PSI - Issue 44

Annalisa Napoli et al. / Procedia Structural Integrity 44 (2023) 2182–2189 Annalisa Napoli, Roberto Realfonzo / Structural Integrity Procedia 00 (2022) 000–000

2185

4

by the authors in the available data; such reduction involved all the FRCM systems with the only exception of the S FRCM one. Specifically, for some datasets, the lack of values on the ultimate strain of the fiber mesh  f,u made their inclusion in the analyses not possible; other datasets were discarded since they were characterized by experimental ̅  values lower than 1 or at most equal to 1, thus making the FRCM confinement ineffective. The pie chart in Figure 1b shows the distribution per FRCM system of the datasets effectively used in the analyses which are 99 against the 116 compiled in the database, involving a total of 227 specimens against the 261 originally collected. The most significant reduction of datasets is for B-FRCM systems with 10 datasets removed, followed by the C-FRCM system (4 datasets removed), by the G-FRCM (2 datasets) and, lastly, by the PBO-FRCM system with 1 dataset removed. In particular, in the case of the B-FRCM system, the datasets used in the analyses are 13 (7 related to prismatic specimens and 6 to cylindrical one), covering the 13% of the database and collecting a total of 37 specimens (17 prismatic and 20 cylindrical). For the C-FRCM system, the usable datasets are 25, covering about 26% of the database and collecting a total of 50 prismatic specimens. For the G-FRCM system, the valid datasets are 26 (25 prismatic + 1 cylindrical), covering the 26% of the database and collecting a total of 53 specimens (49 prismatic and 4 cylindrical). In the case of PBO-FRCM, the datasets useful for the analyses are 10, covering the 10% of the database and collecting 10 prismatic specimens. Finally, in the case of the S–FRCM system, the usable datasets are 25, covering the 25% of the database and involving all the specimens originally collected in the database (N = 69).

a) Total specimens: N = 261

b) Datasets used in the analyses: n = 99

Prismatic specimens

Cylindrical specimens

Fig. 1. (a) Distribution of specimens per FRCM system within the database; (b) distribution of datasets per FRCM system within the database.

By focusing the attention on the usable datasets ( n = 99), noteworthy is the information on the masonry type (natural or artificial) investigated for each FRCM confinement system. In particular, artificial masonry (AM) specimens made of CBs were the most investigated ones; indeed, in the case of C- and PBO- FRCM systems, these represent the totality of collected datasets. Only in the case of the G-FRCM system, a large number of datasets refer to specimens made of natural masonry (NM), covering the following typologies (Napoli and Realfonzo 2022): • tuff units (TUs), typology very common in existing buildings of Southern Italy, n = 5 (prismatic); • limestone (LS), n = 4 (prismatic); • LS – calcarenite (CALC), very common in several regions in the Mediterranean area, which is natural limestone composed by sand and carbonate grains, n = 3 (prismatic) + 1 (cylindrical). TU masonry was the only typology investigated for the S-FRCM systems, while in the case of B-FRCM systems LS and CALC were considered for one and two datasets related to prismatic and cylindrical specimens, respectively. Finally, another important detail is related to the masonry mass density g m which is a parameter believed affecting the FRCM performance. To this purpose, indeed, the nonlinear strength model suggested by DT 215 (2018), similarly to what already done in DT 200 R1 (2013) for FRP confinement applications, accounts for the contribution of the masonry mass density g m (expressed in kg/m 3 ) in the estimate of   ; this parameter roughly considers the influence of voids and porosity of both the constituent materials and the masonry texture itself on the axial strength. However, looking at the 99 datasets considered for the analyses, it is observed that the value of g m is always provided in the scientific papers in the case of NM specimens; in particular, most datasets have g m values lower than 1600 kg/m 3 , with