Investigations confirmed that a circulation system had the largest influence on the loss of heat through distribution. This leads, when permanently in use, (which in practice is often the case) to increased use of energy, and can in decentralized systems of one and two family houses mean a doubling of energy use and around a halving of the original planned solar fraction. Through the use of technical devices such as clock timers, thermostats, current control switches and keypads, energy savings can be achieved; however, a heating concept without any circulatory system and an appropriately short distribution system offers the best solution.

■ Due to the longer circulatory system used in centralized systems and the fixed regulations set down by the DVGW regarding Legionnaire’s Disease, it is not recommended to omit the appropriate protection measures. In any case, the additional energy requirements are not a critical factor due to the steady use of the heating system over day and night, and losses can be limited further by the use of the above technical devices. These findings apply also to conventional energy systems, and are not dependent on a solar heating unit.

■ Further optimization opportunities were provided by the system set up. In many systems, the upper storage temperature limit was set too low at 55 — 65 °C, which causes a reduction in the possible solar energy yield. With the decentralized systems in Gelsenkirchen, the energy yields could be improved on average by 25% simply by setting the temperature limit to 85°C, and with the centralized systems by at least 2.6%. The conditions for this are that the circulation always passes through the thermal hot water system mixer and not only through the latter part of the freshwater storage tank. Only under these conditions can both energy-saving advantages be enjoyed, and sufficient protection against scalding be at all times guaranteed.

Comparing the decentralized and centralized systems

■ The decentralized system showed a 4 to 10 increase in percentage points of the solar fraction in comparison to the centralized system, when in optimal use. In any case the decentralized system demonstrated higher deviations due to the individual differences in the use of the system. Decentralized systems are much more sensitive to these changes in consumption than centralized units, as the latter benefit from many individual users at differing times of day, and a more straightened demand profile. This leads to the conclusion that in practice centralized systems will frequently offer energetic advantages in comparison with non-optimized systems installed in single
family dwellings. This is also supported by the higher performance from the usually better serviced units of a centralized system by a company with the appropriate technical experience. Finally, changes to the installation of the hot water system by non-specialists are usually thus avoided, which often occur in private decentralized units.

■ The economic comparison of the systems has shown that central heating units for several terraced houses with a solar hot water system have proved to offer clear cost advantages in comparison with single family units.