Renewable and Sustainable Energy Reviews 134 (2020) 110331

Contents lists available at ScienceDirect

Renewable and Sustainable Energy Reviews

journal homepage: http://www.elsevier.com/locate/rser

A survey of solar concrete shell collectors for parabolic troughs

Patrick Forman a, *, Sebastian Penkert b, Christoph Kämper a, Tobias Stallmann b, Peter Mark a,
Jürgen Schnell b
a
Ruhr University Bochum, Institute of Concrete Structures, Universitätsstraße 150, 44780, Bochum, Germany
b
Technical University of Kaiserslautern, Institute of Concrete Structures and Structural Engineering, Paul-Ehrlich-Straße, Building 14, 67663, Kaiserslautern, Germany

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

Keywords: Concrete shell collectors offer an alternative to conventional parabolic trough collectors. The principle design
Parabolic trough collector concept is derived from existing barrel rooves that effectively bridge large spans in halls or buildings with
High-performance concrete minimum material usage. The concrete troughs merge the bearing structure and mirroring surface to just one
Solar concrete shell structure
shell of a few centimeters. They are made of high-strength concrete and track the sun via pure axial rotation or
Prototype application
lateral movements that avoid any lifting works. In the present contribution, basic constraints in materials, ge­
Design criteria
Wind loads ometry, and static calculation are derived and converted into a framework of possible designs. This contribution
Photogrammetry thereby presents a survey of concepts that range from small-scale prototypes to full-scale realizations of 140 m2
apertures and large-aperture concepts with a 10 m width. Design concepts with bearing and drive systems as well
as optimization-based form findings are introduced to elaborate shells of minimum weight with solid sections,
stiffeners, and hollow cores.

tanks that enable electricity to be generated during sunless hours. This

electricity generation is also the main difference to photovoltaics, in
1. Introduction
which solar radiation is directly transferred to electricity. Alternatively,
PTCs can also be used for industrial heat production (e.g. Ref. [5,6]) and
Renewable energy production amounts to 2537 GW (1246 GW
in other applications like air heating systems [7,8] or refrigeration
without hydropower) which is equivalent to 26% of the global energy
systems (e. g. Ref. [9]). For industrial heat production, required tem­
consumption [1]. Thereby, solar energy contributs approx. 100 GW.
peratures lie in the range of 100–300 ◦ C and consequently have lower
With a total volume of about 1 GW worldwide [2] – more than 50% of
concentration ratios. Overviews of the applications are provided in Refs.
which is already operational (the rest being under construction or still in
[10–12] whereby an additional historical review is given in Ref. [13].
the planning stage) – concentrated solar power (CSP) remains a
Moreover, hybrid concentrating photovoltaic/thermal systems have
comparatively young technology. Parabolic trough power plants are
been proven to be efficient [14,15], too. Developments on solar thermal
currently the most well-established form of CSP plants. In CSP plants,
collectors in general are provided in Ref. [16,17].
incident solar irradiation is bundled onto an absorber by means of a
PTCs for CSP plants usually consists of spatial steel frameworks [18]
mirror surface, and a heat transfer fluid is heated. CSP can therefore be
that support point-wise mirror elements. The single modules are com­
classified mainly into line- and point-focus systems. These systems differ
bined to collectors of up to 100–150 m in length. The individual modules
in their concentration ratio – that is, the ratio of reflecting aperture area
consist of a main girder and additional secondary bearing structures to
to absorbing receiver area – and thus also in their process temperature.
support the mirror elements with sufficient inherent rigidity. However,
Point-focus systems contain a central receiver – such as solar tower
due to this separation into main and secondary supporting structure,
power plants [3,4] – and generate process temperatures of up to
deformations of both as well as of mirror elements superimpose and
1000 ◦ C. Parabolic trough collectors (PTCs) belong to line-focus systems
could lead to losses in solar concentration. For assembly, special
and achieve temperatures of up to 400–500 ◦ C for solar energy gener­
mock-ups are used to fulfill the necessary precision of the structures.
ating systems. The thermal energy generated in the absorber is used to
Widely used steel framework PTCs include the EuroTrough (with
operate a downstream power plant unit with integrated heat storage

Abbreviations: CSP, Concentrated Solar Power; LACC, Large-Aperture Concrete Collectors; PTC, Parabolic Trough Collector; REPA, Rotation and Expansion
Performing Assembly; SIMP, Solid Isotropic Material with Penalization; (U)HPC, (Ultra-)High-Performance Concretes.
* Corresponding author.
E-mail address: [email protected] (P. Forman).

https://doi.org/10.1016/j.rser.2020.110331
Received 31 March 2020; Received in revised form 12 August 2020; Accepted 29 August 2020
Available online 12 September 2020
1364-0321/© 2020 Elsevier Ltd. All rights reserved.
P. Forman et al. Renewable and Sustainable Energy Reviews 134 (2020) 110331

Nomenclature l Module length (m)

M Moment on foundation (kNm)
αc Endurance-coefficient (− ) N Normal force on foundation (kN)
C Concentration ratio (− ) qb Gust pressure (kN/m2)
c Compliance (Nm) qw Wind load (kN/m2)
cp,net Wind load coefficients (− ) ρ (Pseudo) Density (− )
d Absorber tube diameter (mm) ρc Density of concrete (kN/m3)
e Eccentricity (m) ρair Density of air (kN/m3)
Ecm Mean Young’s modulus (N/mm2) rms Root mean square (− )
fcm Mean compressive strength (N/mm2) SD Slope deviation (mrad)
fct Characteristic tensile strength(N/mm2) σ Stress (N/mm2)
fctm Mean tensile strength (N/mm2) σI 1st principle stress (N/mm2)
fct,fl Flexural tensile strength (N/mm2) vref Reference wind speed (m/s)
fctm,fl Mean flexural tensile strength (N/mm2) w (Aperture) Width (m)
fPT Acceptance function (− ) wk Weighting factor (− )
h Height (m) ϕr Rim angle (◦ )
θ Solar beam spread (mrad) x,y,z Cartesian coordinates (− )
θ1,2 Acceptance half angle (mrad)

aperture dimensions of 5.77 m by 12 m [19] used in solar power plants materials. Since the efficiency of PTC is developed quite far, future po­
Andasol 1–3 [20]), the Ultimate Trough (with 7.5 m by 24 m [21]), and tential in solar energy [31] is focused on the reduction of investment and
the SpaceTube (with 8 m by 14.2 m [22]). The main girder is often non-technical costs [32] with respect to local constraints of CSP plants
designed as a torque box [19,21,23] or torque tube [24] to transmit the (e.g. Ref. [33]). Thereby, the most cost-efficient PTC nowadays is the
torsional loads between the single modules. The Solarlite SL 4600 SkyTrough with costs of <100 $/m2 [34].
module represents a concept that uses alternative materials with an In this contribution, the design and development of PTCs aim to
aperture width of 4.6 m and a length of 12 m [25]. The module consists achieve cost reduction by using free shapeable concrete for the struc­
of a lightweight glass-fiber reinforced composite material with a directly tural design. Doing so, material efforts are minimized applying shells
applied thin glass layer and a minimum use of steel components as a rather than beams or grillages. Additionally, the supporting and the
torque framework. The first known concrete collector – which is already reflector structure are merged to form a low-deformation and preferably
being used in a test facility in Morocco [26] – was built from pre­ lightweight shell, which means that mirror elements with inherent
fabricated, pre-stressed steel-fiber-reinforced concrete elements. The stiffness, can be omitted since either mirror foils or thin glass mirrors are
concrete structure supports a pneumatic mirror system with a receiver directly applied to the shell surface. Therefore, the development of the
and an integrated secondary mirror, which becomes necessary because PTC shell structure alters the load-bearing behavior of barrel vaults,
of a not ideal parabolic shape of the pneumatic system. The collector has which are mostly used in roof constructions (barrel rooves). The chal­
an aperture width of 9.8 m and a length of 45 m [27]. However, the lenge now is to transfer the load-bearing capacity of this stationary
elements exhibit rather large cross sections and thus a pronounced dead construction with mainly static loads to a solar collector that tracks the
load. An overview of the developments of mostly established sun and is hence continuously exposed to varying loads. The essential
steel-framework PTCs is provided in Ref. [28,29], and alternative con­ demands for this task are a high bearing capacity in order to be able to
cepts are summarized in Ref. [30]. Generally, the trend aims toward withstand the specific action effects and an uncracked state to ensure the
higher aperture widths. The use of alternative materials is motivated by low-deformations required for full solar concentration (Fig. 1). To fulfill
reducing costs and generally focuses on assembly, operation, and these requirements, high-performance concrete (HPC) with high

Fig. 1. Principle of barrel rooves (left) and derived principle of concrete shell parabolic collector (right).

2
P. Forman et al. Renewable and Sustainable Energy Reviews 134 (2020) 110331

strength and pronounced durability is chosen. Specific action effects and Table 1
design criteria for concrete shell PTCs are introduced. Moreover, Material properties of Nanodur® concrete (Subscripts: c – concrete, t – tension,
structural and conceptual design options and strategies as well as their m –mean, fl – flexural) acc. to [38–40].
limitations are presented. Actual implementations are shown by means Young’s modulus Ecm 47,600–52,700 N/mm2
of numerical, optimization-based studies and realizations in small- and Compressive strength fcm 130–139 N/mm2
full scales. Furthermore, a visionary design concept for large-aperture Flexural tensile strength fctm,fl 15.1–18.2 N/mm2
concrete collectors (LACC) with an aperture width of 10 m is derived. fct,fl 12.8–15.5 N/mm2
In section 2 constraints of the design of PTC concrete collectors are Axial tensile strength fctm 7.8–9.3 N/mm2
fct 6.6–8.0 N/mm2
derived. They comprise the material properties of HPC, a criterion to
Density ρc 2440–2524 kg/m3
qualify the accuracy, and specific action effects. Section 3 presents the
possibilities of the collector design with respect to motions, statical
systems, dimensions and cross sections. Moreover, recommendations for determined. It can be mathematically transferred to the characteristic
suitable, unfavorable and not possible design options are given. An axial tensile strength with respect to the geometry of the test specimen
overview of built-up prototypes, full-scale realizations and numerical acc. to Ref. [42]. Then, the characteristic value of the axial tensile
investigations are demonstrated in section 4 and discussed, especially, strength fct is derived from the experimental, scattering data. In addi­
with respect to the comparatively higher weight of a concrete structure. tion, it is reduced by an endurance-coefficient of αc = 0.80 [43] that
Finally, conclusions are drawn section 5. takes into account long term action effects. The density of approximately
2500 kg/m3 corresponds to that of normal strength concrete and defines
2. Specific action effects and design criteria the self-weight.

The design and dimensioning of concrete shells and their compo­

nents for PTCs are mainly dependent on the material – in this case, HPC. 2.2. Accuracy analysis
Different load- and deformation conditions thereby have to be consid­
ered. Due to the tracking of the sun, variable load conditions occur The optical efficiency of a collector crucially depends on an undis­
throughout the course of the day. However, the load-bearing capacity of turbed parabolic shape. If the deformations are too large, the direct solar
the PTCs and the concentration of the solar radiation must always be radiation can no longer be concentrated on the absorber tube. To eval­
ensured. uate the concentration of solar rays, an acceptance function fPT was
derived in Ref. [44] that determines the local effectiveness as a function
2.1. High-performance concrete of the collector geometry, the absorber tube’s diameter, and the solar
beam spread of a reflected beam by a function of the acceptance half
Normal-strength concrete has a low modulus of elasticity and low angles θ1 and θ2 (cf. Fig. 1):
tensile strength. This characteristic would lead to components with ⎧ 1 for |θ| < θ1
necessary minimum thicknesses in the range of several centimeters to ⎪
⎪
⎪
⎪
⎨ ( ϕ )
decimeters and, as a consequence, would result in high self-weights. The 2 tan r
fPT = ϕ
cot r 2 − 1 (1)
use of normal-strength concrete therefore does not appear to be suitable ⎪
⎪ 2 Cθ
for θ1 < |θ| < θ2
⎪
in the application of lightweight, slender shells. (Ultra-)High-perfor­ ⎪
⎩
mance concretes (UHPC) [35] exhibit up to twice the modulus of elas­ 0 for |θ| < θ2
ticity and – even more importantly – a much higher tensile strength,
which significantly affects the design. The compressive strength is also with:
greater than that of normal-strength concretes, yet is of minor impor­ sin ϕr
tance here. The high strengths are achieved by a dense packing of the θ1 = (2)
C
aggregates and the cement matrix. As a consequence, it also results in a
high durability [36]. 2 tan ϕ2r
θ2 = (3)
Until the axial tensile strength is reached, UHPC maintains an almost C
linear-elastic material behavior (state I). Afterward, cracks nucleate,
Here, ϕr corresponds to the rim angle of the parabola, C to the geo­
thereby causing brittle failure. To avoid this brittle failure, ductile
metric concentration ratio (defined as the ratio of aperture width w to
behavior is ensured by reinforcements. Thus, when cracks do occur, they
absorber tube diameter d), and θ to the local solar beam spread caused
are bridged by the reinforcement, but the deformations increase
by a disturbed parabola shape. For the design of parabolic shells, for
disproportionately (state II). For the design of PTC concrete shells, the
example, by numerical calculation, deformation caused by specific
first principal stress is limited to a share of the axial tensile strength in
stresses (see Section 2.3) is limited such that a resulting slope deviation
order to ensure a mathematically uncracked state I. However, rein­
of the surface causes a solar beam spread θ smaller than the acceptance
forcement is always necessary to ensure robustness.
half angle θ1 (which is the more conservative case). The absolute
The collectors presented here were developed using a high-
acceptance half angle as well as other geometric values for the small-,
performance concrete based on the bonding agent Nanodur® [37]. In
addition to the attributes already mentioned for UHPC, Nanodur®
Table 2
concrete is characterized by very good workability and self-compacting
Geometric values and acceptance half angles for different collector designs.
behavior. Nevertheless, the realizations shown in this article differ in
their exact concrete mixture and therefore in the resulting material Design Aperture Rim Absorber Concentration Acceptance
width w angle tube ratio C [− ] half angle θ1
properties. The data ranges determined in experimental investigations
[m] ϕr [◦ ] diameter [mrad]
[38–40] are summarized in Table 1. For numerical analysis, the mean d [mm]
Young’s modulus Ecm is decisive in predicting the deformations. In
Small 2.205 70 51 43 21.7
contrast to UHPC, the compressive strength fcm is less than 150 N/mm2, scale
which is the only reason that this concrete is categorized as an HPC only Full 5.77 80 70 81 12.1
[41]. Nevertheless, the numerical material model is restricted by a share scale
of the axial tensile strength which is more than ten times smaller than Large 10 80 95 105 9.4
scale
fcm. Based on bending tests, the mean flexural tensile strength fctm,fl is

3
P. Forman et al. Renewable and Sustainable Energy Reviews 134 (2020) 110331

full-, and large-scale designs shown in Section 4 are presented in Table 2. equivalent to those of the EuroTrough-like module. Results are line-like
Obviously, with increasing aperture width, the acceptance angle de­ distributions of wind load coefficients cp,net (Fig. 2, bottom), which are
creases, and higher accuracy demands consequently have to be fulfilled. multiplied by the gust pressure as a function of a mean reference wind
It must be noted that further uncertainties (e.g., due to tracking errors) speed vref and thus provide the wind load qw:
have to be accumulated to the load-induced solar-beam spread. Never­
qw = cp,net ⋅qb (4)
theless, the acceptance half angle is used as an absolute, maximum limit
3
in the form-finding process. with: qb = 0.5ρair vref ρair = 1.25 kg/m (air density).
For a more precise determination of the surface quality, the slope An evaluation of force and pitching moment coefficients is provided
deviation (SD) of the surface is evaluated [45]. A reference value is given in Ref. [51]. The results reveal good conformity with similar wind tunnel
by the root mean square SDx,rms – significantly in the transversal x tests in Ref. [52].
direction (cf. Fig. 1) since the reflected solar rays are deviated along the In order to investigate influences in corner and edge areas of the
absorber in the longitudinal direction. This value allows for an evalua­ shells as well as effects of row- and field arrangements, additional wind
tion of the efficiency (intercept factor) – that is, the percentage of re­ tunnel tests were performed on large-area models as single models as
flected radiation bundled onto the absorber tube both for the numerical well as in group arrangements with varying flow direction [53]. These
deformation calculations and for photogrammetric measurements [46]. tests are valid for PTC modules with dimensions of approximately 10 m
For the EuroTrough module, slope deviations SDx,rms,i due to different by 30 m and are used for the design of large-aperture concrete collectors.
effects (e.g., collector torsion, mirror shape uncertainties, tracking er­ For the dimensioning, a distinction between three states with indi­
rors) that accumulate to 6.2 mrad and a resulting intercept factor of vidual restrictions was made with respect to an approach in Ref. [54] for
98.7% [47] were determined. To achieve similar efficiency, a maximum steel PTCs. The three states differ in their reference wind speed. In the
slope deviation SDx,rms of approximately 2 mrad caused by operational state, the optical efficiency has to be maintained under
load-induced deformation was derived for EuroTrough-like PTC mod­ moderate wind conditions (vref ≤ 10 m/s). In the transition state, the
ules. For other PTC dimensions, their specific characteristics have to be load bearing capacity (i.e., in this case, an uncracked state I) has to be
adopted from modules with similar dimensions as long as similar mea­ ensured under arising wind (10 m/s ≤ vref ≤ 15 m/s). In the survival
sures exist. state under violent wind loadings (vref ≤ 33 m/s) no operation but pure
survival, i.e. ensuring the bearing capacities, takes place. Table 3 dis­
2.3. Action effects plays the assumed wind speeds for dimensioning and the imposed
restrictions.
PTCs are subjected to various action effects, namely deflection- Time-dependent deformations result from creep and shrinkage.
dependent self-weights and wind loads (varying over the course of the Creep causes increased deformations due to continuous loads (self-
day due to sun tracking) as well as temperature constraints, and de­ weight). However, since parabolic trough collectors track the sun,
formations due to creep- and shrinkage effects using concrete. Tem­ different load situations occur during the day and – if symmetric –
perature constraints are of secondary importance as they almost solely mainly neutralize each other in their resulting deformations. Creep de­
result in residual stresses and cause negligible deformations [48]. formations can theoretically be fully neglected via smart control of the
Self-weights and wind loads have the most significant influence on collector deflection (in the sunless hours). Shrinkage deformations result
the design. The self-weight is determined by the concrete’s density. The from (chemical) drying of the concrete. For UHPC, a large share of these
wind loads mainly depend on the geometry of the shell and the wind deformations occur in the first hours or days, i.e. before further in­
flow. To determine an equivalent, quasi-static wind load on PTCs, wind stallations, e.g. of the absorber, take place. Uniformly, they only cause
tunnel tests were performed in the boundary layer wind tunnel at the volumetric compactions that can be easily compensated by a vertical
Ruhr University Bochum on a soloist model (Fig. 2, top) [48,49]. The adaption of the receiver tube. The same applies for shrinkage gradients
measured wind pressures were evaluated for the central axis by means of that arise from unequal drying conditions, e.g. exposed by different
an extreme value statistic according to Cook [50] and thus satisfy a formwork materials on the upper and lower shell surface. The basic
50-year return period. The geometric boundary conditions were shape of a parabola is still maintained. However, then noticeable

Fig. 2. Wind tunnel model (top) and wind load coefficients cp,net for different collector deflections (bottom).

4
P. Forman et al. Renewable and Sustainable Energy Reviews 134 (2020) 110331

Table 3 the center of gravity remains on a horizontal level and additionally is

Operational conditions and design criteria for wind speeds with corresponding always orthogonal to the support point during sun tracking. Doing so, no
restrictions. torsion due to self-weight occurs and – except from overcoming geo­
State Condition Wind Restrictions metric uncertainties and friction – no mechanical work has to be applied
speed to rotate the collector (Fig. 3, right). Therefore, additional circular
Operational The collector follows the sun vref = 10 Optical segments are located on the back of the shell. The lower segment at the
state under low wind speed m/s efficiency bearing structure has the shape of a shortened cycloid [56].
conditions. In this state, the solar θ ≤ θ1 , Generally, collectors are moved by a central drive, which implies
concentration must be at its Slope deviation
high torsional stiffness of the individual modules since torsional effects
maximum. SDx, rms
resulting from asymmetrical wind loading, bearing friction, and unin­
Transition The collector is driven in the vref = 15 1st principal
state survival state under upcoming m/s stress σI ≤ αc fct tentional torsion due to static unbalance cause axial rotations over the
moderate wind speed conditions. collector’s length. PTCs with high torsional stiffness due to a torque box
No more solar concentration is (or tube) are needed. In addition, the central drive must be able to bear
achieved, but the collector has to the cumulated loads. The individual modules are alternately mounted to
withstand the action effects for
the outsides as single-span girders and connected to form a collector
different deflections.
Survival state The collector is held in survival vref = 33 1st principal (Fig. 4, top).
position (φ = 90◦ , cf. Fig. 2) m/s stress σI ≤ αc fct Concrete shells with desired thicknesses of only a few cm have a
under strong wind speed. The comparatively low torsional stiffness such that either alternative
failure of the collector has to be
(decentralized) drive concepts or strengthening elements, such as outer
prevented.
bracings or hollow core structures, have to be developed. The main
problem of decentralized drives is the continuous synchronization of the
corrections of the vertical receiver position can occur [39,40]. individual modules in the collector assembly. Without synchronization,
the failure of one drive causes the breakdown of the collector. Fig. 4
3. Design basics for solar concrete shell collectors (center and bottom) displays concepts for decentralized drives, whereby
a rocker bearing according to Fig. 3 (right) is assumed such that the
The design of PTC concrete shells is a Pareto-optimal problem as both longitudinal static system is also adjusted to the changed conditions. For
a lightweight load-bearing structure (to reduce self-weight and material the drive of individual modules, a system as a single-span girder with
(costs)) and a low-deformation structure (to meet accuracy demands) cantilever arms on both sides appears to be the most suitable in order to
are required. Both demands contradict. Therefore, the design of para­ balance field lengths and thus to reduce deformations (center). For
bolic concrete shells always compromises both requirements. The decentralized drives that move two modules (bottom), a single-span
mathematical fulfillment of the maximum solar radiation concentration girder with one cantilever arm is the assumed best solution to reduce
additionally restricts the design. The collector dimensions (weight), the the field length and simultaneously transfer the occurring torsion loads
tracking, and the static system (as well as the resulting type of drive) are via the side bearing into the drive.
decisive for the design. For central drive concepts, conventional, already used linear drive
systems seem appropriate, since no additional loads are expected
3.1. Motions and static systems compared to steel framework collectors. When attached in the center of
gravity, the higher self-weight of concrete will not affect the drive.
The tracking of PTCs consisting of several modules is conventionally However, higher friction forces have to be overcome. Disadvantageous
carried out by rotation around the center of gravity so that – despite would here be shadings by reaching into the shell. For systems with
effects of friction and geometric uncertainties – no torsional load occurs rocker bearings, linear drive systems can be used, too, that can be
due to the self-weight. These loads would accumulate over the entire attached on the circular segments. Thereby, no shadings occur, but
collector length. Consequently, the bearing axis and the center of rota­ additional forces have to be absorbed due to the lever arm to the center
tion lie in line. In contrast with conventional steel framework collectors, of gravity. Key challenge for decentral concepts is the synchronization
in the case of concrete shells, the center of gravity is located within the between the drives. The synchronization has to be redundant to avoid
parabola. This tracking method would therefore require reaching into failure of the whole collector when individual drives malfunction.
the parabola to transfer loads, which would cause frequent shadings due Thereby, mechanical systems, e.g. by shafts, for synchronization are
to load-bearing components (Fig. 3, left). For the specific requirements preferred compared to wireless systems via digital radio link.
of concrete collectors with a comparatively high self-weight, an alter­
native bearing system was developed [55] that – in simplified terms –
corresponds to the rolling of two (almost) circular segments. Thereby,

Fig. 3. Principle of the center-of-gravity-oriented rotation (left) and of the rocker bearing with horizontal movement of the center of gravity (right).

5
P. Forman et al. Renewable and Sustainable Energy Reviews 134 (2020) 110331

Fig. 4. Schematic of a central drive system (top) and synchronized decentral drive systems for single modules (center) and pairs of modules (bottom).

3.2. Dimensions and cross-section design Statical systems of single span girders (S1) using central drive con­
cepts must transfer loads and especially torsional moments through
For concrete PTC, the self-weight becomes the crucial load with several modules. Thus, stiffeners (C2) for full-scale (A2) or even for
increasing aperture size following the trend of increased efficiency by small-scale modules (A1) and hollow core structures (C3) for large-
enlarging the aperture width and additionally the module length (e.g. scale modules (A3) are necessary here. For decentral drive con­
Ref. [21,57]). Fig. 5 (left) displays the apertures of selected parabolic cepts, which only drive single or pairs of modules, statical systems
trough collector modules, such as the EuroTrough, as well as a with lower effective span length are chosen (S2 and S3). Addition­
small-scale concrete prototype with a projected area of 7 m2 up to the ally, no loads have to be transferred along several modules. Thereby,
visionary concept of a large-aperture concrete collector (LACC) of 300 one-walled shells (C1) for small-scale (A1) and also for full-scale
m2. With increased apertures, however, the self-weight can lead to modules (A2) seem to be appropriate.
disproportionate deformations due to the wall thicknesses required for • Decentral drive concepts premise a motion by means of rocker
rigidity or solar concentration demands, meaning that single-walled bearings:
(solid) designs of PTC shells (Fig. 5, (1)) no longer make sense. In
order to increase the ratio of rigidity (defined by the moment of inertia I) Since decentral drive systems assume statical systems that possess at
to self-weight (defined by the cross-sectional area A), it is therefore least one indented support arrangement, a bearing in the center of
advisable to use tapered cross-sections similar to T-profiles (i.e., with gravity (M1) is not possible so that only rocker bearings (M2) are
external bracings (Fig. 5, (2))) or hollow structures (i.e., with internal used here.
voids (Fig. 5, (3))).
Some minor trends are:
3.3. Design options and their limits
• Single walled-shells (C1) might be appropriate for full-scale collec­
The concepts presented for tracking, aperture, bearing and statical tors (A2) with decentral drive systems (S2 and S3) due to lower ef­
system, and cross-section can usually be combined in many desired forts in formworks compared with the more rigid bracing-
ways, whereby some combinations are prioritized (++,+), others are strengthened (C2) or hollow-core structures (C3).
not recommended (–,-), and still others are incompatible, i. e. not • Single walled shells (C1) are not recommended for large-scale
possible (n.p.). Fig. 6 provides a rating matrix where the combinations of structures (A3) due to the low ratio of rigidity to self-weight.
the different statical systems (subscript S), motions (subscript M), • Hollow core structures (C3) are too expensive for small-scale col­
aperture areas (subscript A) and cross-sections (subscript C) are lectors (A1). The same applies for full-scale collectors (A2) besides
compared with each other. systems with central drive concepts (S2 and S3). In general, a hollow
Major trends can be identified as follows: core structure (C3) becomes only decisive if sufficient static height of
the cross-section is required to arrange the necessary void formers
• An increasing aperture area requires an increasing rigidity of the within the shell.
shell:
Further details can be obtained from Fig. 6. The implementations and
• This means that small-scale collector modules (A1) are preferably feasibility studies of selected systems and combinations are shown in the
designed as single-walled shells (C1), full-scale modules (A2) with following sections.
outer bracings (C2) and consequently large-scale modules (A3) as
hollow cores (C3). Nevertheless, some exceptions may occur. 4. Examples of the design and manufacturing of parabolic shells

• Central drive concepts require a higher (torsional) stiffness than The conceptual and optimization-based design of parabolic shells
decentralized drive concepts: was performed by means of parameterized Finite Element (FE) models in
the software environment ANSYS. The parabolic shells as well as the
stiffeners were modelled by 4-noded shell elements with three degrees of
translational and rotational freedom each. For large-scale collectors
with integrated hollow cores, the shell elements were divided into
several layers that considered the integrated hollow bodies over
smeared effective stiffness parameters [51]. The material parameters
used were determined experimentally (cf. Table 1). Due to the restric­
tion to a share of the axial tensile strength of the concrete, the numerical
investigations were based on a linear-elastic material model. The design
of the individual models was carried out according to the criteria in
Table 3.
Fig. 5. Left: Typical apertures of parabolic trough modules; Right: Possible The derived designs show primarily the feasibility of PTC made from
cross-section layouts (solid, with stiffeners, hollow core).

6
P. Forman et al. Renewable and Sustainable Energy Reviews 134 (2020) 110331

Fig. 6. Rating matrix for possible concrete collector designs depending on motion, static system, dimensions, and cross-section.

concrete. They compromise from real-scale applications and numeri­ direction (S3). The mean shell thickness is 2.5 cm, which tapers from a
cally derived design studies. The latter demonstrate specific, mainly center thickness of 3 cm to 2 cm at edge, and the focal length corre­
visionary concepts. Already built-up collectors or modules in real-scale sponds to 0.78 m. Fig. 7 (left) displays the collector module with an
are first prototypes or fully operational collectors in full-scale. Next to installed linear drive and absorber tube. In order to avoid tilting or
the constraints in Fig. 6, also demands from manufacturing have to be slipping, specially fitted steel gears were mounted on the rocker bearings
considered due to the individual production of these modules. Thus, and sickles.
simple geometries – and consequently a simple fabrication – might be To determine the optical efficiency, the demonstrator was measured
preferred rather than stiffer but more complex cross-sections. The pro­ in its noon position by means of digital close-range photogrammetry
duction of complex structures, however, is usually shown in experi­ using 200 target points. The deformations could thus be derived in
mental realizations, e.g. individual components or details. relation to the intended geometry. They vary in the range of approxi­
mately ± 5 mm and are characterized by a kind of “dishing” effect,
4.1. Small-scale collector shells which is typical of flat, thin-walled concrete components (Fig. 7, right).
Based on these deformations, an optical efficiency of 94% was deter­
To demonstrate the general feasibility of PTC shells, a small-scale mined [48]. The reduced efficiency can be traced back to the systematic
prototype with dimensions of 2.20 m by 3.20 m was developed as a deformation as well as to an increased waviness of the mirroring due to
single module [38,48]. This module illustrates the combination of sup­ manual application.
porting and reflector structure to a thin-walled shell (C1) with mirroring
by a manually applied PVD-coated aluminum sheet with a thickness of 4.2. Parabolic shells for full-scale collectors
0.5 mm. The sun tracking is achieved by means of rocker bearings (M2),
which are arranged indented from the edges so that the static system of a The design studies and implementations of PTC shells for full-scale
single-span girder with two cantilever arms results in the longitudinal collectors are oriented at the widely used EuroTrough module which

Fig. 7. Small-scale prototype (left) and photogrammetrically measured deformations along with target points (right); Photo: Thomas Brenner.

7
P. Forman et al. Renewable and Sustainable Energy Reviews 134 (2020) 110331

maintains an aperture width of 5.77 m, a module length of 12 m, and a 4.2.2. Structurally optimized ribbed shells for decentralized drive systems
focal length of 1.71 m. Substantial material savings could be achieved by means of bracings.
It therefore seems reasonable to also build up modules for decentralized
4.2.1. Central drive systems drive systems (S3) using stiffeners which now be realized with a bracing
Collector modules as single-span girders are recommended for cen­ pattern designed according to the flux of forces (C2). The bearings are
tral drive systems (S1). For this purpose, the individual modules are adapted from the small-scale prototype (M2). The design is achieved in
supported at the ends in the center of gravity (M1) which lies within the two steps. At first, the basic pattern of stiffeners is identified by means of
parabola cross-section. For bearing, this model was enhanced with outer topology optimization [59] taking into account the most relevant yet
bulkheads according to the principle of barrel rooves, which, however, simplified action effects. Then, this pattern is transferred into bracings
create moderate shading. Tensile ties in the quarter points were neces­ with a rectangular cross-section. The bracings bear the main loads and
sary to counteract buckling for single-walled (C1) or ribbed shells (C2). support the actual parabolic shell structure. They are dimensioned by
The conceptual design with exemplary first principle stress distribution means of shape optimization techniques, while the actual shell is
and visualization is shown in Fig. 8. designed as thin as possible. Ductile behavior (i.e., avoiding brittle
In order to reduce shading by the bulkheads, it is advisable to mirror failure) is ensured by reinforcements.
them so that they serve as secondary reflectors. Alternatively, the outer To generate structures featuring maximized rigidity while using only
bulkhead can be replaced by a spoked wheel-like steel strut structure to a reduced amount of initial material, topology optimization is recom­
generate a hybrid shell, which can significantly reduce shading [39]. mended. It is used to predict optimal material distributions within a pre-
For the transfer of torsional moments, stiffeners are necessary. These defined design space. For identifying a bracing pattern, the shell struc­
can be best designed as cross-bracings on the back of the shell so that the ture is discretized into finite elements (cf. Fig. 11). Using these elements,
axial moment to be transmitted is applied as a pair of forces at the the material distribution is controlled by means of a pseudo-density ρ
corners of the PTC shell (Fig. 9, left). Thereby, torsional loads are (per element i) which varies between 1 and 0 (i.e., material is required
transferred over the entire collector length. To ensure a torsional stiff­ or material is removed) with respect to given loads and boundary con­
ness of 2 kNm/mrad, the required moment of inertia of possible cross- ditions. Intermediate densities are possible but are avoided using the so
bracings can be derived. Based on an equivalent statical system of a called SIMP (Solid Isotropic Material with Penalization) method [60].
cantilever arm for the cross-bracings (Fig. 9), the geometry of The objective here is the minimization of the structure’s compliance c:
EuroTrough-like modules, and a maximum deformation due to 1 mrad
min c(ρi ) (5)
axial rotation, the moment of inertia results to about 2.5 × 10− 4 m4. This ρ
moment of inertia can be designed as a rectangular cross-section with a
The decisive action effects are self-weight, wind load, and torsional
width-to-height-ratio of around w/h = 15/27 [cm/cm]. However, the
load, which are considered as simplified load situations [39,61] for a
height of the bracings can be adapted to its utilization caused by the
qualitative identification of the bracing pattern. The boundary condi­
bending moment, i.e. they possess a maximum cross sectional area at the
tions are adapted from the small-scale prototype. The resulting material
middle of the shell which taper towards the edges. Thereby, the bracings
distributions with corresponding load cases are displayed in Fig. 11
are uniformly stressed due to torsion (Fig. 9, right).
(left). However, parabolic trough collectors are usually simultaneously
In a study on different bracing concepts for the model according to
exposed to several loads that also vary in the course of the day. For this
Fig. 8, all of which have a cross-bracing as primary stiffeners, resulting
purpose, the objective function is extended in terms of a multi-objective
deformations and slope deviations were compared (Fig. 10). The basis
optimization. Although the single load cases are initially considered
was a single-walled shell. The loads were set for a crucial load situation
separately, the sum of their resulting compliance values is minimized:
in operation state with a collector deflection of 45◦ (cf. Fig. 2). It should
be noted that all models underwent a shape-optimization process in ∑
n
min wk ck (ρi ) (6)
which the stiffeners and the shell were optimized in their cross-section ρ
k=1
and thicknesses [58]. The axial tensile strength was restrictive and the
actions were only applied for a single module. All models displayed full Thereby, a material distribution which is robust against all applied
optical efficiency, whereby the single-walled shell possesses the largest loads as well as against every single load case alone is generated. Here,
slope deviation. The volume was reduced by almost up to 30% for the the n = 3 load cases are equally weighted (w1 = w2 = w3). The resulting
model with a diamond-shaped bracing pattern, whereby every model distribution can be identified as a diamond-shaped pattern (Fig. 11,
with bracings lost at least about 20% of its mass. A dimensioning of the right). Material-specific properties of concrete and reinforcement steel
cross-bracing for the transmission of torsional moments has not been can also be controlled for separately in the design space using multi-
included but would have the same effect on every concept. However, the material- and/or hybrid approaches [62–65] yet they are not neces­
before dimensioned bracings for a torsional stiffness of 2 kNm/mrad can sary for an initial identification of possible stiffener’s pattern.
simply be added to the designs. After the basic pattern is known, the bracings are dimensioned.
These bracings possess a constant rectangular cross-section. For a
quantitative design, specific action effects according to Section 2.3 are

Fig. 8. Conceptual design and FE model of the trough-like shell (left) and visualization of the collector module (right).

8
P. Forman et al. Renewable and Sustainable Energy Reviews 134 (2020) 110331

Fig. 9. Cross-bracing concept for the transfer of torsional moments with equivalent statical system.

Fig. 10. Bracing concepts (left) with related deformations (center) and slope deviations (right).

used. In a shape optimization process, only the cross-sectional di­ reinforcement’s bar spacing. The module length of 12 m could thus be
mensions of the stiffeners (height h and width w) were sized, whereby maintained. By means of shape optimization inside ANSYS, the cross-
the objective is the minimization of the volume and the restrictions are sectional dimensions of the stiffeners results to h/w = 19/5 [cm/cm]
set to the maintenance of axial tensile stress and accuracy demands [61]. Additional reinforcement has to be added into the bracings based
according to Table 3. Since the stiffeners carry the main loads, the basic on basic design principles for reinforced concrete structures. Fig. 12
shell which supports the reflector elements is designed as thin as (left) displays a visualization of the developed design, which has an
possible. For a ductile behavior, the shell is reinforced with a conven­ average shell thickness of 2.3 cm. A feasibility study of a test segment
tional mesh Q188, which exhibits diameters of 6 mm and spacings of 15 (Fig. 12, right) that was produced in a reusable concrete formwork
mm, with minimal concrete covering. These reinforcements should be revealed that even complex geometries such as the structurally opti­
installed at a 45◦ angle of rotation so that the diamond-shaped bracing mized shell are possible. However, for an actual implementation of these
pattern orients to the reinforcements. However, this was not possible structures, reusable, high-precision steel formworks are recommended.
without adjusting the aperture width to 5.42 m due to the Although the steel formworks are expensive, they would amortize over

9
P. Forman et al. Renewable and Sustainable Energy Reviews 134 (2020) 110331

Fig. 11. Topology optimization results as pseudo-density distribution (red: ρi = 1, blue: ρi ≈ 0) for simplified load cases (left) and for a multi-objective optimization
with an identified bracing pattern (right) acc. to [61]. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of
this article.)

Fig. 12. Visualization of a full-scale prototype with diamond-shaped bracings and a concrete test specimen of crossing ribs with integrated reinforcement mesh.

the number of modules to be produced.

4.2.3. Solid shells with decentralized drive systems

In the ConSol (“Concrete Solar Collector”) project, a parabolic trough
collector was developed and erected as a single-walled shell (C1) in a
full-scale design [40,56,66]. The collector possesses a decentral drive
system with an output of only 180 W and drives two modules with an
individual weight of approximately 10 t. The shells are supported by
sickles on rocker bearings (M2), which are arranged on the sides of the
drive such that the drive forces can be transmitted via the sickles into the
shell (S2). A special detail of the collector is the concrete gears of the
rocker bearings and sickles, which prevent the shell from tilting or
slipping. These and other highly stressed components, such as the shells
and sickles, are made of high-performance concrete, while the sub­
structures and the supporting structure of the drive are made of (les­ Fig. 13. Exploded drawing of the collector made from two modules [40].
s-expensive) normal concrete C30/37. The entire collector consists of a
total of 25 precast concrete components that were connected on-site or the range of 10–15◦ /h. An angle sensor on one sickle measures the
in advance in the laboratory. Fig. 13 shows the collector as a digital twin current deflection and repositions the collector for deviations greater
– which was especially necessary for possible collision detection – in an than 0.05◦ [40].
exploded view. The collector was designed as a single-walled shell with an average
A specific drive was developed for this collector. It attaches at the thickness of 4.5 cm (6 cm in the center linearly decreasing to 3 cm at the
center of gravity of the shell and moves it on a horizontal lever (cf. Fig. 3, edge) due to the simple manufacturing of the superstructure for this
right) by means of several gear wheels and chains, which are supported prototype implementation. As mentioned, a structure with stiffeners
by their own concrete bearing structure. Further modules and their would be even more appropriate. The formwork of the shell consisted of
drives can be mechanically synchronized by means of a longitudinal a wooden structure with a thin steel sheet covering for the shell surface
shaft. Fig. 14 shows an exploded drawing of the drive as well as the to ensure high surface accuracy. The concreting was carried out with the
actual operating conditions. Tracking during the operating condition is parabolic opening downward. The cover formwork was attached by
carried out automatically with a high accuracy of <0.1◦ and a speed in means of clamping through the shell. The formwork for geometrically

10
P. Forman et al. Renewable and Sustainable Energy Reviews 134 (2020) 110331

costly steel formwork – but one that makes sense for series production –
would also improve the (optical) quality.
Benchmark for this design is the EuroTrough collector which is used
in the CSP plants Andasol 1–3 [20]. For a cost comparison, all plant
components except the concrete collector are considered identical.
Then, the costs – normalized to the aperture area – are derived and
summarized in Table 4 [66]. In total, they amount to 179 €/m2 what
exceeds the reference of the EuroTrough with 160 €/m2. Thereby, the
concrete components constitute the largest amount. However, addi­
tional material savings of approx. >50% can be achieved using struc­
tural optimization (cf. section 4.2.2) for the shell as well as for the
Fig. 14. Drive engine with bearing in an exploded drawing (left) and detail of substructure. Some components like receiver and REPA (Rotation and
the drive engine (right). Expansion Performing Assemblies) are standard components. Further
cost reductions can be achieved for the steel components, mirroring, and
more complicated components – such as gear wheels, sickles, and tracking. Most steel components have been only necessary for transport
bearings – was cut from polystyrene. Conventional steel reinforcements purposes and, therefore, not required in an in situ fabrication of a CSP
were used to ensure the ductile behavior. In the shell, it was applied in a plant. The tracking that includes the drive is a newly developed system
single layer due to its slenderness. The assembly of the collector had to and can also be optimized. Taking these parameters into account, a cost
be carried out with high accuracy as inaccuracies of individual elements reduction to less than 100 $/m2 (cf. [34]) seems possible.
could accumulate for the whole structure. Due to the limited aperture
area, an application to generate electricity would not have been suffi­ 4.3. Large-aperture concrete collector modules (LACC)
cient. The PTC shell’s surface was covered with a mirror material
developed in the project, which consists of an electrochemically pol­ To implement large-aperture collectors as a concrete structure, suf­
ished aluminum strip with a thickness of 0.4 mm with a silver-based ficient structural stiffness is required which is preferable reached by
multi-layer PVD coating and additional sol-gel protection. The spec­ hollow core structures (C3). The static system (S3) and the rocker
ular reflectivity amounts to 92.3% [66]. The assembled collector is bearings (M2) of the small-scale prototype provide the basis for an
shown in Fig. 15. upscaling. A focal length of approximately 3 m is chosen based on a rim
To determine the optical efficiency, the western PTC module (cf. angle of 80◦ and a desired aperture of 10 m × 30 m (A3). To ensure the
Fig. 15) was photogrammetrically measured before applying the mirror required stiffness (especially in the transverse direction), the cross-
sheets. The deformations compared to an ideal parabolic shape are section is shape optimized at first. The cross-section is adjusted to the
shown in Fig. 16 (top). The longitudinal edges of the parabola can be restrictive tensile strength of the concrete, resulting in a tapering shell
seen to have an almost constant deformation of about 30 mm, which is with a thickness of 0.50 m in the parabolic apex (Fig. 17, left). Then, in
significantly higher than the numerically determined deviation of less order to avoid the predominantly high self-weight without a super-
than 5 mm. Since this is an almost constant deformation, it seems to be a proportional loss of rigidity, the homogeneous shell structure is dis­
systemic error that is likely due to one-sided shrinkage. In addition, solved into an inhomogeneous one with hollow cores. The hollow cores
local, high-slope deviations occur at the clamping points, which were can be designed as internal longitudinal and transverse ribs or by means
necessary for formwork construction. These initial deformations, which of void formers as they are already used in structural engineering
are superimposed by the self-weight deformations, result in slope de­ (Fig. 17, right). Void formers are usually available as spheres or flat­
viations SDx in the range of approximately − 40 to 10 mrad (Fig. 16, tened ellipsoids as shown in a flat plate made of HPC in Fig. 17 (bottom).
bottom). From this finding, a SDx,rms value of 10.12 mrad results, The LACC was dimensioned for the internal hollow core structure by
which exceeds the EuroTrough-oriented derived limit of 2 mrad (cf. means of derived equivalent stiffnesses as superimposed membrane and
Section 0) and corresponds to an intercept factor of only 58.6%. With a bending stiffness components [51]. The resulting structures possess
theoretical adjustment of the focal length fromrm 1.71 m to 1.78 m, great rigidity which could be proven in numerical analyses and evalu­
which almost compensates for the shrinkage effect deformation, a solar ations of the surface accuracy both for approaches with spherical void
concentration of 85.6% can be achieved. formers and with internal longitudinal and transverse ribs. While the
The achieved accuracy remains below the expectations and numer­ manufacturing of LACC with void formers appears less extensive, a
ical calculations, but from the authors’ point of view, it is due to the structure with inner ribs has greater potential to save material and thus
manufacturing process and especially to the prototypical character of also self-weight. Further optimization possibilities are seen in the
the concrete collector, which is unique in its dimension and obtained arrangement of the ribs or void formers, which, as already shown in
slenderness. Further investigations should deal with the shrinkage ef­ Section 4.2.2, can be arranged according to the flux of forces. For a
fects, which can be reduced, for example, by heat treatment [37,67]. A practical realization further improvements and developments, e. g.
including drive system and installation parts as well as of structural
components, have to be performed. A prototypical application has not
yet taken place due to the large dimensions.

4.4. Comparison of designs and discussion

The developed concrete PTC concepts illustrate the design and

manufacturing possibilities for different combinations of aperture area,
drive concept, and cross-sectional design. Methodical dimensioning and
design methods were derived and practically realized. Essential
geometrical parameters and the weight of the developed modules in
absolute and relative values are provided in Table 5. As expected, the
weight of the superstructure (including outer sickles) increases with
increasing aperture due to the necessary structural rigidity. However,
Fig. 15. Assembled prototype collector. the mass that is normalized to the aperture area also rises. For the LACC

11
P. Forman et al. Renewable and Sustainable Energy Reviews 134 (2020) 110331

Fig. 16. Deformations (top) and slope deviation (bottom) of the western collector module.

31 kg/m2 [19] as a corresponding reference module. Aiming for cost

Table 4
reduction, the ConSol module can be replaced by shell modules with
Cost overview of the ConSol collector [66].
stiffeners offering up to 45% of material savings for the concrete shell
Component Costs in €/m2 only.
Foundations, sickles, rocker bearing and shell (concrete parts) 87 Nevertheless, an increased weight compared with steel PTCs does not
Mirror and adhesive tape 19 have to be regarded only as negative. First, concrete exhibits a basic
Steel components 16
price that is of one order of magnitude less than the one of structural
REPA 4
Tracking 23 steel. Second, costs can be saved in the realization of the foundations.
Receivers 30 This possibility bases on the dominant vertical load from self-weight
since horizontal forces due to wind now have a comparatively small
influence. While steel PTC modules with dominant horizontal loads have
to be founded on expensive pile foundations, a flat foundation is usually
sufficient for concrete PTC modules. The dimensioning of the foundation
depends on the acting normal force N, mainly from self-weight, and the
moment M resulting from horizontal wind. The ratio M/N between both
defines the eccentricity e that characterizes the offset of N from the
center of gravity. For flat foundations, e must lie within the first core
width (h/6) such that the foundation is entirely compressed and tilting is
prevented. Fig. 18 reveals the global forces acting on the foundation due
to wind and self-weight as well as the resulting loads for a center of
gravity bearing and a rocker bearing. However, for the rocker bearing,
Fig. 17. Design of a large-aperture concrete collector module (left), exemplary eccentricities can be even greater, up to h/2. Doing so, N would be
visualization of the inner structure with void formers or hollow cores (right), located in the center of one foundation and then only affects this one.
and flattened ellipsoid void formers in a plate (bottom). Since the contact point varies with the collector’s position, both, self-
weights and wind, cause offsets from the center.
module with void formers, a maximum of almost 500 kg/m2 is reached. The resulting eccentricities (Table 6) were determined for the full-
Only the structurally optimized shell with stiffeners offers high material scale collectors with self-weights according to Table 5 and wind loads
savings, which can be recognized by a relatively low mass of 80 kg/m2. derived from the pressure distributions in Fig. 2 for the transition and
It almost equals the effective self-weight of the small-scale prototype yet the survival state. For the ConSol collector, a maximum eccentricity of
still is more than 2 times heavier than the EuroTrough with a weight of 1.73 m is derived for a deflection of 180◦ . Thereby, a maximum moment

Table 5
Main characteristics of the presented concrete collector modules.
small-scale prototype shell modules with stiffeners ConSol module LACC modules

central drive decentral drive hollow core void formers

aperture width w [m] 2.205 5.77 5.42 5.77 10

module length l [m] 3.20 12 12 30
focal length f [m] 0.78 1.71 1.71 (1.78) 2.98
aperture area A [m2] 7 69 65 69 300
mean thickness of the shell [cm] 2.5 3.0 2.3 4.5 10.0 17.5
weight of the superstructure in total [t] 0.48 5.80 5.18 9.94 89.8 152
normalized to the aperture area [kg/m2] 68 84 80 144 277 484

12
P. Forman et al. Renewable and Sustainable Energy Reviews 134 (2020) 110331

width-to-height ratio of w/h = 15/27 [cm/cm] are sufficient for a

torsional resistance of 2 kNm/mrad.
• A small scale prototype shows the general feasibility and a full-scale
concrete PTC (ConSol collector) featuring a newly developed drive
system demonstrates the true-to-scale application and its novelty in
slenderness. However, for the ConSol collector, effects from
shrinkage and superimposed inaccuracies from manufacturing lead
to an insufficient accuracy with an intercept factor of 58.6% and
85.6% with a theoretically adjusted focal length, respectively. These
effects have to be further investigated and eliminated, e. g. by means
of heat treatment against shrinkage.
• Serial production of parabolic trough modules as precast elements
make sense for geometrically complex structures with optimization
based stiffeners and hollow cores. Thereby, expensive, yet frequently
reusable steel formworks amortize over the number of modules for a
solar field. Additionally, in an industrialized serial production
Fig. 18. Global horizontal (h) and vertical (v) forces on the collectors and
quality control mechanism can be integrated to guarantee required
resulting foundation loads (M and N) with corresponding eccentricities for
accuracies during the fabrication.
collectors with a center of gravity bearing (left) and a rocker bearing (right) for
an original scale-collector module.
• For concrete PTCs with dominating self-weights, foundations are
mainly loaded by vertical forces so that expensive pile foundations
(due to dominating horizontal wind loads for conventional PTCs) can
Table 6 be replaced by more cost-efficient flat or sleeve foundations.
Eccentricity e of the resulting load N on foundations for full-scale collector
modules. Declaration of competing interest
Deflection Eccentricity e [m] for transition/survival state
Θ [◦ ] The authors declare that they have no known competing financial
Ribbed shell modules ConSol module
interests or personal relationships that could have appeared to influence
Central drive Decentral drive the work reported in this paper.
0 1.43/- 0.76/- 1.11/-
45 1.20/- 0.09/- 0.28/- Acknowledgements
90 0.10/1.32 0.08/1.40 0.03/0.32
135 0.79/- 1.35/- 1.07/-
180 0.54/- 1.85/- 1.73/- The authors thank the German Research Foundation (DFG), Ger­
many for its financial support within the framework of the priority
program SPP 1542 “Concrete Light” as well as the German Federal
results from the superposition of backside wind load with a maximum Ministry for Economic Affairs and Energy (BMWi), Germany for its
lever arm for the self-weight due to the outermost contact point of the financial support of the project “ConSol” within the framework of the
shell. The same applies for the ribbed shell with decentral drive, where 6th Energy Research Programme on the basis of a decision by the
an eccentricity of 1.83 m becomes decisive. However, for both modules German Bundestag. Sincere thanks are given to the partners of the
assumed with sleeve foundations, the rocker bearing of the ConSol ConSol project, especially K. Hennecke and D. Krüger (German Aero­
collector is sufficient with an axial distance between the foundations of space Center), R. Dasbach (Almeco), M. Kintscher and A. Natterer
5.4 m. For a flat foundation of the ribbed shell with central drives, an (Pfeifer Seil- und Hebetechnik), J. Krüger and T. Zippler (Solarlite CSP
eccentricity of 1.43 m leads to required dimensions of approximately Technology) and T. Stanecker (Stanecker Betonfertigteilwerk) as well as
8.5 m. These dimensions might not be appropriate for a flat foundation. to the partners of the DFG funded projects, especially B. Sagmeister
However, when taking into account the self-weight of the foundation, (durcrete), T. Friedrich (Innogration) and R. Weissbach.
which has not been done yet and exhibits up to the weight of one
module, the eccentricities can be significantly reduced. References

5. Conclusions [1] Murdock HE, Gibb D, André T, Appavou F, Brown A, Epp B, Sawin JL. Renewables
2019 global status report. Paris, France: REN21; 2019.
[2] https://www.solarpaces.org/csp-technologies/csp-projects-around-the-world/.
Concrete shells as solar thermal collectors for parabolic trough [Accessed 22 January 2020].
power plants are a economic alternative to established steel framework [3] Pfahl A. Survey of heliostat concepts for cost reduction. J Sol Energy 2014;136(1):
PTCs. The main derivations and findings are: 014501-1–014501-9.
[4] Pfahl A, Coventry J, Röger M, Wolfertstetter F, Vásquez-Arango JF, Gross F,
Arjomandi M, Schwarzbözl P, Geiger M, Liedke P. Progress in heliostat
• Shells made of high-performance concrete bridge large span widths development. Sol Energy 2017;152:3–37.
with minimal material efforts and therefore are most suitable for [5] Coccia G, Di Nicola G, Sotte M. Design, manufacture, and test of a prototype for a
parabolic trough collector for industrial process heat. Renew Energy 2015;74:
PTCs. 727–36.
• The design of the shell’s cross-section differs with the aimed aperture [6] Zou B, Dong J, Yao Y, Jiang Y. An experimental investigation on a small-sized
area. Small-scale modules are preferred to be built as single walled parabolic trough solar collector for water heating in cold areas. Appl Energy 2016;
163:396–407.
shells, EuroTrough-like full-scale modules with bracings and large- [7] Tyagi VV, Panwar NL, Rahim NA, Kothari R. Review on solar air heating system
aperture modules with hollow cores ensuring sufficient structural with and without thermal energy storage system. Renew Sustain Energy Rev 2012;
stiffness with respect to accuracy demands. 16(4):2289–303.
[8] Tyagi VV, Panwar NL, Rahim NA, Kothari R. Review on solar air heating system
• Collectors with decentral drive concepts need lower stiffnesses than with and without thermal energy storage system. Renew Sustain Energy Rev 2012;
ones with central drives. For central drive concepts, cross-bracings 16:2289–303.
for full-scale collectors are recommended to ensure torsional stiff­ [9] Al-Alili A, Hwang Y, Radermacher R, Kubo I. A high efficiency solar air conditioner
using concentrating photovoltaic/thermal collectors. Apll Energy 2012;93:138–47.
ness along the collector length, whereby a cross-bracings with a
[10] Jebasingh VK, Joselin Herbert GM. A review of solar parabolic trough collector.
Renew Sustain Energy Rev 2016;54:1085–91.

13
P. Forman et al. Renewable and Sustainable Energy Reviews 134 (2020) 110331

[11] Hafez AZ, Attia AM, Eltwab HS, ElKousy AO, Afifi AA, AbdElhamid AG, et al. [40] Forman P, Penkert S, Krüger D, Mark P, Schnell J, Hennecke K, et al. A concrete
Design analysis of solar parabolic trough thermal collectors. Renew Sustain Energy solar collector - from design to assembly in full scale. VGB PowerTech 2018;9:
Rev 2018;82:1215–60. 42–8.
[12] Fuqiang W, Ziming C, Jianyu T, Yuan Y, Yong S, Linhua L. Progress in concentrated [41] Schmidt M, et al. Sachstandsbericht „Ultrahochfester Beton“, Deutscher Ausschuss
solar power technology with parabolic trough collector system: a comprehensive für Stahlbeton (DAfStb) 561. Berlin: Beuth Verlag; 2008.
review. Renew Sustain Energy Rev 2017;79:1314–28. [42] Telford T. Design of concrete structures. CEB-FIP-Model-Code; 1990. 1993.
[13] Abdulhamed AJ, Adam NM, Ab-Kadir Mza, Hairuddin AZ. Review of solar [43] Chopradup A. Zur Tragfähigkeit von punktförmig gestützten filigranen
parabolic-trough collector geometrical and thermal analyses, performance, and Fassadenplatten aus faserverstärkten Feinkornbeton. Germany: Technical
applications. Renew Sustain Energy Rev 2018;91:822–31. University of Kaiserslautern; 2010. Phd-Thesis.
[14] Sharaf OZ, Orhann MF. Concentrated photovoltaic thermal (CPVT) solar collector [44] Bendt P, Rabl A, Gaul HW, Reed KA. Optical analysis and optimization of line focus
systems: part II - implemented systems, performance assessment, and future solar collectors. Golden, Colorado, USA: Solar Energy Research Institute; 1979.
directions. Renew Sustain Energy Rev 2015;50. 1566–633. SERI/TR-34-092.
[15] Kasaeian A, Tabasi S, Ghaderian J, Yousefi H. A review on parabolic trough/ [45] Lüpfert E, Geyer M, Schiel W, Esteban A, Osuna R, Zarza E, et al. Eurotrough design
Fresnel based photovoltaic thermal systems. Renew Sustain Energy Rev 2018;91: issues and prototype testing at PSA. In: Proc. of Solar Forum 2001, Solar Energy:
193–204. the power to choose. ASME; 2001. p. 387–92. Washington, DC.
[16] Kalogirou SA. Solar thermal collectors and applications. Prog Energy Combust Sci [46] Pottler K, Lüpfert E, Johnston GH, Shortis MR. Photogrammetry: a powerful tool
2004;30:231–95. for geometric analysis of solar concentrators and their components. J Sol Energy
[17] Evangelisti L, De Lieto Vollaro R, Asdrubali F. Latest advances on solar thermal Eng 2005;127(1):94–101.
collectors: a comprehensive review. Renew Sustain Energy Rev 2019;114:109318. [47] Pottler K, Ulmer S, Lüpfert E, Landmann M, Röger M, Prahl C. Ensuring
[18] Weinrebe G, Balz M. Efficient steel structures for solar thermal power plants performance by geometric quality control and specification for parabolic trough
(German: effizienter Stahlbau für solarthermische Kraftwerke). Stahlbau 2019;88 solar fields. Energy Procedia 2014;49:2170–9.
(6):529–36. [48] Forman P, Müller S, Ahrens MA, Schnell J, Mark P, Höffer R, et al. Light concrete
[19] Lüpfert E, Geyer M, Schiel W, Esteban A, Osuna R, Zarza E, et al. EuroTrough shells for parabolic trough collectors – conceptual design, prototype and proof of
design issues and prototype testing at PSA. In: Campbell-Howe R, editor. Proc. of accuracy. Sol Energy 2015;111:364–77.
the ASME international solar energy conference – forum 2001, solar energy: the [49] Oya J, Kalender-Wevers C, Winkelmann U, Höffer R. Experimental and numerical
power to choose. ASME; 2001. Washington, D.C., USA. investigation of the dust transport on the reflector panels of a parabolic trough
[20] Millennium Solar. The parabolic through power plants Andasol 1 to 3. Self- power plant. In: Proc. of the European-African Conf. on Wind Engineering EACWE
Published Solar Millennium; 2008. 2013; 2013. Cambridge, UK.
[21] Riffelmann K-J, Richert T, Nava P, Schweitzer A. Ultimate Trough – a significant [50] Cook NJ. The designers guide to wind loading of building structures – part 2: static
step towards cost competitive CSP. Energy Procedia 2014;49:1831–9. structures. London: Oxford; 1990. Butterworths.
[22] Marcotte P, Manning K. Development of an advanced large-aperture parabolic [51] Kämper C, Forman P, Stallmann T, Ahrens MA, Mark P, Schnell J. Optimised high-
trough collector. Energy Procedia 2013;49:145–54. performance concrete shells for parabolic trough collectors. J IASS 2017;58(1):
[23] Geyer M, Lüpfert E, Osuna R, Esteban A, Schiel W, Schweitzer A, et al. EuroTrough 105–19.
– parabolic Trough collector developed for cost efficient solar power generation. [52] Hosoya N, Peterka JA, Gee RC, Kearney D. Wind tunnel tests of parabolic trough
In: Steinfeld A, editor. Proc. of the 11th SolarPACES international symposium on solar collectors. In: Subcontract report NREL/SR-550-32282. National Renewable
concentrated solar power and chemical energy technologies. Villigen, Switzerland: Energy Laboratory; 2008.
Paul Scherrer Institut; 2002. Zurich, Switzerland. [53] Winkelmann U, Kämper C, Höffer R, Forman P, Ahrens MA, et al. Wind actions on
[24] Riffelmann K-J, Kötter J, Nava P, Meuser F, Weinrebe G, Schiel W, et al. large-aperture parabolic trough solar collectors: wind tunnel tests and structural
HelioTrough – a new collector generation for parabolic trough power plants. In: analysis. Renew Energy 2020;146:2390–407.
Proc. Of the 15th SolarPaces conference; 2009. Berlin, Germany. [54] Gianuzzi GM, Majorana CE, Miliozzi A, Salomoni VA, Nicolini D. Structural design
[25] Feldhoff JF, Eickhoff M, Keller L, Alonso JL, Meyer-Grünefeldt M, Valenzuela L, criteria for steel components of parabolic-trough solar concentrators. J Sol Energy
et al. Status and first results of the DUKE project – component qualification of new Eng 2007;129:382–90.
receivers and collectors. Energy Procedia 2014;49:1766–76. [55] Weissbach R. Die abrollende Parabolspiegelrinne. Schutzrecht DE102011011805
[26] Good P, Ambrosetti G, Pedretti A, Steinfeld A. A 1.2 MWth solar parabolic trough A1. Industrial property right; 2012.
system based on air as heat transfer fluid at 500 C — engineering design, [56] Penkert S, Forman P, Mark P, Schnell J. Konzeptionierung und Errichtung eines
modelling, construction, and testing. Sol Energy 2016;139:398–411. https://doi. originalmaßstäblichen Parabolrinnenkollektors aus Hochleistungsbeton. Beton-
org/10.1016/j.solener.2016.10.002. Stahlbetonbau 2019;114(11):806–16. https://doi.org/10.1002/best.201900045.
[27] Pedretti A. A 3 MW thermal concentrated solar power pilot plant in Morocco with [57] Farr A, Gee R. The SkyTrough parabolic trough solar collector. In: ASME 3rd
the Airlight Energy Technology. In: Proc. Of the 18th SolarPaces conference; 2012. international conference on energy sustainability, es2009-90090; 2009. San
Marrakesh, Morocco. Francisco.
[28] Fernández-García A, Zarza E, Valenzuela L, Pérez M. Parabolic-trough solar [58] Forman P. Optimierungsstrategien für Parabolschalen zur
collectors and their applications. Renew Sustain Energy Rev 2010;14:1695–721. Solarstrahlenkonzentration aus Hochleistungsbeton. Ruhr University Bochum;
[29] Schiel W. Kollektorentwicklung für solare Parabolrinnenkraftwerke. Bautechnik 2016. Phd-thesis.
2012;89(3):182–91. [59] Bendsoe MP, Sigmund O. Topology optimization. Berlin, Heidelber: Springer-
[30] Bellos E, Tzivanidis C. Alternative designs of parabolic trough solar collectors. Prog Verlag; 2004.
Energy Combust Sci 2019;71:81–117. [60] Rozvany GIN. Aims, scope, methods, history and unified terminology of computer-
[31] Kabir E, Kumar P, Kumar S, Acelodun AA, Kim K-H. Solar energy: potential and aided topology optimization in structural mechanics. Struct Multidiscip Optim
future prospects. Renew Sustain Energy Rev 2018;82:894–900. 2001;21:90–108.
[32] Lilliestam J, Pitz-Paal R. Concentrating solar power for less than USD 0.07 per [61] Forman P, Stallmann T, Mark P, Schnell J. Multi-level optimisation of parabolic
kWh: finally the breakthrough? Renew Energy Focus 2018;26:17–21. shells with stiffeners made from high-performance concrete. In: Hordijk DA,
[33] Sharma C, Sharma AK, Mullick SC, Kandpal TC. Cost reduction potential of Lukovic M, editors. Proc. of the 2017 fib symposium. Maastricht, Netherlands:
parabolic trough based concentrating solar power plants in India. Energy Sustain High Tech Concrete: Where Technology and Engineering Meet; 2017. p. 2503–11.
Dev 2018;42:121–8. [62] Smarslik M, Ahrens MA, Mark P. Toward holistic tension- or compression-biased
[34] Schuknecht N, McDaniel J, Filas H. Achievement of the $100/m2 parabolic trough. structural designs using topology optimization. Eng Struct 2019;199(109632):
AIP Conference Proceedings 2018;2033:030019-1–030019-11. 1–13. https://doi.org/10.1016/j.engstruct.2019.109632.
[35] Schmidt M, Fehling E. Ultra-high-performance concrete: Research, development [63] Gaganelis G, Jantos DR, Mark P, Junker P. Tension/compression anisotropy
and application in europe. In: Proc. of the 7th int. Symposium on utilization of enhanced topology design. Struct Multidiscip Optim 2019;59(6):2227–55. https://
high-strength/high-performance concrete. Washington D.C., USA, SP-228: ACI doi.org/10.1007/s00158-018-02189-0.
Spacial Publication; 2005. p. 51–78. [64] Smarslik M, Mark P. Hybrid reinforcement design of longitudinal joints for
[36] Sagmeister B, Pahl G. Protective layers made of UHPC. Bauingenieur 2019;94: segmental concrete linings. Struct Concr 2019;20(6):1926–40. https://doi.org/
179–83 [In German]. 10.1002/suco.201900081.
[37] Sagmeister B. Maschinenteile aus zementgebundenem Beton. Berlin, Wien, Zurich: [65] Gaganelis G, Mark P. Downsizing weight while upsizing efficiency: an
Beuth Verlag; 2017. experimental approach to develop optimized ultra-light UHPC hybrid beams.
[38] Müller S, Forman P, Schnell J, Mark P. Leichte Schalen aus hochfestem Beton als Struct Concr 2019;20(6):1883–95. https://doi.org/10.1002/suco.201900215.
Parabolrinnen solarthermischer Kraftwerke. Beton- Stahlbetonbau 2013;108(11): [66] Krüger D, Penkert S, Schnell J, Janotte N, Forman P, Mark P, et al. Development of
752–62. a concrete parabolic trough collector. AIP Conference Proceedings 2019;2126:1–7.
[39] Forman P, Kämper C, Stallmann T, Schnell J, Mark P. Parabolschalen aus https://doi.org/10.1063/1.5117626. 120008.
Hochleistungsbeton als Solarkollektoren. Beton- Stahlbetonbau 2016;111(12): [67] Tkocz J, Mark P. Heat treatment and shrinkage steering of precast thin walled
851–61. concrete tubes. In: Derkowski W, et al., editors. Concrete – innovations in
materials, design and structures, fib Symposium 2019. Krakow; 2019, ISBN 978-2-
940643-00-4. p. 2246–53.

14