Adaptation of Type 307 to EAW Wegracal SE 15

For the final adaptation of Type 307 the values fNC and fDEI are determined using new performance data files. These have been prepared from laboratory test measurements of an EAW Wegracal SE 15 chiller carried out by ILK-Dresden. After normalising these data two different files have been prepared (PD1 and PD2). This difference is explained by Fig. 1 where fNC is shown for increasing hot water inlet temperatures tGi while chilled and cooling water inlet temperatures are constant. From the PD1 curve it is seen that the TRNSYS look-up approach interpolates stepwise between the neighbouring supporting points of the data file, i. e. the ‘tines’ of the solid curve. Since the supporting points in the performance file are not allowed to change, i. e. once the first values have been fixed the same values have to be used for all combinations, there is a slight difference to the measured values. Furthermore it can be seen that extrapolation to lower as well as to upper values is not possible. In these cases the TRNSYS simulation gets stuck at the lower or upper end (fNC = 0.89 or 1.19, respectively). This behaviour is important to know for users working with Type 107 based models, since TRNSYS neither interrupts the simulation nor writes a warning to the list file. This could result in a wrong plant dimensioning in the worst case and has to be avoided. For example, feeding the model with tGi = 70°C, tACi = 27°C, and tEi = 18,5°C will return a value of fNC « 0.9 from the look-up approach with PD1. In reality the fNC is considerably lower (cf. Fig. 1 for PD2 and additional independent data points). Thus, the solar cooling system would have been simulated too efficient for this operating condition.

Подпись: 60 70 80 90 100 Hot water inlet temperature tGi / °C Fig. 1. Fraction of nominal capacity as a function of the entering hot water temperature (Type 307) Since the data reading and interpolation procedure in Type 307 is the same as in Type 107 (stepwise interpolation, no extrapolation) a trick has been applied to avoid the above mentioned problems. From a linear fit of the measured values a global slope is achieved for a certain group of independent variables (e. g. in Fig. 1 tGi = var., tACi, tEi = const.). The global slope is a result of the fit over the whole range 80°C < tGi < 95°C at fixed tACi/tEi conditions («fixed temperature lift, see section 2.4).

Obviously it is steady enough to allow extrapolation. This is confirmed by two additional independent points at lower driving temperatures which have not been used in the fit for the global slope. The new interpolation limits (e. g. in Fig. 1 fNCmin = 0.54 and fNC, max = 1.30) have to account for physical and/or safety limits. Up to now, for the EAW Wegracal SE 15 chiller they have been set to the known temperature limits of operating conditions where measurements are available

from ILK test data and from the plant at Ebner Solartechnik. Now the new PD2 performance file includes only the fNC and fDEI values at these borders (i. e. only two points per tGi/tACi/tEi combination). During simulation the interpolation — which is still done stepwise, but just between these end points — merges into a global interpolation.