PSI - Issue 62

Ettore De La Grennelais et al. / Procedia Structural Integrity 62 (2024) 763–772 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

769

7

Fig. 5. The location of accelerometers in a schematic plan of a single bay.

Potential sources of error that could impact the comparison between calculated and observed values are: • Numerical method uncertainties (err FEM ), for the adopted FEM method, a theoretical error of ±5% can be assumed on the actual value. • Statistical uncertainties (err stat ): to ensure the repeatability of measurements across the various spans of the bridge, the following assumptions are considered: a) The same measurement method was consistently used. b) The same technicians were employed for measurements. c) The same measurement equipment was utilized. d) The measurement equipment was calibrated and certified. e) All tests were conducted on a single day. Given these considerations, by observing the values for the N = 4 spans analyzed, there is a small statistical distribution from which it is possible to calculate: Average: f av = ∑ f N i=1 i N Standard deviation: σ=√ ∑ (f i −f av ) 2 N i=1 N−1 Uncertainty, to assess the statistical precision of the observed mean frequency value: u=√ σ 2 N = √ σ N • Various uncertainties (errvar), there are also a series of uncertainties related to various aspects, including: a) Uncertainties about the modeled total mass (e.g., on the carried permanent loads); b) Uncertainties about the modeling of structural elements and the nature of internal degrees of freedom (e.g., considering perfect hinges at the ends of the truss elements rather than evaluating possible rotational rigidities); c) Uncertainties about the environmental test conditions (temperature, irradiation, etc.); d) Uncertainties about the tolerance and precision of the instruments used; e) Uncertainties about traffic conditions during the test, which may vary the considered mass conditions;