PSI - Issue 84

Giulia Rossini et al. / Procedia Structural Integrity 84 (2026) 1183–1190

1189

Table 2: saw cut tests results.

y [cm] µ , [MPa] , [MPa] [MPa] , ,

Prestressing losses [%]

x [cm]

Bridge Test name

1 1 1 1 1 2 2 2 3 3 3 3 3 4 4 4 4 5 5 5 6 6 6

S2-G4 1 1224 37 41.27 -1.42 S2-G4 2 1282 43 62.30 -2.14 S2-G4 3 1342 45 57.54 -1.97 S1-G4 1 515 48 66.52 -2.28 S1-G3 1 488 43 47.20 -1.62 S3-G10 1 126 42.5 58.90 -2.14 S3-G10 2 156 36 73.66 -2.67 S3-G10 3 205 40 68.13 -2.47 S1-G3 1 1056 41.5 27.64 -1.00 S1-G3 2 1026 45 12.62 -0.46 S1-G3 3 998 45 24.63 -0.89 S1-G4 1 1062 43.5 23.16 -0.84 S3-G2 1 1633 25.5 79.39 -2.80 S3-G2 2 1633 51 70.34 -2.49 S3-G3 1 1693 27 63.53 -2.24 S3-G3 2 1693 50 67.99 -2.40 S1-G22 1 797 37.5 62.36 -2.09 S1-G22 2 765 30 73.14 -2.45 S1-G22 3 738 31 51.58 -1.73 S1-G4 1 770 40.5 165.65 -6.11 S1-G4 2 747 32.5 203.69 -7.52 S1-G4 3 710 33 216.52 -7.99 S1-G4 2 1031 46 5.50 -0.20

-1.58 -2.38 -2.19 -2.53 -1.80 -2.38 -2.97 -2.74 -1.11 -0.51 -0.99 -0.93 -0.22 -3.11 -2.77 -2.49 -2.67 -2.32 -2.72 -1.92 -6.79 -8.36 -8.88

-2.93 -2.52 -2.38 -2.65 -2.60 -2.81 -2.79 -2.94 -0.76 -0.99 -1.13 -0.97 -0.52 -2.91 -4.15 -2.53 -3.75 -2.65 -2.59 -2.56 -7.99 -9.02 -8.95

0.48 0.85 0.83 0.86 0.62 0.76 0.96 0.84 1.31 0.46 0.79 0.87 0.39 0.96 0.60 0.89 0.64 0.82 0.99 0.70 0.77 0.83 0.89

0.54 0.94 0.92 0.96 0.69 0.85 1.06 0.93 1.46 0.52 0.88 0.96 0.43 1.07 0.67 0.98 0.71 0.88 1.05 0.75 0.85 0.93 0.99

62 62 62 50 50 50 50 50 50 50 50 50 50 34 34 34 34 36 36 36 20 20

20 Based on the experimental results, prestress losses were estimated and expressed as percentages, by calibrating the analytical model so that the average ratio , for each individual bridge was as close as possible to 1. The prestress losses obtained for Bridges 4, 5 and 6, which do not exhibit significant signs of degradation, are consistent with the prestress losses typically assumed at the design stage and commonly reported in design codes and recommendations. According to Eurocode 2 calculation (UNI EN 1992-1-1, 2024) typical long-term prestress losses for existing prestressed concrete structures generally range between 25% and 35%, depending on material properties, prestressing technology and environmental conditions. In agreement with these indications, prestress losses of 36% and 34% were obtained for Bridges 4 and 5, respectively. These bridges are approximately 40 years old and can therefore be assumed to have undergone the full development of long-term losses. Conversely, Bridge 6, which is only 10 years old, exhibited lower prestress losses, equal to approximately 20%, consistent with the shorter duration of creep, shrinkage and relaxation effects. Conversely, for Bridges 1, 2 and 3, which exhibit pronounced shear cracking patterns, the prestress losses inferred from the comparison between experimental and analytical stresses are markedly higher and fall outside the range of typical design values. These results suggest that the observed cracking patterns, which are anomalous for prestressed concrete members under service conditions, may be associated with prestress losses higher than expected. While other contributing factors cannot be excluded, the findings support the hypothesis that excessive prestress losses may play a key role in the development of these atypical shear cracking patterns. 6. Conclusions This paper presented the application of the saw-cut stress-release technique for the evaluation of residual prestress in six existing prestressed concrete girder bridges subjected to Level 4 investigations according to the Italian Guidelines. The method was applied on girder webs under real in-situ conditions and interpreted through a combined experimental, analytical and numerical framework. The experimental results indicate that, when an appropriate test configuration is adopted and measurements are carefully executed, the saw-cut method can provide reasonable and repeatable estimates of concrete stress. The comparison with analytical stress calculations, supported by finite element models used to estimate internal actions at