First it can be stated that during a draw-off the collector with the higher capacity has a higher performance, just because the energy that was stored in the capacity before is then released.

For the process of charging the storage tank the following can be stated. If the system temperatures at the time when the pump is switched off are higher than when operation began (which will almost always be the case in reality), the energy that has loaded the collector’s capacities during operation is lost. For this process the collector with the lower capacity is superior.

For the switch-on process, the collector with the lower capacity has a clear advantage. It heats up faster, and the threshold temperature is reached earlier.

The effects of fast fluctuations of irradiance are more difficult to judge. Some of the energy that loads the capacities while the irradiance is increasing will probably be recovered when the irradiance decreases again. Nevertheless the collector with the lower capacity is likely to have higher gains.

In a more detailed analysis, TRNSYS simulations for standard systems were carried out. For the period of one year that was investigated, all operation cycles and the corresponding collector gains were divided up into classes of lengths of operation periods. The investigations were carried out for a flat-plate collector (5 m2) and a vacuum-tube collector (3 m2). Both the values of the collector capacity from the calculation procedure
and from the J.3-method were regarded in comparative simulations. Weather data from the ISFH meteorological station in Hanover from 1997 were used. The time resolution is 5 minutes, which is sufficient for the short term dynamics to be investigated in this paper. The calculation time step is 1.5 minutes. Draw-offs take place at 7 (80 litres), 12 (40 litres) and 19 hours (80 litres).

The results for the two vacuum tube collectors (capacities 9 and 40 kJ/m2K) are shown in figure 3. It can be seen that there is a minimum operation period. The “light” collector always runs for at least 1.5 to 3 minutes, whereas the “heavy” collector stays in operation for at least 12 to 16 minutes. This shift is due to the fact that the “heavy” system, once having started, remains in operation for a longer time, exactly because of its higher thermal capacity.

Figure 3: Distribution of the collector yields in intervals of periods of operation of the solar circuit pump (vacuum-tube collector, comparative illustration for two values of the effective thermal capacity)

Due to this shift, all the intervals up to 16 minutes must be considered for a discussion of the short term dynamics. It can be seen that on this time scale the collector with the higher capacity is superior. On the other hand, on the long-term time scale and in total the collector with the low capacity has higher gains.

Where on the time scale of figure 3 can the four dynamic processes discussed above be found? The switch-on process is not investigated here.[3] The heating up of the storage tank takes place on the long term scale. The short-term dynamic processes are spread over all intervals: Draw-offs and fast fluctuations of irradiance may well take place during long intervals of operation, but sometimes they induce an operation period. In the second case, generally a short operation interval will result. Moreover, all of the short operation periods are caused either by a draw-off or by a fast fluctuation of irradiance. So the effects of draw-offs and fluctuations of irradiance on the collector gains can be identified on the short time scale of figure 3.

As discussed above, during draw-offs the “heavier” collector has an advantage, whereas during fluctuations of irradiance the “lighter” collector is superior. Taking into account that on this time scale the “heavier” collector is superior (see figure 3), it turns out that the
influence of the capacity on the collector gains is higher for draw-offs than for fluctuations of irradiance.

Furthermore, the dynamics on the long time scale is more important for the collector gains than the effects on the scale of minutes. The long term is the scale that produces the overall result that collectors with high capacities have lower yearly gains than those with low capacities. This underlines the influence of capacitive losses that are due to differences between the system temperatures at switch-off and at switch-on.

Summarizing, it can be seen from the system simulations that the influence of the effective capacity of the collector on the yearly gains is high for draw-offs and for slow charging of the storage tank, but lower for fast fluctuations of irradiance. Those dynamic processes for which the corresponding thermal capacity has a high influence on the collector gains are therefore well described by the determination procedures calculation and J.2, whereas the J.3 method does not well correspond with these processes. Consequently, for the determination of the collector capacity in the framework of stationary collector tests the calculation or the J.2 method should be preferred. The J.3 procedure does not seem to be appropriate.

Dunkle was the first to investigate the heat and mass transfer relationships in a solar still under steady state conditions in 1961. In 1969, Cooper developed a computer simulation model for analyzing the performance of such a still based on Dunkle’s relationships. Using a Fourier series with one harmonic term to describe the solar radiation and ambient temperature, Baum, Hirschmann and Roefler presented the periodic analysis for a solar still in 1970. Nayak, in 1980, noted that it was impossible to reproduce solar intensity and ambient temperature with one harmonic, and proceeded to increase the number of harmonics to six in order to represent the weather data in their model. However, random components exist in both solar radiation and ambient temperature, which can not be described satisfactorily by a Fourier series. The transient analysis of the solar still was studied by Sodha in 1980. One shortcoming of their approach is the simplification of assuming all coefficients and still parameters are constants. Thus the effects of climatic, design and operation parameters on the still performance are not clearly reproduced, and so design optimization is not possible.

In summary, based on the widely used relations from Dunkle, this study has analyzed the transient performance of the solar still in which all coefficients and still parameters are calculated using equations within the model. The weather data used for simulation will be either from actual measured data or data generated from the computer program developed by the author.

The conservatory-modeled solar pond is composed of six sections, including the plastic-cement greenhouse, the operation pond, the gas-supply system, the water inlet and drainage system, the electricity-supply system, the complementary heat-supply system and so on.

The building area of the plastic-cement greenhouse is 800 m’. The mono-slope of the greenhouse faces the sun. The greenhouse is 80m long and 10m wide. The heat-integration wall is 4.8m high and 0.38m wide. The top of the greenhouse is arc-structured. Dropless PVC membrane is adopted to be the translucidus membrane.

The working pond is round and is brick-concrete structured. The diameter of it is 7.5m and the depth is 1.6m. It is used to study the multifunctional operation, such as over-wintering, breeding young products, developing commodities and so on. There are 8 ponds in all, and the effective utilization water body is 500m.

Blowing machines of 11000 watts and polythene plenum pipeline are used to supply gas and increase hydrogen.

Seawater solar pond and shallow-styled solar pond are combined to preheat the supplied water. Electricity is supplied by the national electrified wire netting. It is equipped with an electric generator of 26000 watts.

As far as the complementary heat-supply system is concerned, the quantity of the water that needs changing should reach 100% to 200% sometimes in the process of breeding young aquatic products. Since the solar system can not fully meet the need of production, a heating boiler of 2600 watts is equipped with.

The system has accomplished the experimental research of ove-rwintering and cultivation for two periods. More than ten varieties of aquatic products are selected in order to carry out the overwintering experiment, such as pagrus major, paralichthys olivaceus, scophthamus maximus, puffers, scylla serrata forskal, portunus trituberculatus, eriocheir sinensis, scapharca subcrenata, venerupis variegata, sea snails and so on. All of them can live through the winter without employing any complimentary energy and the energy-saving percentage of over-wintering achieves 100%.

In the process of the experiment, the heating effect of the solar ponds combining the seawater solar pond, shallow-styled solar pond and conservatory-modeled solar pond is inspiring. The energy-saving percentages of April and May are 50% to 70% and 80% to 100% respectively. See charts 1-2 for the parameters of the working temperature.

Chart 1 Temperature chart for the cultivation of April, 2002 Unit: "C

Pond number, date, temperature, room temperature, air temperature, natural temperature

SHAPE * MERGEFORMAT

With exception of the thermal performance, the solar domestic hot water systems and the solar combisystems were tested according to the same procedures.

3.1 Thermal performance

The solar domestic hot water systems were tested on the basis of EN 12976-2 “Thermal solar systems and components — factory made systems — test methods” according to the DST method (Dynamic System Testing). In addition a separate collector test was carried out and the most important parameters of the store were determined from enhanced DST — measurements with sensors in the collector loop in order to enable a component based system simulation with TRNSYS.

The solar combisystems were tested on the basis of EN 12977-2 “Thermal solar systems and components — custom built systems — test methods” according to the CTSS method
(Component Testing — System Simulation). The parameters of the most important components (collector, store, controller) determined in this way and the hydraulic configuration of the individual systems were transferred into the simulation program TRNSYS.

For the determination of the “usable hot water volume” an additional test sequence on the store was performed according to the “DFS hot water comfort test” /3/.

The energy yield of the thermal solar systems was determined by means of system simulations for the boundary conditions described in chapter 2 on the basis of the system or component parameters, respectively, that were determined during the test.

For the calculation of the fractional energy savings, the energy saved by the thermal solar system was compared with the energy demand of a conventional (none solar) system. The system efficiency is determined by relating the energy saving of the solar thermal system to the available solar radiation. Hence the system efficiency is an indicator how effective the solar energy is used.

For the assessment of the thermal performance, the fractional energy savings, the system efficiency, the usable hot water volume, and for combisystems additionally the space heating buffer volume, are taken into consideration. The assessment concept was intentionally designed in a way that the typical design parameters such as collector area, store volume, usable hot water volume and, if existing, the space heating buffer volume did not affect the results as long as they are varied within sensible limits. Due to this approach the thermal performance of the system is primarily affected by the performance of the different components and their interaction within the complete system.

The requirements on the additional test sequence depend on the typical operation conditions (irradiance, collector temperature, angle of incidence, etc.) of a solar collector is as well as on the boundary conditions of the dynamic simulation (weather data and time step). The choice of the criteria and limiting values is based on the following considerations:

• a reasonable amount of solar irradiation must be taken into account

• a minimum fluctuation of solar irradiance is necessary to provide a check of the effective collector capacity

• to reproduce the fact that most solar collectors are operated together with a thermal store, an increase of the collector inlet temperature during the test sequence is required

• to check the conversion factor and the heat loss coefficients fluid temperatures close to ambient and significantly higher fluid temperatures are needed

• the angle of incidence must cover the typical range of daily operation

• the time step for recording the mean values must be such as that dynamic behaviour can be detected reasonably well

The criteria and limiting values used are summarised in Table 1.

SHAPE * MERGEFORMAT

Embedded systems provide measuring data for data logging and online visualisation to the Internet. Additionally a WAP server (wap = wireless application protocol) can transfer current measuring data in WML format (wml = wireless markup language) to conventional cellular phones which can display the data as text or simple graphics (Fig. 5). Therewith system performance can be checked at any time and anywhere without any PC.

Data analysis

Today control systems with permanent Internet connection can be realised because prices for DSL (Digital Subscriber Line) flatrates or volume tariffs are affordable. Measured data of conventional monitoring systems usually are acquired with larger time intervals to reduce online time with using analog modems. In contrast to that embedded systems with permanent Internet connection allow contemporary analysis of measuring data. System performance is monitored continuously and errors are detected very fast. Broadband Internet enables to transfer big amounts of data from the control systems to more powerful computers. These machines do the archival storage and extense analysis of the system performance of the solar thermal system. The results are used to improve the controller software. Measuring data can be processed for visualisation automatically (Fig. 6). Operators and users of the system can access a web portal to check the solar system performance.

The CPC collector

A design approach is described in this work to use an existing commercial CPC collector (CPC Ao Sol 3000) (portuguese manufacturer of CPC collectors (www. aosol. pt)). In order to reach high temperatures (higher than 100°C) the concentration factor is 1,5 and the
acceptance half angle is 36°, truncated to 54°, which means it collects energy for a minimum 5 hours at full power (in a North-South orientation) without any need to track the sun’s apparent daily motion in the sky. The tilt is fixed and should be close to the latitude of the place where it is going to be installed (39° for the case of Lisbon) although, for security reasons, never less than 10° because of the fact that the heat pipes used are gravity assisted.

The heat pipe

The heat pipe is a closed system with a fluid inside. Basically, it delivers the collected energy either to a storage mass or directly to the cooking vessel. It consists of three parts:

• the evaporator, which in this case coincides with the collector absorber, where the fluid evaporates, absorbing a large quantity of heat;

• the condenser, where the vapour condenses delivering the collected energy;

• the adiabatic region which connects the evaporator to the condenser and which can be long enough to cross a wall enabling to cook inside the house.

The operation of this device is based on a phase change principle. For that is only necessary to ensure a minimum constant tilt along the heat pipe (Farinha Mendes, 1988). The power that can be delivered depends on fill factor (F) and on tilt (p), among other factors. For thermosiphon heat pipes, typical values are in the ranges of 30° < p < 60° (Kobayachi et al., 1984) and 20% < F < 50% of the evaporator volume (Nguyen et al., 1981). Shorter start-up times are achieved with smaller quantities of fluid. It was concluded that in the case of heat pipes without wick structure, p could be 2° with a minimum F of 4,4 % (Farinha Mendes, 1988).

2 — Construction

As shown in Figure 1 the original CPC collector used in this prototype has six troughs (highly reflecting aluminium sheet mirrors) each one concentrating the solar radiation on an inverted V-shaped fin containing a pipe (riser), in turn soldered to larger diameter pipes, at the top and bottom of the collector (headers).

Figure 1 — The original CPC used Figure 2 — Cross section of a CPC through

In the new configuration, both headers were removed in order to turn the risers into the evaporators of six stand alone heat pipes. Figure 2 shows a cross-section detail of a CPC through. The connection between each evaporator and condenser (the adiabatic region mentioned above) is shown in Figure 3. This region is very well insulated with a layer of rock wool over and underneath the whole area. The slope of these pipes is 5° just to ensure the return of the condensed fluid.

The condenser of each heat pipe enters a copper box with thermal oil inside. This design feature not only enables an instantaneously uniform temperature in the copper box but also assures a certain energy storage capacity in case of temporary clouds. Moreover, the heat transfer capacity is maximized, given the fact that the condensers are immersed in

oil. The box will either be in contact with the cooking pan or with the aluminium storage mass as shown in Figure 5.

The copper box dimensions are:

• top and bottom — 300 x 300 mm, based on a medium size pan (~ 0240 mm);

• height — 47 mm — enough to ensure the 5° slope of the condensers.

The thermal oil characteristics (www. shell. pt, 2002) are:

• specific heat — 0,449 Kcal/kg K at 20 °C,

0,709 Kcal/kg K at 320 °C;

• thermal expansion — 0,00076 /°С.

Because of this thermal expansion it is necessary to add an additional volume to the oil container which is done by adding two 250 mm length pipes (internal diameter 26 mm), as shown in Figure 4.

The heat pipes cross the wood container (Figure 3 shows vacuum valves in order to allow for the preliminary heat pipe cleaning).

The aluminium mass (30 kg) was planned to store enough energy to boil 4 kg of water for about one hour. It is divided in six blocks:

• two of 300 x 340 x 25 mm (7 kg) each on the top;

• two of 300 x 340 x 25 mm (7 kg) each underneath;

• one of 300 x 20 x 47 mm (1 kg) on both sides of the oil container.

While the aluminium is absorbing the energy, the empty space is filled with rock wool thermal insulation in order to reduce the heat loss.

Figure 5 — Cross section of the cooker with: a) energy storage, b) direct energy delivery

On the other hand, as shown in Figure 5, when the cooker is delivering energy directly to the pan, the aluminium is previously removed (from the back of the wood container) and the blocks located underneath the oil box will be replaced by a piece of thermal insulation,
partially protected by a stainless steel box. At the moment of the removal, the oil box is held in place both to the condenser pipes and to the wood container with two chains.

Whether the cooker configuration is a) or b) (Figures 5 a) and b)), the whole wood container is insulated with a layer of 60 mm thick rock wool, being both top and back blocks removable to make it possible to insert and remove the pan and the aluminium blocks, respectively.

In order to function properly, every heat pipe must be carefully cleaned. It is necessary to remove every contaminant either in solid, liquid or gaseous state. These contaminants, which might have been left mainly during soldering, can free non-condensable gases which could easily hinder the necessary contact between the fluid and the inside wall of both evaporator and condenser.

Because of the size and weight of the whole set it is not possible to clean all pipes at the same time. The pipes must be cleaned individually before being fixed to the collector and to the copper box.

Cleaning process

Although the cleaning process varies with the type of material and fluid used, and also with how clean the heat pipe needs to be, the usual steps (Chi S. W., 1976; Farinha Mendes J., 1988) are:

• degreasing;

• solid particle removal;

• deoxidizing;

• degassing.

Jorn Scheuren1 (corresponding), Francisco Pujiula1, _ Wolfgang Eisenmann1,
Werner Bohle2, Bernd Hafner2

1Institut fur Solarenergieforschung GmbH Hameln/Emmerthal (ISFH)

Am Ohrberg 1, 31860 Emmerthal, Germany
ph: +49/(0)5151/999-523; fax: +49/(0)5151/999-500
email: j. scheuren@isfh. de; Internet: http://www. isfh. de
2Viessmann Werke GmbH & Co, Viessmannstr. 1, 35107 Allendorf/Eder, Germany

Up to now, comparisons of high-flow and low-flow operation were focused on solar systems with stratification devices in the hot water store. These systems show an improved thermal performance caused by better temperature stratification in the storage. What has been more or less disregarded is a comparison between identical solar domestic hot water (DHW) systems without stratification devices and with dif­ferent flow rates in the collector loop.

For small solar DHW systems with an internal heat exchanger most manufacturers recommend a flow rate of approx. 40 liter per hour and square meter collector area (high-flow operation). Low flow rates (15 l/m2h) are only used in large scale solar systems or small systems with a stratification device.

In cooperation with the company Viessmann two identical solar DHW systems were installed at ISFH and measured simultaneously. Besides, a complete dynamic system test prcedure (DST) and a simulation study were carried out to compare both systems.

Description of the Investigated Systems

Each system consists of 5 m2 collector area and a 300 liter domestic hot water store with an internal heat exchanger. They are designed and recommended for high flow rates and offer no stratification device. In summer 2002 both systems were installed at ISFH test roof II and the measurements were successfully finished one year later. Figure 1 shows a sketch of the investigated domestic hot water systems and their characteristics.

Figure 1: Sketch of the investigated solar system with the most important characteristics.

Different visualisation methods have been used in the past to gather information on the flow inside hot water tanks. Some of these methods, for example applying dyes, could only provide limited information on flow patterns. In the recent years more advanced methods have been developed. One of these methods is a non-intrusive optical method called Particle Tracking Velocimetry (PTV) to measure 2 or 3 dimensional velocity fields in a fluid. Particles are added to the investigated stream and tracked one by one. The method was used by (Shah, 1999, 2001) to investigate the flow pattern in the mantle gap of a mantle tank.

A similar optical method is called Particle Image Velocimetry (PIV). As for PTV, small tracer particles are added into the fluid and illuminated by a laser sheet. The scattered images of the particles are recorded with a camera, based on electronic solid-state images (change couples device, CCD camera). The time delay between two laser pulses needs to be adapted to the mean velocity of the flow and the magnification at imaging. It is assumed that the tracer particles move with the local flow velocity between two illuminations.

For absorbing particles, light is mostly scattered directly by the particles and it is additionally scattered by diffraction around the edges of the particles.

To evaluate PIV recordings, the area of the laser sheet is divided into small sub-areas, called interrogation areas, with a typical size of 64 x 64 pixels.

Assuming a homogeneous flow within the interrogation area, the particle distribution of two succeeding records is identified with statistical methods (cross correlations). A local displacement vector is evaluated for each interrogation area. A velocity vector map is then calculated by taking into account the time delay between the two recordings and the magnification of imaging.

With this method, mean velocity vectors are calculated for areas containing groups of several particles. The method was used for example by (Knudsen et al., 2003) to investigate the flow pattern in the mantle of a mantle tank and the natural flow pattern in the inner tank of the mantle tank caused by heat transferred from the mantle to the tank. The method was also used by (Jordan et al., 2004) to investigate mixing in small Danish marketed hot water tanks during draw-offs with different inlet devices.

The accuracy of the measurements is influenced by the inadequacy of the statistical method and the measurement uncertainty induced by background noise of the CCD recording, particle diffraction patterns, lens aberrations, the density gap between particles and fluid, the number of particles within the volume of interest, etc. (Raffel et al., 1998).

Example temperature measurements

Temperature measurements, carried out during two charging procedures of a storage tank are shown in Figure 3. In both experiments, a volume flow rate in the pipe of about 2 l/min was applied, as well as initial tank temperatures of about 20°C and inlet temperatures of about 40°C. During the measurement the initial temperature reaches its set value of 40°C faster in the measurements shown at the left compared to that shown on the right side.

The reason for this was, that the heating unit was much warmer prior to the beginning of the experiment shown at the left side.

The value of 2 l/min is a typical volume flow rate that develops if a heat exchanger is placed at the bottom of the tank below the stratification pipe and the flow inside the pipe is driven thermosyphonally.

The curves in the figures show the inlet temperature into the store (top curve), the temperatures just below the lower, the middle and the upper pipe opening and the storage temperatures, respectively (from top to bottom). After a duration of 50 min, the temperature in the bottom of the store is about 27°C, and at the top of the tank, about 36°C. The temperature below the lower flap is slightly below the inlet temperature during the entire measurements, whereas the temperatures are much lower at the two higher positions in the pipe in both measurements. This temperature drop is caused by the mixing of cold water sucked through the lower opening with the up-streaming warm water (dominating effect) and by heat losses in the pipe.

During the first about 25 min the difference between the inlet temperature and the temperatures in the upper parts of the pipe is distinctly larger for the measurement with the
more gradually increasing inlet temperature compared to the corresponding temperatures for the more rapid inlet temperature rising. This means, that more cold water is sucked into the pipe for more gradually increasing inlet temperature.

Although the inlet temperatures are the same during the time interval between 20 minutes and 25 minutes, the values of the mixed temperatures in the pipe differ by about 3 K during this time period. This effect is mainly caused by the inertia of flow through the lower flap into the pipe, i. e. that the flow that occurred during the first 20 minutes is continuing to a certain degree. Thus, the current flow does not only depend on current reference conditions, but also on flow patterns, which occurred previously.

The deviations of the thermal stratification in the two experiments are relatively large during the entire period.

The solar assisted district heating system has been in operation since the beginning of June 2000. During the Expo 2000 all dwellings were rented to the Expo-Apartment — Service. After the Expo till the mid of 2001 the regular tenants moved in.

Heat balances

In 2000, the collectors supplied an energy gain of 156 MWh (solar heat exchanger), which was mainly used for the unique heating up of the storage. Therefore, the overall heat consumption of 419 MWh was almost completely covered by the back-up system. Heat balances for the years 2001 to 2003 are represented in Table 1. Figure 6 shows the monthly heat balances of the years 2002 to 2003.

Compared to the results of 2001, a distinct increase of the solar heat supply into network was noticed in 2002 and 2003. This was caused by a greater heat supply of the collectors. Furthermore an improved cooling of the seasonal storage could be reached due to lower return temperatures in the heat distribution network.

In the past the expected energy yield of the solar plant was not achieved. Because of repeated leakages, caused by leaky compensators mounted between collector field and piping, partial collector fields were not in operation at times. As a result the solar gain decreased. The blocking of greater parts of the collector area in spring 2003 is supposed to be the main cause of the reduction of the direct used solar yield by half in comparison with the previous year.

The heat losses of the storage are higher than the predicted 70 MWh/a. In the first years of operation, however, the soil surrounding the storage had to be heated up once. In future, a decrease of the heat losses is expected.