PSI - Issue 64

Cedric Eisermann et al. / Procedia Structural Integrity 64 (2024) 1224–1231 Eisermann et al./ Structural Integrity Procedia 00 (2019) 000–000

1225

2

1. Introduction As the bridge inventory in Germany ages, the structural assessment of existing bridges is becoming increasingly important. At present, existing bridges are mostly assessed using the partial factor method (PFM) according to EN1990:2002. This approach was originally developed for the design of new structures and uses partial factors to account for uncertainties in both material resistances and actions. However, the assessment of existing structures differs from the design of new structures (Gino et al. 2020; Steenbergen et al. 2015). Firstly, the remaining service life of existing structures is typically shorter than the standard design life of 50 to 100 years assumed for new structures. Secondly, information about the actual structural condition and the magnitude of permanent and variable actions already exists or can often be measured through in-situ and laboratory testing or monitoring. As a result, uncertainties in materials and actions can be reduced or partially eliminated (Lara et al. 2021). In the best case, reserves of structural capacity can even be identified, as demonstrated in Gebauer et al. (2024). Hence, a strict application of the design-oriented PFM for existing structures may yield conservative results, potentially leading to unnecessary rehabilitation, strengthening and replacement measures (Enevoldsen 2001; Cremona and Poulin 2017). To address this issue, fib Bulletin 80 (Caspeele et al. 2016) proposes two methodologies: The design value method (DVM) and the adjusted partial factor method (AVFM). According to Gino et al. (2020), both methods are able to adjust partial factors according to updated target reliability levels and the remaining service life for existing structures. In addition, the probabilistic nature of the methods allows in-situ measured variations of materials and actions to be incorporated into the adjustmentprocess. This paper investigates how DVM and AVFM can be used to specify the self-weight related partial factor for existing bridges based on precise geometry measurements. For this purpose, the Nibelungen Bridge in Worms, which serves as the validation structure for the German Research Foundation funded priority programme SPP 100+, is used as an example. As a part of this programme, an as-designed 3D model based on inventory technical drawing as well as an as-is 3D model based on point clouds of the bridge were created. Before the practical application, the basic concepts of DVM and AVFM are briefly introduced. Then, the as-designed model and as-is model of the Nibelungen Bridge are compared and the variation of bridge geometry is determined based on the derived deviations. Subsequently, object-specific dead load partial factors for the structure are defined using DVM and AVFM, and compared with the value proposed by the German Recalculation Guideline (Bundesministerium für Verkehr, Bau und Stadtentwicklung 2011). Finally, conclusions resulted from the model comparison and the application of the different methods are summarized, and an outlook for further research is given. 2. Methods for adjusting the dead load partial factor 2.1. German Recalculation Guideline The Recalculation Guideline (Bundesministerium für Verkehr, Bau und Stadtentwicklung 2011) applies in Germany to the assessment of existing bridges that were not designed and constructed according to current standards. It provides engineers and planners with special regulations that enable better utilization of structural and material reserves, without compromising the reliability level required by EN1990:2002 for new structures. For concrete bridges, the Recalculation Guideline allows a reduction of the partial factor γ G from 1.35 to 1.20 if the dead load distribution is determined through representative and sufficient in-situ measurements of the bridge geometry and concrete density with considering the reinforcement content from inventory plans. Deviations from the planned state are directly incorporated into the object-specific dead load distribution. In this way, the partial factor γ G no longer needs to account for statistical variations in geometry and material properties and only contains the model uncertainty γ Ed,G , conservatively estimated at 1.20 (Maurer et al. 2012). As the Recalculation Guideline does not specify requirements for the accuracy and the extend of the measurements, the reduction must always be evaluated on a case by-case basis.