PSI - Issue 62

Carla Assunta Trifarò et al. / Procedia Structural Integrity 62 (2024) 57–64 C.A. Trifarò et al./ Structural Integrity Procedia 00 (2019) 000 – 000

62

6

3.3. Clustering and further considerations In order to better interpret the available data, a first attempt of clusters of homogeneous structures were defined. Thus, the structures were grouped in seven homogeneous clusters according to the type of structure, year of construction, static scheme, deck material and attention class. The defined clusters are summarized in Table 1.

Table 1. Defined clusters.

Cluster Type of structure

Year of construction Static scheme

Material

Attention class

#1 #2 #3 #4 #5 #6 #7

Bridges and viaducts Bridges and viaducts Bridges and viaducts Bridges and viaducts Bridges and viaducts

‘60 ‘70 ‘80

Simply supported p.c. (post-tensioned) High

Simply supported p.c. (post-tensioned) ≥ Medium -High Simply supported p.c. (post-tensioned) ≥ Medium -High

’60 - ‘00 ’60 - ‘00 ’60 - ‘00 ’60 - ‘00

Simply supported p.c. (pre-tensioned)

≥ Medium -High

Continuous

p.c. (pre-tensioned)

≥ Medium ≥ Medium ≥ Medium

Overpasses Underpasses

Gerber

All All

Slab

In this sense, it is worth mentioning that the clustering procedure was of the “ exclusive ” type and thus a single structure may belong to only a cluster. The probability distribution functions of the obtained D r for each cluster are plotted in the graphs of Fig. 5.

Fig. 5. Probability distributions of D r in the considered clusters

Generally, the following main conclusions can be drawn. The worst conditions have been obtained for cluster #3, namely the one composed by simply-supported post- tensioned decks realized in the ’80 of the last century (being D r,m =10). For this group, a good homogeneity of the results was also obtained, given that C.o.V. resulted 20%. Conversely, for clusters #4 and #5 (i.e., pre-tensioned reinforced concrete decks) the dispersion of the obtained results cannot highlight a clear picture of the situation. This can be related to the fact that these groups include structures realized between the ’60 s and ’00 s. A seemingly controversial outcome was obtained for cluster #1, which includes structures realized more than 50 years ago. For this group the low value of the mean defect index ( D r,m = 5.7) should be traced back to recent maintenance interventions performed on such structures. Finally, as far as the other clusters are concerned (i.e., #2, #6, and #7) similar results to the whole stock were obtained in terms of both mean values and dispersions.