The advantage of one-axis tracking concentrators is to require tracking the sun only in one direction, because in the other direction these collectors do not perform sunlight concentration. The axis of solar trough can either be placed parallel to the North-South direction of Earth rotation axis, or parallel to the East-West direction of Earth rotation plane.

If the solar trough axis is parallel to the North-South direction, the tracking system must follow the sun in its daily excursion and the sun’s altitude over the horizon depends on Latitude and day of the year. With this layout the solar rays impinge on the collector with an inclination dictated by Latitude and hour of the day.

If the solar trough axis is parallel to the East-West direction, the sun should not be tracked in its daily excursion. However the tracking system must follow the displacements in sun’s altitude occurring every day of the year. In this layout only at noontime the sun’s rays are perpendicular to the entrance aperture of solar trough.

The results of the studies presented in this paper are referred to sun’s altitude over the horizon, so they can be used in both placements of solar trough axis. Considering the North-South positioning, the sun’s altitude represents the daily excursion of the sun, which corresponds to the tracking path that the trough collector must follow. Whilst for the East-West placement, the sun’s altitude indicates the various positions of the sun corresponding to each day of the year, for the chosen Latitude.

J. Gottsche1, B. Hoffschmidt1, S. Schmitz1, M. Sauerborn1*, Ch. Rebholz2, D. Ifland2, K.

Badstubner2, R. Buck3 and E. Teufel3

1 Solar-Institute Julich (SIJ), Aachen University of Applied Sciences (AcUAS), D-52428 Julich, Germany

2 Fraunhofer Institute for Reliability and Microintegration (Fh-IZM), D-82234 Oberpfaffenhofen, Germany 3 German Aerospace Centre (DLR), Institute of Technical Thermodynamics, D-70569 Stuttgart, Germany

Corresponding Author, sauerborn@sii. fh-aachen. de

Abstract

To reduce the construction costs of solar tower power plants an innovative new mirror system with mini mirrors is on the way as an uncommon alternative to the currently used heliostats with huge surface (actual up to 120m2). An about (0.6 x 0.6) m miniature test unit has now been con­structed. This first demonstration unit at first consists of a boxed array of 25 small parallel ar­ranged mirrors with a size of (10 x 10) cm each. All mirrors are linked by an agile, crossed stick stage with an integrated drive system to move them simultaneously into all necessary positions. This specimen, developed in cooperation with all partners, will be tested in the specially pre­pared artificial sun of the SIJ in different ways under changing conditions.

In a simulation an optimum field layout was calculated. With this data a theoretical comparison was made between a typical existing huge heliostat and the developed mini-mirror array unit with respect to the efficiency and cost-performance relations for a solar tower power plant. Keywords: mini mirror array, solar thermal tower power plant, heliostat field layout, raytracing

1. Introduction

The heliostat field of a solar tower power plant is with up to 50% a significant part of the investment cost. During the last 30 years of developing heliostats, it has been found that the key to essential de­crease of expenses lies mostly in innovative design modifications which strongly reduce the material input. The main cost driver of all low cost versions is still the steel structure, securing the mirror and the movement system that has to guarantee high accuracies even under high wind loads and extreme thermal stress situations.

The typical cost figure of heliostats depends on the stiffness requirements of the steel and is currently in the area of 130-150€/m2 (Solucar 120 m2). Because of the foreseeable increasing of the raw materi­als price the cost problem will in future gain even more relevance. Figure 1 shows a rough estimate of the maximal cost reduction potential. To simplify the calculation, the cost of the heliostat construction is reduced to the weight of the steel used (steel price actual > 2€/kg) and ignores the production ex­pense for further treatment (pillar, framework), higher techniques (actuator, controller, sensors) and the mirrors. When a wind load must be taken into account, the decrease of the expense is limited to max. 46% of the current mirror price.

In this way, only a heliostat design with a slight or even without sensitivity to wind load will offer the chance to scale down the mechanical layout to reduce the material input substantially. The easiest way to guarantee very small wind loads on the mirror system is to install the system in a box with a plane surface design. To solve this restriction and to keep the heliostat at a suitable size, small mirrors (such as (10 x 10) cm) could be used. The principle design of a Mini-Mirror Array proposed in this work is shown in figure 2.

example: Solucar (Spain)

An additional effect of the reduction of weight and the decrease of the associated friction losses will be the option to use a less robust actuator system for the mirror movement. So in the end the material in­put and the energy requirement of the system can be reduced effectively.

The current study is an extension to the previous experimental study, however, the objective of the current study is to determine how the heat pump unit responds to a wide range of input temperatures. As well, a goal of the current study is to determine if the steady-state model previously developed in the first study can accurately predict the dynamic operation of the system. A varying power input to the heaters was applied ranging from 750 — 1500 W in a sinusoidal fashion, similar to that of a daily solar heat input. The fluid temperatures delivered to the evaporator varied between 15 and 37oC corresponding to the power output of the heaters. The results of the current study are compared to the simulated results of Freeman’s model, and will be used to eventually refine the model to better predict the actual operation of an ISAHP system.

Tmopt program was used to couple the Tmsys simulations software with GenOpt, a generic optimisation program. Trnopt acts as an interface programme between the two software programs and streamlines the optimization process.

The optimisation process has as objective function the maximisation of the production of desalted water. The result is the optimal operating temperature of the fluid at the outlet of the solar field system coupled to the Rankine Cycle. This temperature has been optimised considering two different time periods: hourly or daily. The optimal daily and hourly temperature conditions are determined for both cycles considering two typical days corresponding to a typical situation in winter and in summer. The results for both cases are shown in Table 2 and Table 3.

As an example of the hourly results, Fig. 4 shows the results of the water system and working at a set point using the hourly and daily optimal temperatures in the selected summer day.

(b)

Fig. 4. Results for the optimal operating temperature of the fluid of the solar system. (a) daily (b)

hourly

2. Conclusion

A model for the optimal integration of Rankine cycles and solar thermal plants to drive Reverse Osmosis desalination plants has been presented. This model combines rigorous models for the simulation of the Rankine and solar field subsystems. The objective is to calculate the optimal operation temperature to produce the highest amount of desalted water. An example using trough solar collectors and water and pentane as working fluids was presented.

Using the trough solar collector and the two selected fluids the results in terms of energy efficiency and production of desalted water are very similar operating the system at an optimal hourly temperature at the collector’s outlet or at an optimal daily temperature. However, significant differences where found between the optimal temperatures in winter and summer days. So, this would mean that the set point for the solar field outlet temperature should be changed throughout the year to obtain the best global system performance but this change will have little effect throughout the same day. The performance using water is similar to that of n-pentane due to the high efficiency of the selected trough collector at high temperatures and the high solar radiation available in the selected geographical location. In the future the developed model will be extended to study also other types of solar collectors and working fluids.

Acknowledgements

This work is financially supported by the Ministerio de Educacion y Ciencia of Spain, OSMOSOL project, ref. ENE2005-08381-C03-03.

References

[1] IDA Desalination, Yearbook 2007-2008, Water Desalination Report, Global Water Intelligence, UK.

[2] L. Garcia-Rodriguez, Seawater desalination driven by renewable energies: a review, Desalination, 143

(2002) , 103-113.

[3] L. Garcia-Rodriguez, Renewable energy applications in desalination: state of the art, Solar Energy, 75

(2003) , 381-393.

[4] S. A. Kalogirou, Seawater desalination using renewable energy sources, Progress in Energy and Combustion Science, 31 (2005), 242-281.

[5] E. Mathioulakis, V. Belessiotis, E. Delyannis, Desalination by using alternative energy: Review and state — of-the-art, Desalination, 203 (2007), 346-365.

[6] J. McHarg, R. Truby West Coast researchers seek to demonstrate SWRO affordability. Desalination & Water Reuse, 14 (2004) 10-18.

[7] Osmosol — Desalacion por osmosis inversa mediante energia solar termica, Memoria de proyecto, Proyectos de investigacion, Ministerio de Educacion y Ciencia, 2006. https://www. psa. es/webeng/projects/joomla/osmosol/

[8] S. Canada, G. Cohen, R. Cable, D. Brosseau, H. Price, Parabolic trough organic Rankine cycle solar power plant, NREL/CP-550-37077, Presented at the 2004 DOE Solar Energy Technologies, Denver (USA), 2004.

[9] G. Burgess, K. Lovegrove, Solar thermal powered desalination: membrane versus distillation technologies. Proceedings of the 43rd Conference of the Australia and New Zealand Solar Energy Society, Dunedin, November 2005.

[10] D. Manolakos, G. Papadakis, E. Sh. Mohamed, S. Kyritsis, K. Bouzianas, Design of an autonomous low — temperature solar Rankine cycle system for reverse osmosis desalination, Desalination 183 (2005), 73-80.

[11] D. Manolakos, G. Papadakis, S. Kyritsis, K. Bouzianas, Experimental evaluation of an autonomous low — temperature solar Rankine cycle system for reverse osmosis desalination, Desalination 203 (2007), 366-374.

[12] Delgado-Torres, A. M., Diseno Preliminar de un Sistema de Desalacion por Osmosis Inversa mediante Energia Solar Termica, PhD thesis, Universidad de La Laguna (Tenerife, Spain), 2006.

[13] J. C. Bruno, J. Lopez-Villada, E. Letelier, S. Romera, A. Coronas, Modelling and Optimisation of Solar Organic Rankine Cycle Engines for Reverse Osmosis Desalination, Applied Thermal Engineering (2008), doi:10.1016/j. applthermaleng.2007.12.022.

[14] TRNSYS 16 — A transient system simulation program, version 16, 2004.

[15] S. A. Klein, Engineering Equation Solver (EES), F-Chart Software, http://www. fchart. com

[16] ROSA — Reverse osmosis system analysis software, ver. 6.1.3, Dow Water Solutions, 2006.

[17] A. C. McMahan, (2006). Design and Optimization of Organic Rankine Cycle Solar-thermal Power Plants, Master of Science, Solar Energy Laboratory, University of Wisconsin-Madison (USA).

[18] A. M. Patnode, (2006). Simulation and Performance Evaluation of Parabolic Trough Solar Power Plants, Master of Science, Solar Energy Laboratory, University of Wisconsin-Madison (USA).

[19] J. Lopez-Villada, J. C. Bruno, E. Letelier, S. Romera, A. Coronas, Simulacion con Trnsys de Sistemas Solares Termicos para Desalinizacion mediante Osmosis Inversa, XIV Congreso Iberico y IX Congreso Iberoamericano de Energia Solar, Libro de actas 647-652, Vigo, 2008.

The scroll expander model has been previously proposed by Lemort et al. [5] and partly validated by tests with steam. In this model, the evolution of the fluid through the expander is decomposed into the following steps (as shown in Fig. 3):

• Supply pressure drop (su^ su,1,1)

• Cooling-down in the supply port of the expander (su1,1 ^ su,1);

• Isentropic expansion from the supply pressure down to the adapted pressure imposed by the internal expansion volume ratio of the expander (su,1 ^ ad);

• Expansion at a fixed volume from the adapted pressure to the exhaust pressure (ad ^ ex,2);

• Mixing between suction flow and leakage flow (ex,2 ^ ex,1) and

•

Cooling-down or heating-up in the exhaust port (ex,1 ^ ex).

The model requires only nine parameters (heat transfer coefficients, friction torque, leakage area, pressure drop equivalent diameter). Those nine parameters, defined for a specific type of expander and for a specific working fluid, are determined on the basis of experimental data.

The modelling of the systems in TRNSYS is based on the system models and boundary conditions used in IEA-SHC Task 26 Solar Combisystems [2], which includes both the building (single node in type 56) and heat distribution using a radiator and PID controller modelling the thermostatic valve. For the system with air cooled pellet stove, no radiator was used. The boundary conditions for the systems are defined by the climate, in this study Stockholm, the domestic hot water (DHW)

load and the space heating demand. The DHW load has been modelled with a load profile developed by Jordan et al. [4] assuming a daily hot water demand of about 200 litre (~3100 kWh/year). The space heating demand is modelled by an one zone building model developed for IEA-SHC task 26 giving a yearly heat demand of approximately 12200 kWh (87 kWh/m2) for Stockholm.

Modelling of pellet stoves, burner and boiler were implemented with TRNSYS-component type 210 [8]. This dynamic model can be used to simulate pellet stoves, pellet burners and pellet boilers and gives flue gas losses during operation and in standby mode (leakage losses), as well as heat supplied to water in a mantle and to the surroundings. The model also calculates the CO-content in the flue gas, including the emissions during the start and stop phases. The parameter values used in this study were derived from parameter identification using measured data from the stoves/boilers, and have been verified against measured data [3; 10]. The parameter values for each of the pellet heaters used to simulate the CO-emissions of the pellet heaters can be seen in Figure 1. The model calculates the CO-emissions as the sum of a power dependent part during normal operation and a lumped constant amount per start and stop.

2 4 6 8 10 12 14 16 18 20

Combustion power [kW]

Fig. 1. CO-emissions during operation (graph) and start/stop (table) of the six pellet heating units.

Two variants of operating strategy were chosen for simulations of each system. On/off control using the full power of the heaters and modulation control was used with the measured modulation range for the specific heaters simulated in the systems. For comparison, system 5 has also been simulated with only the boiler or stove as main heat source and without solar heating system (solar collector loop and combistore).

3. Results

Figure 2 shows the CO-emissions for the six systems in kg divided in start/stop emissions, emissions during operation and standby emissions. The latter occur only for the boiler in system 4 which has an option to operate in a standby mode when there is no heat demand. Keeping the

boiler in this standby mode (by constantly combusting a little amount of pellet) increases the CO­emissions dramatically. The assumption here is that the start emissions are the same as if the boiler would not kept in standby. This has not been investigated in detail and the standby operation has not been included in the system simulations. Instead, the standby emissions in Figure 2 have been determined by separate calculations based on measurement of the boiler during standby operation.

From Figure 2 it can bee seen that the amount of emitted CO varies significantly for the different systems. The boiler systems have large start/stop emissions whereas the start/stop emissions for the stove systems are very low. The pellet stove in system 2 emits with 7 kg in on/off mode the lowest amount of CO per year whereas the boiler in system 5 emits 37 kg CO per year if on/off operated. The stove systems (system 1 and 2) emit most CO during operation whereas the combisystems (system 3-6) emit most CO during start and stop when on/off operated. For system 3, 4, 5 and 6 the start/stop emissions decreases drastically if controlled with modulating power. The CO-emissions of system 2 are much higher when operated with modulating power. The CO-emissions for system 1 are almost the same regardless if the stove is operated with on/off or modulating combustion power.

The pellet consumption is not the same for all systems. For a qualitative CO-emission comparison of the different systems it is therefore necessary to express the CO-emissions in a specific form, in kg per MJ pellet (Figure 3).

Together with the specific emissions of each system the limit values for CO from two eco-labels are indicated. The relative high limit value of the Standard EN 303-5 of 1314 mg/MJ is not indicated. It can be seen that only system 2, if on/off controlled, would fulfil the recently proposed limit values for the Svan-mark if the start and stop emissions and realistic conditions are taken into account. None of the stoves and boilers would fulfil the requirements for the Blauer Engel-mark. The dashed area shows the emissions of the stoves and boilers from lab measurements at constant nominal combustion power. These are much lower than the average annual emissions except for the stove in system 1 that has very little start and stop emissions. Note that for system 4 only the emissions for start/stop and normal operation are included but not the emissions for standby. These emissions have been excluded because no measurement data for the pellet consumption during standby were available.

In Figure 4 the annual CO-emissions of the pellet boiler used in system 5, with and without solar heating system, and the CO-emission of system 6 (with a solar heating system) are compared. It can be seen that the CO-emissions of system 5 can be reduced by almost the half by adding a solar system. This is mainly due to the reduction of the number of starts and stops from 3352 (on/off controlled) and 1601 (modulating power) to 1758 (on/off controlled) and 675 (modulating power). For system 6, that uses an Austrian pellet boiler with relatively low start/stop emissions, the annual CO-emissions would be only a third of the boiler used in system 5.

The system has two independent loops, the solar loop that supplies the thermal energy to the desalination unit and operates with osmosed water in order to protect solar collectors from corrosion and scaling and the desalination loop, which is in turn divided in two circuits, the cold and the hot

respectively. The solar loop, installed and designed for a former project (AQUASOL Project: Development of an advanced hybrid solar-gas multi-effect distillation system), is composed of a 500 m2 stationary CPC solar collector field and a 24 m3 thermal storage system based on water. Only half of the solar field is used in the case of MEDESOL.

The desalination loop consists of two separated 2 m3 polypropylene tanks (PP-H) used as hot and cold water reservoirs. Feed solution, prepared with deionised water and marine salt crystals, at both concentrations of 1 and 35 g/l is pumped, depending on the experiment, into the three AGMD modules connected in series or into one of them (in the case of “real sweater experiments” we only used one of the modules) by means of a centrifugal pump. Before entering the modules, feed is heated by way of the solar loop which is connected to the pilot plant trough a 10 plates titanium heat exchanger (HRS spiratube) coated with an advanced non-fouling layer to protect it from seawater aggressiveness. The maximum hot temperature achieved during the experiments was 90-95 °С. Likewise, deionised water is used as refrigerating fluid and can be cooled down if necessary, passing it through an air cooler. After AGMD process, both cold and hot water are returned to their corresponding tanks thus closing both circuits. The inlet and outlet temperatures of each of the modules are monitored using thermocouples type E. Electromagnetic flow meters are used to measure the feed and cooling stream mass flow rates. Conductivity probes are installed in both hot feed circuit and distillate line. Finally, pressure of both hot and cold channels is also measured at the entrance of the AGMD modules avoiding values higher than 0.2 bar, which is the maximum LEP[4] allowed to prevent from membrane wetting.

The industrial steam system at the ALANOD is based on a gas fired steam boiler, a condensate tank and a feed water tank with chemical conditioning. For an average production year, the heat energy of 940 000 m3 gas is used to produce 2000 kg of steam per hour. An ongoing extension of the production facilities with a new aluminium anodizing line will not only increase the steam demand to 4000 kg per hour, but also open opportunities to consider different options for the integration of the solar system. The steam is distributed to the production facility by two steam lines, with 4 bara / 143 °C and 9 bara / 175 °C, respectively.

The 4 bara steam line feeds the heat exchangers of the anodizing bathes (degreasing, brightening, sealing) as well as a hot water storage tank which is used to periodically replace the water content of the sealing bath. The dryers for the aluminium strip drying in the anodizing process are fed from the 9 bara steam line. From the convective and evaporation losses of the different baths and storages their continuous heat demand was estimated (Table 1). In addition, heating of the total storage content after the periodic discharge requires about 8 hours with a thermal power of 390 kW for storage tank 1 and 195 kW for storage tank 2.

Ground space for the solar field is not available at the site. Roof areas of production halls or office buildings could be considered but would require interference with the static of the existing

structures. Therefore, in the context of the present extension, a dedicated platform for the solar field will be erected adjacent to the main production hall.

The sun’s altitude over the horizon depends on Latitude and it varies during the year and during the day. Figure 3 gives an example of sun’s altitude considering the Latitude of Firenze (Italy), which is 43.75° North. The altitude of sun over the horizon is reported versus solar time (in hours), from 6AM to 6PM, considering four representative months. Each month curve is obtained averaging the values corresponding to all the days composing the month.

In the North-South positioning, the purpose of the tracking system is to compensate the hourly altitude variations, keeping the symmetry axis of parabolic mirror in the direction of sun’s rays.

The vertical symmetry axis is indicated as Z axis in Fig. 1. Figure 2 illustrate the perfect alignment between rays’ direction and symmetry axis in the ideal case of vertical rays’ direction (sun’s altitude = 90°).

In the East-West placement, since the sun tracking acts in the direction perpendicular to the previous one, it cannot compensate the inclination angle corresponding to each hourly altitude of sun over the horizon.

Figures 4-5 illustrate the consequences of the different sun’s altitudes (a), reporting the two extreme cases, of December and June, for the Latitude of Firenze. The maximum value of solar rays inclination, with respect to the vertical direction (a=90°), is achieved in December; the
minimum value in reached in June. For Firenze, the incidence angle of sun’s rays varies from 67° (for a=23° in December in Fig. 4) to 21° (for a=69° in June in Fig. 5).

Furthermore, considering the North-South (N-S) alignment,

Figures 4-5 visualise the consequences of monthly altitudes combined to the effect of collector axis misalignment with respect to the N-S direction. This N-S axis misalignment will be discussed in Section 2 introducing a parameterisation based on the angle P, between the horizontal axis of the collector (X in Fig. 1) and the N-S terrestrial axis. With reference to this parameter p, in Figures 4-5 the N-S axis misalignment is P=1° and the absorber does not rotate with Fig. 3. Altitude of sun over the horizon for Firenze (Italy).

the parabolic mirror. The

effect of misaligning only the mirror, keeping the linear absorber in the N-S direction, is evident at the extremes of the solar trough, especially in Figure 5. Most of sun’s rays are missed and they are plotted in the simulation as cut lines, indicating that they are not received by the absorber. In Figure 4 the very high inclination of solar rays becomes the major cause of rays’ loss.

Fig. 5. Ray incidence in June (max. a=69°).
Mirror N-S misalignment P=1°.

In order to complete this visual analysis of North-South axis misalignment, Figures 6-7 present a view of the absorber. In Fig. 6 the absorber is kept aligned in the N-S direction and the mirror axis is rotated of 1° with respect to the N-S direction. Whereas Fig. 7 refers to the case of a rigid rotation of both parabolic mirror and absorber, considering an angle P=0.5°.

In the first case, the result is an inclination of the image focused over the absorber surface. In the second case, the image is parallel to the absorber, but it results transversally and longitudinally shifted. In both cases the rays are missed at the absorber extremes, as noted in Figures 4-5. The angle in Fig. 7 is only p=0.5° because for p=1° the effect of the rigid rotation of absorber and mirror would give an image completely out of the absorber. The figures show only the most
significant portion of the absorber, whose centre is considered located in the parabola focus. The simulations have been carried out considering the worse case for the sun’s altitude, in December, with a=23°.

This preliminary analysis, simulating different sun’s altitudes, suggested to separately study the two causes of energetic losses.

First of all some light is not

received by the entrance aperture of the collector; then part of it is not correctly focused on the absorber, which cannot entirely receive it. This aspect has been studied also introducing an additional lateral mirror to recuperate part of the light that otherwise is not received by the trough aperture. This lateral mirror is obviously vertically placed at the extreme of the solar trough, on the side opposite to the source (right side in Figures 4-5). The configuration considered in this study is characterised by the following parameters: f=800mm, d=50mm, D=60mm, T=3mm.

The light received by entrance aperture and absorber, with and without lateral mirror, is plotted in Fig. 8. The values of received light are expressed in percentage with respect to the incoming sunlight. As previously noted, we prefer to take as reference parameter the sun’s altitude, so the results can be applied to the N-S positioning, as well as, to the East-West placement.

Nevertheless the studies presented in this paper examine only North-South misalignments of collector axis.

In a still running project such an alternative heliostat design was developed in cooperation between SIJ, IZM and DLR. The resulted design (fig. 3) of a longer development time consists, instead of the typical huge construction with large mirrors, of an array of numerous small parallel arranged mirrors.

The array has a filigree movement system with a cost-effective design and it is driven by small actua­tors that are adjustable by computer program on a linked pc. This mirror array is imbedded in a box with a transparency cover and a plane design to guarantee low wind loads.

Fig. 3. (Left side) Sketch of the first realised test version of the MMA with the parallel mirror array affiliated with each other

by an agile, crossed stick stage for movement.

Fig. 4.: (Right side): A model of the mechanical system that interlinks and moves the mini-mirrors.

The first built test device (see figure 4, and 5) of the newly developed MMA has a reduced number of only 25 small mirrors (around (10 x 10)cm) to keep the heliostat at a suitable size. The mirrors are made of light, thin glass (thickness 2mm) and placed in an adequately large box of (60 x 60)cm to en­sure a free movement in all directions. Future MMAs could be around (2 x 2)m with several hundred mirrors. The box is covered with a highly transparent solar glass generally for photovoltaic modules and the glass has a special coating reducing reflection losses of the sunlight.