The main product design specifications related to embodiment design are:

• Being a mobile system, it must assure safe movements and statically determined positions;

• Considering the main purpose of the installation, which is the experimental high accuracy use, an increased stiffness of the structure must be accomplished, in order to protect the most sensitive devices equipping the tracker;

• The material will be mainly construction steel; only some specific parts will be machined from alloy steel with proper treatment;

• The tracker is equipped with a PV panel with a surface 1.48×0.67 m2, approx. 1 m2;

• The tracker’s angles у and a must cover the ranges у = -90°.. .+90° and a = 9°.. .67°, set for the specific location of the system;

• Linear actuators must be used for both movements;

• The temperature field on which the system must work is -5 … 50°.

E. Yazdanshenas1*, S. Furbo2 and C. Bales3

1’ 2 Department of Civil Engineering, Technical University of Denmark
Building 118, DK-2800 Kgs. Lyngby, Denmark

3 Dalarna University, SERC, S-781 88 Borlange, Sweden

Corresponding author, eya@byg. dtu. dk

Abstract

Theoretical investigations have shown that solar combisystems based on bikini tanks for low energy houses perform better than solar domestic hot water systems based on mantle tanks. Tank-in-tank solar combisystems are also attractive from a thermal performance point of view. In this paper, theoretical comparisons between solar combisystems based on bikini tanks and tank-in-tank solar combisystems are presented.

The investigations are carried out for different designs and sizes of the two solar heating system types installed in different houses. The investigations show which types of solar combisystems are suitable for low energy houses, new houses built according to the building codes and old houses.

Keywords: Solar combisystems, bikini tank, tank-in-tank

1. Background

The solar heating market in most European countries and worldwide grows by 20-40% each year. In most European countries the percentage part of solar heating systems, which are solar combisystems is growing. Solar combisystems can cover both a part of the space heating demand and a part of the domestic hot water consumption. One of the most important studies on solar combisystems was done in the framework of Task 26 of IEA Solar Heating and Cooling programme between 1999 and 2002. Around 21 different solar combisystems were investigated numerically in detail in different IEA member countries. A design hand book for solar combisystems was published [1]. In order to gain high energy savings for solar combisystems, it is important to have:

• A small auxiliary volume in the heat storage.

• A low auxiliary set point temperature of the auxiliary volume in the heat storage.

• A low tank heat loss

• A high efficiency of the auxiliary heater

• A good thermal stratification in the heat storage tank

H. Drnck and E. Hahne [2] investigated four different solar combisystems in detail. They found that the most important parameters for a well performing combistore are low heat losses due to a good thermal insulation. They also suggested using small auxiliary volume and low set point temperature of the auxiliary boiler. Moreover, the connections for the auxiliary and the space heating loop should be in appropriate position.

E. Andersen and S. Furbo [3] investigated theoretically three solar combisystem designs in three different houses with different space heating systems. The solar combisystems are initially equipped
with heat exchanger spirals and direct inlets to the tank. A step-by-step investigation was performed demonstrating the influence on the thermal performance of using inlet stratification pipes at the different inlets. Based on TRNSYS simulations, they found increased thermal performance of solar heating systems by using stratifiers instead of internal heat exchanger and direct inlet to the tank. The best performing solar combisystem is based on a tank-in-tank storage with stratifiers both in the solar collector loop and the space heating loop.

E. Yazdanshenas and S. Furbo [4] investigated theoretically a new so called bikini solar combisystem. Three different houses with four different radiator systems were considered. The thermal performance of the bikini solar combisystem was compared with the thermal performance of a solar domestic hot water system based on a mantle tank. The thermal performance of the solar combisystem is higher than the thermal performance of a solar domestic hot water system based on a mantle tank. The investigation also showed that a bikini solar combisystem is promising for low energy houses.

The current paper deals with the comparison of two different solar combisystems: Tank-in-tank solar combisystems and solar combisystems based on bikini tanks. The aim of the paper is to study which of these two solar combisystem designs is suitable for different houses.

In the solar farm, PV-modules (“modules” below) are exposed by rows where modules are installed side by side in parallel. Usually there are several rows, one behind the other. Evidence of how the hindered row is shaded by the first one has been analyzed in the literature [2, 3]. We focused on the co-operation and mutual shading of two-positionally tracked neighbor modules in a

row. Most of the modules perform inside the farm and are surrounded by neighbors from both sides. Performance of such a module defined as an “inner” module was analyzed first. A module at the end of the row, which has a neighbor on one side only, is defined an “outer” module. Peculiarities of their performance will be described later.

In the theoretical analysis, a simplification has been made: the PV-module (or its column on the roof) is considered being infinitely long. It means that we will ignore edge effects on the top of the inner column. At these limitations the gain has its minimal value as the edge effects increase the illumination. Figure 1 shows the 2-D model of a row of modules deflected eastward in the morning with the deflection angle -%, which shows the view from the top along the axes around which the modules are rotated (deflected) twice per day. Around the noon (exact time is not critical) the modules are triggered into the westward position with the deflection angle +%. The processes in the afternoon are the mirror reflection of those in the morning and were not subjected to detailed analysis.

Each module with a width WC has its axis parallel to the basis. The basis may be ground with the zero tilt angle p0=0, a roof with a free tilt angle 0<р0<л /2 or wall p0=n /2. To simplify the analysis we suggest that the tilted basis is looking south with the azimuth y0=0, but that is not obligatory. The inner modules are installed at the distance DR from each other, the outer module is installed on the distance DC from the edge of the base (roof, wall). An important parameter in the analysis below is the relative distance (density) dR= DR /WC. Illumination on the upper (outer) edge of the inner module appears when the sun has reached the position characterized by the clock angle юF. Then the sun is shining along the plane of a virtual envelope, parallel to the basis, it joins together all the upper edges of the farm. At the clock angle ®G, the whole area of the inner module is illuminated and after that ю >®G the module performs like a tracked stand-alone module. During a transient shading process, characterized by the angle ^=(®G ^F ), the shadow of the neighboring module will move across the module. Converted solar energy E from the direct component is proportional to the share of the illuminated area AC (ю). As AC (®)<AC, the tracked module inside the farm produces always less energy than the stand-alone tracked module. Consequently, the gain of a solar farm with the two-positional tracking will be somewhat lower, compared to that of the two-positionally tracked stand-alone module.

2. Approach

The task of the study is to calculate the hourly, daily, monthly or seasonal energy yield for a module exposed in the two-positional regime. This calculated (and experimentally measured)

energy yield will be compared with the energy yield of a fixed module, which is installed in optimal conditions. Improved efficiency, i. e. the profit (“gain”), is defined as the ratio of the energy produced in (two) deflected positions to the energy produced by the collector that is optimally exposed and fixed in this position. This is a south-faced collector with the tilt angle that warrants the most uniform energy yield during the season. For the latitude around 60° N, the tilt angle should be 45°<p <60°. We refer to the value of 45°.

Gain may be defined for an hour as follows: hourly gain = hourly energy yield of the deflected collector ETx divided by the hourly energy yield ET of the fixed collector, kWh/kWh.

Also, gain may be defined as the ratio of the corresponding irradiances kWm-2/kWm-2.

Gain = GTxj GT, (1)

where GT is the irradiance on the tilted module and GTx is the same in the deflected position.

Daily gain = energy yield in the two-position deflected collector per day divided by the energy

yield of the fixed collector per day, kWh/kWh.

Monthly and seasonal gains are defined analogously to the daily gain. The main goal is to maximize the seasonal energy yield, although the hourly gain is also of interest, considering co­operation of solar PV farms with the grid.

Gain is the function of several variables: geographical location due to the latitude Ф, season due to the changing declination 5, solar clock angle ю, initial tilt angle p0, initial azimuth y0, deflection angle x, and solar radiation that varies by site and time. In view of these circumstances, a general analytical solution would hardly be possible especially due to the radiation data presented as table functions. Therefore, calculations must be performed by help of computer simulation. Radiation data of the beam Gb and the diffuse Gd component have to be considered separately as they are absorbed by the module in the farm differently. In the analysis of a PV-farm performance, in addition, we have to consider the variable dR characterizing the density of the modules in the farm. Geographical (Ф ) and constructional data (p0, yo, x) can be considered as constant for each analysis, radiation data Gb and Gd are tabular functions, 5 and ю are continuous variables sampled for each step of the calculation. Auxiliary variables in the computation process are presented as functions of the independent time variable ю.

A test according to the DC procedure requires the set-up of the complete system in an indoor test facility and its operation for eight days according to a well defined test sequence [1-3]. It is to be used for small factory made systems as well as for small custom built systems with a collector area smaller than 15-20 m2 and a heat store volume of 1500-2000 litres.

The major performance indicator of the solar combiystem given by the DC test is the final energy used by the auxiliary heater, the same as in the CTSS method. For this case, it is essential that the solar combisystem is always tested in combination with the auxiliary heater. This feature is considered to be favourable, because many problems in system operation appear due to improper control strategy for the coupling of the solar and auxiliary parts of the system. In case of absence it is also possible to use a specific well defined laboratory heater.

The prediction for the annual final energy use is only possible when the test conditions during the sequence correspond more or less to the annual conditions. Hence, this test method is restricted to one specific climate zone in Europe and one specific building type with its characteristic insulation parameters, heat distribution system and domestic hot water profile.

From a simple energy balance for the thermally driven chiller, the specific primary energy consumption of the dry cooler and the cooling tower per unit of heat rejected, PEspec, cooling tower, including the electricity needed for the circulation pump of the cooling water cycle, can be expressed as follows:

In this equation Espec, coolin tower is the specific electricity demand of the cooling tower per unit of cooling energy (heat rejection) including the circulation pump of the cooling water cycle and it is expressed in kWhel/kWh

cooling, [13].

In addition to the above defined specific energy consumption, the total cooling energy produced by the chiller, the electricity consumption of the fans and the water consumption were calculated with TRNSYS and reported in Table 2. The simulations were conducted for an entire cooling season, (middle of June to middle of September).

An infra-red camera was used to check for any defects in the thermal insulation around the tank or any other hidden thermal losses. It was found that there were no losses except from one end of the solar tank, where the electric booster is situated, as shown in Figure 9. Because of that, no insulation is usually placed in that area but just a plastic cover. Otherwise, the insulation around the tank was homogenous and no hot spots were identified.

4.2 Statistics of days with insufficient hot water

Data analysis showed that the number of days with insufficient hot water and that would have required boosting by the electric back-up heating element, ranged between 5 days for a conservative user to 15

or more days for an average user. One has to bear in mind the parameters of this study i. e. 150 litres evacuated-tube solar heating systems, being utilised by couples.

5. Conclusions

The results of this study showed that sizing solar systems should take into consideration, not only the number of persons in the residence, but also their washing habits (showers or baths), their specific requirements (connection of solar hot water to kitchen, washing machine, etc..), the number of days for hot water storage, as well as the inclusion of the FUPF, which caters for any future increase in hot water demand.

Another important conclusion that was reached is that savings do not come automatically with installation of a solar system. In fact, misuse of the solar heating system could at times, lead to higher electricity bills. As a result of close interaction with the users, it was concluded that there was a strong deficiency in the provision of information and education on solar heating systems, even from the side of the installers. It is recommended that a national information campaign be made for the different entities concerned (domestic users, suppliers, installers, large-users, etc.). Only then would misconceptions and doubts be eliminated and more citizens opt for solar heating.

It is also recommended that dealers should start providing instruction and troubleshooting manuals, be more accessible and quick to respond to after-sales services and where possible, offer maintenance contracts. Only then would customer satisfaction be improved and more solar systems be installed.

6. Acknowledgements

Our gratitude goes to the Housing Authority, the Housing Construction and Maintenance Department and to all apartment owners, who supported us during this study.

References

[1] National Statistics Office (2007), Results of the National Census 2005, http://www. nso. gov. mt/statdoc/document file. aspx? id=20961, accessed on 30th June 2008.

[2] R. Farrugia, E. Mallia, M. Fsadni & C. Yousif (2006) Barriers and Incentives for the Widespread Application of Renewable Energy in Malta, Proceedings of the World Renewable Energy Congress — IX, 19th-25th August 2006, Florence, Italy.

[3] V. Buhagiar, F. Camilleri, J. Cilia, R. Piscopo, C. Yousif (2002), Recommendations for Energy Efficient Housing at Tal-Ftieh, Birkirkara, April 2002, unpublished report to the Housing Authority, Floriana, Malta.

[4] The Housing Authority (2004), The Policies of the Housing Authority, Energy Saving, Article 5, http://www. housingauthoritv. com. mt/Publications/the policies. pdf, accessed on 30th June 2008.

[5] The Malta Environment and Planning Authority (2007), Development Control Policy and Design Guidance 2007, Part 13, ISBN 978-99932-83-68-3.

[6] C. Fernandez Vazquez (2008), Performance Analysis of Water-in-Glass Evacuated-Tube Solar Heating Systems at a Housing Project in Malta, unpublished final-year dissertation, Institute for Energy Technology, University of Malta, Malta, in collaboration with Valladolid University, Spain, under the Erasmus Student Exchange European Programme 2007/08..

[7] Enemalta Corporation, http://www. enemalta. com. mt/page. asp? p=926&l=1, accessed 30 July 2008.

Until now, assessments of real energy savings achieved by individual solar thermal systems are very scarce, especially for solar combisystems. This is an obstacle for a good global evaluation of the impact of such systems in national or European energy policies. Methodology based on a common European and /or international agreement is still lacking.

In order to increase the quality of solar combisystems, one major item is to build up a test method that is able to produce test results representative of the yearly energy savings gained by a solar combisystem.

In principle the thermal performance or the primary energy savings respectively, can be determined with two methods.

One is the so-called CTSS method [5] (Component Testing — System Simulation) which is based on a separate test of the major components followed by a simulation of the system performance. First, the collector, the store and the controller are tested separately. With the parameters determined for the single components the thermal performance of the whole system can be calculated with an appropriate simulation programme. On this basis, a prediction of the thermal system performance for reference conditions and variable boundary conditions (e. g. different

meteorology and load profiles) is possible. This method is preliminary standardised in ENV 12977 part 1 to 3.

The other test method is the so-called AC/DC and/or CCT test method [1-4, 8]. This method is based on a test of the complete system (except the collector array). In order to determine the system performance, the system is operated during several days according to specified test sequences and the energy consumed and delivered by the system is recorded.

Three systems will be tested in different institutes: two of them will be both tested by CTSS method and one of the black box approaches in order to verify that results are quite the same, and the third one, that can not be tested by the CTSS method will be tested only by a black box approach.

The consortium will take care that the tested systems will also be monitored in order to compare test results to monitoring results.

A round robin test will be performed between the different laboratories involved in this project.

After the oil crisis in the 1970’s the Swedish government encourage households to convert oil based heating systems to electrical heating, for example by offering subsidies. In 2006 more than 700 000 single-family houses used pure electrical heating systems, including heat pumps. That constitutes more than 40 percent of the total Swedish detached houses [3]. Including combined heating systems the number is even higher. In 2006 subsidies for converting direct electrical heating as well as oil based heating systems to district heating, biomass or heat pumps were re-introduced. The subsidy for oil based systems ended in 2007, while the one for direct electrical heating will be available until 2010. Prior to those, there were conversion subsidies from 1997 to 1999 and 2001 to 2003 [4]. The fluctuating allowance system has caused an unpredictable market for heating systems.

Heat pumps have gained significantly increasing market shares since the middle of the 1990’s. Half of the European heat pumps are installed in Sweden [5] and in 2005 the total official number of installed heat pumps reached 600 000 [6]. Air-to-air heat pumps are however not included in these figures since the statistics are kept secret due to competition, but they are simple solutions with low investment costs, frequently used mainly in electrically heated buildings without water-based central heating systems. The market introduction of heat pumps in Sweden is considered a success, although this trade has also suffered from the unstable market the last years.

The use of wood pellet in the residential sector has increased more than seven times in the period 2000 to 2005, according to the pellet industry [7]. So far, burners have mainly been installed in existing (oil) boilers, but at present there is a trend towards boilers and systems adjusted to pellet. The total number of installed pellet boilers in detached houses was more than 80 000 in 2005 [8], which has increased further since then. During 2007 the number of sold pellet units, however, decreased by 80 percent compared to the previous year. Several companies had to reduce their staff, some were bought by larger companies and some even went bankrupt.

The solar heating trade in Sweden is small. There used to be 10-15 different companies, but over the last years numerous new companies were founded, mainly focusing on import and often vacuum tube collectors, and several large well-established suppliers of heating systems started to promote solar collectors as well. A total of 350 000 m2 glazed solar collectors are installed, from which about 30 000 m2 were installed in 2006 [9]. Since 2000 there is a subsidy for installation of solar heating

systems, which will end in the end of 2010. This can be applied for separately from the conversion subsidy. There is no statistics available on sold combined solar and pellet systems.

2.1. Loads

The maximum wind speed for the Bra§ov region is 30 m/s resulting a maximum wind pressure pm = 580 N/m2 [3] and a maximum wind force Wmax = 580 N on the 1 m2 surface of a panel. The wind direction can be considered towards the front of the panel (fig. 2, a. c) or towards the back of the panel (fig. 2, d. f). There can be considered different assumptions on the distribution of the wind pressure on the panels:

• Uniform pressure [3], corresponding to the bigger wind force (fig. 2, a, d);

• Trapezoidal distribution, approximating the distribution presented in [4], reversed for opposite wind direction (fig. 2, b, e);

• Triangular distribution, following the possible distribution presented in [5], reversed for opposite wind direction (fig. 2, c, f).

For all six wind load cases presented in fig. 2, the load can be reduced to a single wind force W and a moment M, placed on the axis of the panel with values according to Table 1.

Snow loads on sloped panels decrease as the tilt angles and slopes increase. Most PV or thermal collectors have glass surfaces that help the snow-shedding process. Since the tilt angles of a tracker during snow period are relatively big, there is a very small chance of snow load on panel. It can be neglected for Bra§ov region and it will not be considered in this paper.

The weight of the panel together with all the parts (frame) directly attached to the panel is G = 250N. The weight of the other parts is not considered for the preliminary calculus.

The seismic load is critical in earthquake-prone regions and the procedure for designing for seismic loads is standardized, mostly depending on the characteristics of the region. In this paper, the seismic load is not considered.

Schematic sketches of the investigated solar combisystems are shown in Fig. 1. In the solar combisystem based on a bikini tank, solar heat is transferred from the solar collector fluid to the domestic water by means of the mantle around the lower part of a hot water tank. Heat is transferred from the domestic water to the water in the space heating system by means of a mantle around the upper part of the hot water tank. The domestic hot water is directly tapped from the tank.

In tank-in-tank solar combisystem, a domestic hot water tank is integrated in the space heating heat store. Solar heat is transferred by the internal heat exchanger spiral placed at the lower part of the tank. The space heating system is connected directly to the tank. Domestic hot water tank is tapped from the domestic hot water tank in the store.

b) Tank-in-tank solar combisystem

Fig. 1. Schematic sketches of the investigated solar combisystems

The thermal performance of the solar combisystems is calculated with the simulation program TRNSYS 16 [5]. The solar collector areas used are 3-8 m2. The collector tilt is 45° facing south. The
mass flow rate in the solar collector loop is 0.15 kg/min. m2 for the bikini solar combisystem and 1.2 kg/min. m2 collector for tank-in-tank solar combisystems corresponding to low flow and high flow systems. A differential thermostat control with one sensor in the outlet of the solar collector and one sensor in the lower part of the lower mantle for the bikini solar combisystem is controlling the pump in the solar collector loop. For the tank-in-tank solar combisystem, the pump is controlled by a temperature sensor located at the solar collector outlet and a sensor installed at the level corresponding to 1/3 of the heat exchanger spiral height from the bottom of the heat exchanger spiral. Both systems have the start/stop temperature differences of 10/0.5 K.

Moreover, if the temperature of the top tank is higher than 98°C the pump in the solar collector loop stops its function. The solar collector efficiency, n and the incidence angle modifier, k0 are given by:

П = 0.772 ■ k0 — 2.907- (Tm — Ta)/G — 0.015 ■ (Tm — Ta)2/G (1)

k0=1-O.128- (1/cos0-1) (2)

Weather data of Copenhagen, Denmark from Meteonorm is used. The heat storage volume is 300 l and the auxiliary volume in the heat storage is 150 l. The auxiliary heater inlet and outlet relative height is 1 and 0.5 for both tanks. The relative height of the space heating inlet and outlet is 0.2 and 0.8 for the tank-in-tank system. Both bikini and tank-in-tank dimensions are shown in Fig. 2a and Fig. 2b. For the tank-in-tank systems, seven different models are considered. Model 1-4 are for tanks with a 120 litre DHW tank and Model 5-7 are for tanks with a 160 litre DHW tank with different heights and diameters of the DHW tank.

The heat exchange capacity rate of both top and bottom mantle is kept constant as 200 W/m2.K for the bikini tank [4]. In tank-in-tank system, the heat transfer coefficient between the DHW tank and the outer tank is assumed to be 120 W/m2.K [6]. The heat exchange capacity rate for the internal heat exchanger spiral in the solar collector loop is 50 W/K. m2 collector area. The tank relative height is 0 at the bottom of the tank and 1 at the top of the tank.

Fig. 2.a. Schematic sketch of the tank-in-tank store and dimensions for store variations

Fig. 2.b. Schematic sketch of the bikini tank and dimensions for store variations

For simulation of the bikini tank and tank-in-tank solar combisystem, the non-standard component multiport store-model, Type 340, is used [7]. Three different houses with space heating demands of 5000 kWh/year (SH108), 9500 kWh/year (SH216) and 16000 kWh/year (SH360) are used for the simulations. The house area is 150 m2 Fig. 3 shows the heat demand throughout the year (left) for the houses and the flow and the return temperature of the space heating system (right).

2. Results of calculations