A schematic diagram is shown in Figure 2, depicting the main components of the ISAHP test rig. Table 1 lists the instrumentation and monitoring equipment used in the experiment, whose part numbers correspond to the schematic diagram.

The main components of the apparatus include: a nominal 1/3 HP single speed compressor, a thermostatic expansion valve, two flat plate counter-flow heat exchangers acting as the evaporator and condenser of the heat pump, a standard residential hot water tank (270 L), a variable speed pump and an auxiliary heater for simulating the solar collector heat input. R-134a was used as the working fluid for the heat pump cycle, and a 50 / 50% glycol/water solution by volume was used in the collector loop.

The collector loop operates in a similar fashion to that of a typical solar domestic water heating system. First, the glycol solution is pumped through the collector, or in this case the auxiliary heater, in a closed loop. The glycol solution absorbs energy through the “collector”, and a heat exchanger is used to extract the heat from the glycol, which acts as the evaporator of the heat pump loop. The R-134a superheated gas exits the evaporator and passes through the compressor, increasing in both pressure and temperature. The refrigerant then releases its heat and condenses through the natural convection heat exchanger, which delivers hot water to the storage tank. The R-134a liquid then passes through the thermostatic expansion valve, reducing its pressure before re-entering the evaporator.

2. Experimental Procedure

To compare the dynamic operation of the ISAHP system with the steady-state computer model [4], a 6 hour test with varying glycol temperature was performed. Due to stratification, the temperature at the bottom of the tank remained constant throughout the test providing a constant input temperature for the condenser on the water side. A description of the experiments performed is given below.

2.1. Test Sequence

For the simulated solar day test, the heat out of the auxiliary heater was varied in a sinusoidal profile similar to that occurring on a clear sunny day. The following procedure was followed for the test:

• Prior to testing the storage tank was filled with water at mains temperature to ensure that the entire tank was at constant temperature.

• During this time the collector loop fluid was brought to the desired initial temperature for the test, and the loop flow rate was set to 77 kg/h (0.021 kg/s). This flow rate was used in the simulations previously undertaken, and is based on recommendations by Fanney and Klein [14].

• The data acquisition (DA) system was initialized with the compressor running, and the program delivering the power profile to the heaters was commenced. Data was recorded every 5 seconds for the duration of the experiment.

• COP and natural convection flow rate were then calculated based on the collected data.

Deyou LIU*, Su GUO, Dingsheng WAN, Feng WANG1

‘Department of Power Engineering, Hohai University, Nanjing 210098, Jiangsu Province, China
* Corresponding Author, liudyhhuc@163.com

Abstract

In this paper, based on the geometrical optics theory, the function of heliostat field boundary line was deduced, and the simplified formula to calculate the receiver depression angle with the only independent variable of latitude was given. Moreover, the calculated heliostat field data were compared for different receiver depression angle by a case study, and the rationality of the formula was demonstrated. From the results, the following conclusions were drawn: For the conditions of the same heliostat dimension and arrangement, it is optimal when the receiver depression angle is at around the value calculated by the formula of receiver depression angle developed in this paper. Now, the installed capacity of plant is mainly dependent on the height of tower and the area of receiver aperture, and almost has nothing to do with the receiver depression angle in a certain range around its optimum value.

Keywords: solar energy, solar power towers, field boundary line, receiver depression angle

1. Introduction

In Solar Power Tower plants, the heliostat field layout is one of the most important parts in the plant design, including many optimization design works, such as the layout of the heliostats, determination of field boundary line and heliostat dimension, etc., as reported earlier [1-3]. Field boundary line refers to the envelope line of mirror centre coordinates of all heliostats in the field, and the heliostat at any location in this boundary line can reflect all rays into the receiver aperture. Determination of field boundary line is not only the basis of field layout design, but also the foundation of selection of plant location and design of installed capacity.

There are a lot of correlative factors effecting the heliostat field boundary line, for instance, the longitude and latitude of the plant location, the installation elavation of receiver, depression angle of the plane of receiver aperture (Receiver Depression Angle for short), shape and dimension of receiver aperture, the installation elavation and the dimension of heliostat mirror, tracking error and rocking error of heliostats, and so on, which has been reported earlier [4,5]. Under the circumstance of ignoring tracking error and rocking error of heliostats, this paper established numerical calculation method of field boundary line and simplified calculation formula of receiver depression angle with the only independent variable ‘latitude’ aiming at receivers with rectangle aperture. Finally, taking a certain plant project as the calculation case, for the condition of the configuration parameters of receiver and heliostats were determined, the influences of different receiver depression angle on the heliostat field boundary line were compared and analyzed, and rationality and validity of the developed formula were certified.

Physical model of the relationship between receiver and heliostats is shown in Fig.1. In Fig.1, x axis is towards due east, y axis towards due north (Northern Hemisphere) or due south (Southern

Hemisphere) and z axis towards vertically sky, which has been reported earlier [1]. Oi (xi, yi ,0) is the mirror center coordinates of Heliostat i. We assume that the mirrors of heliostats is rectangle, with the height Ht and the width Hw. E(0,0, Ht) is the center coordinates of receiver aperture,

and the vertical distance between point E and point Oi is Ht. aR is the receiver depression angle, that is the angle between the normal of receiver aperture plane and horizontal plane.

The condenser and the evaporator are modeled by means of the є-NTU method for counter-flow heat exchangers. The heat exchanger is subdivided into 3 zones, each of them being characterized by a heat exchange coefficient U. Figure 4 shows the 3-zones heat exchanger in the case of an evaporator.

The AU heat transfer coefficient is calculated by considering two resistances in series (secondary fluid and refrigerant sides). The nominal heat transfer coefficients are identified for a nominal flow rate.

The coefficients are then obtained by: h = h ■ M 0,8

n n

This relationship can be justified by the Reynold’s analogy for a turbulent flow through a pipe [2] by assuming that the fluid properties, not included in this expression, remain unchanged.

Pump model

The pump is modeled by its swept volume, its isentropic efficiency (nPP, s) and its motor efficiency (nPP, m)- The two latter are supposed to be constant. The pump electrical consumption and the fluid flow rate are given by :

sh, pp

Л pp, m Л pp, s.^ pp, m

3. Experimental study

Figure 2 and 3 show that the amount of emitted CO varies significantly for the different systems. There are several reasons for these differences. Each pellet heater has its own CO-characteristic for

the operation with a particular combustion power and during start and stop (Figure 1). The pellet heaters vary also in their nominal power from 6 kW to 20 kW which of course together with the size of the buffer volume, the boiler control mode and the way the heat is transferred to the building influence the number of starts and stops.

The boiler systems have large start/stop emissions whereas the start/stop emissions for the stove systems are very low. This is due to the much lower total number of starts and stops of the stoves using the complete building as heat storage. This is a simplification which provides that the heat can be freely distributed to the building. A more advanced multi zone building model for simulations of stove systems has been used by Persson. Perssons studies showed that stove systems have similar number of starts and stops as boiler systems if the convective and radiative part of the heat from the stoves can not be freely distributed to the complete building (see table 3.2 in [9]).

In Figure 3 the specific CO-emissions for each system are compared with the CO-limit values for two Eco-labels. All systems are below the rather high limit value of the Standard EN 303-5 of 1314 mg/MJ which is not indicated in this figure. However, 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 at nominal combustion power. Testing institutes usually use a mixture of the measured CO-emissions at nominal and minimal load to specify CO-emissions of the tested pellet heater. The comparison shows that this leads to a drastic underestimation of the real annual CO­emissions due to the fact that the start/stop emissions are not included. It is therefore suggested to revise the methods to determine the CO-emissions for pellet stoves and boilers in the current norms

nd eco-labels and to include an estimation of total annual emissions based on the operation of the boiler and the average load.

Combining solar and pellet heating systems can reduce significant CO-emissions compared to operating a single pellet heating system. This combination prevents the summer operation of the pellet heater with low efficiency and high emissions. Simulations for one system have shown that the CO-emissions can be reduced by almost the half compared to a single pellet heating system using the same boiler.

5. Acknowledgement

We are grateful to the Nordic Energy Research and the Dalarna University College for their financial support for this work within the REBUS project.

6. References

[1] BAFA, "Richtlinien zur Forderung von Mahnahmen zur Nutzung erneuerbarer Energien vom 12. Januar 2007." Bundesamt fur Wirtschaft und Ausfuhrkontrolle — Bundesministerium fur Wirtschaft und Technologie [BMWi]. 2007.

[2] C. Bales, "Reports On Solar Combisystems Modelled in Task 26 [System Description, Modelling, Sensitivity, Optimisation], Appendix 6: Generic System #11: Space Heating Store With DHW Load Side Heat Exchanger[S] And External Auxiliary Boiler." IEA-SHC Task 26 Solar Combisystems, Paris, France. 2003.

[3] F. Fiedler, "Combined solar and pellet heating systems — Study of energy use and CO-emissions," PhD thesis, Malardalen University, Vasteras. 2006.

[4] U. Jordan, and K. Vajen, "Influence of the DHW profile on the Fractional Energy Savings — A Case Study of a Solar Combisystem." Solar Energy, 33-42. 2002.

[5] S. A. Klein et al., "TRNSYS 16.0 Transient Simulation Program." SEL, University of Winsconsin, Madison, WI, USA. 2005.

[6] Nordic-Ecolabelling, "About Swan-labelled Boilers for solid biofuel, Version 2.0, Background for ecolabelling", Nordic Ecolabelling, Stockholm. 2006.

[7] Nordic-Ecolabelling, "Swan labelling of solid biofuel boilers Version 1.5." Nordic Ecolabelling. 2006.

[8] S. Nordlander, T. Persson, F. Fiedler, M. Ronnelid, and C. Bales, "Computer modelling of wood pellet stoves and boilers connected to solar heating systems." Pellets 2006, Jonkoping, Sweden.

[9] T. Persson, "Combined solar and pellet heating systems for single-family houses — How to achieve decreased electricity usage, increased system efficiency and increased solar gains," Doctoral Thesis, KTH — Royal Institute of Technology, Stockholm, Sweden. 2006.

[10] T. Persson, F. Fiedler, and S. Nordlander, "Methodology for identifying parameters for the TRNSYS model Type 210 — wood pellet stoves and boilers." Solar Energy Research Center, Hogskolan Dalarna, Borlange, Sweden. 2006.

[11] T. Persson, F. Fiedler, M. Ronnelid, and C. Bales, "Increasing efficiency and decreasing CO-emissions for a combined solar and wood pellet heating system for single-family houses." Pellets 2006 Conference, Jonkoping, Sweden.

[12] RAL, "Der Blaue Engel, Grundlage fur Umweltzeichenvergabe, Holzpelletheizkessel RAL-UZ 112." RAL Deutsches Institut fur Gutesicherung und Kennzeichnung e. V., St. Augustin. 2003.

[13] SP, "SPs Certifieringsregler for P-markning av Pelletsbrennare och Pelletspannor, SPCR 028." Sveriges Provinings — och Forskningsinstitut, SP. 1999.

[14] Thur, S. Furbo, and L. J. Shah, "Energy savings for solar heating systems." Solar Energy, 80[11], 1463­1474. 2006.

The AGMD system consists of three commercial rectangular modules (i. e., Scarab development AB, Sweden) with twenty membranes each (2.3 m2 membrane per module).The membranes are fixed to polypropylene cassettes, each cassette consists of plastic frames containing two parallel membranes, feed channels, and two condensing plates, by stacking the cassettes the cold channels are created between the condensing plates of nearby cassettes. The membrane modules are then made up of nine feed channels and nine permeate channels, and the size of the module is 63 cm wide and 73 cm high, with a stack thickness of 17.5 cm. A view of the modules in the pilot plant is shown in fig.4. The membrane material is 100% expanded PTFE spun-bonded with polypropylene support (GoreTM MicroFiltration Media) with a porosity of 80% and a thickness of 0.2 mm. The width of the air gap is 2 mm. These modules have been improved to produce ultra pure water suitable for the semiconductor industry. The modules do not have internal heat recovery that is why the next step in MEDESOL project to be carried out is to evaluate different configurations combining the outlets and inlets of the three modules, to maximize heat recovery.

The SOLITEM PTC 1800 collector will be used in the pilot installation. This parabolic trough collector system is assembled from modules with an aperture width of 1,8 m and a length 5 m each. The concentrator consists of aluminium sheet kept in shape by specially manufactured aluminium profiles. ALANOD Mirosun reflector sheets are glued onto this structure. Additional stiffness is provided by a torsion tube mounted at the back of the concentrator. Up to 4 modules installed in series form a row, 6 rows can be connected to a single drive unit to be tracked via a rope and pulley arrangement. A sun position calculation gives the input for an approximate tracking of the collectors while sun sensors installed in several locations of the collectors provide data for a precise positioning (fine tuning). Mounted in the focal line is a stainless steel absorber tube with 40 mm diameter, which has been galvanically coated with a selective surface. A non-evacuated glass envelope reduces the convective heat losses from the absorber. Flexible hoses with thermal insulation and a metal fabric envelope connect the moving absorber tubes with the stationary field piping.

To obtain reliable performance data for design and performance calculations, a single row has been installed at the DLR test facilities in Cologne for a measurement campaign in summer 2008 (Figure 1). For the design Design considerations

Referring to the N-S positioning, the horizontal axis of parabolic trough (X in Fig. 1) should be oriented in the North-South direction, which corresponds to the axis of Earth rotation. The realisation of a precise alignment is quite complex; moreover the correct positioning should be maintained for long periods with the lower possible misalignment errors.

Errors in N-S axis positioning often occur during the installation of solar plants; while they rarely happen after some times of system working. The solar trough axis, instead of being perfectly aligned to the terrestrial axis, can form a few degrees angle with the N-S direction. The main consequences of this misalignment are collected energy losses, which are studied for a one-axis tracking collector with parabolic profile. In these studies, the reference parameter for the N-S misalignment is represented by the angle p between the horizontal axis of solar trough (X in Fig. 1) and the N-S direction. By a practical point of view it could be useful to visualise the N-S
misalignment angles p considering the associated displacements of trough extreme, presented in Table 1 for a solar trough of length 5 m.

The initial study examines the effect of N-S misalignment for three sun’s altitudes a, pertaining to the following months: a=23° in December, a=60° in August and a=69° in June. The effects of N-S axis misalignment are expressed using the collection efficiency, which is a crucial quantity to be considered in the application of sunlight exploitation. It is obtained as ratio between the light focused on the absorber and the light captured by the entrance aperture of collector.

The results are plotted in Fig. 9 as collection efficiency versus angle of N-S misalignment. The losses of collection efficiency are caused by the N-S misalignment combined with the effect of the different sun’s altitudes, corresponding to the different months. Obviously the maximum effect of collection losses is achieved in December; while the minimum losses are obtained in June, when a reaches its maximum value. For the data in Fig. 9 the absorber is a metal pipe surrounded by a glass tube and the layout parameters are: f=800mm, d=50mm, D=60mm,

T=3mm.

The successive study refer to different configurations for the solar trough, considering an angle of N-S axis misalignment P=15°.

The results in Figures 10-12 refer to a simplified absorber, consisting in a linear plate of width w and length 5 m. The examined layouts combine different focal lengths f of the parabolic mirror with two values of absorber width w, as follows:

1) f=500mm, w=50mm;

2) f=600mm, w=50mm;

3) f=700mm, w=50mm;

4) f=800mm, w=50mm;

5) f w 100mm. Fjg. 10. Absorbed light for N-S misalignment angle of 15°.

The results are summarised

in Fig. 10, plotting light received by the absorber versus suns altitude. the light received by the

absorber in Figures 10-12 is expressed in percentage with respect to the ideal case of sun’s altitude a=90°.

Finally, considering only two configurations of solar trough, the analysis focuses on the various errors in positioning the horizontal axis of collector. The considered North-South misalignment angles P are 5°, 10°, 15°. The layouts taken in to account are characterized by the following parameters: f=800mm, w=50mm and f=1000mm, w=100mm. Figures 11-12 show the results for the examined solar troughs. The curves show a significant dependence on sun’s altitude and the solar trough with f=1000mm and w=100mm experiences lower energetic losses.

The exceptional artificial sun (fig. 6) of the SIJ is an indoor placed sun simulator which can simulate the sun’s position according to the altitude angle and the azimuth angle in different seasons of the year and to any location on earth. This high-precision sun simulator imitates the sun’s path with an accuracy of < 0.5mrad and creates a beam of parallel light with divergence < 4mrad.

The test facility is composed of a circular ring on which the unit can be rotated 360° degrees, and a light source which directs its light to a parabolic disk, which then provide a parallel beam of light. This light unit can be moved along a curved bracket to change its elevation. The artificial sun illuminates targets up to a diameter of 500mm. It is steered by two computer-operated electro motors witch can be

moved stepwise or continuously in slow or high speed. The exact position for the movement control is detected by separate supplementary position encoders on each axis.

An analysis was conducted on the operation of an ISAHP with a varying temperature input. The mains water temperature, and therefore initial storage tank temperature was approximately 20°C. Figure 3 shows the power delivered to the glycol through the heaters, as well as; heat transferred to the refrigerant through the evaporator; heat rejected to the natural convection loop through the condenser; and power consumed by the compressor.

As shown in Figure 3, the heat transferred from the glycol through the evaporator exceeds the heat delivered to the glycol through the auxiliary heaters. This is due to a large amount of glycol in the auxiliary heaters which are initially at a higher temperature. The heat pump unit extracts the extra energy until the heat transferred through the evaporator matches the heat input to the glycol through the heaters. Finally, the heat transfer through the evaporator is observed to lag the heater input, again due to the volume of glycol in the auxiliary heaters.

A temperature probe installed within the storage tank, with thermocouples spaced 0.15 m apart, was used to determine the level of stratification of the tank. Figure 4 shows the temperatures measured in the tank throughout the test, with T11 being the temperature measured at the top of the tank, and T20 measured at the bottom. During the test the tank was observed to stratify well, and throughout the test the temperature at the bottom of the tank, T20, remains constant at approximately 21°C. Therefore, over the course of the test, the temperature of the water entering the natural convection heat exchanger/condenser was effectively constant at 21°C.

The natural convection flow rate was also calculated over the length of the test. Figure 5 shows the natural convection flow rate of water through the condenser, plotted with the temperature of glycol delivered to the evaporator on the same graph. Consistent with the previous constant temperature tests, it was observed that higher glycol temperatures delivered to the evaporator increased the flow rate, and an increase in tank temperature decrease the flow rate. This resulted in the natural convection flow rate remaining constant for the first half of the test, and then decreasing as the glycol temperature declined, and tank temperature continued to increase.

In order to compare the dynamic operation of the heat pump with the steady-state model, the measured compressor power and heat pump COP were plotted with the simulated results on the same figure. To assist in validating the computer model, the simulation was run under the same conditions as the experiment. The model predicted the power consumed by the compressor, the

evaporator and condenser heat transfer rates, and the COP of the system. Figure 6 shows the heat transfer rates through the condenser and evaporator and Figure 7 shows the COP and compressor power curves throughout the test. Both simulated and experimental values are displayed.

COP values measured throughout the test ranged from 2.4 to 3.2, and the measured compressor power ranged from 484 to 635 W. The amount of heat transferred through the condenser to the natural convection loop ranged from 1300 to 2000 kW.

Similar to the previous tests [13] run at constant glycol temperatures, the simulation program was found to overestimate the COP and heat transferred through both heat exchangers, but predicted the compressor power consumption reasonably well. Both heat transfer rates, and the COP, were over-predicted by the simulation by approximately 12%. These results were previously attributed to an over estimation of the heat exchanger effectiveness values. Neglecting this offset, the simulation, based on a steady-state governing equations, predicts the trend of the heat pump system well.

Solar rays reflected by heliostats at any location in the heliostat field should totally go into the receiver aperture. Therefore, the field boundary line should be determined by the condition that the facula reflected to receiver aperture by heliostat should be in the range of aperture.

Assume that, the receiver aperture plane is notated as Plane a, and the plane which is perpendicular to OiE and point E(0,0,Ht) is the foot of a perpendicular is notated as Plane bt. From the solar

ray divergence angle of 32′, which has been reported earlier [6], the facula on Plane bt reflected

Function of facula on Plane bi reflected by Heliostat i is deduced by formula (2) and (3),

■B|-VB.2 -4AC <y<-B, + ,/B,2 -4A, C,

2 A

2 A

z = Xx + УУ + H

And 0.009308 I 2 2 ,, 2 0.009308 / 2 2 ,, 2 xє[———— 2— Vx, + У + Ht, +—2— + У + Ht ]. Put the facula on Plane bt into some area elements dsj ( (j=l,2,3,…is the elements’ number). Assume that dsj is relatively small enough, and its coordinates is expressed as Sj (x]bi, y]bi, z]bi), so that:

(6)

If the shape of receiver aperture is rectangle, it can be assumed that the height and width of receiver aperture are Lr and Wr respectively. When element ds1ai is in the range of receiver aperture, its coordinates need to satisfy the following condition.

-1W < x}. <1W 2 r a 2 r L. , L. -ysinaR < ya, <уsinaR

(7)

When all elements ds]ai are in the range of receiver aperture, Heliostat i is in the range of heliostat field, otherwise Heliostat i is out of in the range of heliostat field. According by this, calculating and distinguishing all locations in a large area around the tower, field boundary line can be educed.