The comparisons of simulated and measured outlet temperatures of the solar collectors and temperature inside the hot tank are shown in Fig. 7. We chose to compare the values from 9 AM to 5 PM when the outlet temperature of solar collectors exceeds hot tank temperature. In this case, the mean absolute error is 10.6% on the outlet temperature collectors and 2.8% on the hot tank. Our solar collector model based on efficiency curves slightly overstated the performance of solar collectors, while the evolution of the hot tank temperature determined by our simulation seems fairly close to our experimental platform value.

4.1. Analysis of the outlet temperatures of generator, condenser and cold tank

The comparisons of simulated and measured outlet temperatures of the generator, the condenser and temperature inside the cold tank are shown in Fig. 8. We chose to compare the values during the functioning period of the absorption chiller (from 11:10 AM to 4:20 PM). The mean absolute error on the outlet temperature of the generator is about 5% while the mean absolute error of the condenser is about 6%. These errors are quite high but in general, our model follows quite well the actual evolution of these temperatures. In contrast, the mean absolute error on the cold tank temperature is about 40%. This value is very high, explained by the fact that our model of absorption chiller is very simplified.

4.2. Analysis of the generator, the evaporator and the cooling tower powers

Looking at the temporal evolution of the powers, visible on the fig. 9 and 10, and calculating mean absolute error on the generator (16.6%), the evaporator (30.2%) and the cooling tower (10, 6%) we reach the same conclusion as seen before. Therefore it is imperative to establish a more accurate absorption chiller model particularly for start-up (peak powers on the actual values of the generator and the cooling tower) and stop (real refrigerated production enduring 10 minutes after shutdown).