**4. Simulation Results**

In order to evaluate the benefits of the green roof for a longer period of time, commercial software TRNSYS 17 was used as theenergy simulation tool of the experimental setup. During the simulation, conventional and green roof were modeled based on their constructive characteristics and energy relationships. TRNSYS model was defined with 6 layers with different thickness: gravel or green roof (0.07 m), polystyrene insulation (0.05 m), membrane (0.001 m), concrete (0.075 m), building slab (0.17 m) and air chamber (0.5 m). The estimated U-value for the traditional roof was considered 0.518 <sup>W</sup>/m2·◦K, while for the green roof, estimation was 0.409 <sup>W</sup>/m2·◦K.Solar absorption of the roof was modelled with 0.8 for the gravel and 0.2 for the green roof.

Figure 13 shows the input data for the simulation, external conditions, building characteristics and temperature set points, and the deduced outputs, temperatures in the different layers of the roof and cooling and heating demand along the year.

**Figure 13.** Parameters of TRNSYS Model.

Initially, rooftop models developed in TRNSYS were validated using collected experimental data. Validation was carried out in two stages, using a down-top approach: first, it was checked that modelled temperatures were similar to the experimentally measured at the rooftop layers along a week period. Then, the model was calibrated using the data from experimental daily profiles in the summer period.

In the case of the week validation, temperatures of the different layers of the roof were determined with the model and results (Figure 14) were compared with the corresponding experimental data. Maximum simulated temperatures of 59 ◦C were obtained during summertime in the gravel, while registered data during the same time period revealed a temperature range of 45 ◦C to 55 ◦C.

**Figure 14.** Simulated temperatures of the initial rooftop.

One-day validation was performed using as a reference two similar days in terms of environmental parameters (temperature, humidity etc.) and use-of-space (workday) during the experimental campaign, one for each of the two different types of roof considered, on 24th July 2017 (conventional roof) and 31st July 2018 (green roof), in order to deduce the parameters used in the simulation. Comparison of the simulation results, detailed in Figure 15, with the experimental data plotted in Figure 11, shows a good enough agreement, thus, the simulation for the entire year could be addressed using these parameters. In order to complete the simulation task, a selected time window during the day with stable conditions

was identified. This time span was from 11:00 to 13:00. During this time span, TRNSYS simulations were addressed considering a set-point of 22 ◦C in heating demand (October to March) and 26 ◦C for cooling needs (April to September).

**Figure 15.** One-day validation of the models for conventional and green roof.

Figure 16 shows temperature evolution in the building outdoors (T4) and indoors (T5), as well as the temperature below the growth medium (T1). As can be observed, top roof temperatures decreased both in summer (40 ◦C to 35 ◦C) and winter (10 ◦C to 5 ◦C), which requires to compare the benefits for cooling in summer versus the negative effects in heating during winter period.

**Figure 16.** Results for temperatures evolution along the year.

Recorded data for cooling energy demand in both scenarios (conventional and green roof) showed an energy savings of approximately 25% in cooling energy demand, decreasing the maximum peak power demand by 33%. Heating energy demand in both scenarios (conventional and green roof) is presented in Figure 17. In this case, the results show that energy heating demand increased12% in the green roof scenario. Moreover, the maximum energy peak due to heating also rose6% with the green roof in comparison to the conventional rooftop, due to the reduction of solar heat gain reaching the building.

**Figure 17.** Results of the heating demand in both rooftops.

As a summary, results of the simulation are presented in Tables 4 and 5, which show total energy consumption, saving in cooling and heating mode and the average power saved along the operation time.


**Table 4.** Annual energy consumption for the conventional and green roof.

**Table 5.** Simulation results in energy savings (green roof vs. conventional rooftop).


These results are compatible with the experimental values detailed at Table 2, which indicate a net gain in energy saving for the entire year, in the order of 19%. In contrast, an energy demand increase of 6% is noted, due to the requirement for additional heating in the winter period.
