*3.1. February–March 2020 Results*

In the first week of monitoring (from 10 February 2020 to 14 February 2020), the HVAC-1 and HVAC-2 systems were loaded with the same heat transfer fluid (water–glycol 30% vol). During this period, a short database was acquired, to be used as reference data for the experiment. Then, on 17 February, nanofluid was loaded in the HVAC-1 and the next 7 working days were used to balance both the heat pumps by performing preliminary tests. After such balance, energy consumption comparison between both the machines was restarted on 26 February and continued until 9 March.

Figure 5 shows the hourly COP comparison between HVAC-1 and HVAC-2, while Figure 6 shows the average daily COP comparison and the daily COP ratio between HVAC-1 and HVAC-2.

**Figure 5.** Hourly COP comparison between HVAC-1 and HVAC-2 (10 February 2020–9 March 2020).

**Figure 6.** Average daily COP comparison and daily COP ratio between HVAC-1 and HVAC-2 (10 February 2020–9 March 2020).

In the first 3 days of monitoring (from 26 February to 28 February), after nanofluid loading, the mean increase in performance was 9.36%. However, the best performances were achieved in the last two weeks of experimental data acquisition, when the average increase in performance was 10.8%. On 12 March 2020, the HVAC systems were shut off.

During the acquisition period, the density of the nanofluid was measured weekly by sampling from the system in operation. Since it was 1079 g/L over the test period, a constant concentration of 2% of nanoparticles was ensured inside the system fluid.
