PSI - Issue 84

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

1184

lower than that assumed at the design stage. Since prestress plays a key role in both serviceability and shear resistance, unexpected prestress losses may substantially reduce the safety margins of existing structures (Mantelli, 2023). Several investigations documented severe prestress losses in existing bridges, often attributed to creep, shrinkage, relaxation, corrosion and construction-related deficiencies (Czaderski & Motavalli, 2006; Fernández Ruiz & Muttoni, 2009). In response to these issues, the Italian Guidelines for the classification and management of existing bridges (MIT - Ministero delle Infrastrutture e dei Trasporti, 2022) introduce a multilevel assessment framework, explicitly requiring advanced diagnostic investigations (Level 4) when structural deficiencies or uncertainties are identified. Within this framework, the evaluation of residual prestress is recognised as a critical aspect, especially for post tensioned bridges, where deterioration mechanisms may affect internal tendons without clear external evidence. Given these vulnerabilities, numerous studies (Azizinamini et al., 1996; Bagge et al., 2017; Bellini & Mazzotti, 2024; Kralovanec et al., 2022; Kralovanec & Prokop, 2021) have underlined the importance of in-situ measurements of prestressing stress to better quantify unexpected losses induced by relaxation, shrinkage, creep, or corrosion. Various destructive and non-destructive techniques have been developed for this purpose (AICAP, 2016; Bellini & Mazzotti, 2024; Kesavan et al., 2005; Martinello, 2021; Moravčík & Kral’ovanec, 2022; Trautner et al., 2011; Zanini et al., 2022). Among them, the saw-cut stress-release method has gained increasing attention due to its relative simplicity, limited invasiveness and ability to provide direct in-situ stress measurements (Bellini & Mazzotti, 2024; Lupoi & De Benedetti, 2021; Mantelli, 2023; Martinello, 2021; Romano & Mazzotti, 2022). Previous laboratory studies demonstrated that the saw-cut method can provide reliable and repeatable results when applied to full-scale beams and simplified specimens under controlled conditions (Mantelli, 2023; Romano & Mazzotti, 2022). Nevertheless, its application under real in-situ conditions remains challenging, due to uncertainties related to boundary conditions, load distribution and incomplete knowledge of prestressing details. 1.1. Scope and motivation This research focuses on the saw cut method application on six prestressed girder bridges subjected to detailed Level 4 (L4) investigations in accordance with the Italian Guidelines. In particular, Bridge 1,2 and 3 have been selected since they present notable signs of degradation, such as corrosion, which may lead to prestress reduction, or atypical shear cracking patterns that are inconsistent with the expected behaviour of prestressed structures under the actual loads, reinforcement layout, and assumed prestress levels. For each bridge, an extensive diagnostic campaign was carried out, including both material and structural investigations as prescribed by the L4 methodology. Finite element (FE) models were developed to support the interpretation by estimating load distribution and the bending moment at the tested sections, while analytical calculations were used to compute the expected concrete stresses for comparison with measurements. The overall goal is to highlight both the practical potential of the method and the main sources of uncertainty affecting in-situ residual prestress assessment. 2. Experimental program and saw cut method Based on the findings reported in (Mantelli, 2023), pairs of cuts are carried out at a distance of 6 cm, with a depth of 3 cm and a minimum length of 15 cm. This configuration had previously demonstrated to provide reliable and repeatable results in the laboratory, both on full-scale PRC beams and on simplified RC specimens used for the test validation. Moreover, the outcomes remain on the safe side of the assessment. The experimental procedure consisted of gluing two resistive strain gauges (SGs) to the web surface between the two cuts. The two strain gauges, SG_A and SG_B, are each 30 mm long and have a nominal resistance of 120 Ω , as in (Mantelli, 2023). Their signals were continuously recorded throughout the tests. For each test, sufficient time was allowed for signal stabilization (at least 5 minutes). To ensure redundancy and improve reliability, two strain gauges were always installed, and their average reading was used in the analysis. The method is based on measuring the strain variations caused by the local perturbation of the stress field induced by the cuts. Assuming linear-elastic material behaviour, these strain variations can be directly related to the in-situ stresses that existed prior to cutting.