In Fig. 2 the cooling capacity of the ILK test data is shown as function of AAt*. The modified characteristic temperature difference AAt* has been calculated from the arithmetic means of the measured inlet and outlet temperatures assuming a constant Duhring parameter (5=1.15). In addi­tion Fig. 2 depicts the results of a linear data fit with constant and variable characteristic parameters sE and rE. For the determination of the variable parameters sX and rX the data have been grouped into five classes between the minimum and maximum temperature lift included in the ILK data base. In Fig. 2 only the groups with lowest and highest temperature lift (AtACE = 13±2 K and AtACE = 29±2 K) are high-lighted with grey symbols to illustrate the variation of sE, rE and, AAtmin, respectively. If only these two groups are used, the coefficients sE1 and sE0 for the linear dependency of sE(AtACE), in equation (6) can be found from the upper two equations in Fig. 2. However, for the parameter list of Type 177 all five groups of AtACE-intervals have been used as supporting points in a linear least square fit to determine all the values sX,1, sX, o and rX1, rX,0.

As seen from Fig. 2 the general trend of the cooling capacity is predicted correctly when cal­culated by the characteristic equation with constant (solid line) or variable (closed symbols) parameters. If the variable parameters are used, also the scatter of the measured data is reproduced slightly better. Nevertheless, a relatively large deviation remains. This may be an effect of measurement uncertainties and / or different methods and time intervals of chiller evacuation during the test procedure. These effects are not included in the method of characteristic equations.