PSI - Issue 84

Silvia Manarin et al. / Procedia Structural Integrity 84 (2026) 231–239

233

established in the literature and described in detail by Perséy (1985), provide distinct fitting strategies tailored to different surface characteristics. In the present test campaign, all three algorithms were applied in order to enable a comparative analysis and to identify the method that appears to yield the highest accuracy. Once the instrument has been connected to the power supply, but prior to the activation of the X-ray source, the emitted laser beam is used to align the instrument with the longitudinal axis of the strand or wire under investigation. When the correct distance from the specimen is reached, the main control unit display will indicate a green numerical value approaching zero. Only at this stage, after deactivating the laser, may the measurement procedure be initiated via the PC using the dedicated EDGE software. 3. Definition of the Accuracy Parameter In this experiment, as fully described by Manarin (2025), the ratio between the effective stress (measured by a calibrated load cell) and the stress estimated by the XRD system is defined as: = (1) where is the effective applied stress [MPa] and the stress estimated by the XRD system [MPa]. A value of θ ≈ 1 indicates high measurement accuracy. Values of θ > 1 suggest that the XRD device tends to underestimate the real stress, while θ < 1 indicates an overestimation. Ensuring that θ remains close to 1 is of critical importance: indeed, an overestimation may mask possible degradation of prestressing strands, which is particularly dangerous if no visible cracking is detected, potentially resulting in missed diagnosis of structural deterioration. Conversely, underestimation is not necessarily a conservative scenario; it may lead to an overestimation of the residual load-carrying capacity and could induce premature, unnecessary maintenance interventions, leading to avoidable economic costs. To assess the consistency and reliability of the measurements, the following statistical indicators are calculated for each set of experimental conditions: • Mean value of θ , indicating the average measurement accuracy; • Coefficient of Variation (CoV) to quantify the dispersion of θ values around their mean, the CoV is calculated as follows: = (2) where SD θ is the standard deviation and μ θ is the mean of the θ values. A low CoV indicates high measurement repeatability and limited dispersion, which is essential for a reliable structural assessment. The evaluation of both θ and CoV is particularly critical in engineering applications that require the estimation of residual prestress. Inaccurate or highly scattered measurements may lead to an underestimation of the shear forces associated with tendon curvature and an overestimation of the residual load-bearing capacity, potentially resulting in non-conservative safety assessments or, conversely, unnecessary and costly strengthening interventions triggered by an apparent prestress loss. For these reasons, ensuring both accuracy and consistency of XRD measurements is fundamental for the condition assessment and safety verification of existing prestressed structures. 4. Experimental Validation of XRD Stress Measurements on Prestressing Wires To assess the performance of the XRD system, each measurement was compared against a reference stress value applied during mechanical testing. Initially, tests were conducted on single isolated wires and subsequently extended to the outer wire of a seven-wire strand.