Cross-comparison of Non-Linear Seismic Assessment Methods for Unreinforced Masonry Structures in Groningen

Abstract: A large amount of low-rise unreinforced masonry structures (URM) can be found in Groningen, the Netherlands. More and more induced earthquakes with short duration have been detected in this region due to gas exploitation. Local unreinforced masonry (URM) buildings were initially not designed for withstanding seismic actions, so that unexpected damage may occur due to their vulnerability, raising insecurity among residents. Existing low-rise masonry buildings in Groningen can be divided into different categories based on their characteristics. Two types of residential masonry buildings that fulfil the prerequisites for performing non-linear seismic assessment are chosen to be studied in this thesis project, including the terraced house and the detached house. The seismic assessment of structures requires the use of both a discretization method and a seismic assessment method. The discretization method is used to translate the mechanical model into a finite element model used for the numerical analysis. Several methods have previously shown to be applicable for seismic assessment, but this work investigates the implications of using a continuum model (CM) and an equivalent frame model (EFM) approach to discretization in the general-purpose finite element package described in n DIANA-FEA-BV (2017). The continuum model approach adopted was in a previous work by Schreppers et al. (2017) validated against experimental results and is as such deemed representative of the physical behaviour of the mechanical models investigated. An equivalent frame model approach to be used with DIANA is proposed in the work by Nobel (2017). The continuum model approach uses continuum elements with a constitutive model developed for the seismic assessment of masonry structures. This constitutive model captures both shear and flexural failure mechanisms. The equivalent frame model approach uses a combination of numerically integrated beam elements and nodal interfaces, each with a distinct constitutive model, thus decoupling the description of the flexural and shear behaviour. This approach aims to capture the macro-behaviour at the structural level. The applicability of the proposed equivalent frame model approach is evaluated by how well it replicates the validated continuum model approach results. The two discretization methods described are evaluated using two types of seismic assessment methods. The first seismic assessment method used consists of first performing a quasi-static non-linear pushover analysis (NLPO) on the model. This results in the pushover curve, which describes the global behaviour of the model under an equivalent lateral load based on the fundamental mode shape of the structure. The pushover curve is then used with the N2-method described in EN1998-1 (2004) to assess at which peak ground acceleration (PGA) that the model reaches the near-collapse (NC) limit state. The second seismic assessment method consists of performing dynamic non-linear time-history analyses (NLTH). This method uses recorded accelerograms to impose the inertial forces. The PGA for the accelerogram where the near-collapse limit state is reached is compared to the PGA from the use of the N2-method. The applicability of the pushover analysis in conjunction with the N2-method is evaluated by how well it replicates the PGA found from the time-history analyses and by how well it replicates local failure mechanisms.  Therefore, the main objectives of this project can be described by the following two questions: i. To what extent can the equivalent frame method be applicable as a proper discretization method for pushover analyses and time-history analyses of low-rise unreinforced masonry residential buildings in the Groningen region? ii. To what extent can the non-linear pushover method be adopted to assess the seismic behaviour of low-rise unreinforced masonry residential buildings in the Groningen region? The applicability of the equivalent frame model showed to vary. For describing local failure mechanisms its applicability is poor. Further work on connecting the edge piers to transverse walls is needed. For seismic assessment using the N2-method the applicability of the equivalent frame model approach is sensible. The conservative displacement capacity counteracts the fact that it is worse at describing local unloading, which produced a larger initial equivalent stiffness of the bi-linear curves in comparison to the continuum model. For seismic assessment using the time-history signals, its applicability is possible. While it could show different behaviour in terms of displacement and damping forces, it still showed a similar PGA at the near-collapse limit state for the cases at hand. The seismic assessment of the terraced and detached houses by the N2-method is similar to the seismic prediction by applying time-history analyses. However, there are still some variations in the initial stiffness, force capacity and displacement capacity between these two assessment methods due to the assumptions and limitations in this study. Overall, considering the pros and cons of the quasi-static pushover method, it is deemed applicable during the seismic assessment of the unreinforced masonry structures in the Groningen area.