The first generation prototype system (illustrated in figure 2) was installed in the laboratory of AEE-INTEC in Gleisdorf/Austria. System simulations showed that for a long-term heat storage application (storage periods of several months) a system consisting of several adsorbers and one evaporator/condenser would yield the best performance. This was the reason to install a system assembled of two adsorbers. All containers have a nominal volume of 1.25 m3, filled with 1.1 m3 of silica gel, and are equipped with internal heat exchangers. For the adsorber heat exchangers, a spiral plate heat exchanger was chosen. The evaporator/condenser contains two plate fin heat exchangers of different size located in the upper (condenser) and lower (evaporator) part of the container.

The hydraulic system consists of two distribution manifolds. There are two heat sources (the solar plant and the electrical flow heater) and three heat sinks (low temperature heat delivery system in a test apartment, hot water boiler in the test facility and a rain water cistern) available. A solar thermal plant with an aperture area of 20.4 m2 is on-hand for the test plant as the primary source of energy, the electrical flow heater is used as additional energy source. Figure 3 shows the four individual components in schematic terms as well as the heat flows between components.

DHW storage
tank

Figure 2: Sorption storage tank plant at the AEE INTEC in Gleisdorf/Austria
(EV/CO: evaporator/condenser, DHW: domestic hot water).

Figure 3: Schematic presentation of hydraulic circuits.

A simulation study shows the impact of the weather chronology on the thermal performance of a typical office room without internal and solar heat gains but with a typical air change during the day. Consequently, the mean room temperature is identical to the mean ambient air temperature since the room is in thermal balance with its environment. The weather is simulated for periodic steady state conditions whereby a mean summer day (Ta, m=19.4 °C) and a warm summer day (Ta, m=26,3 °C) alternate in certain time intervals:

— 2 average days followed by 1 warm day,

— 10 average days followed by 5 warm days and

— 40 average days followed by 20 warm days.

As in each simulation the average period is twice longer than the warm period, the mean outdoor and indoor temperature in each simulation is 21.7 °C, but Fig. 2 shows that the temperature distribution differs: While short-term variations are highly attenuated, the thermal inertia is not large enough to mitigate variations with a longer cycle period. Thus, the room cannot compensate for the temperature variations at very long cycle periods and the room temperature is coupled more directly to the ambient air temperature. Due to this close correlation, the slope of the regression line increases with the duration of the cycle period.

frequency [1/d] ; 60 days

The sketches on the right side show the cycle periods for each simulation run and the corresponding Fourier series for 60 days. Consequential, the daily amplitude is identical.

The low frequencies (= small numerical value) have a stronger impact (= higher amplitudes) when the chronology is characterised by longer cycle periods.

Fig. 3 shows the regression lines for the three office buildings for the summer 2002 and 2003. Additionally, the mean room and ambient air temperature in the summers of 2002 and 2003 are marked by grey dots. The stability index R2 is 0.6 for Pollmeier and 0.7 for Fraunhofer ISE and Lamparter. Though the room temperatures are higher in summer 2003, the excess temperature is smaller than in summer 2002. For better clarity, the line ambient = room air temperature is plotted.

In other words: On the one hand, the hourly room temperatures are not only higher but also exceed the comfort criteria more often in the summer of 2003 than in the summer of 2002.

On the other hand, the mean room temperatures increase less than expected.

The differences between indoor air temperature measurements and simulations (residuals) are represented in Figure 9. The indoor air temperature measured has been averaged at 1 h time step for comparison with simulation outputs.

* CeFsiu$

2 1,5 1

0,5 0

-0,5 -1 -1,5

72 172 272 372 472 572

Time (hour)

Figure 9: Residuals time evolution From April 28th 2003 to May 19th 2003.

The mean value of the residuals is 0.37 °C and its standard deviation is 0.45 °C. Residuals time evolution is not stationary. It present trends which follow outdoor temperature trends. The model globally underestimates the indoor air temperature during the warmer periods and overestimates it during the colder periods. It reveals a deficient model behaviour. Besides residues are lower during colder days than during warmer ones, it indicates that the model has performed better for colder periods.

The power spectral density of the residuals (see Figure 10) shows that 96% of the variance is concentrated on the frequency range [1, 1/10 h-1] and present peaks at 1/24 h-1 and 1/12 h-1 frequencies. It reveals that model disagreements with measurements appears mainly in the frequency range in which the building is manly excited, as expected according to [5].

In order to create a well-designed illumination fuzzy controller, the first set of preliminary experiments was done. The observed variables were inside luminous efficacy and inside illumination resulting from weather conditions during the year. To find out the optical response of the test chamber the experiments were made with fully open roller blind. The figures show the inside illumination and luminous efficacy as response to the global and reflected solar radiation. The first example in Fig. 5 is measurement in April.

Inside luminous efficacy is of about 20 lm/W or less by global solar radiation with maximum by 650 W/m2. Inside illumination is extremely high (up to 12000 lx) when the sky is clear. This is because of the south orientation of the window in the small test chamber with white inside painting. Thus, it is obvious that shading is necessary during the highest solar radiation period. The experiment in Fig. 6 shows the solar radiation in September with optical response of the test chamber.

Luminous efficacy in Fig. 6 is in the range of 2-14 lm/W. On the first day of experiment it is high, up to 14 lm/W, despite of the cloudy sky conditions (solar radiation is of about 100 W/m2). By clear sky conditions solar radiation is more than 700 W/m2, and luminous efficacy is less than 13 lm/W.

Fig. 7 shows the luminous efficacy in wintertime conditions, when the sky is mostly overcast and the sun has the lowest elevation.

Using shading devices in dim wintertime is mostly senseless, while capturing solar radiation in the living space is desired because of energy gain for providing both inside thermal and optical effect. Inside luminous efficacy during the dim winter days is between 20 lm/W and 5 lm/W when solar radiation is between 100 and 350

W/m2. The resulting inside illumination is satisfactory — more than 400 lx. It is interesting to note that in the evenings and in the mornings the luminous efficacy is two times higher than it is at midday. High luminous efficacy on overcast days with low solar radiation derives from high grade of the diffuse component of the daylight.

The observed optical response of the test cell during the year was the basis for the design of the fuzzy controller. Fuzzy controller was progressively optimised. In Fig. 8 the examples for well-designed “winter” and “summer” fuzzy controller are shown. In wintertime regime the direct solar radiation is desired, and this fact was considered in “winter” fuzzy designing. In the summertime regime shading is necessary. Therefore, during the summer period the “summer” fuzzy controller selected higher percentage of shading on window than “winter” fuzzy controller during the winter period for equal amount of solar radiation.

blind movement was relatively moderate and continuous.

Fig. 9 shows the oscillatory movement of the roller blind on the second experiment day as response to changeable solar radiation. Luminous efficacy follows the roller blind movement during the day, it is about 10 lm/W, but in the mornings and evenings it increases extremely.

During the day the inside luminous efficacy is low, less than 8 lm/W (Fig. 10). It is because the window was shaded as response to high global solar radiation with the maximum of 900 W/m2. With the

automatic roller blind movement the desired inside illumination level is maintained.

Fig. 11 shows the inside temperature profile when the inside illumination is controlled during the conditions represented in Fig. 10. It is evident that the inside temperatures are in the tolerable range despite the high solar radiation. Without shading the inside temperature will be at least 20 K higher than the outside air temperature. Therefore,
maintaining the inside illumination on suitable level with proper shading excludes the excessive inside temperatures.

6

Conclusions

The available luminous flux and illuminance in the building are closely related to the available solar radiation. We used the quantitative influence of the available solar radiation for the fuzzy design and the optimization of the

fuzzy control system, which enables optimal dynamic response of the roller blind according to the desired inside illumination and the outside conditions. To design the fuzzy system some preliminary experiments were done, where the observed variables were luminous efficacy and the inside illumination. Luminous efficacy is defined as the ratio of the luminous flux to radiant flux and tells the relationship between the optical and the thermal effect of the available solar energy. Maintaining the desired inside illumination level in the range of 500 — 1500 lx means the inside luminous efficacy between 5 lm/W (summer shaded window) and 14 lm/W (winter unshaded window). The system for the automatically adjustable window geometry is executed in the test chamber with the fuzzy control system, which makes decisions similar to human thinking process. The design of the fuzzy controller is based on setting up a set of linguistic control rules derived from the experimental optical knowledge. The controller was adjusted through experimentation. The two well-modified illumination fuzzy controllers, one for wintertime regime (direct solar radiation inside is the desired priority) and one for summertime regime (the direct solar radiation must be excluded as much as possible), are presented in the paper. The controlling performance is satisfactory and assures the inside daylight illumination with moderate continuous movement of the roller blind in the area, where the desired value oscillates up to ± 50 lx. Such illumination the fuzzy control system enables the optimal use of the available solar energy for improving the optical and thermal inside comfort, and can be applied in any building. The particularity of fuzzy control system is that it must be designed and optimized according to the site and its weather conditions in relation to the desired internal conditions, i. e. experimental designing.

One of the results of analyzing monitored projects in the framework of IEA SHC Task 28 "Sustainable Solar Housing” offer valuable lessons how to design housing with extremely low non-renewable energy consumption. Compact, tight and highly insulated buildings with an average envelope U-Value lower than 0.5 W/m2K facade area and a high efficient intelligent home systems are the key. At such levels, a very efficient ventilation system is also a must. Active and passive solar strategies can still contribute a valuable fraction of the remaining energy demand. All together, their are numerous combinations of strategies to achieve the same result. High performance has been demonstrated in very diverse climatic regions in Europe with much success.

The subsequent reduction in primary energy leads to both drastic reductions in the environmental impact of a house and impressive savings in the household budget of the occupants.

Many thanks all colleagues for their support, in contributing monitoring and analytical data. Also thanks to the Projekttrager PTJ Julich for the support and the German Ministry of Economics for the financial funding of the project monitoring as well as the Swiss Federal Office of Energy for funding the Operating Agent of this IEA Task.

Literature

S. R,. Hastings, Sustainable Solar Housing, Solar 2000, Gleisdorf, A

K. Voss, A. BUhring, M. Ufheil, Solarenergienutzung und Energieeinsparung im

Geschosswohnungsbau — Erfahrungen und Ergebnisse aus realisierten und geplanten Projekten, Soarthermischwes Symposium, Staffelstein 2002

C. Russ, S. R. Hastings, K. Voss, Zukunft fur Zuhause, Sonnenenergie 2, 2004, S.31 — 35

IEA Task 28 „Sustainable Solar Housing“, Internal working dokuments for building characterisation and monitoring of the demonstration projects Gemis 4. 1 Globales Emissions Model Integrierter Systeme, Februar 20O2

Workplane illuminances monitored within a scale model placed in two different considered locations (cf. Figure 1.b) overestimate daylighting performances of the test module. Figures 7 and 9 illustrate the observed discrepancies, which remain constant and close to 35 — 40% in relative terms for all profile positions. It shows that moving the model from one location to the other has no impact on the remaining divergence (0 — 1% percent-point reduction) : this indicates that the different sky view factors and external reflected component are apparently not responsible for the remaining discrepancy (cf. Table 3).

a. b. c.

Figure 7: Comparison of relative divergence between illuminances monitored in the test

module and in scale model 2 a. 2.2 m., b. 4.2 m., c. 6.2 m from window side.

The parameters that quantify the redirecting or diffusing properties of a daylighting glazing system can be measured using a goniophotometer with a light beam perpendicularly incident on the sample. The luminance coefficient q, defined as the ratio between the luminance L of the sample surface and the incident illuminance on the sample as a function of the observation angle є, with 0° < e< 90°, is the main parameter. The observation angle is 0°, when the observation is normal to the sample. The goniophotometer available at Istituto Elettrotecnico Nazionale Galileo Ferraris, Torino [2, 3], was used to perform investigations on the samples 12 and 13.

If the following conditions are satisfied, an absolute (a reference standard is not required) and accurate measurement is performed:

• all the half-plane (or half-space if the sample is not isotropic) in the transmission configuration is analysed with short steps for the observation angle;

• the detector exhibits such a linearity that a sufficiently accurate ratio is obtained between the measured luminance and illuminance.

As a consequence of the rotational symmetry of the samples, they were supposed to be isotropic and measurements were performed on the horizontal half-plane (only one angular coordinate, e, is needed to establish the observation angle). The collimated light
beam was in the opposite half-plane (transmittance configuration) perpendicularly incident on the sample.

Due to the samples characteristics, the observation angle є step was varied as follows:

• between -2° a +2° to the sample normal the measurements have 0.5° as resolution step;

• between (+)2° and (+)10° the step is 1°;

• between І0° and 25° the step is 2.5°;

• between 25° and 60° the step is 5°;

• between 60° and 90° the step is 10°.

These intervals were chosen considering the low scattering properties of the selected samples. It was, hence, expected to collect the highest amount of transmitted energy, close to the normal direction, typical of materials with regular behaviour. The resolution then increase with the increase of the observation angle.

In figure 4 the goniophtometric measurements on the samples 12 (continuous line) and 13 (dotted line) are presented. In particular the graph reports the luminance coefficient as a function of the observation angle. No experimental data are available for sample 6, since it was not available during the test campaign. As expected, no redirecting components, nor diffusing behaviour of the samples, are evident in figure 4 and the two glazings confirmed their mostly regular properties.

A rapid growth of interest in superinsulation has been noted in 1970-80s and has been connected with the development of cryogenic engineering, space engineering, aviation, surface and underwater sea fleet. An interest in superinsulation was also resumed in the end of 1990s and in the beginning of the 21st century. The development of hydrogen power industry in the symbiosis with nuclear power industry is a part of prospective national programs of a number of developed states. The placement of a nuclear reactor in the world ocean water area for hydrogen production and liquefied hydrogen transportation to an island
is one of the prospective projects of the Japanese power industry development. In order to store reserves of liquid hydrogen, oxygen and other liquefied gases, effective cryogenic reservoirs and pipelines will probably be required.

The volatility of most effective reservoirs is 0.8-1% per day of the total amount of liquid being stored.

Giant dimensions of present-day reservoirs determine a large amount of expenses on manufacture of the internal and external reservoir shells. In order to optimise expenses, the external shell is manufactured from a low-alloyed steel. The latter circumstance leads to increased gas releases of inter-lattice hydrogen into the heat-insulating cavity. The hydrogen content in the casing metal of most spread modern cryogenic reservoirs is 9.5-11 cm3/100 grams of metal. Zeolites being most widely used in Russian cryogenic engineering in cryoadsorption pumps comparatively well absorb hydrogen in the range of 20.2K and considerably worse at higher temperature.

As a result of long-term reservoir utilisation without a possibility to conduct of obligatory process of technological TIP blowings of heat-insulating cavities (HIC) and KSN regeneration, the amount of residual hydrogen in HIC achieves significant values. The reason of the hydrogen concentration increase in HIC can be both the inter-lattice hydrogen of structural materials and hydrogen inflowing (or diffusing) through microleaks from the internal vessel at the storage of hydrogen therein as well as from the atmosphere. The consequence of it is a considerable increase of the cryogenic liquid volatility. The situation being considered is much more related to the emergency categories and manifests itself to the full extent quite rarely. However engineers dealing with the operation of such systems can rather often observe the phase of appearance of several occurrences of this situation at the normal functioning of the reservoir too, especially at the final stage of the routine maintenance interval.

In order to provide for the new uses, the existing buildings had to be radically altered and extended. However, the local planning authority required that the views of the outside of the building must remain largely unchanged. Both the ‘coach house’ and ‘horseshoe’ buildings had to be converted for modern office use with, in addition, exhibition, catering, conference, meeting, and main plant spaces.

The conversion of the coach house was relatively straightforward: the building fabric was upgraded to meet contemporary office use and the courtyard was enclosed by inserting a new steel structure. The conversion of the horseshoe was more complex. The construction between the two towers, except for the timber roof structure, was entirely demolished, the ground floor was lowered, the upper level floor and the roof reinforced, and the outer external wall rebuilt. The ground floor was extended into the courtyard by 5m and a new single storey link, incorporating the main entrance, was placed between, and connecting, the two wings of the horseshoe. Turf was planted on the roof of the new office space.

A third entirely new building was introduced close to the northern perimeter of the site. So as not to intrude in the landscape, this building was partly sunk into the ground and the excavated earth banked up against the north wall. This building provides storage for the harvested biomass crop. Its roof comprises the hybrid photovoltaic/thermal array.

The only sorption system for heating and cooling that is well established in the (Western European) market is the Robur pumped ammonia/water system. However this system is far bigger than the average house system. From sorption systems for houses only costs of prototypes, of future projections and of early market introduction (Nefit) are available. We based a cost curve on the Robur system (catalogue price of 11.810 euro at 36.6 kW condenser/absorber power = 330 euro/kW) and on information from the early market introduction and prototypes. In this way we came to a cost curve in which the price per kW is decreasing with increasing condenser/absorber power (see figure 8).

Table 1 gives the values of other cost parameters that were used for the calculation of the simple pay back time (pay back time without interest, depreciation and inflation).

In figure 9 we can see that with this cost structure a sorption system without solar system can only be cost effective for the existing houses (average and big) and not for the more energy efficient newly build houses (minimum and reference). In newly build houses the space heat demand is not high enough to make this system feasible. For the market segment of average and big existing houses a condenser/absorber power higher than 3 to 4 kW is not attractive.

When we add a solar system (figure 10) the simple pay back time becomes longer than 15 years for all options. In these calculations no tax credits or subsidies were accounted for. When there is already a solar system with the right dimensions available in the house, the pay back time of adding the sorption system becomes comparable to figure 9.

Again the system is more attractive for the average and big house (existing houses) and less for the minimum and reference (newly build). There is now a clear optimum for the average house (at 2 to 3 kW) and for the big house (at 3 to 4 kW).

Technology

There are numerous sorption processes that could be used for a solar sorption heating and cooling system. The most common sorption process that is used for cooling (LiBr/water) is however not suitable as a heat pump, because it cannot operate below 4 oC. The other common sorption process is ammonia/water. Recently in Austria the Solarfrost development was launched. This ammonia-water-hydrogen absorption/diffusion heat pump is especially designed for driving temperatures below 100 oC. The system can thus be driven by a standard solar hot water system. Other developments in ammonia/water in the Netherlands
are the Nefit diffusion/absorption heat pump and the Remeha pumped ammonia/water heat pump. These companies do not consider the adaptation of their heat pump for solar cooling. Another group of sorption processes that can be used is the sorption of vapour in solid materials like water/silica gel (Sortech) or water/zeolite (Vaillant)

In existing buildings the weight and the height of the system are important success factors (especially in the Netherlands, were we do not have cellars). In this respect the pumped ammonia system and the Sortech solid sorption system are potentially better than the Solarfrost diffusion/absorption system, because of the large pipe diameters at high pressure that are needed for the self pumping ammonia system. Moreover the diffusion/absorption system does not have the possibility to adjust from heating operation in winter to cooling in summer, because the condenser/absorber temperature is fixed by the hydrogen pressure.

A solar sorption cooling/gas driven heat pump is technically a feasible system (see also the work of IEA task 25 [Henning, 2004]). Because of the large number of circulation pumps that are needed (at least a solar pump, a generator pump, an evaporator pump and a condenser/absorber pump) parasitic power consumption is a point of attention in further developing the solar sorption system.

Safety

In The Netherlands there are no regulations for the application of ammonia in houses (at less than 2.5 kg of ammonia). However the application of ammonia at 15 to 25 bar can be hazardous when leakages occur in closed spaces (like attics or cellars). The water/silica gel of the Sortech system or the water/zeolite of the Vaillant system is in no way hazardous. The system pressure will stay below the ambient pressure (at condenser/absorber temperatures below 100 oC). Silica gel or zeolite is not hazardous.

Developments

Ecofys is one of the partners in the European Modestore project. In this project a technical test of the Sortech solid sorption system is performed in Germany, Finland, Austria and The Netherlands. The first system has just been installed in The Netherlands this spring and monitoring will be performed during the summer of 2004 and the winter 2004/2005.

Acknowledgement

This project was supported by the NEO (New Energy Research) programme that is implemented by NOVEM (Netherlands Agency for Energy and Environment) commissioned by the Dutch ministry of economic affairs.

Conclusions

• A sorption system without solar system can be cost effective for average and big existing houses. In newly build houses the space heat demand is not big enough. Sorption with solar for heating and cooling is not (yet) cost effective for the Dutch cost structure (without subsidy of tax credit).

• Optimal condenser/absorber power for cooling and heating is around 3 to 4 kW for the market segment of average and big existing houses. This is equivalent to an evaporator power of around 1 to 1.5 kW.

References

1. BAK, Survey natural gas use in houses in the Netherlands, Energiened, Arnhem, 2001.

2. Hennig H.-M. (ed.) Solar-Assisted Air-Conditioning in Buildings, Springer Verlag, Vienna, 2004.

3. Herold K. E. Absorption Chillers and Heat Pumps, CRC press, New York, 1996.