Usually a storage tank with 250 litre covers the hot water demand of at least one day. So it is necessary to heat it up only once a day. This is the basic assumption of the control strategy: The heat source with a better relation of thermal energy gain to electric consumption is the first choice, in this case the solar collector. Fortunately, there is some knowledge on the availability of solar energy. The expectation of solar gains is high in the morning, lower in the afternoon an zero at night. So it would be good to heat up the storage with the heat-pump in the late afternoon. The hot water consumption in the evening and in the morning will cool down the lower part of the storage and allow a maximum of solar gains.

But as an other complication also the heat-pump has a relatively low power and needs several hours to heat up the storage. It is necessary to have more knowledge of the temperature profile in the storage.

The control strategy was developed in computer simulations with MATLAB Carnot [1]. In the simulations, it has been shown that three temperature sensors are sufficient for the knowledge of the temperature profile. The strategy defines a time at which the storage has to be at set point temperature (e. g. 5 p. m.). The 1.5 kW heat-pump may rise the temperature in the storage with a rate of approx. 5 K/h. A temperature ramp is calculated
and compared to the average of the 3 storage temperature sensors (see figure 2). On a day with no solar gains the average temperature will cross the ramp early and the heat — pump will start to heat the storage. With high solar gains the average temperature will always be above the ramp.

When the temperature at the top sensor drops below the set-pint the heat-pump and eventually the electric heating are switched on immediately.

In the computer simulations the control strategy has a solar fraction between 40 and 50 %, depending on the hot water demand (5 m2 flat plate collector, weather data Wuerzburg, 200 litre per day at 45 °C).

Results

The strategy was tested on the test rig under different conditions. The first tests were carried out with a prototype of the compact unit. The control program was running under MATLAB on a normal PC and the compact unit was controlled with an IO-card.

The hot water demand was varied from 100 to 300 litres per day. The control strategy works well with a hot water consumption up to 200 litres. At higher values the losses and thermal conduction in the storage may force the heat-pump to heat the storage more than once a day and the solar gains are reduced. The control strategy is not very sensitive to variations of the taping profile unless the daily consumption is not more than 200 to 220 litres.

In summer 2002 the first compact unit was installed in a passive house. The house has a surface of 180 m2 and a heating demand of 14.1 kWh/(m2y) (PHPP result [2]). A 3 m2 vacuum tube collector (type Vitosol 200) is installed. The consumption is relatively low with less than 100 litres per day. The solar fraction in 2003 was about 53 %.

Figure 3 shows the temperatures in the storage during a day with moderate solar gain.

Conclusions

The test rig results and the first data from field tests show a good behaviour of the system. For the compact unit with a small heat-pump it is a very convenient solution since heat — pump and solar collector have a better performance if they are allowed to work on the lower cold part of the storage tank.

The control strategy may be a solution for a simple solar system with a relatively small storage tank if a solar fraction of about 50 % is accepted.

The strategy may also be applied in standard systems with a bivalent storage tank (two separated heat exchangers). In this case the strategy may help to reduce the charging of the storage with the backup heating in the morning when the expectation of solar gains is still high. For a backup heat source with a higher power the knowledge of the temperature profile in the storage is less important. The two sensors which usually installed are sufficient.

[1]