PSI - Issue 44

Ubaldo Saracco et al. / Procedia Structural Integrity 44 (2023) 721–728 Ubaldo Saracco et al./ Structural Integrity Procedia 00 (2022) 000–000

724

4

3.3. Assessment of the progress of corrosion damage As previously introduced, this paper also predicted the thickness reduction of the bridge members due to the corrosion phenomena. Several analyses were conducted using forecasting models provided by scientific literature and current legislation. The reason behind the use of these models is based on the lack of in-depth data regarding the thicknesses of the bridge elements and, therefore, of their decrease after about 95 years from the construction of the artwork. Since there were no indications regarding possible maintenance interventions and use of protective paints on the structure, they were not considered in the analysis process. Finally, being a restricted asset, it was not possible to conduct invasive investigations with sampling of elements, but only visual investigations and some considerations on the types of corrosion in progress. The used prediction models for corrosion can be traced back to the following three standard and literature references: EN ISO 9223 (CEN 1992a) and EN ISO 9224 (CEN 1992c) standards . ISO 9223 standard proposed a classification system for environmental corrosivity, while ISO 9224 standard provided formulae for calculating the corrosion rate according to the exposure classes (C1-C5) provided for by the previous code. The first provisions delivered by ISO 9224 standard refers to exposure periods into a corrosive environment not exceeding 20 years. Nevertheless, the accuracy of the formula for time periods greater than 20 years (eq. 1) cannot be verified because the lack of sufficient experimental data: d w (t) i = r av ∙ 10 + r lin ∙ (t – 10) ∙ t for t ≥ 10 years (1) where d w (t) i is the corrosion depth for the time interval considered (μm) , r av is the average annual corrosion rate (μm/year) and r lin is the steady- state corrosion rate (μm/year). Despite the lack of accuracy for t>20 years, eq. ( 1) was herein used considering 125 years as useful life of the bridge. Therefore, for this case study, a degree of exposure C3 was chosen and the following minimum and maximum corrosion rates were assumed as follows: 8.3 < r av < 17 [μm/year] and 4.9 < r lin < 10 [μm/year] When selecting the highest values, the thickness reduction value (in mm) is as follows: - Upper limit d(t) = 1.32 mm ISO 9224 (2012) standard and Kotes et al.’s paper (2018) These two references report a second formulation for times exceeding - though not overly - 20 years through the following formula: d w(t > 20) = r cor ∙ [20 b + b(20 b-1 )(t-20)] (2) in which r cor is the first year of exposure’s corrosion rate, assumed -on the safe side - faster than r av and r lin . The r cor value, referred to a category C3, is 25 < r cor < 50 [μm/year], whereas 20 b and b(20 b-1 ) are reported in ISO9224 standard (2012). The chosen values refer to a B1 class Carbon Steel and led to the following result: - Upper limit d w(t > 20) = 0.8958 mm Rizzo et al.’s paper (2019) This model aims to extend the calculation of the corrosion forecast to a time period greater than 20 years considering three alloy types: mild carbon steels, weathering steels and wrought irons. The paper defines the following equation, which provides the best approximation of the fitted data: ( ) = + + + ∀ t > 0 (3) where d w (t) is the thickness loss (mm/year); t is the exposure time interval (years); p 1,…,4 are the interpolating damage curves coefficients. From eq. (3), with t = 125 years, d w (t) = 0.4708mm is obtained. • • •