Case Studies in Thermal Engineering 33 (2022) 101994

Contents lists available at ScienceDirect

Case Studies in Thermal Engineering

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

Particle-to-fluid direct-contact counter-flow heat exchanger:

Simple-models validation and integration with a particle-based
central tower system
Shaker Alaqel a, b, *, Nader S. Saleh a, b, Rageh Saeed a, b, Eldwin Djajadiwinata a, b,
Muhammad Sarfraz c, d, Abdulelah Alswaiyd a, Hany Al-Ansary a, b, Obida Zeitoun a, b,
Syed Danish d, Zeyad Al-Suhaibani a, Abdelrahman El-Leathy a, e, Sheldon Jeter c,
Ahmad Khayyat f
a
Mechanical Engineering Department, King Saud University, P.O. Box 800, Riyadh, 11421, Saudi Arabia
b
K.A.CARE Energy Research and Innovation Center at Riyadh, Saudi Arabia
c
Georgia Institute of Technology, School of Mechanical Engineering, 771 Ferst Drive, Atlanta, GA, 30332, USA
d
Sustainable Energy Technologies Center, King Saud University, P.O. Box 800, Riyadh, 11421, Saudi Arabia
e
Mechanical Power Engineering Dept., Faculty of Engineering, El-Mataria, Helwan University, Cairo, 11718, Egypt
f
Research & Development Center, Saudi Aramco, P.O.Box 11953, Dhahran, 31311, Saudi Arabia

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

Keywords: This paper presents validation studies of two theoretical models of a direct-contact counter-flow
Particle-to-fluid heat exchanger particle-to-air heat exchanger. It also describes the integration of the proposed heat exchanger
Direct-contact heat exchanger with a particle-based central tower system. Two basic models are analyzed: (1) the mixing model,
Falling-bed heat exchanger which treats falling particles as an isolated particle falling downward in unbounded-atmospheric
air, and (2) the simple-equilibrium model (SEM), which assumes that both the media reached
thermal equilibrium in every segment along the exchanger. The mixing model predicted the
outlet temperature quite accurately for the particles with 0.5 mm diameter. For CARBOEAD
particles of 0.5 mm diameter, the predicted heat rate was within ±5% of the experimental results.
The SEM overpredicted the outlet temperature regardless of particle size. A simple modification
was added to the SEM, which led to improve the model predictions greatly. An integrated con­
ceptual system design is also presented and discussed. The use of a double dump valve introduces
many advantages towards a practical integrated system. Finally, employing particle strainers at
the exchanger outlet is crucial to reduce/eliminate risks about the swept particles.

1. Introduction
Heat exchangers are commonly used in thermal renewable energy systems today. For example, solar energy can be transferred from
solar rays to the working medium of power generation units by means of heat exchangers. Furthermore, heat exchangers can be used in
other renewable energy systems to provide the required thermal energy for desalination, cooling, and heating [1]. In particular,
particle-based direct-contact heat exchangers are an attractive alternative to conventional heat exchangers in a wide range of
high-temperature processes, such as calcination [2,3]; hot water production [4,5]; and drying [6].

* Corresponding author. Mechanical Engineering Department, King Saud University, P.O. Box 800, Riyadh, 11421, Saudi Arabia.
E-mail address: [email protected] (S. Alaqel).

https://doi.org/10.1016/j.csite.2022.101994
Received 11 February 2022; Received in revised form 29 March 2022; Accepted 30 March 2022
Available online 6 April 2022
2214-157X/© 2022 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
S. Alaqel et al. Case Studies in Thermal Engineering 33 (2022) 101994

Nomenclature

CFD computational fluid dynamics

DCHX direct-contact heat exchanger
PBCSP particle-based concentrating solar power
PBCT particle-based central tower
PBHX particle-based heat exchanger
PHU particle handling unit

Symbols
Ap particle surface area (m2)
CD drag coefficient
Cr Air to particle heat capacity ratio
cpa air specific heat (kJ/kg-◦ C)
cpp particle specific heat (kJ/kg-◦ C)
cpCB CARBOBEAD specific heat (kJ/kg-◦ C)
dp particle mean diameter (m)

Greek
g gravitational acceleration (m/s2)
kf fluid thermal conductivity (W/m.K)
LDCHX length of the DCHX (m)
mp hold-up mass inside the DCHX (kg)
msp mass of a single particle (kg)
ṁsp air mass flow rate (kg/s)
ṁP particle mass flow rate (kg/s)
NP number of particles inside the heat exchanger
P air pressure (bar)
Pr Prandtl number
Q̇a heat gained by the air (kW)
R Gas constant (kJ/kg-◦ C)
RePT terminal particle Reynold number
T temperature (◦ C)
t particle residence time (s)
Va air velocity (m/s)
VPter Particle terminal velocity (m/s)
∀p particle volume (m3)
U overall heat transfer coefficient (W/m2.K)
α Void fraction
μf fluid dynamic viscosity (Pa-s)
ρf air density (kg/m3)
ρp particle material density (kg/m3)
τ thermal response time

For power generation applications; typical “falling cloud” heat exchangers use solid particles, which are dropped from the top of a
tube and fall against up-flowing atmospheric air that is fed through the bottom of the tube. A variety of design variants have been
considered, such as the addition of various pressurized tanks to permit operations at elevated air pressures [7]. However, realistic
design considerations have not yet been considered that could overcome numerous drawbacks of such heat exchangers. For example,
falling cloud heat exchangers failed to take into consideration the problem of the particles being carried upward by the up-flowing air,
when the velocity of the air increased beyond the terminal velocity of the particles. In practice, up-flowing particles can cause severe
damage to downstream components such as turbine.
Further, those falling cloud heat exchangers (or designs) typically include relatively basic feeding mechanisms for particles,
resulting in highly uneven particle distribution [7,8]. Therefore, impractically long residence time (or long heat exchanger) is required
for maximum heat exchange between the particles and air. Furthermore, these designs have mainly focused on the heat exchange itself
rather than optimizing air flow. As a consequence, large eddy current, large scale turbulence, and other undesirable conditions exist
within heat exchange chambers [9]. A related problem is the positioning of the air outlet near the distribution point of particles, which
disrupts the particle distribution due to the abrupt pressure drop at the point where the fluid has its greatest velocity.

2
S. Alaqel et al. Case Studies in Thermal Engineering 33 (2022) 101994

[10] descibed extensively the main components and the added enhancment, featured towards efficient and reliable direct-contact
heat exchanger (DCHX). Experimental results confirmed that the proposed design overcomes the issues revealed in earlier
direct-contact heat exchanger designs. Compared to earlier studies, a shorter heat exchanger can be employed thanks to the new
enhanced features. This includes the particle distributor module and the tangential configuration for the air inlet. Results showed that,
for tests with excessive particle flow rate, more than 80% of the temperature rise was achieved within 0.35 m above the air inlet,
attributed to the tangential inlet’s swirl flow pattern. With such performance, the predominant feature of DCHX over the indirect ones
was manifested. The overall temperature rise along the heat exchanger was not significantly influenced by a change in operating air
pressure from 1 barg to 3 barg.
However, to scale up such devices, modeling the performance of DCHX is very critical in order to have the required size and hence
accurate prediction of cost. Several performance models have been reported in the literature for different DCHX designs. To reduce the
analysis complexity, researchers carried out investigations on the heat transfer and flow characteristics of a falling curtain by
considering several assumptions including: steady one-dimensional flow, uniform distribution of particles over each cross-sectional
area, neglecting particle-to-particle interaction, insignificant particle-to-wall interaction, uniform-spherical particle size, neglecting
the radiative exchange between particles & walls and considering the pressure of the mixture as that of the gas.
[11] analyzed the concept of heat recovery from a hot gas by using solid particles in a direct-contact circulating bed. The par­
ticulates included in the study were stainless steel and ceramic balls; 1, 1.6, and 2 mm in diameter with 0.8–2 solid loading. At the same
operating conditions, they demonstrated that the small particles were able to recover heat more effectively than the large particles.
However, small particles had a higher tendency to be carried over by the flowing gas [12]. continued the previous work with more
emphasis on particle-to-particle and particle-to-gas radiation exchange. The finite difference method was used to solve the radiative
heat transfer equations. For large scale applications, the results showed that the gas radiation significantly affected both media
temperatures.
A steady one-dimensional two-fluid model was developed by Refs. [13,14] to simulate the gas-solid counter-flow direct-contact
heat exchanger. The model showed a good agreement with experimental data reported by Ref. [8]. Accordingly, the model was
used to predict the effect of solid loading ratio and particle size on the heat exchanger’s performance. They reported that higher the
solid loading and smaller the particles enhanced is the heat recovery process. On another occasion [13,14], used the aforementioned
model to study the implications of integrating multiple gas inlets in a conveying duct on the heat recovery process. Results showed that
(except for fine particles) distributing the gas over the conveying duct leads to higher heat recovery.
[15] presented an analytical solution for the heat balance between a countercurrent gas and solid particles in a packed column.
Particle-to-particle conduction and heat losses from the column were neglected in the study. Results were validated against the
experimental data obtained by Ref. [16]. Two packing media were considered: metallic Pall-15 and ceramic Pall-25 rings. Both studies
confirmed that there was a certain gas velocity for which maximum heat exchanger efficiency was achieved. Besides, they reported
that increasing the packing height enhanced exchange efficiency.
The flow and heat characteristics of packed beds have been reviewed and studied by Ref. [17]. He presented correlations to predict
the convective heat transfer coefficient, pressure drop, bed effective thermal conductivity, and wall to fluid heat transfer for both
stagnant and flowing gas. The studies were conducted on a heated particle submerged in a colder packed bed. Thus, the effect of point
conduction, bypass, and natural convection (especially for stagnant gas) had been considered in predicting the heat transfer
coefficient.
[18] used the finite difference method to solve the two-dimensional heat equation for a hot-packed bed in cross-flow with
cold-vertical air. The general shape of the packing medium was assumed spherical (ash 2.224 mm in diameter). A set of non-linear
differential equations was formulated and solved by using the steady-two fluid model technique. In order to simplify the problem,
the bed temperature was assumed homogeneous in all directions. Moreover, particles were considered isothermal in the analysis and
allowed to move in the horizontal direction only. Computational results showed a good agreement with the experimental observations
obtained by the authors.
[8,19], used the mixing model to predict the air temperature along the DCHX. The main assumption was that the particles while
falling along the DCHX exhibit isolate sphere behavior falling in unbounded air. Radial mixing of air temperature and momentum
occurs at an infinite rate. A maximum amount of particle surface area is available for heat exchange. Additionally, the highest possible
temperature difference is maintained, thus ensuring the highest possible heat transfer rate. However, in their experimental setup,
particles exhibit jet flow rather than well-dispersed spheres [8]. conducted experiments in which two feeding mechanisms were used,
namely, bare tube and with basket distributor to break down the particle jet into many streams (Jets). Results show that the predictions
of the mixing models overestimate the air temperature along the heat exchanger. The use of basket distributor enhances the particle
distribution in the radial direction, hence more particle-to-air uniformity is achieved, better predictions of the mixing model were
reported. This enhancement in the model prediction could be improved further if better radial particle distribution is achieved.
In the current work, mixing model is tested against new experimental data [10]. Unlike the previous studies, an efficient particle
distributor was designed and tested in which the particle mainstream is split into several separated streams, and each stream can be
directed to a different zone inside the DCHX. Therefore, better particle-to-air mixing is established, each particle approaches the
behavior of an isolated sphere which meets the basic assumption of the mixing model. In addition, a simple equilibrium model as it
introduced by Ref. [20] is tested against the experimental findings. Although many conservative assumptions are adopted to simplify
the models, yet the models’ predictions showed good agreement with the current experimental results. Finally, a comprehensive
particle-based CSP integrated system using the proposed DCHX is presented.

3
S. Alaqel et al. Case Studies in Thermal Engineering 33 (2022) 101994

2. Proof-of-concept apparatus
The DCHX system (Fig. 1a) was designed and built to heat an upwardly moving-pressurized air stream by freely-falling hot solid
particles. The system consists of three main components: the heat exchange chamber, DCHX, and the particle handling unit (PHU). The
heat exchange chamber is simply a variable-area (tapered contour) duct with a circular cross-section that is 2 m high, and ranges in
diameter from 0.2 m at the lower to 0.4 m at the upper end. The PHU includes the parts/components responsible for delivering and
collecting the particles to and from the heat exchange chamber, such as the feeder, defeeder, particle distribution module, particle

Fig. 1. (a) Direct-contact heat exchanger system, (b) Particle distributor of [8]. and (c) Current particle distributor.

4
S. Alaqel et al. Case Studies in Thermal Engineering 33 (2022) 101994

valves, and particle lift. Communication between the feeder and the upper end of the heat exchange chamber occurs through the
particle distribution module.
A particle distribution module is responsible for delivering and uniformly distributing the hot particles across the cross-sectional
area of the heat exchanger, and is composed of a manifold assembly and distributor pipes. The main particle stream emerging from the
feeder is split in the manifold assembly into four streams destined for the distributor pipes, each of which has a ball at the lower end to
scatter the incoming particle stream (Fig. 1c). A defeeder is connected to the lower end of the heat exchanger to collect the cooled
particles.
The exchange chamber employs several enhancing features to acquire optimal performance including: a tapered-shaped design, a
tangential inlet for the up-flowing air, and the disengagement zone above the particle terminal ports (at the top of the DCHX). The
tapered-shaped design is intended to prevent the upflowing air from surpassing the particle terminal speed. Ascending air temperature
increases due to heat exchange with the hot falling particles, leading to an increase in its ascension velocity. Suppressing the air
velocity from achieving the terminal velocity of the particles is essential to avoid particle carry-over. The tangential inlet arrangement
for the air inlet facilitates cyclonic flow at the bottom of the DCHX, thus stimulating heat exchange and eliminating eddy currents and
large-scale turbulence as reported by Ref. [9]. An extended zone above the particle distributor pipes, referred to as the disengagement
zone, is employed to avoid an abrupt change in the flow conditions imposed by a change in particle containment area. Due to shift in
the area between the upper end of DCHX and the air exit pipe, air accelerates rapidly, and particle carry-over opportunity increases.
The disengagement zone maintains stable flow conditions near the particle terminates, helping the particle feeding process, so particles
leaving the particle distributor pipe tend to fall rather than migrating with the discharge air.
[10] reported that more than 80% of the particle heat could be captured at the first quarter of the DCHX (above the air inlet). This
was inferred to the aggressive air-to-particle mixing induced by the cyclonic flow pattern originated by the tangential inlet. Also, no
swept particles were noticed. However, the absence of swept particles could be related to the reduced air mass flow rate (air velocity)
used in the tests reported by Ref. [10]. Further, the current “lab-scale” DCHX was tested in batch mode, whereas a commercial DCHX
needs to be run in a continuous flow mode. Therefore, to permit a safe-reliable integration of the current DCHX with an air-breathing
power cycle, the DCHX must be operated in a continuous mode with more design constraints to prevent particle carry-over. Section 4 of
the paper outlines the design features that smooth the way for this endeavor.

2.1. DCHX models

2.1.1. Mixing model
In the current particle-based direct-contact heat exchanger, CARBOBEAD CP particles (particle curtain) fall against up-flowing
pressurized air. To reduce the complexity of the current model, the air velocity and pressure are assumed uniform in every section
along the heat exchanger although air velocity depends on air temperature, which experience spatial variation due to heat exchange
with the falling particles. The main assumption of the current model is that each particle moves downward as an isolated sphere that
fall isothermally through a stagnant unbounded air. Particle rotation & lateral movement, particle-to-particle & particle-to-wall in­
teractions, radiative exchange and wall losses are all neglected. As a result, the falling particles could attain its terminal velocity very
quickly, thus particle velocity along the DCHX is assumed equal to the particle terminal velocity. Therefore, the estimation of the
particle terminal velocity, particle Reynold number, and heat transfer coefficient (or particle Nusselt number) are achievable. The
particle terminal velocity is estimated by following [21]. He used the correlation proposed by Ref. [22] to develop a dimensionless
term (∅1 ), which can be calculated based on the particle and fluid properties as:
( )
4 ρP − ρf ρf gdp3
∅1 = (1)
3μ2f

where ρf is the air density (obtained at Tair and operating pressure), μf is the fluid dynamic viscosity (obtained at Tair ), dp is the particle
diameter, and ρP is the particle material density.
The drag coefficient (CD ) is related to ∅1 as:
( 2
)
432 1 + 0.047∅1 + 0.517
3

CD = − 1
(2)
∅1 1 + 154∅ 31

Therefore, the terminal Reynold number is defined based on both ∅1 and CD as:
( )0.5
∅1 ρf VPTer dp
ReP T = = (3)
CD μf
Solving for the particle terminal velocity (VPTer ), Equation (3) can be rearranged as:
( )0.5 ( )
∅1 μf
VPTer = (4)
CD ρf dp
The convective heat transfer coefficient can be estimated from Ref. [23] correlation based on the terminal Reynold number:

5
S. Alaqel et al. Case Studies in Thermal Engineering 33 (2022) 101994

hdp
(5)
1 1
Nu = = 2 + 0.6ReP 2T Pr3
kf

where, kf and Pr are the thermal conductivity and Prandtl number estimated at the film temperature (Tf = 0.5 ​ *(Tp + T∞ )).
The elemental heat rate gained by the air (equal to that released from the falling particles) can be estimated as:
dq̇i = UdAdTi (6)

where U is the overall heat transfer coefficient (for fine particles it can be assumed equal to the heat transfer coefficient), dA is the
elemental heat transfer are (particles’ surface area), and dTi is the temperature difference between the average particle temperatures
and the average air temperatures. The total heat transfer area (at every second) can be estimated based on the particle terminal velocity
and the heat exchanger length such as:
A = N p Ap (7)

where A is the heat transfer area (the total surface area of the spherical particles suspended by drag within the heat exchanger), Np is
the number of particles inside the heat exchanger, and Ap is the surface area of a single particle (area of a sphere). Np can be estimated
as:
/
Np = mp msp (8)

where mp is the hold-up mass inside the DCHX whereas, msp is the mass of a single particle. The hold-up mass can be deduced from the
time (t) needed for a unit mass to travel across the DCHX:
t = LDCHX /VPTer (9)
Then, the hold-up mass can be addressed as:
mp = tṁp (10)

The elemental heat rate gained by the air can be used to estimate the local air temperature as:
T∫
P,i+1
( )
dq̇i = ṁa cpa Ta,i − Ta,i− 1 = − ṁP cpP dT (11)
TP,i

The temperature profile of both the media can be determined by combining Equations 6 and 11. An iterative numerical method

Fig. 2. Schematic diagram showing the iterative procedure of the mixing model.

6
S. Alaqel et al. Case Studies in Thermal Engineering 33 (2022) 101994

(shooting method) can be used to determine the temperature profiles and net heat exchange rate between both the media. The DCHX is
divided into infinitesimal n segments with n + 1 sets of endpoint temperatures for both streams. Each segment has an increment of
UdA, which is equal to UA/n. The following four steps show how the iterative procedure is performed (shown in Fig. 2):
i. An initial guess for the temperature profiles for both streams is estimated. The specific heat of the CARBOBEAD CP was then
updated using the first guessed temperature profile. The specific heat for air was assumed constant and calculated at average
inlet particle-air temperature.
ii. Then, for each segment, the average temperature of both the media was determined by averaging the endpoint temperatures,
which permits the estimation of dq̇i as shown in Equations 6. The temperature values across each segment (Equation (11)) were
determined.
iii. The new temperature values were then used to update the particle specific heat, and calculate new set of dq̇i .
iv. The iterative procedure continued until the calculated endpoint temperatures stopped changing. The total heat transfer rate was
then calculated by taking the sum of dq̇i values after the iterative process is realized.

As mentioned earlier, the basic assumption behind the current model is that each particle is considered as an isolated particle. To
validate the accuracy of this assumption, the void fraction is to be estimated. The higher the void fraction (approaching unity as
reported by Ref. [24], the more accurate is the model. The void fraction (α) is defined with knowledge of the ratio between the total
suspended particles’ volume (Np ∀p ) to that of the heat exchanger (∀HX ); hence:
/
α = 1 − Np ∀p ∀HX (12)

Finally, the model can be used to predict the effectiveness of a heat exchanger (ε), which is defined as the ratio between the actual
heat transfer to the maximum possible heat transfer rate. The maximum possible heat transfer rate is expressed as:

Q̇max = Cmin (TP i − Ta i ) (13)

Cmin = min(ṁa cpa , ṁP cpP ) (14)

Accordingly, the effectiveness (ε) of the heat exchanger is expressed as:
∑
(dq̇i )
ε= (15)
Cmin (TPi − Ta i )

2.2. Simple-equilibrium model

As shown in Fig. 3a, the DCHX is split into several infinitesimal segments in which energy balance could be performed across each
segment (control volume). The basic assumption behind the simple-equilibrium model (SEM) is that the particles within each segment
are well-mixed with air; a thermal equilibrium between both media is reached, i.e., both media exit the control volume with equal
temperature (Ta,i = TP,i ). Therefore, the total thermal conductance (W/K) is approaching infinity i.e., UAs → ∞ (or the LMTD ap­
proaches zero), where U represents the overall heat transfer coefficient and As is the total surface area of the suspended particles [20].
Further, by neglecting radiative heat transfer and wall losses, a heat balance between both media across each control volume can be
illustrated as:

Fig. 3. Schematic of the simple equilibrium model.

7
S. Alaqel et al. Case Studies in Thermal Engineering 33 (2022) 101994

ṁa cpa (ΔTa ) = − ṁP cpP (ΔTP ) (16)

where ṁa and ṁP are the mass flow rates of the air and particles, cpa and cpP are the specific heats of the air and particles, and ΔTa and
ΔTP are the elemental temperature difference across the control volume. Arranging Equation (16) for segment (i) can yield to:
( )
Cr Ta,i − Ta,(i− 1) = TP,(i+1) − TP,i (17)

The heat capacity ratio (Cr ) is defined as the ratio between the air heat capacity (ṁa cpa ) to the particle heat capacity (ṁP cpP ). The
specific heat of the particles (CARBOBEAD CP) can be estimated using the formula proposed by Ref. [25] at average particle
temperature:

cpCB = 0.365T 0.18 (18)

Hence, according to the basic assumption of the SEM, TP,(i+1) = Ta,(i+1) and TP,i = Ta,i , Equation (17) can be rearranged to the form:
( )
Cr 1
Ta,i = Ta,(i− 1) + Ta,(i+1) (19)
1 + Cr Cr
The temperature profiles of both the streams can be generated by applying Equation (19) in every control volume i.e., a system of N
algebric equations will be genrated. However, for the air, at stage N the air temperature Ta,(N+1) will appear. Therfore, to reach a
solution, Ta,(N+1) should be predefined. To simplify the predefination of Ta,(N+1) , additional control volume is added as shown in Fig. 3b.
It can be noticed that, according to the basic assumption of the SEM, the air temperature exiting the control volume (N +1) is equal to
the Tp,0 . As will be shown later, this additional assumption has a significant influence in predicting the outlet air temperature against
the experimental data. It has been found that adding an additional constraint for the outlet air temperature could significantly enhance
the model predictions. For instance, by considering that Ta,(N+1) = 0.9Tp,0 , the air outlet temperature prediction is significantly
enhanced. This additional constraint (Ta,(N+1) = 0.9Tp,0 ) was obtained by fitting curves to experimental results. In this paper, the SEM
after employing this constraint is referred to as an Enhanced Simple-Equilibrium Model (ESEM). The physical interpretation of the
aforementioned constraint is that the upper limit of air’s exit temperature is 10% lower than the particle’s inlet temperature. It should
be noted that deriving and sloving the SEM, encoutered several essential assumptions which include:
a) Air and particle exit each segment at same temperature
b) No carryover particles between segments
c) Particles are very small (neglect the particle interal resistance)
d) Neglect radiation exchnage between segments and neglect wall losses
e) Ta,(N+1) = Tp,0 [20]
f) Constant heat capacity (evaluated at average temperature)

2.3. Results
Tests were performed using two patches of CARBOBEAD CP particles, namely, 0.3 and 0.5 mm in diameter. For better data
visualization, the test results are arranged based on the value of the outlet air temperature (descending order), as shown in Tables 1 and
2.
As shown in Fig. 4; the air and particle temperatures were measured at various locations in the test apparatus. Thermocouples of
type Omega K sheathed/ungrounded (KMQSS-IM300U-600) were used with an accuracy of 0.2 ◦ C. The particle inlet temperature was
measured right before the particle entrance into the DCHX. To increase the measurement accuracy of particle outlet temperature (Tp o ),
a conical basket mesh (size 10) was installed in the converging section at the bottom of the DCHX (under the air inlet). By doing so,

Table 1
Conditions of the tests performed on 0.5 mm CARBOBEAD CP particles.

Test P [abs] ṁp × 103 [kg/s] ṁa × 103 [kg/s] Tai [◦ C] Tpi [◦ C] Tao [◦ C] Tpo [◦ C]

1 1.97 180.20 38.53 39.03 247.66 226.80 203.13

2 2.10 67.05 36.16 39.00 236.74 195.30 138.98
3 3.00 60.01 37.56 40.04 239.88 191.90 126.00
4 3.90 43.79 35.12 39.05 255.63 190.70 111.55
5 2.00 80.11 40.03 39.08 197.63 169.50 117.07
6 1.95 43.03 25.13 40.01 180.45 145.70 103.69
7 2.10 63.37 34.81 40.01 175.21 144.60 105.39
8 3.90 47.81 34.63 40.01 178.92 138.80 89.98
9 4.10 40.22 25.87 40.01 174.86 136.40 90.91
10 3.05 40.16 25.94 40.01 175.21 136.00 90.60

8
S. Alaqel et al. Case Studies in Thermal Engineering 33 (2022) 101994

Table 2
Conditions of the tests performed on 0.3 mm CARBOBEAD CP particles.

Test P [abs] ṁp × 103 [kg/s] ṁa × 103 [kg/s] Tai [◦ C] Tpo [◦ C] Tao [◦ C] Tpo [◦ C]

1 2.60 125.20 37.53 37.04 257.64 237.51 190.23

2 2.91 108.90 42.38 37.04 259.39 230.60 176.57
3 2.60 108.60 42.39 37.01 249.96 223.24 166.80
4 1.70 23.56 17.85 37.04 193.37 152.58 91.13
5 1.70 34.72 43.74 37.01 213.98 133.91 68.26
6 1.70 23.09 27.14 37.05 202.47 130.18 67.98

particles were brought to rest within the basket before exiting the DCHX, enabling good contact between the particles and the ther­
mocouples. Two thermocouples were used to measure Tp o ; the arithmetic average of the two readings represents Tp o .
The air temperature was measured along the DCHX using twenty-one thermocouples. The air inlet temperature was measured with
a thermocouple inserted into the air inlet duct (tangential inlet). The air outlet temperature was measured using a thermocouple
mounted after the 90◦ elbow in the air outlet tubing, where rapid mixing of the discharged air can be accomplished. There were four
thermocouples mounted along the centerline of the DCHX to determine the air temperature. Besides, 15 thermocouples were located
inside the DCHX to acquire the air temperature’s radial variation at four elevations. These thermocouples were mounted at 0.35, 0.75,
1.4, and 2.1 m above the air inlet (Fig. 4). With sheathed and ungrounded thermocouples, the electrostatic charge’s influence produced
by the friction of the falling particles was prevented.
As mentioned earlier, the primary assumption of the mixing model is that each particle falls as an isolated particle. The minimum
void fraction estimated as shown by Equation (12) is 0.9998, which pertains to the maximum particle flow rate (test 1 in Table 1). As
this void fraction approaches unity, the condition reported by Ref. [24] for isolated particle behavior is satisfied. The measured air heat

Fig. 4. Locations of temperature and pressure instrumentation in the test apparatus.

9
S. Alaqel et al. Case Studies in Thermal Engineering 33 (2022) 101994

Fig. 5. A comparison of measured air heat transfer rate to predictions of mixing model for (a) 0.5 mm particles, and (b) 0.3 mm particles.

transfer rate (Q̇a ) calculated by Equation (20) is plotted against the analogous mixing model predictions for both 0.5 and 0.3 mm
CARBOBEAD particles, as shown in Fig. 5.

Q̇a = ṁa cpa (Tao − Tai ) (20)

The solid line in Fig. 5a represents the ideal agreement line, while the dashed lines represent the ±5% deviation. Most of the
estimated values fall within - 5%, and the maximum deviation from the ideal line falls within ±7%. However, as can be seen in Fig. 5b,
the mixing model generally overpredicts the measured heat transfer rates (the dashed lines represent the ±10% deviation). This is
expected since (as will be shown later) the mixing model overpredicted the air outlet temperature in case of 0.3 mm particles.
Fig. 6 shows the predicted air temperature along the DCHX using both models against the experimental measurement for three
different tests. It appears that the SEM overpredicts the air temperature except for that at 0.35 m above the air inlet. This might be
attributed to the contribution of the tangential inlet configuration. The produced cyclonic flow pattern significantly augments the
particle-air heat transfer. It should be noted that the SEM assumes that both media reach a thermal equilibrium, which might explain
the overestimation of air temperature for the other locations. On the contrary, for most locations, the mixing model underestimates the
air temperature along the DCHX except for test 1. In the mixing model, particles were assumed to fall at a constant velocity (terminal
velocity), which is the “upper” limit velocity. Therefore, the interaction time between the falling particles and the up-flowing air is
reduced, which leads to underestimating the air temperature along the DCHX. However, as shown in Fig. 6, by imposing that the air
outlet temperature’s value cannot exceed 90% of the particle inlet temperature, the model predictions are greatly enhanced. As
mentioned earlier, with this constraint, the SEM is called the Enhanced Simple-Equilibrium Model (ESEM).

10
S. Alaqel et al. Case Studies in Thermal Engineering 33 (2022) 101994

Fig. 6. Comparison of measured air temperatures to predictions of mixing model, SEM, and ESEM for 0.5 mm particles.

It should also be noted that, at each level (shown in Fig. 6), the air temperature represents the arithmetic mean of the temperature
readings at that level. Air temperature across the DCHX’s cross-sectional area experiences spatial variation due to the unstable flow
conditions and particle distribution. For instance, at the first level (0.3 m above air inlet), the recorded air temperature at the vicinity of
the DCHX’s wall shows a higher value than the other readings. This might explain the discrepancy between the predicted values by the
mixing model and experimental data at this level. However, the spatial variation of air temperature reduced significantly at the
location where the outlet air temperature was measured. As shown in Fig. 6, the air outlet temperature was measured downstream the
90◦ elbow, where the excellent-rapid mixing enables a representative temperature measurement.
The predicted air outlet temperature obtained by the two models was plotted against the measured one as shown in Fig. 7. For the
0.5 mm particles (a), the simple mixing model predicted the air outlet temperature quite accurately. The maximum absolute deviation
from the experimental results was within 7% which occurred in tests with high particle mass flow rates and high air pressure. The
deviation was 7% and 6.5% in test number 1and 9 respectively. Whereas, the deviation was less than 7% in the remaining tests (varies
from 1.1% to 6%). This might suggest that the simple mixing model can be considered as a “fairly” accurate (maximum deviation is less
than 7%) tool to estimate the air temperature exiting the DCHX for the 0.5 mm CARBOBEAD CP particles. However, as shown in
Fig. 7b, the mixing model overpredicted the air outlet temperature with the 0.3 mm CARBOBEAD particles. The deviation margin from
the experimental data falls between 8 and 12% which could be attributed to the influence of heat exchanger length (as shown in
equation (9)). Smaller particles respond faster to the change of thermal conditions than larger particles; on the other hand, larger
particles travel faster than the smaller particles. This implies that with the current falling path (DCHX length), the particle residence
time for the 0.3 mm CARBOBEAD particles increased. These two reasons could lead the mixing model to overestimate the air outlet
temperature for such particles.
On the other hand, the SEM overpredicts the air outlet temperature compared to the experimental observations. This should be
expected since the SEM represents an ideal situation in which the falling particles are brought to thermal equilibrium with the uprising
air. This implies that the particle residence time is exceptionally high for thermal equilibrium to be reached in every segment inside the
DCHX. The contribution of the additional constraint added to the SEM is shown in Fig. 7. It appears that the prediction of the air outlet

11
S. Alaqel et al. Case Studies in Thermal Engineering 33 (2022) 101994

CARBOBEAD 0.5mm
250

200

150

100

50

0
1 2 3 4 5 6 7 8 9 10
Test ID[-]

SEM 1-D Model Experimental ESEM

CARBOBEAD 0.3mm
300

250

200

150

100

50

0
1 2 3 4 5 6
Test ID[-]

SEM 1-D Model Experimental ESEM

Fig. 7. A comparison between the predicted and measured outlet air temperatures (a) 0.5 mm particles, and (b) 0.3 mm particles.

temperature is greatly enhanced.

Nonetheless, it should be noted that the SEM gives no credit to the particle size nor the air pressure. Besides, the SEM requires no
tedious procedure to obtain an estimate for the air outlet temperature. For that, the SEM offers a simple approach to predict the highest
possible air outlet temperature regardless of particle size and air pressure.

3. Integration with particle-based CSP system

3.1. Particle-based central tower integrated system
Fig. 8 shows a conceptual particle-based central tower system (PBCT) employing the proposed DCHX. The PBCT system comprises
of a particle heating receiver (PHR) within the cavity at the top of the tower, high and low-temperature thermal energy storage tanks,
particle to working-fluid heat exchanger, an airpower cycle (gas turbine), and a particle circulation mechanism. Operation mechanism
of PBCT integrating DCHX is detailed in Refs. [26,27].
Integrating high-pressure DCHX with the TES at atmospheric pressure as shown in Fig. 8, is a major obstacle to achieve continuous
operation. It should be noted that, the proof-of-concept DCHX was tested in batch operation. In previous work [10], mentioned
multiple lock hoppers’ arrangements to facilitate system integration. The multiple lock hoppers are widely used in real-life applica­
tions. However, there are concerns pertaining to complexity, heat loss, and parasitic loads. This section proposes less complicated, low
parasitic loads, and cost-effective integrated system via the use of particle flow controllers.

12
S. Alaqel et al. Case Studies in Thermal Engineering 33 (2022) 101994

Fig. 8. Conceptual particle-based central tower integrated system using DCHX.

3.2. Flow controllers in particle-based direct-contact heat exchanger

As described in Ref. [26], the particles leaving the PHR cavity may reach temperatures of >1000 ◦ C before entering the HT-TES
tank. The HT-TES tanks are designed to store the particles at high temperatures for an extended period of time. Particles entering
the DCHX are at very high temperatures. Therefore, the DCHX should be able to receive and regulate the high-temperature particles
very efficiently.
To achieve a fully integrated system, selecting the flow control valves and their location in the particle loop is very important. The
DCHX must be reliable to operate at high temperature (up to 1000 ◦ C) and pressure (~4 bar absolute), which represents the operating
conditions of the integrated PBCT system described in Ref. [26]. Therefore, the flow control valves should ensure no air leakage during
the inflow and outflow of the particles to and from the DCHX, respectively, to keep the system at constant pressure. Fine particles may
trap between the stationary and moving part of the control valve during the opening and closing, thus obstructing the valves from
reaching their full-close position. Thus, introducing loss of pressure in the system. Therefore, the selected valves must have some
self-cleaning mechanism so that the particles do not get stuck during their operation. Furthermore, the valves should be fabricated
from abrasive resistant material to ensure long service life.
Different common types of valves were studied based on the optimum selection criteria. Their high-level comparison is provided in
Table 3.
For the current particle-based direct-contact heat exchanger (DCHX) system, double-dump valve (DDV) with the airlock system
(shown in Fig. 8) and rotary paddle wheel, have been identified as suitable candidates that can be operated at high temperatures and
pressures. A high-temperature double dump valve (HT-DDV) can be installed upstream the DCHX to permit communication with the
HT-TES bin to facilitate particle feeding. Similarly, a low-temperature double dump valve (LT-DDV) is located downstream the DCHX
to assist discharging process. The location of the two DDVs is shown on the left of Fig. 8. With this arrangement, the DCHX system will
be shorter more reliable, cost effective, and exhibit lower parasitic load than that proposed in earlier work [10].
It should be noted that the DDVs are commercially available and highly customizable to meet the needs of the operating conditions.
For instance, Fig. 9b shows a DDV (also called double flap airlock valve) made by PLATTCCO CORPORATION, which is rated to 980 ◦ C

Table 3
Comparison of common types of particulate flow valves based on the set selection criteria.

Valve Type Controlled Particle Flow High-Temp Resistant Low Air-Leakage Abrasive Resistant Self-Cleaning Mechanism

Rotary Paddle Wheel Yes Yes Yes Yes Yes

Butterfly Valves Yes Yes Yes Yes No
Gate Valves Yes Yes Yes Yes No
T-Valve Yes Yes Yes Yes Yes
Pinch Valve Yes No Yes Yes No
Double-Dump Valve Yes Yes Yes Yes Yes

13
S. Alaqel et al. Case Studies in Thermal Engineering 33 (2022) 101994

Fig. 9. (a)Double Dump Valve basic operation, (b) PLATTCO R-series double flab valves.

at 2.75 barg. Fig. 9a shows the basic operation of a DDV in which it can isolate the high-pressure vessel from a low-pressure one
ensuring minimal air leak. This feature is very critical since smooth particle charge/discharge is crucial for smooth-continuous
operation of the plant.

3.3. Particle strainers

In particle-based systems, particles experience attrition; thus smaller particles are continually produced. Inside the DCHX, if the
velocity of the uprising air exceeds the terminal velocity of the fragmented particles, some particles will be carried to downstream
equipment. To prevent them from reaching the downstream components, particle strainers in the discharge pipe are necessary. Conical
strainers (Fig. 10a) are widely used in particle-based industries and can be designed to provide 100%–300%1 open area, i.e., very small
pressure drop. Moreover, strainers might get clogged with time and cause a considerable pressure drop. To prevent strainer degra­
dation, frequent cleaning and inspection is highly recommended. However, strainer cleaning requires a complete shutdown of the
whole direct-contact system. Therefore, achieving a smooth operation requires a parallel manifold design (Fig. 10b) for the air
discharge system.
Fig. 10b shows the parallel layout of the DCHX discharging system. It is composed of a parallel manifold assembly, set of air valves,
air strainers, and pressure transducers (PT). The instantaneous performance of the strainer can be monitored by measuring the pressure
drop across it. The increase in pressure drop across the strainer relative to the pressure drop recorded for a new strainer indicates the
clogging percentage. Through a set of valves, air can be directed to another discharge branch to enable strainer maintenance when
needed. The valves can be operated manually or automatically by connecting them and the pressure transducers to a control unit.
These multiple paths for the outlet air can provide reliable integration of the DCHX with an air-breathing power cycle.

Fig. 10. (a) Conical air strainer, (b) schematic diagram of the proposed parallel discharge system.

1
According to manufacturer specs, https://www.dannenbaumllc.com/portfolio-items/.

14
S. Alaqel et al. Case Studies in Thermal Engineering 33 (2022) 101994

4. Conclusion
Two basic models were analyzed, namely, mixing model and a simple-equilibrium model (SEM). The models were tested against
experimental data in which a good particle distributer was employed. Unlike previous designs, the novel design introduced in the
current work improves the particle-air mixing, which enhances the particle-to-air heat exchange. Therefore; the performance of the
current heat exchanger approaches the theoretical predictions of a direct-contact heat exchanger.
CARBOBEAD CP particles (0.3 and 0.5 mm in diameter) were used in the experiments. The basic assumption behind the mixing
model is that each particle travels downward as an isolated sphere falling through an unbounded atmospheric air. Due to the good
particle-to-air mixing in the current design, results showed that the mixing model predicted the air outlet temperature accurately for
the 0.5 mm CARBOBEAD particles.
The primary assumption behind the SEM was that the particles within every segment (sub-control volumes) along the heat
exchanger are well-mixed with air. A thermal equilibrium between both media is reached, i.e., both media exit the control volume at
equal temperature. Results showed that, for both particle sizes, the SEM overpredicted the air outlet temperature. This can be
attributed to the SEM distinct feature in which a thermal equilibrium between both media was assumed, which implies that the total
thermal conductance (W/K) is approaching infinity. The enhanced SEM, on the other hand, shows a good agreement with the
experimental results. Nevertheless, the SEM provides an uncomplicated procedure to estimate the air outlet temperature regardless of
particle size, particle shape, and air pressure.
A particle-based central tower integrated system using DCHX is proposed and described. DDVs will help isolate the DCHX from the
low-pressure tanks, ensure smooth particle feeding/discharge, simplify the system control, reduce the parasitic loads, and introduce a
cost-effective integrated system. Additional precautionary design for the air outlet is introduced to eliminate the risk of particles
carryover. This will help to protect the downstream component (such as the turbine blade) from getting eroded by the swept particle.

Author statement
Shaker Alaqel: Conceptualization, Methodology, Visualization, Writing - Original Draft, and Formal analysis.
Nader S. Saleh: Review & Editing, Data Curation, Visualization, and Formal analysis.
Rageh Saeed: Review & Editing.
Eldwin Djajadiwinata: Review & Editing.
Muhammad Sarfraz: Visualization, Review & Editing.
Abdulelah Alswaiyd: Review & Editing.
Hany Al-Ansary: Visualization, Investigation, Review & Editing Resources, Supervision, Project administration, and Formal
analysis.
Obida Zeitoun: Review & Editing, Investigation.
Syed Danish: Review & Editing.
Zeyad Al-Suhaibani: Review & Editing.
Abdelrahman El-Leathy: Project administration.
Sheldon Jeter: Investigation, Resources, Visualization, Project administration.
Ahmad Khayyat: Funding, 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.

Acknowledgment
The authors extend their appreciation to the Deanship of Scientific Research at King Saud University for funding this work through
research group No. RG-1440-087.

References
[1] M.E.H. Assad, M.A. Nazari, Heat exchangers and nanofluids, in: Design and Performance Optimization of Renewable Energy Systems, Academic Press, 2021 ,,
pp. 33–42.
[2] C. Klett, M. Missalla, R. Bligh, Improvement of product quality in circulating fluidized bed calcination, in: Minerals, Metals and Materials Society/AIME, 2010,
420 Commonwealth Dr., P. O. Box 430 Warrendale PA 15086 USA. [np]. 14-18 Feb.
[3] N.A. Madlool, R. Saidur, M.S. Hossain, N.A. Rahim, A critical review on energy use and savings in the cement industries, Renew. Sustain. Energy Rev. 15 (4)
(2011) 2042–2060. .
[4] M.T. Johansson, M. Söderström, Options for the Swedish steel industry–Energy efficiency measures and fuel conversion, Energy 36 (1) (2011) 191–198. .
[5] A.M. Pantaleo, J. Fordham, O.A. Oyewunmi, P. De Palma, C.N. Markides, Integrating cogeneration and intermittent waste-heat recovery in food processing:
microturbines vs. ORC systems in the coffee roasting industry, Appl. Energy 225 (2018) 782–796. .
[6] A. Reyes, G. Diaz, F.H. Marquardt, Analysis of mechanically agitated fluid-particle contact dryers, Dry. Technol. 19 (9) (2001) 2235–2259. .
[7] L. Catalano, R. Amiranta, P. Tamburrano, S. Copertino, Analysis of the complementary energy losses of a high temperature gas to gas heat exchanger based on a
solid intermediate medium, WIT Trans. Eng. Sci. 75 (2012).
[8] M.J. Frain, D.P. Schmidt, A.W. Fiveland, An experimental investigation of the influence of gas and solid particle interaction on the heat transfer effectiveness of a
falling bed heat exchanger, J. Heat Trans. ASME (2005), 127/1077.
[9] M.S. Sagoo, The development of a falling cloud heat exchanger – air and particle flow and heat transfer, Heat Recov. Syst. (1981), 1/133–138.

15
S. Alaqel et al. Case Studies in Thermal Engineering 33 (2022) 101994

[10] S. Alaqel, N.S. Saleh, E. Djajadiwinata, R. Saeed, A. Alswaiyd, H. Al-Ansary, Z. Al-Suhaibani, A novel particle-to-fluid direct-contact counter-flow heat exchanger
for CSP power generation applications: design features and experimental testing, Renew. Energy 170 (2021) 905–926. .
[11] N. Gat, The circulating balls heat exchanger (CIBEX), Int. J. Thermophys. 1 (1987) 105–111.
[12] J. Park, S. Baek, S. Kwon, Analysis of a gas-particle direct- contact heat exchanger with two-phase radiation effect, Numer. Heat Tran. 33 (1998) 701–721.
[13] K.S. Rajan, S.N. Srivastava, B. Pitchumani, B. Mohanty, Simulation of countercurrent gas–solid heat exchanger: effect of solid loading ratio and particle size,
Appl. Therm. Eng. 27 (2006) 1345–1351.
[14] K.S. Rajan, S.N. Srivastava, B. Pitchumani, B. Mohanty, Simulation of gas–solid heat transfer during pneumatic conveying: use of multiple gas inlets along the
duct, Int. Commun. Heat Mass Tran. 33 (2006) 1234–1242.
[15] E. Saatdjian, J. Large, Heat transfer in a countercurrent gas-solid packed column, Journal of Heat transfer, ASME 110/385 (1988).
[16] M. Tibourtine, Hydrodynamique et Transfert Thermique dans un echangeur garni a pluie de particules, These de Docteur Ingenieur, Universitede Technologie de
Compiegne, France, 1981.
[17] E. Achenbach, Heat and flow characteristics of packed beds, Exp. Therm. Fluid Sci. 10 (1995) 17–27.
[18] A. Liu, S. Liu, Y. Duan, Z. Pan, Numerical simulation of heat transfer in a gas solid crossflow moving packed bed heat exchanger, J. Therm. Sci. 10 (3) (2000).
[19] L.A. Catalano, F. De Bellis, R. Amirante, M. Rignanese, An immersed particle heat exchanger for externally fired and heat recovery gas turbines, J. Eng. Gas
Turbines Power 133 (3) (2011) .
[20] H.J. Green, C.M. Leboeuf, M.S. Bohn, Technical and Economic Evaluation of a Solid-Particle/air Direct-Contact Heat Exchanger (No. SERI/TR-252-2663), Solar
Energy Research Inst., Golden, CO (USA), 1986.
[21] R.P. King, Introduction to Particle-Fluid Flows, Elsevier Butterworth-Heinemann, 2002.
[22] R. Turton, O. Levenspiel, A short note on the drag correlation for spheres, Powder Technol. 47 (1) (1986) 83–86. .
[23] W.E. Ranz, W.R. Marshall, Evaporation from drops, Chem. Eng. Prog. 48 (1952) 141.
[24] C.T. Crowe, J.D. Schwarzkopf, M. Sommerfeld, Y. Tsuji, Multiphase Flows with Droplets and Particles, CRC press, 2011 ,.
[25] C.K. Ho, M. Carlson, P. Garg, P. Kumar, Technoeconomic analysis of alternative solarized s-CO2 Brayton cycle configurations, J. Sol. Energy Eng. 138 (5) (2016)
.
[26] H. Al-Ansary, A. El-Leathy, S. Jeter, M. Golob, C. Nguyen, E. Djajadiwinata, H. Gandayh, Design features of the world’s first commercial concentrating solar
power plant using the particle heating receiver concept, in: ASME 2019 13th International Conference on Energy Sustainability Collocated with the ASME 2019
Heat Transfer Summer Conference, American Society of Mechanical Engineers Digital Collection, 2019. .
[27] M. Sarfraz, R. Yeung, K. Repole, M. Golob, S. Jeter, H. Al-Ansary, A. Alswaiyd, Proposed design and integration of 1.3 MWe pre-commercial demonstration
particle heating receiver based concentrating solar power plant, in: Energy Sustainability vol. 84881, American Society of Mechanical Engineers, 2021 ,, June.
V001T02A005.

16