Microsoft Word - 211.docx


 CHEMICAL ENGINEERING TRANSACTIONS  
 

VOL. 61, 2017 

A publication of 

 

The Italian Association 
of Chemical Engineering 
Online at www.aidic.it/cet 

Guest Editors: Petar S Varbanov, Rongxin Su, Hon Loong Lam, Xia Liu, Jiří J Klemeš 
Copyright © 2017, AIDIC Servizi S.r.l. 

ISBN 978-88-95608-51-8; ISSN 2283-9216 

Design and Experimental Testing of a Combustion Nozzle for 

Use with a Solar Reactor 

Jens Heylena, Nesrin Ozalpb,* 

aSolar Thermal Technology Laboratory (STTL), KU Leuven, 2860 Sint Katelijna Waver, Belgium   
bMechanical and Industrial Engineering Department, University of Minnesota Duluth, 55812 Minnesota, USA 

 nozalp@d.umn.edu 

It is important to have homogenous temperature distribution inside solar receivers and reactors. This paper 

presents a methodologic adaptation of a burner nozzle for use with a solar receiver to test and compare the 

performance of mixing effect of the nozzle with that of used in combustion system. Design steps are described 

per understanding of the influence of burner nozzle and then creation of a new type of injection port for a solar 

reactor via adaptive modification. Influence of the nozzle design on the temperature distribution inside a solar 

reactor is studied per experimental testing of the solar receiver with the new nozzle using a 7 kW high flux solar 

simulator. Finally, a CFD simulation of the flow pattern inside the solar receiver is given. The results show that 

it is possible to create different flow patterns inside a solar reactor and enhance the mixing of reactants. The 

results also show that nozzle provides more homogenous temperature distribution inside the solar reactor. This 

research demonstrates that implementation of a technique used in a mature field like combustion into solar 

thermal technology may inspire adaptation of other applications from combustion field to solar thermal field 

which would accelerate developments in emerging solar thermochemical process technology. 

1. Introduction 

Solar thermochemistry is an emerging field focusing on the use of concentrated solar radiation for production of 

fuels and commodities. Solar thermochemical processes take place in solar reactors. Each reactor has its own 

configuration, size and performance. To maintain production efficiency, it is crucial to have homogenous 

temperature distribution inside solar reactor. This requires that the heat losses are minimized, and the mass 

and heat transfer between the reactants promoted. Another important aspect is the concentration of the solar 

radiation. According to the use of solar radiation as a heat source in the reactor, there are two types of solar 

reactors: directly and indirectly irradiated reactors. In indirectly heated solar reactor concept, solar radiation 

heats the external opaque walls of reactor. Heat transfer by conduction and convection from the walls to the 

reactants provides high temperature process heat for endothermic reactions. A typical example of an indirectly 

heated reactor is a catalytic tubular reformer designed and tested by Alonso and Romero (2015). That solar 

reactor concept consists of a tube in which a gas is forced to flow through a catalyst and the solar radiation is 

exposed along the external walls of the tube. As for direct solar reactors, incoming concentrated solar power is 

directly used to heat the reactants where the radiation enters reaction chamber through a transparent window. 

This type of solar reactors are capable of operating at higher temperatures because the reactants are directly 

exposed to solar radiation. Solar reactors types can also be defined as entrained, fluidized and stacked. 

Selection of solar reactor type depends on the desired application. For example, fixed bed reactors are 

recommended for solar catalytic reactions (Ozalp et al., 2013). On the other hand, fluidized beds could be best 

for reactions requiring high thermal transfer properties. Homogenous heating at the reaction medium is favoured 

by the close contact between gas and fluidized particles, and higher temperatures are expected to be achieved. 

It is possible to make the same product through different thermochemical routes with different reactor concepts. 

However, homogenous distribution of temperature inside solar reactor is the common target of all solar reactor 

concepts. 

                               
 
 

 

 
   

                                                  
DOI: 10.3303/CET1761246

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Please cite this article as: Heylen J., Ozalp N., 2017, Design and experimental testing of a combustion nozzle for use with a solar reactor, 
Chemical Engineering Transactions, 61, 1489-1494  DOI:10.3303/CET1761246  

1489



Swirling flows are commonly used in industrial applications especially in combustion burners for enhanced 

burning rate. A swirling flow creates different turbulence flow patterns and enhanced heat transfer inside a 

reaction chamber. This paper presents the influence of this effect and demonstrates a new type of injection port 

for a solar reactor exhibiting similar enhancements achieved in combustion chambers.  

2. Adaptive design of the nozzle 

The original nozzle design is based upon a concept developed by Vanierschot (2007) for a burner, which was 

successful per experimental observation at the KU Leuven Mechanical Engineering Department and was 

available for detailed inspection during the course of this research. The most important feature of the nozzle is 

its capability to create a swirling flow inside the combustion chamber providing intense mixing of the reactant 

gases. It was expected that adoption of this technology into solar reactor would yield similar flow pattern and 

homogenous temperature distribution. Figure 1 shows the schematic of burner nozzle designed by Vanierschot 

(2007). 

 

Figure 1: Cross section of the burner nozzle (Vanierschot, 2007) 

The nozzle is composed of two parts. There is an inner hollow rod surrounded by an outer part. This design 

provides two gas streams. The primary gas stream is injected in an annular channel between the inner rod and 

the outer part. The second stream is injected inside the inner hollow rod. At the end of the rod, there are 24 

holes with diameter of 0.5 mm to allow radial injection. This radial injection is the main parameter to control flow 

pattern inside the reactor. The part outside is a stepped conical nozzle with a slot. Because of the slot, sub-

pressure of corner recirculation zone behind the sudden expansion increases. Characteristic dimension of the 

nozzle is Do, which is the outer diameter of the annular channel, and the length L is 11.9 mm and α = 20°. 

   

Figure 2: Exploded view of the nozzle designed for solar reactor (left), cross section of the nozzle (right) 

To make sure that the nozzle is compatible with the solar reactor, a few adjustments had to be made. To begin 

with, a flange was attached to the outer nozzle. Then a new method of centralizing the inner rod and the outer 

part had to be found. To centralize the inner rod with the outer nozzle, the back of the outer part was made 

1490



completely closed except for a hole where the inner rod could fit through. To provide the annular channel for air, 

two injections were symmetrical placed on the sides. The last change on the outer nozzle was the dimensions. 

To scale the nozzle, the dimension Do was changed from 27 mm to 29.3 mm. Consequently, all the dimensions 

depending on Do were changed as well. Manufacturing limitations required some changes on the inner rod as 

well. For example, it was too complex to drill 24 holes with a diameter of 0.5 mm on the rod. Therefore, the 

number of the holes were reduced and the diameters were made bigger. The new inner rod became 16 holes 

with diameter of 1 mm equally spaced at 3 mm from the end of the rod. An exploded view of the nozzle designed 

for solar reactor is shown in Figure 2. To hold the inner rod in its place, the inner rod needed to be welded to 

the outer nozzle. These changes were made in a way so that the flow field is not disturbed. 

3. Numerical study of the flow field and temperature distribution inside solar reactor 

In order to understand the effect of the nozzle design on the flow field inside the solar reactor, CFD simulations 

were done. The first objective of the simulations is to see whether the design changes and the scaling of the 

original nozzle design disturb the flow pattern. After simulating the gas flow inside the reactor, radiative heat 

transfer is taken into account. Simulation results are compared with experiments for validation. All simulations 

were done using Ansys Fluent 16.1. For CFD simulations, finite volume method was used where the flow field 

is divided to finite number of small flow fields. For each element, three conservation equations (mass, 

momentum and energy) are numerical solved. A complete 3D simulation was made because the new nozzle 

design is not completely axisymmetric. Conservation of mass: 

𝜕𝜑

𝜕𝑡
+ ∇ ∗ (𝜑𝑢) = 0                                                                                                                                               (1) 

Conservation of momentum: 

𝜕(ρ𝑣)

𝜕𝑡
+ ∇(φ𝑣 ⊗ 𝑣) = −∇p + ∇𝜏 + 𝜑                                                                                                                    (2) 

Conservation of energy: 

𝜑
𝑑ℎ

𝑑𝑡
=

𝑑𝑝

𝑑𝑡
+ ∇(𝑘∇𝑇) + 𝛷                                                                                                                                     (3) 

Since the expected flow field is swirling and has recirculation zones, the realizable k-ε model was used. The 

model has a good performance for flows involving: rotation, strong pressure gradients and recirculation (Ozalp 

and Devanuri, 2010). The turbulence kinetic energy, k-equation: 

𝜕

𝜕𝑡
(𝜑𝑘) +

𝜕𝑦

𝜕𝑥𝑗
(𝜑𝑘𝑢𝑗) =

𝜕

𝜕𝑥𝑗
[(𝜇 +

𝜇𝑡

𝜎𝑘
)
𝜕𝑘

𝜕𝑥𝑗
] + 𝐺𝑘 + 𝐺𝑏 − 𝜑𝜀 − 𝑌𝑀 + 𝑆𝑘                                                                     (4) 

The rate of dissipation, ε-equation: 

𝜕

𝜕𝑡
(𝜑𝜀) +

𝜕

𝜕𝑥𝑗
(𝜑𝜀𝑢𝑗) =

𝜕

𝜕𝑥𝑗
[(𝜇 +

𝜇𝑡

𝜎𝜀
)

𝜕𝜀

𝜕𝑥𝑗
] + 𝜑𝐶1𝑆𝜀 − 𝜑𝐶2

𝜀2

𝑘+√𝜈𝜀
+ 𝐶1𝜀

𝜀

𝑘
𝐶3𝜀𝐺𝑏 + 𝑆𝜀                                                 (5) 

where: 

𝐶1 = max [0.43,
𝜂

𝜂+5
],    𝜂 = 𝑆

𝑘

𝜀
 ,  and  𝑆 = √2𝑆𝑖𝑗𝑆𝑖𝑗                                                          

In order to understand the effect of swirling flow on temperature distribution inside the reactor, radiative heat 

transfer simulations were done per following governing equation: 

𝜕

𝜕𝑡
(𝜌𝐸) + ∇ ∗ (�⃗�(𝜌𝐸 + 𝑝)) = ∇ ∗ (𝐾𝑒𝑓𝑓∇𝑇 − ∑ ℎ𝑗𝐽𝑗⃗⃗⃗𝑗 + (𝜏𝑒𝑓𝑓̿̿̿̿ ̿̿ ∗ �⃗�)) + 𝑆ℎ                                                                 (6) 

Ansys Fluent features several radiation models, however, the Discrete Ordinates (DO) radiation model is the 

only model that allows radiation through semi-transparent walls. It is possible to model gray radiation as well as 

non-gray radiation by using a gray-band model. The use of the non-gray radiation model is necessary for 

materials that have absorption coefficients that vary stepwise across spectral bands. Behavior of gases which 

absorb and emit energy at different wavenumbers is not included in the non-gray radiation. The field of modeling 

non-gray gas radiation is still evolving. By using the Discrete Ordinates model, it is also possible to include 

scattering, anisotropy and particulate effects in the simulations. The Discrete Ordinates model considers 

radiation transfer given in Eq(7) as a field equation in the s⃗ direction. 

∇ ∗ (I(r⃗, s⃗)s⃗) + (α + σs)I(r⃗, s⃗) = an
2 σT

4

π
+

σs

4π
∫ I(r⃗, s⃗′)
4π

0
ϕ(s⃗ ∗ s⃗′)dΩ′                                                                     (7) 

1491



The discrete ordinates model solves this equation for each discrete solid angle that is associated with a direction 

vector s⃗. The angular space 4π is discretized in Nθ×Nϕ  solid angles with magnitude ωi for each octant. The 

control angles θ and φ are the polar and the azimuthal angle respectively. For 3D simulations, 8 octants are 

solved which makes a total of 8NθNϕsolved directions. The finer the discretization, the more precise the 

simulation is. A disadvantage of the fine control angles is the increase in computing time because the transport 

equation is solved for each direction. The boundary conditions of finite volume analyses are used to calculate 

the discrete values for each variable in the equations. They should define how radiation and flow enters and 

leaves the computational domain of the reactor. Heat transfer in solar reactor is approximated by two different 

types of boundary conditions: opaque and semi-transparent walls. The reactor walls are made of stainless steel 

so they are considered as opaque. The reactor walls are external to the domain because they only have a fluid 

zone on one side. Although radiation does not transmit through an opaque wall, there are still energy losses at 

the reactor walls due to external convection and radiation. Conduction through the reactor walls can be 

simulated by using shell conduction. The incident concentrated solar radiation enters the reactor through a 

quartz window, which has to be modeled as a semi-transparent boundary condition. The incident radiation flux 

enters the fluid cell zone through a semi-transparent wall. The transmitted radiation is refracted and dispersed 

specularly and diffusely. The boundary conditions which control the pheromones are the refractive index and 

the diffuse fraction. When the refractive index is not equal to 1, a part of the incident radiant is reflected. The 

incident radiation magnitude is modeled as radiative heat flux (W/m²). The solid angle over which the radiation 

is distributed is specified as the beam width. Simulations were done for two different geometries. First, reactor 

without the nozzle, and next, the reactor with nozzle. Once the reactor flow field was defined, it was divided into 

a finite number of small flow fields. Meshing of the flow field geometry was done via workbench as shown in 

Figure 3 for both geometries. 

      

Figure 3: Mesh of the reactor with no nozzle (left), and mesh of the reactor with nozzle (right) 

Flow rate of the secondary gas stream is the manipulating parameter to create a certain flow pattern with the 

nozzle. During the simulations, flow rate of the secondary gas flow needs to be increased to obtain different flow 

patterns. To obtain that, a transient analysis was done and a velocity profile was used at the secondary inlet. 

Based on the simulations, velocity of the primary gas flow was set at 1.3 m/s. During tests, the inlet velocity of 

the secondary gas flow should be small compared to the inlet velocity of the primary gas flow. Air is used as 

inlet gas for the primary inlet. Nitrogen is used for the secondary inlet. 

4. Numerical results  

A comparison was made between experimentally obtained flow field per Particle Image Velocimetry (PIV) results 

of Vanierschot (2007) and the CFD simulated flow field after 10 s of duration. As seen in Figure 4, flow field per 

CFD looks similar to the experimentally obtained flow field which indicates that this flow pattern could be created 

by the nozzle inside a solar reactor like it does in burner.  

 

Figure 4: Experimental flow field by Vanierschot (2007) (left), and CFD simulation results (right). At t = 10 s 

1492



After 30 seconds, nitrogen velocity reaches to 0.2 m/s. As seen in Figure 5, CFD simulated flow looks similar to 

experimentally obtained flow which reconfirms that flow pattern similar to the burner can be obtained inside the 

solar reactor by using the nozzle. 

 

Figure 5: Experimental flow field by Vanierschot (2007) (left), and CFD simulation results (right). At t = 30 s 

Results of the heat transfer simulations are shown in Figure 6 for reactor setup with no nozzle, and reactor with 

nozzle. Both simulations were done with the same boundary conditions to see the impact of the nozzle. 

       

Figure 6: Radiative heat transfer simulations for old reactor setup with no nozzle 

By visually comparing both simulation results, it is clear that the nozzle has an impact on the temperature 

distribution inside the solar reactor. Point marks A, B, and C are used to make comparison with reactor 

temperatures experimentally measured by using 7 kW high flux solar simulator. 

5. Experimental results 

Experiments were done using the setup seen in Figure 7. A 7 kW high flux solar simulator was used as the high 

temperature process heat. The reactor was placed at the focal point of the simulator and the temperature was 

measured at points A, B and C. Experiments were repeated for reactor with no nozzle and reactor with nozzle.  

 

Figure 7: Photo of the experimental setup and the reactor 

1493



First experiment was done using reactor without nozzle. The input current was be set at 155 A which represents 

solar radiation intensity at noon whereas input current of 115 A and 135 A represent solar radiation intensities 

in the morning and afternoon, respectively (Menon et al., 2013). The second experiment was done with the flow 

rates obtained from the simulations. Since the nozzle needs two different inlet flow rates, an additional air flow 

was provided on top of nitrogen flow where air velocity was 1.3 m/s, and nitrogen velocity was 0.3 m/s. The data 

of this experiment was used as reference. The third experiment was done with a total flowrate of 10 L/min. 

However, the air/nitrogen ratio was kept the same as in experiment #2. This experiment was done with a total 

flowrate of 10 L/min. Air and nitrogen flowrates were the same as in experiment #3. The changed parameter for 

this experiment is the input current which was set at 115 A. That experiment was done with a total flowrate of 

10 L/min. The air and nitrogen flowrates were the same as in experiment #3. The changed parameter for this 

experiment is the input current which was set at 135 A. Table 1 gives an overview of the experimental data 

measured. Temperature differences can only be compared between the first three experiments due to the power 

level. It is seen that the nozzle has favourable impact on the temperature distribution inside the reactor. 

Table 1: Summary of the experiment results 

Result type 

 

Setup Temperature    

at A [°C] 

Temperature    

at B [°C] 

Temperature      

at C [°C] 

Simulation Without nozzle 155 355 316 

Experiment 1 Without nozzle 155 322 289 

Simulation With nozzle 155 342 307 

Experiment 2 With nozzle 115 228 203 

 

In general, we can conclude that the temperatures in point C are in the same range, however, there is a 

significant difference between the temperature in point A and B. It should be noted that the simulated model is 

based on a flow field surrounded by shell conduction layers which represents the opaque reactor walls. It seems 

that the amount of heat transferred by conduction through the boundary shells is not comparable by the amount 

that is transferred by the stainless-steel walls of the solar reactor. To obtain more accurate simulation results, 

the reactor can be modelled with less simplifications. For example, the reactor can be fully 3D modelled instead 

of the shell layers, however, this would increase the computation time drastically. 

6. Conclusions 

A nozzle type inlet design inspired from a burner was tested for a solar reactor. It was an iterative design process 

based on CFD simulations. Once the nozzle was fully designed, numerical radiative heat transfer simulations 

estimated the effect of the nozzle design in comparison to the same reactor without nozzle. Due to its geometry, 

the nozzle creates different flow patterns inside the reactor. The goal was to make the temperature distribution 

inside the reactor more homogenous. At the end, the nozzle was tested by experiments and the CFD simulations 

were compared with experiments. Although the nozzle design was successful and made the temperature 

distribution inside the reactor more homogenous, the difference between the temperature in front of the reactor 

and the temperature at the back of the reactor was higher than targeted. However, it was concluded that it is 

possible and helpful to use nozzle or similar apparatus from traditional combustion systems as a starting point 

or preliminary work in solar thermochemical systems. 

Acknowledgment 

This research was funded by the Impulse Fund project # IMP/14/049 of KU Leuven. 

References 

Alonso E., Romero M., 2015, Review of Experimental investigation on Directly Irradiated Particles Solar 

Reactors, Renewable and Sustainable Energy Reviews, 41, 53–67. 

Menon A., Asadullah F., Ozalp N., 2013, A New Solar Reactor Aperture Mechanism Coupled with Heat 

Exchanger, Chemical Engineering Transactions, 35, 751–756. 

Ozalp N., Ibrik K., Al-Meer M., 2013, Kinetics and Heat Transfer Analysis of Carbon Catalysed Solar Cracking, 

Energy, 55, 74–81. 

Ozalp N., Devanuri J., 2010, CFD Analysis of Multi-Phase Turbulent Flow in a Solar Reactor for Emission-free 

Generation of Hydrogen, Chemical Engineering Transactions, 21, 1081–1086. 

Vanierschot M., 2007, Fluid Mechanics and Control of Annular Jets with and without Swirl, PhD thesis, 

Department of Mechanical Engineering, KU Leuven, The Netherlands. 

1494