There are two basic differences in the system configuration of solar combi-systems in a way how the boiler and the thermal solar collectors are integrated in the system with a lot of variations. With a serial integration of the thermal solar system as shown in figure 2 only the solar collectors charge the buffer store. If the temperature of the buffer store is too low the boiler delivers the remaining necessary heat.

With a parallel integration (see figure 3) the boiler is charging e. g. the upper third of the buffer store and keeps it on a certain temperature level if the solar collectors do not deliver enough heat. An advantage of the parallel integration is that the boiler does not start as often as in serial integration which improves its efficiency. However the higher temperatures in the buffer store can diminish the solar fraction and lead to higher heat losses of the buffer store.

Which of the two systems is better depends on certain criteria, e. g. the heat load of the connected housing area. This will be considered in the project as well. Both serial and parallel integration can be found in the chosen systems. An overview of these systems with the main features is given in table 1.

In the following, the methodology of determining the energy payback time is explained by an example of two thermal solar systems. Both systems are domestic hot water systems (SDHW-systems) with the same design parameters: 5 m2 collector area, 300 l total store volume including an auxiliary volume of 150 l. The two systems differ only in materials used for the collector and the supporting frame. The fractional energy savings are equal for both systems investigated.

2.1 Cumulative Energy Demand for Production

The cumulative energy demand (KEAp) comprises the energy required for the production of the goods at all phases, including extraction, mining of raw materials, semi­manufactured products and the production process itself. For all following calculations the values are taken from an extensive database called "Okoinventare fur Energiesysteme” from Switzerland.

Table 1 shows the impact of the use of different materials for the collector on the cumulative energy demand. The basis is system 1 with a collector that has a low cumulative energy demand. System 2 varies only in some collector materials used so that the impact on the energy payback time can be shown. The two systems only differ in absorber coating technique, the casing and insulation of the collector and in the material of the supporting frame. It can be seen that the absorber coating technique has only a minor influence on the cumulative energy demand of the whole system. Concerning the cumulative energy demand the relevant components of the collector are the absorber and the casing material.

The cumulative energy demand has to be determined for each component of the thermal solar system as shown in Table 2. It has to be considered that the store volume of the conventional heating system is reduced by using a thermal solar system. Therefore both systems are credited with the cumulative energy demand of a conventional hot water store of 135 litres (store credit in Table 2). The collector of system 1 is integrated in the roof, saving a large number of rooftiles. The cumulative energy demand for the saved rooftiles is therefore also credited to the thermal solar system.

In addition the cumulative energy demand of the transport of the thermal solar system from the manufacturer to the place of installation has to be considered. It was assumed that a distance of 300 km from the manufacturer to the wholesale dealer is covered with a truck and that a distance of 100 km from the wholesale dealer to the place of installation is covered by a delivery van. The cumulative energy demand for transportation is directly coupled with the total weight of the thermal solar system (including package).

With respect to the integrated mounting mode a credit for the rooftile transport has to be granted. With a general approach that the average transport distance is 400 km and that the transport is carried out by truck, the transport credit amounts to 205 kWh.

Concerning assembly and installation of the thermal solar system no general data base is available. Since the effort of installation varies depending on the kind of thermal solar system, it is calculated with a general approach of 10% of the cumulative energy demand for production of the materials and for the transport.

Table 3 shows the impact on the energy payback time. It can be seen that both systems only differ in the cumulative energy demand for the production that comprises materials used, transport, assembly and installation of the system. All other influences on the energy payback time such as cumulative energy demand for operation and for maintenance and the primary energy saved by the solar system are equal for both systems investigated.

System 1 with a minor cumulative energy demand for production has an energy payback time of 1.4 years. The cumulative energy demand for production of system 2 is 43 % above the value of system 1. This results in an increase of the energy payback time to 2.1 years.

1.2 Cumulative Energy Demand for Operation

The cumulative energy demand for operation includes the electrical power consumption of the solar loop pump and the electrical power consumption of the controller. The power consumption in [W] is multiplied by the respective operating hours of the pump and the controller. For the determination of the cumulative energy demand the resulting electrical power consumption has to be multiplied by the primary energy equivalent for electrical power.

01

Calculations with MANTLSIM were carried out in order to investigate how the thermal performance of a small low flow SDHW system is influenced by the mantle tank design. The mantle tank design analysis is carried out with the mantle tank, Danlager 1000 marketed by Nilan A/S, as the standard reference tank. The design analysis is performed in such a way that only one parameter has been changed at a time in the calculation. Table 3 gives data for the standard reference system.

The circulation pump in the system is controlled by a differential thermostat, which measures the temperature difference between the outlet from the solar collector and the bottom of the mantle. The differential thermostat has start/stop set point at 10/2 K.

All the calculations in this chapter are carried out with weather data from the Danish Test Reference Year [11]. The daily hot water consumption is 0.100 m3 heated from 10°C to 50°C, which is tapped from the tank in three equally large parts at 7 am, 12 am and 7 pm. The yearly hot water consumption is 1674 kWh. The auxiliary energy supply system heats the top 0.082 m3 of the tank to 50.5°C and the indoor air temperature is 20°C.

The tank parameters that are investigated are the mantle inlet position, the mantle height, height/diameter-ratio of the tank, auxiliary volume and insulation of the tank.

Figs. 3-7 show calculated yearly net utilised solar energy of the system with the differently designed mantle tank. The standard reference system is marked in the figures.

Fig. 3 shows the calculated yearly net utilised solar energy of the system as a function of the mantle inlet position. The figure shows that the thermal performance of the system increases for the mantle inlet position moved down from the top of the mantle to a relative position of 0.35 from the mantle top, and that the thermal performance decreases if the inlet position is moved further down. The net utilised solar energy can be increased by 2.5% by moving the inlet port down to a relative position of 0.35. These results are in good agreement with the experimental results from the previous section.

Fig. 4 shows the net utilised solar energy as a function of the mantle height. The highest thermal performance is obtained with a mantle height of 0.25-0.30 m. The thermal performance can be increased by 5% by reducing the mantle from 0.43 m to 0.27 m. This is not in agreement with earlier theoretical investigations showing that the top of the mantle is best situated just below the level of the auxiliary volume, because this position maximises the heat exchange area without the auxiliary energy supply system heating the solar collector fluid in the mantle [8], [14]. If the top of the mantle is located above the level of the auxiliary energy supply system then the auxiliary energy supply system will heat up the mantle fluid and the thermal performance of the system will decrease.

The main reason for the new results is that the simulation model now takes the heat flow in the water in the inner tank into consideration. The heat flow in the water in the inner tank is caused by the upward fluid velocities along the tank wall during supply of heat from the collectors. Therefore the model calculates the thermal stratification which is built up in the hot water tank during periods with supply of heat from the collectors, not only in the mantle level of the tank, but also above the mantle. Another reason is that the mixing, occurring in
the mantle caused by differences between the temperature of the incoming solar collector fluid and the temperature of the solar collector fluid which is already in the mantle, now is taken in consideration by the simulation model. Therefore the simulation model now calculates the heat, which in periods with relatively low solar collector fluid inlet temperatures to the mantle is transferred downwards in the tank. This mixing will equalize temperature differences in the tank resulting in a decreased thermal performance of the system. With a reduced mantle height the influence of this mixing on the thermal performance of the system will be reduced.

The reasons for the increased thermal performance of the system by reducing the mantle height are a reduced tank heat loss due to the smaller mantle surface area and the increased insulation thickness, a decreased equalization of temperature differences in the tank in periods with relatively low solar collector fluid inlet temperatures to the mantle and the fact that the heat transfer area for the heat transfer from the solar collector fluid to the domestic water is not strongly decreased by reducing the mantle height. This is the case because heat by thermal conduction is transferred from the tank wall surrounded by the mantle to the tank wall above the mantle. Consequently, the heat transfer area used for transferring solar heat to the domestic water is a large part of the tank wall. However, the heat flow model used in MANTLSIM was developed based on CFD-calculations where the mantle covered either the lower half of the tank or all the tank height. The mantle height that gives the highest thermal performance in this study is when the mantle is covering less than one-third of the mantle height. There is a risk that when the mantle gets too small the model is not calculating the natural convection flow in the inner tank correctly and is over-predicting the effect of natural convection.

On the other hand, if these results are true, it opens a new perspective in the mantle tank design because less material can be used making the mantle tank cheaper, less heavy and therefore easier to install if only the bottom third of the tank should be covered by the mantle. There is therefore a need to verify by experiments that the small mantle height is able to create the degree of thermal stratification above the mantle calculated by the model.

Fig. 5 shows the net utilised solar energy as a function of the height/diameter ratio of the tank. It appears that the thermal performance increases with increasing height/diameter ratios. The net utilised solar energy increases by about 4% if the height/diameter ratio is increased from 2 to 5. When the height/diameter ratio is increased, the thermal stratification in the inner tank will also be increased, because the distance between top and bottom of the tank is increased and because the cross section area of the tank is decreased. The heat loss from the tank is also increased when the height/diameter ratio is

increased, but this is for height/diameter ratio between 2 and 5 overshadowed by advantage caused by the thermal stratification. If larger height/diameter ratios than 5 were investigated, the net utilised solar energy would at some point start to decrease because the increasing heat loss gets more and more dominating.

Fig. 6 shows the net utilised solar energy as a function of the auxiliary volume. It is seen that the thermal performance increases with decreasing auxiliary volumes. This result is expected as a larger auxiliary volume will increase both the auxiliary energy demand and the heat loss, and thus decrease the net utilised solar energy. The storage capacity for the solar energy is increased with decreasing auxiliary volume, and this also influences the result. The net utilised solar energy increases by 19% by decreasing the auxiliary volume from 82 l to 39 l. The auxiliary volume should be as small as possible while still meeting the demand of hot water.

Fig. 7 shows the net utilised solar energy a a of the insulation thickness of the sides of the tank. The figure shows that the thermal performance of the system increases with increasing insulation thickness at the sides.

A number of parameter variations have been carried out to reveal how the different tank parameters influence the thermal performance of low flow SDHW systems. In the following it will be elucidated how to improve the design of the Danlager 1000 mantle tank by relatively simple geometrical changes. The change of the design is made in such a way that one parameter is changed at a time in the calculations. Also here the data from Table 3 are used for the reference system in the calculations.

The following tank parameters are changed: height/diameter ratio of the tank, mantle height, insulation, thermal conductivity of the tank material and the wall thickness of tank and mantle.

Fig. 8 shows the net utilised solar energy and the solar fraction as a function of the different changes in the mantle tank design. The first column shows the thermal performance of the system with the Danlager 1000 heat storage with mantle inlet position at the top of the mantle.

The first change is the height/diameter ratio of the tank, which is changed from 2 to 4. The total volume and the ratio between the auxiliary volume and the total volume are kept constant. The second change is the mantle height, which is decreased from 0.72 m to 0.55 m. The third change is the insulation of the tank. It is assumed that the tank should fit into a cabinet with dimensions 0.6×0.6×2.0 m3, and by increasing the height/diameter ratio the outer diameter is reduced, and the side insulation is increased by further 0.05 m. The fourth change is the tank material, which is changed from normal steel to stainless steel. The fifth, and last, change is the wall thickness of tank and mantle, which is reduced from 3 mm to 2 mm.

The change concerning the insulation gives the most significant improvement, while the change of the wall thickness gives the smallest improvement. The net utilised solar energy is increased from 802 kWh/year to 1009 kWh/year by applying the mentioned changes in the design. It is an improvement of 26% of the net utilised solar energy.

2. CONCLUSIONS

An improved simulation model for low flow SDHW systems has been developed and validated by means of experiments.

Calculations with the model have shown that marketed mantle tanks can be strongly improved by relatively simple design changes: By increasing the height/diameter-ratio, by reducing the mantle height, by increasing the insulation thickness on the sides of the tank and by using stainless steel instead of steel as tank material.

The thermal advantage foreseen by the calculations by decreasing the mantle height is of great interest, since the cost of the mantle tank can be decreased by reducing the mantle height. However, it must be mentioned that the model is not validated for small mantle heights. It is therefore recommended to start investigations with the aim to elucidate if the model simulates the thermal behaviour of mantle tanks with small mantle heights in the right way.

[1]

At first, an algorithm which is capable of generating a fractal hydraulic network on a given area with fluid in — and outlet was developed (Fig. 2). Based on this algorithm, the com­puter programme FracTherm was written. The appearance of the structures is strongly in­fluenced by the net parameters chosen. This first step is a pure geometric process which does not yet include hydraulic or thermal calculations.

In order to assess the fractal structures with respect to their energy efficiency, it is neces­sary to carry out hydraulic as well as thermal simulations. For this purpose the structures are exported into a file format which can be read by the simulation environment ColSim that was originally developed for dynamic simulations of solar thermal systems [8]. ColSim was extended by the ability to calculate multiple branched hydraulic networks. This hy­draulic solver also takes pump characteristic curves into consideration.

Net parameters

Net generating algorithm

raclherm

Fig. 2: Generating fractal hydraulic networks with FracTherm

The results of the hydraulic simulations can be visualised within FracTherm. Fig. 3 shows a calculated volume flow distribution drawn on top of a fractal structure. The height as well as the colour indicate the volume flow of each branch.

The collector efficiency factor F is a common measure to evaluate the thermal efficiency of a solar absorber. Analytical approaches exist to calculate F for a conventional absorber fin with a tube attached to it [1, 2]. In order to make use of these formulae, the complex fractal structure is discretised into small absorber fins i. Afterwards, a collector efficiency

factor F is calculated for each absorber fin i, taking the local flow situation and the result­ing heat transfer coefficient into consideration (Fig. 4). Finally, a total collector efficiency factor F can be determined, which can act as a means of comparison between fractal hy­draulic structures and conventional ones.

Fig. 5 shows the calculated distribution of the collector efficiency factor F for an absorber with the dimensions 2500 mm x 2000 mm (working fluid: water; specific volume flow:

51.2 l/(m2h)). It can be seen that a uniform F distribution at a high level can be obtained; the total F amounts to 0.97, which is a rather high value (according to [3], measured val­ues range between 0.81 and 0.97). The detail in Fig. 5 (circle) reveals the influence of the fin width: F values in corners are lower due to the longer distance between the perpendic­ular edges and the curved fluid channel.

FracTherm also allows to visualise the fluid temperature distribution (Fig. 6). The temper­ature gradient between fluid in — and outlet can be recognised easily. Current values such as fluid in — and outlet temperatures, collector efficiency, hydraulic power and total volume flow are shown dynamically during a simulation (e. g. while pump stages are switched).

Uta-Maria Klingenberger, Elmar Bollin, Sascha Himmelsbach Fachhochschule Offenburg, Badstr. 24, 77652 Offenburg, Germany Tel.: +49 781 205122, klingenberger@fh-offenburg. de

The support programme “Solarthermie-2000”, Part Program 2, was developed in 1993 to apply the technology of solar thermal systems for large hot water demands like in hospitals or bigger residential buildings. Large-scale solar thermal system with collector areas greater than 100 m2 should be investigated, demonstrated and standardised. Since 1999 the University of Applied Sciences Offenburg (Fachhochschule Offenburg) is part of this scientific programme. The paper will show the operational experiences, measured results and some reasons for malfunction of the last 5 years.

1. The Research Programme Solarthermie-2000

Since 1993 more than 60 large-scale solar thermal systems for domestic hot water generation were installed as demonstration plants in the frame of the Solarthermie-2000 programme. Financed by the German Ministry of Economics all these systems got subsidies for solar system installation (~50 %) and system monitoring (100 %).

The partial program 2 of the Solarthermie-2000 project, which was focused on domestic hot water heating, aimed:

• to build up to 100 best practice large-scale solar thermal systems all over Germany,

• to build high energy-efficient and cost-efficient plants with specific cost of solar heat below 0,13 €/kWh,

• to transfer know-how concerning solar thermal systems to universities, contractors and solar consultancies,

• to demonstrate solar thermal systems as reliable, sustainable and even cost-effective measure to compete with conventional energy.

At the Fachhochschule Offenburg a group of three scientists is responsible for expertise, planning supervision, data acquisition, monitoring and optimisation of presently six large — scale systems /1/. The scientists supervise the projects in all project phases including conception, planning, call for tenders, set into operation and the detailed long-term monitoring over a two years period of system operation. Meanwhile the detailed monitoring has been finished for five systems (Fig. 1). All plants have met the guaranteed values during the two years of operation within a tolerance of 10 %. One system is still in the monitoring phase: Wellness centre albtherme Waldbronn with a combination of solar potable water heating and pool heating /2/.

Contractors of theses large-scale systems have to guarantee the annual solar yield. This solar yield is part of the call for tenders and is calculated on the base of a dynamic system simulation. If the warranty was not fulfilled, the contractor has to optimise the system.

The economic aspect is an important part of the program. The predicted specific costs have to be below 0,13 €/kWh (based on 20 years lifetime and 6 % interest). So the sizing is based on a small cost minimum which is simulated by a load of 70 litre of domestic hot water at 60 °C per day and m2 collector area.

2. Monitoring

A very complex measurement instrumentation is installed in each plant to evaluate the solar thermal system and the operation. The data flow with the aid of remote monitoring is shown in figure 2.

At each solar thermal system approximately 50 sensors are installed. Measured values are:

• temperature (°C)

• irradiance (W/m2)

• flow rate (m3/h)

• state of the pumps and valves (0 or 1)

• electrical energy consumption (kWh)

The measured values are acquired and stored every 10 seconds. With this data the values like power (kW), energy (kWh), hours of operation and volumes (m3) are calculated. These data are readout every day from the Fachhochschule Offenburg. Additionally there is the possibility to readout the instantaneous values online.

Finally the measured data become characteristic data of the plant like:

• solar yield (energy provided by the solar thermal system to the domestic hot water system),

• system efficiency (how much of the irradiation on the collector area can be used for the hot water),

• Solar fraction (how much is the contribution of the solar thermal system to the domestic hot water consumption),

• hot water consumption

• solar energy costs.

These values enable an evaluation of the solar thermal system operation.

3. Results

Out of five years of monitoring period there are plenty of data. A short overview of the most important and characteristic values will be given below.

3.1 Costs

The specific costs of the six plants (without considering the subsidies and additional costs for measurements) are shown in Figure 3. The biggest contribution to the system costs have the collectors. The kind of substructure has an significant influence on the system costs. For example the solar roof of the hospital in Singen allowed a very cost effective system. Mounting on a flat roof requires an extensive construction for the fixing of the collector, which makes the whole system more expensive like in Freiburg (Fig.3).

Significant advantages of using the proposed Solar Rankine system for RO desalination are:

• Use of market available components (expanders, heat exchangers, diaphragm pumps) widely applied in heating and cooling industry available at low cost implies lower investment cost compared to other solar thermal technologies.

• Ideal exploitation of low temperature energy sources. This concerns the efficient usage of thermal wastes of industries, diesel generators and geothermal energy fields as well as thermal energy coming from biomass or MSW.

• Mechanical work is driven directly to RO pumps so there is a direct efficiency gain from no transformation to electrical power.

• It can be easily standardised due to most of the system components are market available.

• Rankine cycle approaches the efficiency of Carnot cycle meaning maximum efficiency.

• Continuous and safe operation at low temperatures.

• The working fluid (HFC-134a) is not corrosive and environmentally friendly.

• Low maintenance cost since except the fact that most of the components are market available, they are characterized by long life time (e. g. expander >150000 hrs)

• Due to low O&M cost, simplicity in construction and reliable operation it fits perfectly for applications in isolated, not grid connected areas.

• Compared to PV-RO desalination system this system prevails in the following:

o Water storage is used instead of batteries (electrical energy storage) o It is environmentally friendly since no batteries are used (problems of old batteries disposal).

o The absence of batteries implies less maintenance (usually 5 years batteries’ replacement). This is critical for installations in isolated areas where the maintenance is the key factor for long life-time. o Since no electricity is produced, the system is safer for the end users. o No qualified staff is needed for O&M. o The fresh water cost is expected to be lower.

• Compared to thermal systems operating in the same temperature range the proposed one is characterised by a much higher efficiency and much less product water cost.

• Variable pressure working conditions in RO system, implies higher energy availability and higher efficiencies in the whole system, making it more flexible to the variability conditions of RES. (solar)

1. Acknowledgements

The system will be developed under the project COOP-507997, partially financed by European Commission.

The duct heat store, which is used for seasonal heat storage, is described in detail in [1], [2] and [3]. In figure 2 temperatures in the centre of the first and second extension of the store are shown for different depths. The store was heated up from about 10 °C and has not yet reached the designed maximum temperature of about 85 °C due to its long heating-up time of 5-8 years. Nevertheless discharging started in 2003 and a steady state operation with storage efficiencies of about 70 % (for the completed store) is expected within the next 3-4 years.

In figure 3 the temperatures in the centre of the first extension in different depths are shown for 2003. The highest temperature in this part was approx. 58 °C and was reached in October. At the end of October discharging started and the temperatures decreased. In the second extension, see figure 4, the maximum temperature was approx. 65 °C. The temperatures in the first extension are lower since this part was not charged in 2002 in order to reach a fast temperature adaptation of the various store parts. The temperatures below the store (depth 30 to 40 m) slightly increased due to heat losses of the store. For economical and constructional reasons the store is only insulated on top.

Until now some minor problems related to the duct heat store occurred. In 2003 a sludge trap was installed since corrosion deposits were found in distribution pipes. Since the duct heat store is directly connected to the heat distribution system the borehole heat exchangers must be prevented from clogging.

The temperatures in the various parts of the store vary (referring to same depth). The temperatures between the first and second extension vary because of different operating times. The analyses of the measured data also show different temperatures in the north and south part of the second extension of the store. Assuming equal hydraulic transmissivity in the ground, no groundwater flow and equal amounts of charging heat, temperatures should be the same. The most likely explanation for this observation is a poor hydraulic adjustment of the double-U-pipes.

Figure 4: Temperatures in the seasonal duct heat store in 2003 (second extension)

comparison of the numerical and ex­perimental fluid flow data is shown in Fig. 8. They correspond to case Th = 30 0C, considering studies on cells c13 and c26. The upper — index * in the velocities indicates that they have been normalized using the reference value of, while

the length was normalized with L. Fig.

8 contains two sets of data. Velocity profiles at the central vertical section fo each observation window are shown at the bottom. At the top there are maps of local differences and between the numerical and experimental data in

the two observation windows. These maps of differences were obtained by interpolating the experimental and numerical data at the nodes of a regular mesh of 16 60 nodes, and by calculating the differences at each of the nodes. Third-order accurate interpolations were used to minimize additional errors in the data processing.

For cell c13, it was observed that Дм*[%] values were higher than Дм*[%] values. The maximum values of local differences Дм* were localized in small regions close to isother­mal walls, while the Дм* distribution was uniform. For cell c26 results of Дм*[%] were lower than Дм*[%], which corresponds to the local distribution of Дм* (Fig. 8b), where their max­imum values were localized in small regions close to isothermal walls. As for for Дм*, their maximum values were localized in extensive regions located in the central zones of each cell.

For Tft = 30 0C case, the maximum value of local differences was 10.1 %, for Tft = 70 0C case was 10.2 %, and for Tft = 120 0C case was 11.4%. These maximum values were localized in cell 26 and for the Дм* map. For two cells, and for all cases, the experimental u* — profiles are perfectly reproduced by the numerical profiles.

5. — Conclusions

Except in some special cases, it is demonstrated that the numerical model proposed to solve reduced computational domains with periodic boundary condi­tions, perfectly reproduces cases over whole computa­tional domains. A final validation test was carried out by comparing numerical data with experimental heat transfer and fluid flow measured in an ad-hoc experi­mental setup. The fluid flow was measured by a Digital Particle Image Velocimetry device.

References

[1] J. Cadafalch, C. D. Perez-Segarra, R. Consul and A. Oliva. Verification of finite volume computations on steady-state fluid flow and heat transfer. Journal of Flu­ids Engineering. 124(1):11-21, March 2002.

[2] K. M. Kelkar and D. Choudhury. Numerical Prediction of Periodically Fully Developed Natural Convection in a Vertical Channel with Surface Mounted Heat Gener­ating Blocks. International Journal of Heat and Mass Transfer. 36(5):1133—1145, 1993.

[3] B. A. Meyer, M. M. El-Wakil and J. W. Mitchell. Natural convection heat transfer in small and moderate aspect ratio enclosures; An application to flate — plate collec­tors. American Society of Mechanical Engineers, New York. 1978.

[4] C. D. Perez-Segarra, A. Oliva, M. Costa and F. Es — canes. Numerical experiments in turbulent natural and mixed convection in internal flows. International Jour­nal of Numerical Methods for Heat and Fluid Flow. 5:13-33, 1995.

[5] M. Quispe, J. Cadafalch, M. Costa and M. Soria. Com­parative Study of Flow and Heat Transfer Periodic Boundary Conditions. Proceedings of the ECCOMAS 2000. Barcelona, September, 2000.

[6] PJ. Roache. Perspective: A method for uniform report­ing of grid refinement studies. Journal of Fluids Engi­neering. 116:405-413, 1994.

[7] R. Scozia and R. L. Frederick. Natural Convection in Slender Cavities with Multiple Fins Attached to an Ac­tive Wall. Numerical Heat Transfer, Part. A. 20:127­158, 1991.

[8] J. P Van Doormal and G. D. Raithby. Enhancement of the SIMPLE method for predicting incompressible fluid flows. Numerical Heat Transfer. 7(6):147-163, 1984.

The project was carried out on the Altiplano in northern Argentina. This region is situated at an average altitude of 3.600 meters. The extreme dryness and strong solar radiation cause daily ambient temperatures to vary over 30°C. In winter, the temperatures at night drop to -2O °C, while daytime solar energy levels amount up to 8 kWh/m2. At 2200 kWh per square meter this region has one of the highest amounts of annual solar insolation en­ergy in the world. Figure 2 shows the temperature and the insolation over one year. An important consideration for the installation of solar air systems is the by 30% reduced den­sity of air at altitudes of 3600 m.

2. Heating and hot water production

The two heating systems described here were installed in two kindergartens. The areas of the buildings were 100 and 120 m2 respectively. Both buildings were built in the traditional way using a clay construction. The roof is made of corrugated metal sheets with a false ceiling. The false ceiling gives only a rudimentary protection against heat loss. Figure 2 shows the simulated natural room temperature in the kindergartens. As one can see, room temperatures can fall to a minimum value of zero degrees in wintertime.

With the help of the building simulation program LACASA, which runs under MALAB/SIMULINK, calculations were made to determine the energy demand. By using a local annual weather data set, an annual heat demand of 115 kWh per square meter was

calculated. This figure assumes a room temperature of 18°C with an average annual am­bient temperature of 9.5°C. The daily heat demand was found to be around 40 kWh per day in yearly average. Figure 2 shows the simulated energy demand for two houses with different floor base. The maximum heat demand occurs in wintertime with 0,7kWh/m2d. The solar air heater consists of a single glazed collector with the air flowing over the ab­sorber. On the underside the absorber is insulated only by an air-layer. The modular con­struction of this system allows the connection in series and parallel of as many modules as necessary. One module has an aperture area of 2.33 m2. Figure 4 shows a cut through this simple collector construction. The total surface areas of the installed collectors are 18 and 29 m2 respectively. The

Figure 5 shows a picture of an installed collector and the characteristic efficiency curve. The curve shows, that 40% efficiency can be obtained, when the difference between the average collector temperature and the ambient temperature is 40°C. The curve is 5 to 10 percent better than similar known constructions. One reason for this can be seen in the lower convective losses due to the reduced air density in the high mountain region.

The air current moving through the collector is directly controlled by the amount of solar radiation. This is managed by directly connecting the ventilator to a PV panel with 70W. Further on this ensures that the unit operates independent of an electrical power source. With solar radiation levels at 1000 W/m2, an air current of 0.4 m3/s is achieved. To provide long lifetime, brushless ventilators are used. These can be connected in parallel to adapt to the collector area.

The design of the unit must take account of the heating energy demand as well as the en­ergy demand for the daily warming of around 300 litres of hot water. On a typical winter day, insolation values of 5kWh/ m2 can be expected on site. This means that on a collector surface of 29 m2, a total of 145 kWh is irradiated. Assuming an average collector efficiency of 40%, a daily heating supply of 58 kWh is available for use. Figure 6 shows the gained daily energy per square meter for different orientations of a solar air heater at a geographi­cal latitude of 21° south. As can be seen in the figure 6 the best orientation of the collector can improve the yearly energy gain by almost 100 percent.

As the first results were very promising, the collector developed during the project was adopted by a local Argentinean company which was also involved in the project. The solar air heater can also be used for other purposes as for example, the drying of fruit, vegeta­bles and tobacco, as well as being used in the mines for drying dissolved minerals. One module is sold for 350 Pesos, which equals about 45 Euro per square meter.

The air heated in the solar heater flows through a valve, into an air-water heat exchanger and then through the ventilator which presses the air though an insulated chimney into the pebble bed. The exit on of the pebble bed is connected to the collector opening, thus clos­ing the circuit (see Fig. 7). The valve prevents automatically the discharge of the storage during times without insolation.

The air-water heat exchanger is designed so that it is able to extract up to 6 kW from the air current. This is sufficient to heat 300 litres of hot water each day to a temperature of 60°C. The water tank is fixed under the roof above the heat exchanger, so that the trans­port of water takes place through natural convection. In order to reduce costs, a normal car radiator was installed. The indoor installation ensures, that the the water circuit never can freeze and so damage the installation. Measurements have shown satisfying results, since hot water with up to 60°C was produced. For each 100 liters of warm water about 2 m2 of collector area should be added.

The thermal storage for the energy carried in the air comprises of a pebble bed with a vol­ume of 11 m3. The equivalent diameter of the stones is around d= 0.15m. This presents a compromise between pressure loss and heat transfer. The thermal storage system is inte­grated in the house beneath the floor and heat is transferred by radiation and convection to the room. Additional air vents can boost the heating power through direct convection into the room. When the pebble bed storage is heated up to dT= 40°C, the energy content is around 110kWh.

Since the target region of the system is a high mountain region, the by 30 percent reduced air density had to be taken into account for the system design. While the lower density has advantages since it reduces heat losses through natural convection it rises pressure losses of the air system on the other hand. On 3700 m the convective heat losses are low­ered by 26 percent while the required electric power to achieve the same massflow as on sea level rises by 20 percent. So special emphasis has to be put on an aerodynamically optimised system since the power for driving the massflow comes from PV-modules and so is an important cost factor.

Figure 8 shows the simulated room temperature of a building with 100 m2 equipped with a solar hot air system. In this case a collector area of 28 m2 is used, the storage has a vol­ume of 16m3. Compared with the unheated building, the room temperature on yearly aver­age is 12°C higher. Instead of falling down to 4°C the minimum average temperature lies at 13°C in June.

Based on various reports from Denmark and Austria [2-7] on monitoring projects during the last years, realistic hot water demand profiles were elaborated. The evaluation of monitoring projects of in total 66 one family houses and apartments in Denmark and Austria gives us an overview of realistic hot water consumption.

For both evaluations we can see the typical distribution of the consumption over the year with a maximum in winter time and a minimum in summertime due to the changes of the cold water temperature over the year and due to the behaviour of the inhabitants. Looking on the energy consumption, it is interesting to see that the apartments in the passivhouses have about half the consumption of the consumption of the one family houses. The main reason is that water saving equipment typically is installed in passivhouses.