In the following, the disagreement between the corresponding effective capacities of the

J.3-procedure and of a charging process of the storage tank is demonstrated by an illustrative example. (The first hint for this inconsistency was given in [1].)

Typical values of the physical thermal capacities Cphys of the components of a dewar-type vacuum tube collector are given in table 1.

When a step change of irradiance is applied, increasing G from zero to 1000 W/m2, and starting from Tabs = TF = Ta, then ATF « 8 K, and ATabs « 30 K (this corresponds to habsF = 30 W/irFK and a thermal power of about 660 W/m2, in agreement with a typical conversion factor ^0 = 0.66). The calculation for the resulting thermal capacity CJ3 is given in table 2. The result is CJ3 = 21.4 kJ/irFK.

These results are now applied for the calculation of a storage charge period. On a sunny day, it takes about 4 hours to heat up a 300 litre storage tank by 20 K (4 m2 collector, mean irradiance 800 W/irF, mean efficiency 0.6).

The energy needed to load the capacities of the collector during this period is calculated as follows. As discussed in section 5, the amplitude ATabs is smaller or approximately equal to the amplitude ATF. For the sake of simplicity it is assumed here that both components are heated up by 20 K. In reality, the physical capacities Cphys of the components are heated (and not any model capacities). So the result is AEcol, phys = 20 K ■ (4.5+4.5)kJ/m2K = 180 kJ/irF. In contrast to this, a simulation model that uses the capacity CJ3 calculates AEcol, J3 = 20 K ■ 21.4 kJ/mFK = 428 kJ/mF. By this, the energy that loads the capacities of the components is strongly overestimated. The difference AEcol, J3 — AEcol, phys = 248 kJ/mF of energies stored in the capacities corresponds to an extra time that the simulated collector needs to achieve the temperature rise of 20 K.

With a collector thermal power of 0.6 ■ 800W/m[2] this time delay is 517 s (approximately

8.5 minutes). So the collector with CJ3 needs 3.6 % longer to increase the system temperatures by 20 K than the realistic one.

Furthermore, it has to be kept in mind that the measured J.3-capacity of this collector was even higher than in our example above (40 instead of 21.4 kJ/m2K). Here the error of the energetic description of the storage charging process amounts to 620 kJ/m2, which corresponds to a delay of 1292 s. Hence the collector gain in the period under consideration is underestimated by 9%.

Moreover, it has to be kept in mind that the process of charging the storage is a significant and typical one, since it is a real everyday process for solar thermal systems.