Professional Documents
Culture Documents
Computational Methodology
4.1 Introduction
systems, CFD analysis will be carried out to estimate the flow and heat transfer in the
absorber alone. Three dimensional steady state conjugate heat transfer analysis was
conducted to investigate numerically the flow and heat transfer characteristics of the
absorber fluid using the finite volume based CFD code Ansys Fluent 15.0.
the absorber fluid depends mainly on the incident heat flux, in the form of focused solar
rays reflected from the trough collector. The percentage of the rays reflected by the
parabolic trough collector, from the incident solar rays and thus the heat flux
and sophisticated instrumentation. The present thermal problem at hand has surface
curvatures in the parabolic trough, absorber tube and poses large challenge in assessing
However, spatially varying heat flux distribution on the absorber outer surface as
received from the reflector; need to be specified as a thermal boundary condition for a
61
more accurate modeling of the solar thermal systems. For STA, this heat flux will be
varying around its circumference. Typically the heat flux will be maximum at the bottom
of the absorber tube and gradually decreases to the top side of the tube. For the modified
absorbers HTA and SPTA, the reflected heat flux will be hitting both the absorber tube
Ray-Tracing techniques are widely employed for estimating the heat flux from
solar rays received on the system under study. A direct method of calculating the heat
flux using an in-built ray tracing algorithm within the chosen CFD code or heat flux
specification as a boundary condition is needed. Hence an attempt has been made in the
present work to estimate the solar heat flux incident on the absorber using a novel
approach, where a MCRT (Monte-Carlo based Ray Tracing) code was coupled with a
FVM (Finite Volume Method) based CFD analysis solver. In the present numerical study,
the distribution of heat flux focusing on the absorber tube outer surface was numerically
estimated using a general purpose ray trace code SolTrace, with the solar intensity at
occur in the field erection of the PTC systems. Any small change in the distance between
reflector and absorber will alter the quantity of reflected solar rays focusing on the
To assess the influence of the parameters like (i) absorber distance from the
reflector, (ii) absorber tube diameter and (iii) absorber fluid flow rate on the flow and
thermal characteristics of the STA, numerical analysis was carried out. Off design
62
positioning of the absorber assembly towards or away from the reflector, were
numerically simulated. The influence of tube diameter and coolant flow rate was studied
numerically, by modifying the absorber tube geometry and changing the boundary
conditions in the solver. For the modified absorbers HTA and SPTA, the influence of the
parameter (iv) absorber sheet width was also considered, in addition to the parameters
considered for STA. The computational methodology adopted and the geometrical
parameters studied in the present numerical work are discussed in detail in the following
sections.
In the present study, the concentrated solar irradiation heat flux distribution as
reflected by the PTC was obtained by using a general purpose ray trace code SolTrace. It
follows a sequential coupled method to obtain the concentrated heat flux distributions on
the absorber tube surface. A brief introduction to the ray-tracing code is given in the
following section.
Laboratory (NREL) is a general purpose ray trace code based on the Monte-Carlo
method. The code is capable of modeling the concentrating solar power optical systems
and analyzes their performance. Although originally intended for solar applications, the
code can also be used to model and characterize many general optical systems. SolTrace
can model parabolic trough concentrators as well as dishes, towers or other unique
geometries (linear power towers, solar furnaces, etc.). It models optical geometries as a
63
series of stages composed of optical elements that possess an extensive variety of
Fig.4.1 shows the screenshot of the SolTrace software. The code utilizes a ray-
tracing methodology developed by Spencer and Murty [97]. The system under study for
the solar thermal / optical analysis need to be modeled within the software.
The sun, sun position, shape of sun intensity, optical elements to be tested and the
absorber are modeled within the SolTrace.Fig.4.2 shows the components of the SolTrace
software such as Sun, Optical Properties, System Stages and Trace Options.
64
Fig. 4.2 Components settings window in SolTrace software
Rays falling from the sun are traced through the system while encountering
various optical interactions was modeled. Some of these interactions are probabilistic in
nature (e.g. selection of sun angle from sun angular intensity distribution) while others
surface and resultant redirection). SolTrace has the advantage over other codes such as
Comsol, Star ccm+ Trnsys since SolTrace has been more user friendly.
Therefore SolTrace can provide accurate results for complex systems that cannot
be modeled otherwise. Accuracy of ray prediction increases with the number of rays
traced and larger ray numbers means more processing time. Also, complex geometries
translate into longer run times. Fewer rays (and therefore less time) are needed to
determine relative changes in optical efficiency for different sun angles on a given solar
concentrator than that are needed to accurately assess the flux distribution on the receiver
65
4.2.2 Inputs to SolTrace
The PTC system of investigation was modeled in the SolTrace code with the solar
trough collector and the single tube absorber, as independent elements, positioned
relative to each other in the same coordinate system similar to the experiments. The sun
position was specified with respect to the PTC system to model the impinging solar rays
using global coordinate system approach. The stage properties modeled in the STA with
the parabolic trough collector and the absorber tube is shown in Fig. 4.3(a). The element
surface and aperture modeling details for the parabolic trough surface and the absorber
66
(c) Aperture definition for the PTC and the absorber – STA
Two characteristics completely define the “sun” as the light source in SolTrace
code: (1) the angular intensity distribution of light across the sun‟s disk (referred to as the
sun shape) and (2) the sun‟s position. A predefined sun profile was chosen with a peak
Total number of solar rays traced for the present study was 1 million in SolTrace.
Fig.4.4 shows only a sample set of 100 rays (1-100 no‟s) traced in the PTC system to
avoid clutter. Fig.4.4 (a) shows the front view of the rays traced from the sun hitting the
reflector and the reflected rays hitting the absorber tube. It is to be noted that the picture
aspect ratio is not to scale and reproduced here as obtained from the SolTrace code. The
solar rays were made to hit the reflector from the top towards the bottom. The
day. Fig. 4.4 (b) shows the isometric view of the PTC system with the solar rays. The
67
shadow cast by the absorber tube on the trough can be visualized as a dark region, in
which the effective ray impinging and subsequent reflection was blocked.
Fig.4.5 shows the incident and reflected solar rays for the HTA / SPTA assembly.
The absorber plate alone has been used to trace the heat flux incident on the absorber
assembly. The heat flux value on the absorber plate bottom surface has been used to map
the absorber tube assembly and absorber plate in HTA. Similarly this heat flux was
mapped to the serpentine tube and absorber plate assembly for the SPTA. It can be
observed that the shadow cast by the HTA on the parabolic reflector is wider than that of
68
Incident rays Absorber tube
Reflected rays
Parabolic trough
Wr
Incident rays
Lr
Shadow of
absorber tube
Fig. 4.4 Sample rays traced for the PTC system using SolTrace - F = 42.85 cm
69
Absorber plate
Shadow of
absorber plate
Fig. 4.5 Sample set of solar rays traced in SolTrace for STA / HPTA - F = 42.85 cm
70
4.2.3 Working Procedure
coordinate system. A stage is loosely defined as a section of the optical geometry which,
once a ray exits the stage, will not be re-entered by the ray on the remainder of its path
through the system. Complete system geometry may consist of one or more stages. It is
The motivation behind the stage concept is to employ efficient tracing and
therefore save processing time and allow for a modular representation of a system. The
other significant benefit of stages is that they can also be saved and employed in other
system geometries without the need for recalculating element positions and orientations.
Segmentation of the geometry via stages may also be useful in assignment of different
coordinate systems for specific geometries and can make adjustment of locations far
more straightforward.
interaction type, an aperture shape and, if appropriate, a set of optical properties. The
location and orientation of stages are defined within the global coordinate system
whereas the location and orientation of elements are specified within the coordinate
system of the particular stage in which they are defined. Stages can be one of two types:
optical or virtual. An optical stage is defined as one that physically interacts with the rays
and can potentially altering their trajectories. Conversely, a virtual stage is defined as one
that does not physically interact with the rays. The virtual stage is useful for determining
71
ray locations and directions and incident power or flux at various positions along the
optical path without physically affecting ray trajectory and has no optical properties.
Beyond this, optical and virtual stages are identical in how they are defined and
used. Stages can be duplicated and moved around as groups of elements and saved for
use in other system geometries. It is not possible to mix virtual and optical elements
system, the stage coordinate system and the element coordinate system. These are
illustrated in Fig. 4.6. Each element in a stage has a local coordinate system (i.e. location
and orientation) defined relative to the stage coordinate system. Each stage has a
coordinate system defined relative to the global coordinate system. The direction of the
72
The sun direction can be input in either vector form, or in time-of-day, day-of-
year format with latitude specified. For the present numerical analysis, the vector form of
specification for the sun was used. Light rays are generated from the sun and then traced
sequentially through each stage in the geometry. The position and direction of each ray in
The heat flux distribution on the absorber tube as estimated by SolTrace is shown
in Fig. 4.7.It is to be noted that the heat flux distribution on the absorber surfaces were
estimated for an incident beam flux of 800 W/m2, as observed in the experimental
measurements. The two dimensional plan view of the STA tube outer surface heat flux is
shown in Fig. 4.7(a) for the tube diameter Do=25.4 mm. The abscissa represents the
circumferential perimeter length of the absorber tube and the ordinate represents the tube
axial length. The three dimensional heat flux distribution on the STA is shown in
Fig. 4.7 (b). Since most of the reflected sun rays are focusing to the bottom of the tube,
higher heat flux was observed in this region and decreases towards the tube top. At the
tube top, the heat flux is mostly from the direct beam radiation.
The spatially varying three dimensional heat flux distribution thus obtained was
mapped on to the absorber surfaces in the CFD solver Fluent. The temperature
distribution on the absorber tube and the heat gained by water were estimated from the
CFD analysis. Figs.4.7 (c) and (d) show the corresponding plan view and the three
dimensional map of the heat flux distributions on the absorber sheet bottom surface of the
absorber sheet used in HTA / SPTA. Figs.4.8 and 4.9 show the heat flux map for the STA
73
and HTA / SPTA at different absorber focal positions (F), as obtained from SolTrace.
Fig.4.10 shows the heat flux distribution on STA for different tube diameters whereas
Fig. 4.11 shows that on HTA/SPTA absorber sheets for different widths. For the CFD
analysis with a particular absorber positioned at the chosen focal length, the
corresponding three dimensional heat flux map was used as the thermal boundary
condition.
Tube
Top
Tube
Tube Tube bottom
Tube Top Bottom
bottom
Tube axial length, Lr
(c) HTA/ SPTA – Plan view (d) HTA / SPTA – Isometric view
Fig. 4.7 Heat flux distribution on the absorber from SolTrace - F = 42.85 cm
74
Tube bottom
Tube top
(a) F = 41.35 cm (b) F = 41.85 cm
(g) F = 44.35 cm
75
Absorber
plate mid
(g) F = 44.35 cm
76
Tube bottom
Tube top
(a) Do = 10.0 mm (b) Do = 12.7 mm
(c) Do = 25.4 mm
Fig. 4.10 Heat flux distribution on STA for different tube diameters
Absorber
plate mid
(c) Wa = 15.0 cm
Fig. 4.11 Heat flux distribution on HTA / SPTA for different absorber widths
77
4.3 Physical Domain
The physical model or domain is the setup that was used in the experiments. The
physical model consists of a parabolic trough collector, an absorber tube with a glass
cover. Apart from the PTC, the single cylindrical copper tube in the STA or the absorber
tube assembly and the absorber sheet in the HTA form the physical model, whereas the
corrugated tube will be considered for the SPTA. For the experiments conducted on the
conventional and modified absorbers, the experimental setup details were discussed
already in Chapter 3.
The computational model or domain is the CAD model built with reference to the
physical model tested in the experiments and the details of which is discussed in this
section. The water carrying copper tube in the STA or the set of copper tubes and the
absorber plate in the HTA or the serpentine tube and the absorber plate in the SPTA
forms the computational domain for the present study. The internal volume of the
absorber tube where the working fluid flows will be treated as the „fluid domain‟ in the
numerical analysis. The solid thickness of the absorber tube and the absorber sheet will
be modeled as „solid domain‟. The solid and the fluid domains of the absorber system
were modeled in a conjugate fashion, to have the un-interrupted exchange of heat flux
across the domains. The solar heat flux estimated from the SolTrace code was specified
as a boundary condition on the absorber tube outer surface or absorber sheet, tube
assembly. The PTC trough, air domain between the trough and absorber and the ambient
78
air surrounding the absorber are not considered for the present computational study as
Figure 4.12 shows the computational domain of STA ie., the absorber tube used
for the present CFD analysis where one end of the computational domain is shown. The
STA is a simple copper pipe of chosen diameter Do, with thickness t and of length Lr. The
solid copper tube domain and the water domain inside the tube were modeled in a
conjugate fashion, where a common surface will be used to represent both the fluid and
solid domains in the pre-processor, which enables an accurate thermal prediction in the
solver.
Tube top
Tube bottom
Cold water
inlet
79
4.4.2 Header Tube Absorber (HTA)
Water is passed through the copper tubes and carries the heat gained by the
absorber system. Fig.4.13 (a) shows the isometric bottom side view of the modified
absorber HTA, where the appropriate value of heat flux from SolTrace was mapped as a
thermal boundary condition. Fig. 4.13(b) shows the modeling strategy adopted presently
in the region of absorber tube and absorber plate junction. To avoid any possible mesh
issue due to sharp corner and to avoid a line contact, the tube geometry is made to merge
with the plate to have contact over an area. This will ensure the replication of closure of
the gap between the plate and the tube in the experiments, by brazing and black paint.
Absorber Area
sheet contact
80
4.4.3 Serpentine Tube Absorber
Absorber sheet
Absorber
Absorber sheet
fluid inlet
hexahedral finite volume elements throughout the computational domain. Finer near wall
meshing is developed using the boundary layer mesh feature of the pre-processor with the
first node at 0.1 mm and with a default growth factor of 1.1 for 10 layers. This is done to
capture the flow and thermal features happen close to the wall more accurately. Grid
independency tests are carried out for the STA taking the surface averaged fluid outlet
temperature as the assessment criteria. Similar meshing strategies have been adopted for
81
4.5.1 Single Tube Absorber – STA
Fig. 4.15 Meshed single tube absorber (STA) tube (solid domain only shown)
elements were generated with three different mesh element sizes to form coarse, medium
and fine meshes with 0.65, 0.97 and 1.3 million cells respectively. The generated mesh of
Fig.4.16 shows the variation of fluid outlet temperature obtained from the
computations on STA with different mesh densities. It was observed that the temperature
value increases from a mesh size of 0.65 to 0.97 million. The variation of this value from
0.97 to 1.3 million cells is comparatively less, which indicates that the effect of mesh
density is beyond 0.97 million cells got insignificant improvement in the output and
hence it was decided to continue with the mesh settings of 0.97 million cells for the
present study.
82
50
49
Tf,o (oC)
48
47
46
0.65 0.97 1.3
Computational mesh size (in million cells)
Fig. 4.16 Absorber fluid outlet temperature at different grid sizes – STA
The end view of the computational mesh with 0.97 million cells with fine mesh
Solid
Fluid
Coarse
Fine
Fig. 4.17 End view of the meshed single absorber tube with 0.97 Million cells
83
4.5.2 Header Tube Absorber – HTA
Hexahedral
cells
Tetrahedral
cells
Tube-03
Tube-02
Absorber
Tube-01
sheet
Absorber
fluid inlet
Fig.4.18 shows the meshed computational domain for the HTA where hybrid
meshing, a combination of tetrahedral and hexahedral cells, were employed. The straight
tube regions and most of the absorber plate regions are meshed with unstructured
hexahedral elements whereas the more complex regions like pipes intersection are
Fig.4.19 shows the meshed SPTA computational domain which has been adopted
fully with hexahedral cell. The corrugated tube was meshed with a surface mesh first in
the absorber fluid inlet face and the mesh was swept along the axis of the tube to create a
84
Fig. 4.19 Meshed SPTA assembly
The fundamental governing equations for flow and heat transfer are the
continuity, momentum (Navier-Stokes) and energy equations along with the equations for
modeling the turbulence quantities. The governing equations are presented in this section.
Based on the mass flow rate and tube dimensions, the calculated Reynolds numbers are
found to fall well below the critical value and hence the flow is treated as laminar in the
present computations of single tube absorber. The conservation equations for laminar
flow in an inertial (non-accelerating) reference frame are presented here. The equation for
⁄ ( ⃗) (4.1)
Equation 4.1 is the general form of the mass conservation equation and is valid
for incompressible as well as compressible flows. The source Sm is the mass added to the
continuous phase from the dispersed second phase (e.g., due to vaporization of liquid
85
droplets) and any user-defined sources. For the present computational analysis, the source
( ⃗) ( ⃗ ⃗) ( ̅) ⃗ ⃗ (4.2)
Where p is the static pressure, ̅ is the stress tensor, and ⃗ and ⃗ are the gravitational
body force and external body forces (e.g., that arise from interaction with the dispersed
phase), respectively. ⃗ also contains other model-dependent source terms such as porous-
̅ *( ⃗ ⃗ ) ⃗ + (4.3)
Where is the molecular viscosity, I is the unit tensor, and the second term on the right
The CFD code solves the energy equation in the following form
Where keff is the effective conductivity (k+kt), where kt is the turbulent thermal
conductivity, defined according to the turbulence model being used), h is the sensible
enthalpy for ideal gases, ⃗ is the diffusion flux of species j and Sh is the energy source
term. The first three terms on the right-hand side of the equation represent energy transfer
86
due to conduction, species diffusion, and viscous dissipation, respectively. More details
of the governing equation can be found in Ansys Fluent 15.0 users guide (2013).
For the modified absorber designs – HTA and SPTA, the flow rates specified for
the in combination with the smallest pipe diameters gave Reynolds numbers in the
turbulence regime. Hence turbulence modeling was adopted for these cases. Further to
[74], the sst-k model was chosen for the present computations with HTA and SPTA.
The details of the governing equation for the turbulence model can be found in Ansys
Fluent 15.0 users guide [98] and hence not reproduced here.
Three dimensional steady state approach was used to minimize the computational
effort, time and cost of computing. To investigate the flow and heat transfer in the
absorber, the code Ansys Fluent 15.0 was used. An instantaneous solar intensity has been
given as one of the input to the solTrace code. The heat flux from SolTrace was mapped
interpolation technique within the solver. The absorber fluid outlet temperature from the
computations for the given boundary conditions was estimated. Hence the steady state
The boundary conditions used for the present conjugate thermal computational
fluid dynamic analyses are taken from the present experiments. For an ambient
87
temperature of 30oC, the radiation loss to the ambient is negligible. Hence, the thermal
energy loss to the ambient due to convective and radiative heat transfer, from the
absorber tube, was not considered in the present computational analysis. The summary of
the boundary conditions (BC) used in the present computations are listed below.
The case with absorber tube axis position at 42.85 cm was considered as the baseline. For
the design and off-focal absorber tube axis positions, the heat flux estimates on the
absorber tube surface were obtained from SolTrace and mapped to the CFD analysis.
The heat flux values estimated from SolTrace were exported to excel and
converted as profile files which was later read into the CFD solver Fluent. These heat
flux values were assigned to the absorber tube surfaces in the solver, by importing the
88
profiles in to the solver and assigning the corresponding heat flux values to the
Figure 4.20 shows the heat flux from SolTrace code fed into Fluent software and
mapped to the surface of the absorber.Fig.4.20 (a) shows the contour of heat flux mapped
on to STA outer surface. As observed in the heat flux distribution from SolTrace, the
bottom side of the absorber tube has highest heat flux value and it decreases gradually
(c) SPTA
89
Figs. 4.20 (b) and (c) show the corresponding distribution in the HTA and SPTA,
for the absorbers positioned in the design focal distance of 42.85 cm. Similarly, the heat
flux of the corresponding absorber surfaces were mapped from SolTrace to Fluent, for
different absorber position ( F), different tube diameters ( Do ) and for different absorber
sheet widths ( Wa ).
The convergence criteria for the computational solution are determined based on
scaled residuals for the equations of continuity, momentum equations and turbulence
quantities specific to the respective models. The scaled residuals for solution convergence
have been set to 10-5 for all governing equations and turbulence quantities. The solution
has been considered to be converged when all the scaled residuals are less than or equal
to this preset value. Computations are carried out until the steady state has been reached.
For few cases, the convergence could not be achieved to the desired accuracy. In those
cases, the iteration has been continued further to a stage that the results do not vary even
90