PSI - Issue 29

Fabio Sciurpi et al. / Procedia Structural Integrity 29 (2020) 16–24 F. Sciurpi et al. / Structural Integrity Procedia 00 (2019) 000 – 000

22

7

10 15 20 25 30 35 40

0 10 20 30 40 50 60 70 80 90 100

Temperature ( ° C)

θ Room X

θ Sensor Xs RH Sensor Xs

Relative Humidity (%)

0 5

RH Room X

01/05/2012 00.00

12/05/2012 09.45

23/05/2012 19.30

04/06/2012 05.15

15/06/2012 15.00

27/06/2012 00.45

08/07/2012 10.30

19/07/2012 20.15

31/07/2012 06.00

11/08/2012 15.45

23/08/2012 01.30

03/09/2012 11.15

14/09/2012 21.00

26/09/2012 06.45

07/10/2012 16.30

19/10/2012 02.15

30/10/2012 12.00

10/11/2012 21.45

22/11/2012 07.30

03/12/2012 17.15

15/12/2012 03.00

26/12/2012 12.45

06/01/2013 22.30

18/01/2013 08.15

29/01/2013 18.00

10/02/2013 03.45

21/02/2013 13.30

04/03/2013 23.15

16/03/2013 09.00

27/03/2013 18.45

08/04/2013 04.30

19/04/2013 14.15

Fig. 3. Trend of mean daily temperature and RH measured in Room X compared with the values in the showcase in the same room (Xs).

In this case, the lack of thermal insula tion of the room is reflected in higher and lower temperatures inside the display case. Although theshowcasesare not sealed, they reduce RHvariations respect those of theroomandguarantee a damping of RH changes grea ter than temperature ones. The RH trend inside the cases is rela ted to the indoor temperature trend rather than to the outdoor RHvariations. In Table 3, for the year monitored, maximum daily gradient of temperature (Δθ 24max ) and RH (ΔRH 24max ) for the analyzedenvironments are reported, togetherwith the PI of the thermo- hygrometric parameters analyzed (θ, RH, Δθ 24 , ΔRH 24 ).

Table 3. Maximum daily gradients and PI values of the thermo-hygrometric parameters. Room Δ θ 24 max (°C) Δ RH 24 max (%) PI θ (%) PI RH (%) PI Δ θ 24 (%) PI ΔRH 24 (%) X 2.3 9.1 30.8 40.6 49.7 49.6 Xs 1.6 1.6 30.0 86.7 74.1 98.4 XXII 2.8 7.5 52.8 62.8 32.0 73.8 XXVIII 3.0 7.5 95.3 60.9 64.8 68.5 Outdoor 22.6 63

The energy model has been calibrated considering two sample thermal zones: Room X and XXII which are both not provided with a ir conditioning system. Therefore, the quality of the model has been assessed adopting indoor temperature as calibrationcontrol variable. In particular, hourly values of indoor a ir temperature collectedduring the above-mentionedmonitoringcampaign have been compared with simulatedvalues of the same parameter, by means of three error indices: MBE, CV(RMSE) and Pearson index (r). Since va lida tioncriteria formodels calibrated by means of air temperature are not available (Roberti et a l., 2015), “La Specola” calibration output has been compared with acceptability ranges recommended formodel ca libratedon the basis of hourly values of energy consumption (Coelhoet a l., 2018; Giuliani et a l., 2016). For example ASHRAE Guideline 14/2002 (ASHRAE, 2002) suggest that MBE should be less than ±10% and CV (RMSE) below 30%. As regards as Person Index, it presents an optimal va lue of 1 corresponding to a perfect direct correla tion between measured and simulated va lues trends; however, a va lue bigger than 0.5 can be considered a minimum limit of acceptability for a good correlation (Pernetti et a l., 2014). During the itera tive simulation phase, the model has been optimized, varying the value a ttributed to the infiltration air change ra te, that have been regarded as the uncertain parameter mostly influential on simulation results. Considering an infiltration a ir change ra te of 1.0 h -1 , calibration