrainment ratio of the ejector by using computational fluid dynamics (CFD)
simulation. Toghyani et al. [13] conducted a theoretical analysis to examine the influence of fuel cell stack parameters, including stack
temperature and relative humidity, on ejector performance. Their findings indicated that the hydrogen recirculation ratio decreased
with increasing temperature and humidity. Yang et al. [14] investigated numerically the ejector performance under different stack
operating conditions, revealing that the pressure drop affected the entrainment ratio. Kuo and Hsieh [15] also demonstrated that the
hydrogen entrainment ratio could increase with higher stack pressure and temperature. Furthermore, Bian et al. [16] investigated the
flow separation characteristics of a hydrogen recirculation ejector across a broad range of stack power levels. Their study revealed that
flow separation can indeed occur within the ejector when its geometry is not properly configured, resulting in a decline in entrainment
performance.
The performance of passive ejectors can be influenced by other components in hydrogen recirculation systems, including the
hydrogen injector, separator, and purge valve [4]. The presence of increased resistance loss in the water separator can decrease the
entrainment ratio of the ejector [17]. In an experimental study conducted by Wang et al. [18], the dynamic performance of the ejector
was investigated during the nitrogen purge process. The results revealed that the dynamic operation of the purge valve could improve
the ejector performance under low power conditions. In addition, the hydrogen injector is responsible for supplying pure hydrogen to
the primary inlet of the ejector [19]. Jung et al. [20] examined the stack performance of a 100 kW fuel cell system when the injector
and ejector operated in conjunction. Experimental results indicated that the pulse flow created by the injector facilitated stable voltage
output during low stack power conditions (below 20 kW). Singer et al. [21] found that the pulsing opening of the hydrogen injector
affected the anode pressure fluctuation in PEM fuel cell systems. Additionally, Hwang [19] conducted experimental tests to evaluate
the performance variation of a fuel cell stack equipped with an integrated injector-ejector unit. The results demonstrated that the
entrainment ratio ranged from 40 % to 50 % at stack powers of 1.45 kW. These studies collectively emphasize the interdependence
between the ejector and other components in the hydrogen recirculation system.
Several systematic models have been proposed to investigate the performance of ejectors under dynamic operating conditions and
their interactions with other system components. Nikiforow et al. [22] and Yin et al. [23] developed the one-dimensional (1-D)
thermodynamic system model, analyzing the impact of stack operating parameters and control strategies on ejector performance. Kuo
et al. [24] and Tri et al. [25] implemented a 1-D MATLAB/Simulink model to study ejector-driven hydrogen recirculation systems.
Huang et al. [26] developed a 1-D system model based on the ejector, revealing significant dynamic fluctuations in the secondary flow
due to varying primary flow from the hydrogen injector. However, these 1-D system models lack the ability to conduct a detailed
analysis of the internal flow characteristics of ejectors. Besagni et al. [27] established a lumped-parameter CFD model to analyze the
influence of ejector performance on stack performance. Their CFD model assumed a steady-state condition and did not consider the
effects of dynamic components. Han et al. [28] developed a multi-component CFD model that integrated anode flow channel, water
separator, and purge valve. Nevertheless, their model did not account for the dynamic influence of the hydrogen injector on ejector
performance.
In summary, understanding the interactions and dynamics characteristics of key components is crucial for PEM fuel cell systems.
However, few works have focused on the influence of the hydrogen injector operation and the transient flow characteristics of the
ejector. Consequently, this study proposes a dynamic CFD model for the integrated injector-ejector unit. The dynamic mesh method is
utilized to accurately simulate the actual dynamic working process of the hydrogen injector. By studying the global performance and
local characteristics of the injector-ejector unit under transient conditions, a comprehensive understanding of its dynamic behavior can
be obtained.

Fig. 1. Schematic diagram of a typical ejector-driven hydrogen recirculation in a PEM fuel cell system.

2
J. Han et al. Case Studies in Thermal Engineering 63 (2024) 105225

2. Injector-ejector unit in the PEM fuel cell system

2.1. Hydrogen recirculation system
Fig. 1 illustrates the schematic diagram of a typical ejector-driven hydrogen recirculation in a PEM fuel cell system. The hydrogen
demand of the stack is met by extracting hydrogen from a high-pressure hydrogen cylinder, which can reach pressures of up to 70 MPa.
A pressure-reducing valve is utilized to regulate the hydrogen pressure supplied to the injector-ejector unit. To ensure the efficiency
and durability of the fuel cell stack, an excess supply of hydrogen is typically required. Additionally, the unutilized hydrogen must be
recycled since the fuel cell stack is unable to fully consume all the hydrogen supplied to it. On the cathode side of the fuel cell stack,
nitrogen, and liquid water pass through the proton exchange membrane to the anode side [29]. Consequently, the anode discharge
comprises a mixture of hydrogen, water vapor, nitrogen, and liquid water. To separate the liquid water from the mixture, a water
separator is employed, while the purge valve is utilized to eliminate excess nitrogen.
The injector-ejector unit discussed in this study is specifically designed for a commercial 100 kW fuel cell system. Table 1 presents
the operating conditions of the fuel cell stack. The stack comprises 370 individual cells, with a maximum stack power output of 101
kW. The anode inlet pressure increases proportionally with the stack power. The anode recirculation pressure drop is the combined
effect of pressure losses in the fuel cell stack, recirculation piping, and water separator. The average volume fraction of nitrogen is
approximately 8 %. Water vapor is considered to be saturated and its concentration is determined by the ratio of the saturated water
vapor pressure to the recirculation flow pressure [28].

2.2. Hydrogen injector and hydrogen recirculation ejector

The hydrogen injector is utilized to adjust the primary mass flow rate in PEM fuel cell systems. Fig. 2(a) presents the schematic
diagram of a typical hydrogen injector. The control terminal regulates the current through the solenoid coil, causing the main spool to
move axially under the influence of electromagnetic force [30]. This movement alters the gap between the main spool and the valve
seat, thereby adjusting the mass flow rate through the valve gap. The performance characteristic of the hydrogen injector is the
relationship between the hydrogen supply flow and the valve gap, which can be theoretically calculated based on isentropic flow
principles, as shown in Eq. (1) [31].
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
√ ( )k+1
√
√ k 2 k− 1
mp = 2πDn Ln pp (1)
Rg Tp k + 1

where mp, pp, and Tp represent the mass flow rate, pressure, and temperature of the supplied hydrogen, respectively. Dn and Ln refer to
the throat diameter and gap length of the injector, respectively, as depicted in Fig. 2(a). Rg and k denote the gas constant and isentropic
factor of hydrogen, respectively.
As illustrated in Fig. 2(b), the ejector is a simple device that utilizes a high-pressure motive flow, known as the primary flow, to
compress a low-pressure induced flow, referred to as the secondary flow [32]. Upon passing through the primary nozzle, the fluid
velocity increases while the pressure decreases. The high-velocity primary flow expands at the primary nozzle exit, inducing the
secondary flow into the suction chamber. The two flows subsequently mix in the mixing chamber before entering the diffuser. The
performance indicator of the hydrogen recirculation ejector is the hydrogen entrainment ratio:
ms
ωH2 = (2)
mp

where ωH2 represents the hydrogen entrainment ratio, and mp and ms denote the hydrogen mass flow rate of primary and secondary
flows, respectively.

3. Numerical model
3.1. Governing equations
The fluid flow within the injector-ejector unit is characterized by transient turbulent flow. The effects of gas gravity and heat

Table 1
The operating condition of a 100 kW PEM fuel cell system.

P [kW] Vcell [V] Icell [A] mp [g/s] panode [kPa] Δp [kPa]

8 0.82 28 0.11 133.3 3.1

17 0.80 56 0.21 142.8 4.3
24 0.78 84 0.32 152.2 5.4
31 0.76 112 0.43 161.6 6.6
39 0.75 140 0.54 171.1 7.7
53 0.73 196 0.75 189.9 9.8
66 0.71 252 0.97 208.7 11.2
78 0.69 308 1.18 227.6 12.5
90 0.67 364 1.40 246.5 13.8
101 0.65 420 1.61 265.3 15.1

3
J. Han et al. Case Studies in Thermal Engineering 63 (2024) 105225

Fig. 2. Schematic diagram of (a) the hydrogen injector and (b) the hydrogen recirculation ejector.

transfer between the fluid and solid walls are neglected [31,33]. The working medium is a multispecies mixture with a compressible
density. Furthermore, the potential phase transition of condensation within the ejector is disregarded, as it has minimal impact on the
global performance of the ejector [34]. With these assumptions in mind, the governing equations for the CFD modeling can be
expressed as follows [28].
∂ρ
+ ∇ • (ρv) = 0 (3)
∂t

∂(ρv)
+ ∇ • (ρvv) = − ∇p + ∇ • τ (4)
∂t

∂(ρE)
+ ∇ • (ρEv) = − ∇ • (pv) + ∇ • (τ • v) + ∇ • q (5)
∂t

∂(ρY)
+ ∇ • (ρvY) = − ∇J (6)
∂t

where ρ, v, p, and T represent the density, velocity, pressure, and temperature, respectively, τ denotes the stress tensor, E refers to the
total energy, λ represents the thermal conductivity, h denotes the enthalpy, Y represents the mass fraction, J denotes the diffusion flux.
The SST k-ω model is employed as it has been proven effective in accurately predicting the performance of supersonic ejectors [35,
36]. The pressure-based coupling algorithm is utilized to solve all governing equations. For gradient discretization, the least squares
cell-based scheme is employed, while the PRESTO scheme is utilized for pressure discretization. The second-order upwind scheme is
chosen for the other discrete items [31,37].

3.2. Geometry and mesh method

The fluid domains of the hydrogen injector and ejector exhibit approximately two-dimensional (2-D) symmetrical structures, as

Fig. 3. 2D mesh of the integrated injector-ejector unit.

4
J. Han et al. Case Studies in Thermal Engineering 63 (2024) 105225

shown in Fig. 2. Some numerical studies have opted for a 2-D axisymmetric model instead of a full three-dimensional configuration due
to its ability to achieve satisfactory accuracy while significantly reducing computational costs. This approach has been successfully
employed by researchers such as Besagni et al. [32], Huang et al. [38], and Nikiforow et al. [39]. Consequently, this study also adopts a
2-D axisymmetric geometry.
Fig. 3 displays the 2-D mesh employed for the integrated injector-ejector unit. The meshes consist of quadrilateral elements with a
boundary layer grid height of 0.001 mm. The minimum valve gap size is set at 0.01 mm, ensuring that at least 10 layers of grids are
present within the valve gap. In the numerical calculations, the y + value is maintained below 30 for all cases. This ensures that the
wall mesh is sufficiently refined for accurate modeling using the SST k-ω turbulence model [35]. Furthermore, the dynamic mesh
method, employing height-based layering schemes, is utilized to simulate the movement of the main spool. The wall of the main spool
is designated as a dynamic wall, while the wall of the valve seat remains static. Additionally, the design and optimization of the
structural parameters of the ejector pose significant complexities due to the interplay between multiple factors, including working
conditions and fluid physical properties [40]. This study focuses on investigating dynamic performance rather than optimizing geo­
metric parameters. Therefore, the geometric parameters utilized in this research are derived from our previous study [31], as sum­
marized in Table 2.

3.3. Boundary conditions

Fig. 4 presents the boundary conditions of the numerical model, which are determined based on the operating conditions of the
ejector in the fuel cell system. The pressure inlet boundary condition is applied to both the primary and secondary inlets of the injector-
ejector unit, while the pressure outlet boundary condition is assigned to the outlet. Specifically, the pressure at the primary inlet is
maintained at a stable value. The pressures at the secondary inlet and outlet are calculated using Equations (7)–(9) [41]. The tem­
perature at the primary inlet remains constant at 298 K, while the temperature at the secondary inlet is set equal to the anode tem­
perature due to negligible heat losses [31]. Additionally, the walls of the injector-ejector unit are assigned no-slip and adiabatic
conditions, indicating zero velocity and no heat transfer, respectively.
po = panode (7)

ps = panode− Δp (8)

panode = C1 mp + C2 (9)

where panode represents the anode pressure, ps denotes the pressure at the secondary inlet, po represents the pressure at the outlet, and
C1 and C2 denote the fitting coefficients between anode pressure and mass flow rate. Specifically, the values of C1 and C2 are 87.79 and
123.91, respectively. Fig. 4 shows that the fitting formula aligns perfectly with the data presented in Table 1.

3.4. Uncertainty analysis

Ensuring an adequate number of meshes is crucial to predict accurate simulation results. Nevertheless, increasing the number of
meshes also leads to higher computation costs. Additionally, for transient CFD modeling, the time step plays a role in determining the
accuracy of the results. Therefore, an uncertainty analysis is carried out to evaluate the mesh number independence and sensitivity to
the time step. Fig. 5 presents the variation in hydrogen mass flow rate resulting from the movement of the hydrogen injector for
different numbers of meshes and time steps. The valve gap size transitions from its minimum to maximum value over a period of 1.0 s.
Subsequently, the mass flow rate exhibits a linear variation in response to the dynamic changes in the valve gap.
The initial mesh numbers for the three levels of mesh refinement are 41,247, 83,764, and 160,208, respectively. It is important to
note that, as the moving wall progresses in time, the number of meshes changes dynamically as well. For instance, a medium number of
meshes has a count of 83,764 at the minimum gap and 122,704 at the maximum gap.
The results in Fig. 5(a) demonstrate that the mass flow rate for the three mesh sizes is nearly identical, particularly for the medium
and fine meshes. Hence, a medium number of meshes can satisfy the requirement of mesh independence. The sensitivity analysis of the
time step in Fig. 5(b) reveals that a time step of 1 × 10− 4 s provides sufficient accuracy, representing one ten-thousandth of the real

Table 2
Key geometric parameters of the injector-ejector unit [31].

Geometric parameters Symbol Values

Primary nozzle diameter of the ejector Dt 1.8 mm

Suction chamber length of the ejector Ls 20.0 mm
Nozzle axis position of the ejector NXP 12.0 mm
Convergent section length of the ejector Lc 20.0 mm
Mixing chamber diameter of the ejector Dm 6.0 mm
Mixing chamber length of the ejector Lm 36.0 mm
Diffuser length of the ejector Ld 60.0 mm
Outlet diameter of the ejector Do 14.4 mm
Inlet diameter of the injector Din 5.0 mm
Minimum valve gap size of the injector Ln 0.01 mm
Throat diameter of the injector Dn 2.0 mm

5
J. Han et al. Case Studies in Thermal Engineering 63 (2024) 105225

Fig. 4. Boundary conditions of the integrated injector-ejector unit.

Fig. 5. Uncertainty analysis of the numerical model: (a) mesh independence study, and (b) time step sensitive analysis.

6
J. Han et al. Case Studies in Thermal Engineering 63 (2024) 105225

period [28].

4. Results and discussion

4.1. Theoretical and experimental validation
The CFD model is validated through a combination of theoretical analysis and experimental validation. The performance indicators
of the injector-ejector unit are the primary mass flow rate and entrainment ratio, as expressed in Equations (1) and (2). The experi­
mental data is from our previous experimental study [42]. To compare the numerical results with theoretical or experimental data, the
relative deviation (Er) is employed, defined by Eq. (10).
X1 − X2
Er = (10)
X2

where X1 represents numerical results, X2 denotes theoretical or experimental results.

Fig. 6(a) presents a comparison between the numerical and theoretical primary mass flow rates. It can be observed that the mass
flow rate predicted by the CFD model closely aligns with the theoretical results. However, the numerical mass flow rate is slightly lower
than the theoretical values. The discrepancy observed can be attributed to the fact that Eq. (1) is derived under the assumption of
isentropic flow, neglecting the impact of friction losses. In reality, high-speed gas flow experiences significant friction losses [43],
resulting in a slightly lower actual flow rate compared to the theoretical flow rate.
Fig. 6(b) displays the comparison of the numerical and experimental entrainment ratios. The observed trend of the entrainment
ratio predicted by the CFD model is in agreement with the experimental data. Specifically, the entrainment ratio increases as the
pressure decreases when the primary flow pressure exceeds 0.40 MPa. Conversely, it rapidly decreases as the primary pressure de­
creases below 0.40 MPa. The average relative deviation across all test cases is 4.4 %. Nevertheless, when the primary flow pressure is
0.30 MPa, the maximum relative deviation between the numerical and experimental entrainment ratios reaches 12.1 %. This deviation
can be attributed to the ejector operating in a subcritical state with unstable performance under low flow rates [44]. Overall, the CFD
model accurately predicts the performance of the injector-ejector unit, as demonstrated through theoretical and experimental
validation.

Fig. 6. Validation of the numerical model: (a) mass flow rate, and (b) entrainment ratio.

7
J. Han et al. Case Studies in Thermal Engineering 63 (2024) 105225

4.2. Operating performance curve of the hydrogen injector

The hydrogen injector is employed to regulate the mass flow rate by making precise adjustments to the valve gap. The relationship
between mass flow rate and valve gap size or opening is depicted in the operating performance curve [30,45]. However, it is important
to note that Eq. (1) reveals the dependency of the mass flow not only on the valve clearance size but also on the throat diameter and
inlet pressure.
Fig. 7(a) demonstrates the effect of the throat diameter on the injector’s mass flow rate. Maintaining a constant valve gap size, an
increase in the throat diameter leads to higher mass flow rates. For instance, with a valve gap size of 0.15 mm, the mass flow rate
reaches 1.12 g/s for a throat diameter of 2.00 mm, in contrast to 0.67 g/s for a throat diameter of 1.20 mm. Therefore, a larger throat
diameter should be considered for applications requiring higher mass flow rates. Nevertheless, a larger throat diameter results in a
narrower range of valve gap size for regulation. For example, at a throat diameter of 2.00 mm, the control range of valve gap size is
0.00–0.23 mm, compared to 0.00–0.16 mm at a throat diameter of 2.80 mm. This narrower gap range poses challenges in terms of high
control precision [46]. Consequently, an optimal throat diameter of 2.00 mm is determined for the hydrogen injector in a 100 kW
stack, balancing the mass flow range and the valve gap size range.
Fig. 7(b) illustrates the influence of the inlet pressure on the injector’s mass flow rate. As the inlet pressure rises, the mass flow rate
increases accordingly. For instance, at an inlet pressure of 1.60 MPa, the mass flow rate measures 0.89 g/s, whereas at an inlet pressure
of 2.00 MPa, it rises to 1.12 g/s. Consequently, higher inlet pressure is advantageous for designing an injector with a larger flow rate.
Nevertheless, high inlet pressure can result in a significant pressure difference between the injector’s inlet and outlet [47], necessi­
tating careful considerations in the structure design of the hydrogen injector.

4.3. Operating performance curve of the ejector

The operating performance curve of the ejector provides valuable insights into the hydrogen entrainment ratio across various stack
powers [31], as depicted in Fig. 8. Notably, a discernible linear relationship exists between the primary flow rate and stack power. For
instance, at a stack power of 17 kW, the primary flow rate records 0.21 g/s. The primary flow rate steadily increases to 1.61 g/s as the

Fig. 7. Operating curve of the hydrogen injector: (a) effects of nozzle throat diameter, and (b) effects of inlet pressure.

8
J. Han et al. Case Studies in Thermal Engineering 63 (2024) 105225

stack power reaches 101 kW. A similar trend is observed in the case of the secondary flow rate, which also rises with increasing stack
power. The ejector works under stable critical conditions at rated stack power. However, it is noteworthy that the secondary flow rate
becomes negligible when the stack power falls below 17 kW. This occurrence can be attributed to small flow velocity at the outlet of the
primary nozzle due to insufficient primary mass flow [42]. Consequently, the shear entrainment effect of the primary flow fails to
surpass the flow resistance of the secondary flow [48].

4.4. Dynamic characteristics of the injector-ejector unit

The global performance and local fluid flow characteristics of the ejector can be influenced by the dynamic adjustment of the
hydrogen injector. Fig. 9 presents the global performance of the injector-ejector unit under dynamic conditions, generated by the
periodic movement of the main spool. Operating with a primary inlet pressure of 2.0 MPa and a throat diameter of 2.00 mm, the
hydrogen injector’s behavior is analyzed over a specific cycle lasting from 2.0 s to 4.0 s.
As seen in Fig. 9(a), the main spool progressively moves away from the valve seat from 2.0 s to 3.0 s, resulting in a linear increase in
the valve gap size. The primary mass flow rate exhibits a similar linear increase over time. At 3.0 s, the valve gap size reaches its
maximum value of 0.23 mm, and the mass flow rate reaches its peak of 1.62 g/s. Subsequently, between 3.0 s and 4.0 s, the main spool
gradually moves closer to the valve seat, leading to a corresponding decrease in the primary mass flow rate.
The secondary flow rate and the hydrogen entrainment ratio of the ejector demonstrate significant variability in response to the
periodic changes in the primary flow, as illustrated in Fig. 9(b). Notably, it is crucial to recognize that the fluctuations in the secondary
flow rate do not entirely mirror those of the primary flow rate. For instance, the primary mass flow rate decreases from 3.0 s to 3.3 s,
during which the secondary flow rate increases slightly due to fluid inertia [49]. Another noteworthy scenario arises between 3.6 s and
4.2 s, during which the primary flow rate remains relatively low, while the secondary flow rate is substantial. As a result, a notably high
hydrogen entrainment ratio is observed. The average entrainment ratio over a 2.0 s period (e.g., 4.0 s–6.0 s) is 1.48, which is
significantly higher compared to the value of 1.08 observed under steady conditions (see Fig. 8).
These findings suggest that the pulsating behavior of the primary flow may positively impact the ejector’s performance. Jung et al.
[20] introduced a concept to enhance ejector performance by generating pulsed primary flow, particularly under low stack power
conditions. Their experimental tests demonstrated improved stack voltage under such low-power conditions. Similarly, Hwang [19]
proposed a similar scheme for pulse ejectors. However, it is essential to emphasize that these studies primarily centered on monitoring
the stack’s performance, without in-depth exploration into the dynamic behavior of the ejector itself. Therefore, the present study
contributes significantly by shedding light on the dynamic interactions within the injector-ejector unit, providing robust support for
existing research on the advantages of employing pulsating primary flow to enhance the entrainment performance.
The dynamic movement of the main spool exerts a direct and significant influence on the flow field distribution within the injector-
ejector unit. Fig. 10 showcases the transient flow characteristics, illustrating the distribution of the Mach number, pressure, and
temperature during dynamic operation. Specifically, Fig. 10(a) highlights the behavior of the Mach number within the ejector as the
primary mass flow rate gradually increases from 2.2 s to 3.0 s. The larger the primary flow rate, the more significant the shock wave. A
noteworthy observation in the hydrogen injector is the appearance of a shock train. This occurrence is particularly evident at 2.2 s
when the Mach number at the outlet of the valve gap surpasses 1.0, indicating the presence of strong perturbed compressional waves
[50]. Furthermore, at 3.0 s, the shock within the injector weakens as the larger valve gap causes a decrease in velocity. This effect leads
to a different velocity distribution, aligning with the findings of Ye et al. [47]. Their study revealed that the local fluid velocity was
notably higher when the valve opening was 5 % compared to when it was 60 %.
The pressure contours within the ejector reveal significant changes in response to the increase in the primary mass flow rate from
2.2 s to 3.0 s, as seen in Fig. 10(b). The average pressure within the ejector exhibits a corresponding rise during this period.
Furthermore, the presence of a shock train at the primary nozzle exit becomes more pronounced as the primary mass flow rate in­
creases [51]. The measured data indicates that the length of the first shock wave varies with time, measuring approximately 2.15 mm

Fig. 8. Operating performance curve of the hydrogen recirculation ejector.

9
J. Han et al. Case Studies in Thermal Engineering 63 (2024) 105225

Fig. 9. Transient global performance of the injector-ejector unit: (a) valve gap size and mass flow rate of the injector, and (b) secondary mass flow rate and hydrogen
entrainment ratio of the ejector.

at 2.5 s, 2.92 mm at 2.8 s, and 3.20 mm at 3.0 s, respectively. Notably, this increase in shock wave length aligns with the experimental
findings reported by Han et al. [34].
Fig. 10(c) shows the temperature contours of the injector-ejector unit at different times. It is noted that as the primary flow passes
through the primary nozzle of the ejector, potential energy is converted to kinetic energy, resulting in a decrease in its temperature. For
example, at 2.5 s, the fluid temperature in the primary jet region reaches a minimum of 179 K. Conversely, the temperature of the
secondary flow is as high as 353 K. Due to this significant temperature difference, the water vapor in the secondary flow may condense.
However, our previous studies [52] have shown that condensation significantly impacts ejector performance only when the primary
flow temperature is below 273 K.
Furthermore, the high-temperature secondary flow and low-temperature primary flow mix in the mixing chamber and are dis­
charged from the ejector outlet. The temperature of the discharged mixture is 318 K, which is lower than the stack’s operating
temperature of 353 K. It should be noted that the lower exhaust temperature of the ejector will dissipate some of the heat generated by
the stack, potentially reducing the thermal load on the heat management system.
The dynamic behavior of the injector-ejector unit, characterized by the random adjustment of the main spool, can be effectively
simulated using the developed CFD modeling. Fig. 11 illustrates the transient characteristics of the injector-ejector unit under random
dynamic conditions. The dynamic adjustment of the hydrogen injector involves three distinct states: increasing, stabilizing, and
decreasing the valve opening, spanning from 0 s to 5.0 s.
It can be seen from Fig. 11(a) that a close correspondence is observed between the change in the primary flow and the varying valve
gap size. Specifically, from 1.5 s to 2.0 s, the primary flow rate increases from 0.97 g/s to 1.62 g/s as the valve gap size widens from
0.13 mm to 0.23 mm. Following this, between 2.0 s and 3.0 s, the primary flow rate stabilizes at 1.62 g/s, maintaining a constant valve
gap size of 0.23 mm. Subsequently, from 3.0 s to 3.5 s, the valve gap size rapidly adjusts from 0.23 mm to 0.05 mm, resulting in a linear
decrease in the primary flow rate to 0.35 g/s.
With the changes in the primary flow, corresponding variations in the secondary flow also occur. Generally, the secondary flow rate
increases or decreases in tandem with the primary flow changes. Nevertheless, it is crucial to note that the alterations in the two flow
rates are not perfectly synchronized. For instance, during the period from 0.5 s to 1.5 s, while the primary flow remains stable at 0.97
g/s, the secondary flow continues to increase. Similarly, between 3.5 s and 4.5 s, the primary flow remains stable, while the secondary
flow experiences a continuous decrease.

10
J. Han et al. Case Studies in Thermal Engineering 63 (2024) 105225

Fig. 10. Transient local fluid flow characteristics of the injector-ejector unit.

Fig. 11(b) and (c) depict the temperature and Mach number along the axis of the injector-ejector unit at different times, respec­
tively. It is noteworthy that the change in temperature is closely related to the change in Mach number. For instance, at 3.0 s, a clear jet
shock wave appeared at the primary nozzle exit of the ejector. Concurrently, accompanied by the drastic change in temperature, the
maximum temperature change amplitude reaches 130 K. In the mixing chamber and diffuser, the fluid velocity decreases while the
temperature gradually increases. Moreover, it is observed from Fig. 11(b) that the temperature of the fluid discharged by the ejector
ranges between 321 K and 333 K. The temperature variations of the fluid entering the fuel cell stack may have a potential effect on the
stack’s operation [53]. Therefore, the change in exhaust temperature should be considered when designing the entire fuel cell system.
Fig. 11(c) illustrates prominent shock waves within the primary nozzle outlet of the ejector, particularly evident at 2.0 s and 3.0 s.
However, no shock waves are observed within the ejector at 4.0 s or 5.0 s, as the Mach number at the ejector nozzle outlet remains
below the critical value of 1.0. Of particular interest, a pronounced shock wave is observed inside the hydrogen injector at 4.0 s. This
occurrence can be attributed to the small valve gap size at this time point, resulting in a supersonic flow at the valve gap [47].
Conversely, at 2.0 s, despite the relatively large primary mass flow rate, no supersonic shock train is formed in the injector. This is
because a larger valve gap ensures that the Mach number at the valve gap remains below the critical value of 1.0. Furthermore, it is
noteworthy that shock waves may appear simultaneously in both the injector and ejector, as exemplified at 1.0 s.
Additionally, it is essential to consider the potential consequences of supersonic speeds, as they may generate flow-induced noise.
Consequently, when designing the hydrogen injection valve, measures should be implemented to effectively reduce such noise. For
instance, Ariyadi et al. [45] proposed the modification of pipeline structures as a viable approach to mitigate flow-induced noise.

5. Conclusions
This study delves into the dynamic characteristics of the integrated injector-ejector unit in a PEM fuel cell system. By leveraging a

11
J. Han et al. Case Studies in Thermal Engineering 63 (2024) 105225

Fig. 11. Transient characteristics of the injector-ejector unit under random dynamic conditions.

sophisticated CFD model employing dynamic mesh methods, the actual working process of the hydrogen injector is accurately
captured, enabling a thorough analysis of the global performance and local fluid flow characteristics. The key findings and conclusions
drawn from this investigation are summarized as follows.
1) The developed CFD model exhibits high reliability in predicting performance indicators of the hydrogen injector and ejector. The
model shows relative deviations within 5.0 % for the numerical mass flow rate compared to theoretical analysis, and an average
relative deviation of 4.4 % for the numerical entrainment ratio compared to experimental tests.
2) The operating performance curve of the hydrogen injector is determined by the CFD model. The mass flow rate consistently in­
creases with a larger throat diameter and higher inlet pressure. For the injector equipped with a throat diameter of 2.00 mm and an
inlet pressure of 2.0 MPa, the linear mass flow rate ranges from 0 to 1.62 g/s.
3) Linear control of the primary mass flow rate is achieved by adjusting the valve gap. The secondary mass flow rate exhibits dynamic
variations in response to the primary flow, with a lag of 0.3 s due to fluid inertia. The hydrogen entrainment ratio increases by 37 %,
from 1.08 to 1.48, with the implementation of pulsed primary flow.
4) The velocity distribution inside the injector-ejector unit undergoes significant variations due to the movement of the main spool of
the injector. The injector-ejector unit exhibits a double shock phenomenon at a primary mass flow rate of 0.97 g/s. Moreover,
changes in the primary flow rate can cause significant differences in the temperature distribution within the ejector at different
instances. The temperature of the gas discharged by the ejector ranges between 321 K and 333 K.
The findings contribute to understanding the dynamic behavior of the ejector and hydrogen injector. The developed dynamic CFD
model provides insights into local fluid flow characteristics and global performance under dynamic operating conditions. These results
contribute to the application of PEM fuel cell systems by enabling improved performance and control strategies of hydrogen supply and

12
J. Han et al. Case Studies in Thermal Engineering 63 (2024) 105225

recirculation components.
Nomenclature
Symbols
Dn diameter of the injector nozzle throat [m]
Er relative deviation [− ]
k isentropic factor [− ]
Ln valve gap size [m]
mc mass flow rate of consumed hydrogen [kg s− 1]
mi mass flow rate at position i [kg s− 1]
P stack power [kW]
pi pressure at position i [Pa]
Ti temperature at position i [K]
v velocity [m s− 1]
Greek letters
ω entrainment ratio [− ]
ρ density [kg m− 3]
1
λ thermal conductivity [W m− K− 1 ]
Subscripts
anode anode of the fuel cell stack
cell single fuel cell
H2 hydrogen
n throat of the injector
o outlet
p primary inlet
s secondary inlet
Abbreviations
CFD computational fluid dynamics
PEM proton exchange membrane

CRediT authorship contribution statement

Jiquan Han: Conceptualization, Investigation, Software, Writing – original draft, Writing – review & editing. Yuhang Chen:
Investigation, Writing – review & editing. Zihui Pang: Investigation, Writing – review & editing. Jianmei Feng: Conceptualization,
Methodology, Writing – review & editing. Anna Diao: Resources, Writing – review & editing. Yanchen Yao: Project administration,
Writing – review & editing. Xueyuan Peng: Project administration, Resources, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to
influence the work reported in this paper.

Data availability

Data will be made available on request.

Acknowledgments
The authors also thank the support from the High-Performance Computing Platform of Xi’an Jiaotong University.

References
[1] K. Jiao, J. Xuan, Q. Du, Z. Bao, B. Xie, B. Wang, et al., Designing the next generation of proton-exchange membrane fuel cells, Nature 595 (2021) 361–369,
https://doi.org/10.1038/s41586-021-03482-7.
[2] A. Alaswad, A. Omran, J.R. Sodre, T. Wilberforce, G. Pignatelli, M. Dassisti, et al., Technical and commercial challenges of proton-exchange membrane (PEM)
fuel cells, Energies 14 (2021) 144.
[3] Y. Wang, D.F. Ruiz Diaz, K.S. Chen, Z. Wang, X.C. Adroher, Materials, technological status, and fundamentals of PEM fuel cells – a review, Mater. Today 32
(2020) 178–203, https://doi.org/10.1016/j.mattod.2019.06.005.
[4] J. Han, J. Feng, P. Chen, Y. Liu, X. Peng, A review of key components of hydrogen recirculation subsystem for fuel cell vehicles, Energy Convers. Manag. X 15
(2022) 100265, https://doi.org/10.1016/j.ecmx.2022.100265.
[5] H.K. Mukhtar, S. Ghani, Improving the performance of a commercial absorption cooling system by using ejector: a theoretical study, Case Stud. Therm. Eng. 45
(2023) 102967, https://doi.org/10.1016/j.csite.2023.102967.
[6] B.M. Tashtoush, M.A. Al-Nimr, M.A. Khasawneh, A comprehensive review of ejector design, performance, and applications, Appl energ 240 (2019) 138–172,
https://doi.org/10.1016/j.apenergy.2019.01.185.
[7] B. Elhub, S. Mat, K. Sopian, A.M. Elbreki, M.H. Ruslan, A.A. Ammar, Performance evaluation and parametric studies on variable nozzle ejector using R134A,
Case Stud. Therm. Eng. 12 (2018) 258–270, https://doi.org/10.1016/j.csite.2018.04.006.
[8] Y. Liu, Z. Tu, S.H. Chan, Applications of ejectors in proton exchange membrane fuel cells: a review, Fuel Process. Technol. (2020) 106683, https://doi.org/
10.1016/j.fuproc.2020.106683.
[9] H. Ding, Y. Dong, Y. Zhang, Y. Yang, C. Wen, Energy efficiency assessment of hydrogen recirculation ejectors for proton exchange membrane fuel cell (PEMFC)
system, Appl energ 346 (2023) 121357, https://doi.org/10.1016/j.apenergy.2023.121357.
[10] X. Wang, S. Xu, C. Xing, Numerical and experimental investigation on an ejector designed for an 80 kW polymer electrolyte membrane fuel cell stack, J. Power
Sources 415 (2019) 25–32, https://doi.org/10.1016/j.jpowsour.2019.01.039.

13
J. Han et al. Case Studies in Thermal Engineering 63 (2024) 105225

[11] Y. Yin, M. Fan, K. Jiao, Q. Du, Y. Qin, Numerical investigation of an ejector for anode recirculation in proton exchange membrane fuel cell system, Energ convers
manage 126 (2016) 1106–1117, https://doi.org/10.1016/j.enconman.2016.09.024.
[12] Z. Hailun, W. Sun, H. Xue, W. Sun, L. Wang, L. Jia, Performance analysis and prediction of ejector based hydrogen recycle system under variable proton
exchange membrane fuel cell working conditions, Appl. Therm. Eng. 197 (2021) 117302, https://doi.org/10.1016/j.applthermaleng.2021.117302.
[13] S. Toghyani, E. Afshari, E. Baniasadi, A parametric comparison of three fuel recirculation system in the closed loop fuel supply system of PEM fuel cell, Int J
Hydrogen Energ 44 (2019) 7518–7530, https://doi.org/10.1016/j.ijhydene.2019.01.260.
[14] Y. Yang, W. Du, T. Ma, W. Lin, M. Cong, H. Yang, et al., Numerical studies on ejector structure optimization and performance prediction based on a novel
pressure drop model for proton exchange membrane fuel cell anode, Int J Hydrogen Energ 45 (2020) 23343–23352, https://doi.org/10.1016/j.
ijhydene.2020.06.068.
[15] J. Kuo, C. Hsieh, Numerical investigation into effects of ejector geometry and operating conditions on hydrogen recirculation ratio in 80 kW PEM fuel cell
system, Energy 233 (2021) 121100, https://doi.org/10.1016/j.energy.2021.121100.
[16] J. Bian, Y. Zhang, Y. Liu, L. Gong, X. Cao, Structural optimization of hydrogen recirculation ejector for proton exchange membrane fuel cells considering the
boundary layer separation effect, J. Clean. Prod. 397 (2023) 136535, https://doi.org/10.1016/j.jclepro.2023.136535.
[17] T. Ma, M. Cong, Y. Meng, K. Wang, D. Zhu, Y. Yang, Numerical studies on ejector in proton exchange membrane fuel cell system with anodic gas state
parameters as design boundary, Int J Hydrogen Energ 46 (2021) 38841–38853, https://doi.org/10.1016/j.ijhydene.2021.09.148.
[18] X. Wang, Y. Lu, B. Zhang, J. Liu, S. Xu, Experimental analysis of an ejector for anode recirculation in a 10 kW polymer electrolyte membrane fuel cell system, Int
J Hydrogen Energ (2021), https://doi.org/10.1016/j.ijhydene.2021.10.140.
[19] J. Hwang, Passive hydrogen recovery schemes using a vacuum ejector in a proton exchange membrane fuel cell system, J. Power Sources 247 (2014) 256–263,
https://doi.org/10.1016/j.jpowsour.2013.08.126.
[20] Jung SK, Noh YG, Jeon US. A development of the fuel cell system that the jet-pump is applied. American Society of Mechanical Engineers Digital Collection2013.
https://doi.org/10.1115/FuelCell2013-18004..
[21] G. Singer, G. Gappmayer, M. Macherhammer, P. Pertl, A. Trattner, A development toolchain for a pulsed injector-ejector unit for PEM fuel cell applications, Int J
Hydrogen Energ (2022), https://doi.org/10.1016/j.ijhydene.2022.05.177.
[22] K. Nikiforow, J. Pennanen, J. Ihonen, S. Uski, P. Koski, Power ramp rate capabilities of a 5 kW proton exchange membrane fuel cell system with discrete ejector
control, J. Power Sources 381 (2018) 30–37, https://doi.org/10.1016/j.jpowsour.2018.01.090.
[23] X. Yin, X. Wang, L. Wang, B. Qin, H. Liu, L. Jia, et al., Cooperative control of air and fuel feeding for PEM fuel cell with ejector-driven recirculation, Appl. Therm.
Eng. 199 (2021) 117590, https://doi.org/10.1016/j.applthermaleng.2021.117590.
[24] J. Kuo, W. Jiang, C. Li, T. Hsu, Numerical investigation into hydrogen supply stability and I-V performance of PEM fuel cell system with passive Venturi ejector,
Appl. Therm. Eng. 169 (2020) 114908, https://doi.org/10.1016/j.applthermaleng.2020.114908.
[25] D. Truong Le Tri, H.N. Vu, H.L. Nguyen, Y. Kim, S. Yu, A comparative study of single and dual ejector concepts for anodic recirculation system in high-
performance vehicular proton exchange membrane fuel cells, Int J Hydrogen Energ (2023), https://doi.org/10.1016/j.ijhydene.2023.03.234.
[26] P. Huang, J. Kuo, C. Wu, Design and evaluation of dual passive hydrogen recovery subsystem for 10 kW PEMFC, Int J Hydrogen Energ (2023), https://doi.org/
10.1016/j.ijhydene.2023.01.337.
[27] G. Besagni, R. Mereu, F. Inzoli, P. Chiesa, Application of an integrated lumped parameter-CFD approach to evaluate the ejector-driven anode recirculation in a
PEM fuel cell system, Appl. Therm. Eng. 121 (2017) 628–651, https://doi.org/10.1016/j.applthermaleng.2017.04.111.
[28] J. Han, B. Zhao, Z. Pang, J. Feng, X. Peng, Transient characteristics investigation of the integrated ejector-driven hydrogen recirculation by multi-component
CFD simulation, Int J Hydrogen Energ 47 (2022) 29053–29068, https://doi.org/10.1016/j.ijhydene.2022.06.236.
[29] O.S. Ijaodola, Z. El- Hassan, E. Ogungbemi, F.N. Khatib, T. Wilberforce, J. Thompson, et al., Energy efficiency improvements by investigating the water flooding
management on proton exchange membrane fuel cell (PEMFC), Energy 179 (2019) 246–267, https://doi.org/10.1016/j.energy.2019.04.074.
[30] F. Liu, S. Cao, W. Zhou, D. Zhao, L. Zheng, M. Shao, Transient flow analysis on opening process of pneumatic gas proportional valve with two-solenoid valve,
Flow Meas. Instrum. 89 (2023) 102291, https://doi.org/10.1016/j.flowmeasinst.2022.102291.
[31] J. Han, J. Feng, T. Hou, X. Peng, Performance investigation of a multi-nozzle ejector for proton exchange membrane fuel cell system, Int J Energ Res 45 (2021)
3031–3048, https://doi.org/10.1002/er.5996. STUD.
[32] G. Besagni, N. Cristiani, Multi-scale evaluation of an R290 variable geometry ejector, Appl. Therm. Eng. 188 (2021) 116612, https://doi.org/10.1016/j.
applthermaleng.2021.116612.
[33] Z. Liu, Z. Liu, K. Jiao, Z. Yang, X. Zhou, Q. Du, Numerical investigation of ejector transient characteristics for a 130-kW PEMFC system, Int J Energ Res 44 (2020)
3697–3710, https://doi.org/10.1002/er.5156.
[34] J. Han, Z. Pang, J. Feng, G. Besagni, R. Mereu, F. Inzoli, et al., Experimental and numerical study on the ejector containing condensable species in the secondary
flow for PEM fuel cell applications, Appl. Therm. Eng. (2023) 121091, https://doi.org/10.1016/j.applthermaleng.2023.121091.
[35] G. Besagni, N. Cristiani, L. Croci, G.R. Guédon, F. Inzoli, Computational fluid-dynamics modelling of supersonic ejectors: screening of modelling approaches,
comprehensive validation and assessment of ejector component efficiencies, Appl. Therm. Eng. 186 (2021) 116431, https://doi.org/10.1016/j.
applthermaleng.2020.116431.
[36] W. Sun, X. Ma, Y. Zhang, L. Jia, H. Xue, Performance analysis and optimization of a steam ejector through streamlining of the primary nozzle, Case Stud. Therm.
Eng. 27 (2021) 101356, https://doi.org/10.1016/j.csite.2021.101356.
[37] M. Haida, J. Smolka, A. Hafner, M. Mastrowski, M. Palacz, K.B. Madsen, et al., Numerical investigation of heat transfer in a CO2 two-phase ejector, Energy 163
(2018) 682–698, https://doi.org/10.1016/j.energy.2018.08.175.
[38] P. Huang, J. Kuo, C. Wu, Simulation and experimental measurements of 10-kW PEMFC passive hydrogen recovery system, Int J Hydrogen Energ 48 (2023)
16790–16801, https://doi.org/10.1016/j.ijhydene.2023.01.136.
[39] K. Nikiforow, P. Koski, H. Karimäki, J. Ihonen, V. Alopaeus, Designing a hydrogen gas ejector for 5 kW stationary PEMFC system – CFD-modeling and
experimental validation, Int J Hydrogen Energ 41 (2016) 14952–14970, https://doi.org/10.1016/j.ijhydene.2016.06.122.
[40] H. Zhu, J. Liu, J. Yu, P. Yang, Artificial neural network-based predictive model for supersonic ejector in refrigeration system, Case Stud. Therm. Eng. 49 (2023)
103313, https://doi.org/10.1016/j.csite.2023.103313.
[41] P. Pei, P. Ren, Y. Li, Z. Wu, D. Chen, S. Huang, et al., Numerical studies on wide-operating-range ejector based on anodic pressure drop characteristics in proton
exchange membrane fuel cell system, Appl energ 235 (2019) 729–738, https://doi.org/10.1016/j.apenergy.2018.11.005.
[42] J. Feng, J. Han, T. Hou, X. Peng, Performance analysis and parametric studies on the primary nozzle of ejectors in proton exchange membrane fuel cell systems,
Energy Sources, Part A Recovery, Util. Environ. Eff. (2020), https://doi.org/10.1080/15567036.2020.1804489.
[43] Y. Zhu, Y. Li, New theoretical model for convergent nozzle ejector in the proton exchange membrane fuel cell system, J. Power Sources 191 (2009) 510–519,
https://doi.org/10.1016/j.jpowsour.2009.02.014.
[44] G. Besagni, N. Cristiani, L. Croci, G.R. Guédon, F. Inzoli, Multi-scale evaluation of ejector performances: the influence of refrigerants and ejector design, Appl.
Therm. Eng. 186 (2021) 116502, https://doi.org/10.1016/j.applthermaleng.2020.116502.
[45] H.M. Ariyadi, J. Jeong, K. Saito, Computational analysis of hydrogen flow and aerodynamic noise emission in a solenoid valve during fast-charging to fuel cell
automobiles, J. Energy Storage 45 (2022) 103661, https://doi.org/10.1016/j.est.2021.103661.
[46] J. Zhao, P. Yue, L. Grekhov, X. Ma, Hold current effects on the power losses of high-speed solenoid valve for common-rail injector, Appl. Therm. Eng. 128 (2018)
1579–1587, https://doi.org/10.1016/j.applthermaleng.2017.09.123.
[47] J. Ye, J. Cui, Z. Hua, J. Xie, W. Peng, W. Wang, Study on the high-pressure hydrogen gas flow characteristics of the needle valve with different spool shapes, Int J
Hydrogen Energ 48 (2023) 11370–11381, https://doi.org/10.1016/j.ijhydene.2022.04.073.
[48] O. Lamberts, P. Chatelain, N. Bourgeois, Y. Bartosiewicz, The compound-choking theory as an explanation of the entrainment limitation in supersonic ejectors,
Energy 158 (2018) 524–536, https://doi.org/10.1016/j.energy.2018.06.036.

14
J. Han et al. Case Studies in Thermal Engineering 63 (2024) 105225

[49] P.J. Resto, E. Berthier, D.J. Beebe, J.C. Williams, An inertia enhanced passive pumping mechanism for fluid flow in microfluidic devices, Lab Chip 12 (2012)
2221–2228, https://doi.org/10.1039/C2LC20858J.
[50] Y. Zhang, A. Musa, Influence of coaxial air jet on mass diffusion of hydrogen jet injected through lobe-injector at supersonic flow, Int J Hydrogen Energ 47
(2022) 35886–35896, https://doi.org/10.1016/j.ijhydene.2022.08.140.
[51] Y. Zhou, G. Chen, X. Hao, N. Gao, O. Volovyk, Working mechanism and characteristics analysis of a novel configuration of a supersonic ejector, Energy 278
(2023) 128010, https://doi.org/10.1016/j.energy.2023.128010.
[52] J. Han, Y. Chen, J. Feng, L. Wang, X. Peng, Effects of primary flow temperature on phase change characteristics in hydrogen recirculation ejector for PEM fuel
cell system, Int J Hydrogen Energ 68 (2024) 1133–1143, https://doi.org/10.1016/j.ijhydene.2024.04.338.
[53] J. Xu, C. Zhang, Z. Wan, X. Chen, S.H. Chan, Z. Tu, Progress and perspectives of integrated thermal management systems in PEM fuel cell vehicles: a review,
Renew. Sustain. Energy Rev. 155 (2022) 111908, https://doi.org/10.1016/j.rser.2021.111908.

15