Katrin Beyer

Related publications (243)

Seismic testing of adjacent interacting masonry structures – A pre-diction and a post-diction using equivalent frame approach

Katrin Beyer, Igor Tomic

The SERA project AIMS (Seismic Testing of Adjacent Interacting Masonry Structures) comprised a blind prediction competition related to a shake-table test of an aggregate of two buildings under two horizontal components of dynamic excitation. Each group was provided with a complete set of construction drawings, material properties, testing sequence and the list of measurements to be reported. This paper deals with the prediction submitted by the EPFL team. The prediction was based on an equivalent frame model with a newly developed macroelement able to simulate both the in-plane and out-of-plane behaviour of unreinforced masonry piers. This paper describes the model and presents results of the pre-diction and preliminary results of the post-diction. For the post-diction the same model as for the pre-diction was used and only the seismic input was updated to match the actually applied shake table motion. The paper concludes with future work needed to calibrate the model for the post-diction.

2022

The impact of uncertainties on post-earthquake response predictions of a building with unreinforced masonry walls

Katrin Beyer, Mathias Christian Haindl Carvallo, Ian Smith

In this paper, the impact of several sources of uncertainty on the seismic response of an experimentally tested building is assessed for the case of a post-earthquake scenario. The building comprises unreinforced masonry and reinforced concrete walls. It is modeled using an equivalent frame model and latin hypercube sampling is used to generate randomized nonlinear models of the building. A sensitivity analysis is performed in order to identify the model parameters that contribute the most to the response. To deal with several sources of uncertainty and take advantage of the information of the building at the end of the fifth ground motion sequence of the experimental campaign, a data-interpretation methodology called error-domain model falsification is implemented. This methodology involves discarding models that do not reasonably match the measured response in terms of post-earthquake frequency and damage grade, which is determined from the maximum roof displacement. The models that are not falsified are used to predict the response of the building for the final four runs, which is assessed here in terms of the maximum roof displacements. The falsification procedure significantly reduces the number of models that reasonably explain the observed behavior and improves the reliability of response predictions for all ground motions following falsification. Results are accurate in the sense that the experimentally obtained response lies always within the range of predicted responses.

2022

, ,

We present a method for segmenting cracks in images of masonry buildings damaged by earthquakes. Existing methods of crack detection fail to preserve the continuity of cracks, and their performance deteriorates with imprecise training labels. We address these problems by adapting an approach previously proposed for reconstructing roads in aerial images, in which a Convolutional Neural Network is trained with a loss function specifically designed to encourage the continuity of thin structures and to accommodate imprecise annotations. We evaluate combinations of three loss functions (the Mean Squared Error, the Dice loss and the new connectivity-oriented loss) on two datasets using TernausNet, a deep network shown to attain state-of-the-art accuracy in crack detection. We herein show that combining these three losses significantly improves the topology of the predictions quantitatively and qualitatively. We also propose a new continuity metric, named Cracks Per Patch (CPP), and share a new dataset of images of earthquake-affected urban scenes accompanied by crack annotations. The dataset and implementations are publicly available for future studies and benchmarking (https://github.com/eesd-epfl/topo_crack_detection and https://doi.org/10.5281/zenodo.6769028).

2022

Blind predictions of shake table testing of aggregate masonry buildings

Katrin Beyer, Andrea Penna, Igor Tomic

In many historical centers in Europe, stone masonry is part of building aggregates, which developed when the layout of the city or village was densified. The analysis of such building aggregates is very challenging and modelling guidelines missing. Advances in the development of analysis methods have been impeded by the lack of experimental data on the seismic response of such aggregates. The SERA project AIMS (Seismic Testing of Adjacent Interacting Masonry Structures) provides such experimental data by testing an aggregate of two buildings under two horizontal components of dynamic excitation. With the aim to advance the modelling of unreinforced masonry aggregates, a blind prediction competition is organized before the experimental campaign. Each group has been provided a complete set of construction drawings, material properties, testing sequence and the list of measurements to be reported. The applied modelling approaches span from equivalent frame models to Finite Element models using shell elements and discrete element models with solid elements. This paper compares the first entries, regarding the modelling approaches, results in terms of base shear, roof displacements, interface openings, and the failure modes.

2021

Investigating the cracking of plastered stone masonry walls under shear–compression loading

Katrin Beyer, Michele Godio, Amir Rezaie

Cracks are the most important source of information about the damage that occurs to unreinforced masonry piers under seismic actions. To predict the structural state of unreinforced masonry piers after an earthquake, research models have been developed to quantify important features of crack patterns. One of the most used crack features is the width, but this can be influenced by several parameters such as the axial load ratio, the shear span ratio, and the loading protocol, which have not been fully studied in previous research studies. In this study, we use experimental data to investigate the evolution of cracking in stone masonry piers during the application of cyclic shear–compression loading. The data consists of gray-scale images taken during quasi-static shear–compression tests performed on six plastered rubble-stone masonry walls subjected to constant axial force and cycles of increasing drift demand. Through the combined use of digital image correlation and a pre-trained deep learning model, crack pixels are identified, post-processed, and quantified based on their width. The dependency of the crack width on the axial load ratio, the shear span ratio, and the loading protocol at the peak force and ultimate drift limit states of the piers is clarified by a displacement vector field analysis, histogram of the crack width, and the concentration of deformation in the cracks. We show that, as opposed to flexural cracks, diagonal shear cracks do not fully close when moving from the applied drift demand to the residual drift measured upon removal of the lateral load. Furthermore, we provide the maximum residual crack width at peak force and ultimate drift limit states. This study will improve the decision making abilities of future models used to quantify earthquake-induced damage to stone masonry buildings.

2021

Modelling historical masonry aggregates using the Equivalent Frame approach

Katrin Beyer, Igor Tomic, Francesco Vanin

Many historical centres in Europe, including the city centre of Zagreb, feature masonry building aggregates, which developed as the building layout of the city was densified. Adjacent buildings in aggregates often share a structural wall and the building that was constructed second connects to the wall of the first building through a dry joint. When compared to the response of single freestanding buildings, several factors make the seismic response of masonry building aggregates more complex: It is not unusual for adjacent units to have different material properties, area and distribution of openings, roof and floor heights and orientations, and construction details.

2021

Numerical Simulation of Unreinforced Masonry Buildings with Timber Diaphragms

Katrin Beyer, Ivana Bozulic, Igor Tomic, Francesco Vanin

Though flexible diaphragms play a role in the seismic behaviour of unreinforced masonry buildings, the effect of the connections between floors and walls is rarely discussed or explicitly modelled when simulating the response of such buildings. These flexible diaphragms are most commonly timber floors made of planks and beams, which are supported on recesses in the masonry walls and can slide when the friction resistance is reached. Using equivalent frame models, we capture the effects of both the diaphragm stiffness and the finite strength of wall-to-diaphragm connections on the seismic behaviour of unreinforced masonry buildings. To do this, we use a newly developed macro-element able to simulate both in-plane and out-of-plane behaviour of the masonry walls and non-linear springs to simulate wall-to-wall and wall-to-diaphragm connections. As an unretrofitted case study, we model a building on a shake table, which developed large in-plane and out-of-plane displacements. We then simulate three retrofit interventions: Retrofitted diaphragms, connections, and diaphragms and connections. We show that strengthening the diaphragm alone is ineffective when the friction capacity of the wall-to-diaphragm connection is exceeded. This also means that modelling an unstrengthened wall-to-diaphragm connection as having infinite stiffness and strength leads to unrealistic box-type behaviour. This is particularly important if the equivalent frame model should capture both global in-plane and local out-of-plane failure modes.

2021

Pareto-like sequential sampling heuristic for global optimisation

Katrin Beyer, Mahmoud S. M. Shaqfa

In this paper, we propose a simple global optimisation algorithm inspired by Pareto’s principle. This algorithm samples most of its solutions within prominent search domains and is equipped with a self-adaptive mechanism to control the dynamic tightening of the prominent domains while the greediness of the algorithm increases over time (iterations). Unlike traditional metaheuristics, the proposed method has no direct mutation- or crossover-like operations. It depends solely on the sequential random sampling that can be used in diversification and intensification processes while keeping the information-flow between generations and the structural bias at a minimum. By using a simple topology, the algorithm avoids premature convergence by sampling new solutions every generation. A simple theoretical derivation revealed that the exploration of this approach is unbiased and the rate of the diversification is constant during the runtime. The trade-off balance between the diversification and the intensification is explained theoretically and experimentally. This proposed approach has been benchmarked against standard optimisation problems as well as a selected set of simple and complex engineering applications. We used 26 standard benchmarks with different properties that cover most of the optimisation problems’ nature, three traditional engineering problems, and one real complex engineering problem from the state-of-the-art literature. The algorithm performs well in finding global minima for nonconvex and multimodal functions, especially with high dimensional problems and it was found very competitive in comparison with the recent algorithmic proposals. Moreover, the algorithm outperforms and scales better than recent algorithms when it is benchmarked under a limited number of iterations for the composite CEC2017 problems. The design of this algorithm is kept simple so it can be easily coupled or hybridised with other search paradigms. The code of the algorithm is provided in C++14, Python3.7, and Octave (Matlab).

2021

RC U-shaped walls subjected to in-plane, diagonal, and torsional loading: New experimental findings

Katrin Beyer, Ryan Hoult

Although reinforced concrete U-shaped walls are popular in construction practice internationally, there is a paucity of experimental research investigating the seismic performance of such salient elements. The present paper summarizes an experimental campaign on two slender U-shaped reinforced concrete walls detailed with a single-layer of reinforcement. State-of-the-art instrumentation was used to capture the three-dimensional displacement field of the wall surfaces using digital image correlation techniques. Experimental findings are presented, including strain profiles, equivalent plastic hinge lengths, longitudinal strains at the base, cracking distributions and widths, and out-of-plane deformations. Approximately half of the yielding zone length was found to be equal to the equivalent plastic hinge lengths, which were found to decrease as a function of drift. The longitudinal strains at the base of these walls showed some shear lag effects when subjected to in-plane or diagonal loading. For most directions of loading, the largest crack widths were found to be associated with flexural-shear or shear cracks. When subjected to a pure torque, the vertical strain distribution at the base of the wall correlated with the theoretical distribution for an open section governed by warping torsion. The out-of plane deformations were primarily concentrated within a small region towards the ends of the flanges prior to the local buckling failures that were observed experimentally.

ELSEVIER SCI LTD2021

Seismic testing of adjacent interacting masonry structures

Katrin Beyer, Andrea Penna, Igor Tomic

In many historical centres in Europe, stone masonry buildings are part of building aggregates, which developed when the layout of the city or village was densified. In these aggregates, adjacent buildings share structural walls to support floors and roofs. Meanwhile, the masonry walls of the façades of adjacent buildings are often connected by dry joints since adjacent buildings were constructed at different times. Observations after for example the recent Central Italy earthquakes showed that the dry joints between the building units were often the first elements to be damaged. As a result, the joints opened up leading to pounding between the building units and a complicated interaction at floor and roof beam supports. The analysis of such building aggregates is very challenging and modelling guidelines do not exist. Advances in the development of analysis methods have been impeded by the lack of experimental data on the seismic response of such aggregates. The objective of the project AIMS (Seismic Testing of Adjacent Interacting Masonry Structures), included in the H2020 project SERA, is to provide such experimental data by testing an aggregate of two buildings under two horizontal components of dynamic excitation. The test unit is built at half-scale, with a two-storey building and a onestorey building. The buildings share one common wall while the façade walls are connected by dry joints. The floors are at different heights leading to a complex dynamic response of this smallest possible building aggregate. The shake table test is conducted at the LNEC seismic testing facility. The testing sequence comprises four levels of shaking: 25%, 50%, 75% and 100% of nominal shaking table capacity. Extensive instrumentation, including accelerometers, displacement transducers and optical measurement systems, provides detailed information on the building aggregate response. Special attention is paid to the interface opening, the global behaviour in relation to the interface separation, interstorey drifts and out-of-plane displacements.

2021