Computational Modeling and Multi-Fidelity Uncertainty Analysis for Masonry Walls Under In-Plane and/or Out-of-Plane Loading: Development and Applications

• Author / Creator

  Zeng, Bowen

• Unreinforced and reinforced masonry walls are essential elements in low-to-medium-rise residential and commercial buildings. They play an indispensable role in both vertical and lateral load-resisting systems, withstanding diverse loads such as seismic loads acting along and perpendicular to the wall during earthquakes, wind loads perpendicular to the wall, and eccentric gravity loads transmitted from floor or roof diaphragms. These loads can be categorized into in-plane (IP) loading and out-of-plane (OOP) loading, which occur simultaneously in most scenarios. Nevertheless, predicting their structural behaviors, including the failure modes and load-resisting capacity under IP and/or OOP loading, poses a considerable challenge considering the composite nature of masonry walls and the intricate interactions among their components, particularly when pertinent uncertainty in material properties is taken into account. In response to this challenge, this thesis consists of two parts: (1) development of 3D constitutive models for mortar joints to facilitate high-fidelity simulations of masonry walls and their applications to study the IP-OOP behavior interaction; and (2) development and application of algorithms and estimators for uncertainty analyses (e.g., mean and variance estimation of load resistance, reliability analysis) of masonry walls using low-fidelity models (e.g., design-code models) assisted by high-fidelity models via multi-fidelity approaches.
  To be specific, the first part of this thesis work aims at advancing finite element (FE) modeling techniques for masonry walls and enhancing the understanding of the interactive behavior of masonry walls under combined IP and OOP loading. At the core of this part is the development of two innovative constitutive models for cohesive interfaces to simulate mortar joints of masonry walls within the micro modeling framework. One model is implemented in a computational plasticity-based framework, while the other is developed within a damage plasticity-based framework, enabling the simulation of both monotonic and cyclic behaviors of masonry mortar joints, respectively. These two models are capable of capturing various failure modes, including tensile cracking, shear sliding, and compressive crushing, which allows for high-fidelity representations of masonry walls under both IP and OOP loading scenarios. A significant application of the developed models is to explore the IP-OOP interaction behavior of masonry walls. Two groups of numerical analyses are conducted, with the focus on the unreinforced and reinforced masonry walls, respectively, to investigate the effects of various design parameters, such as aspect ratio (height-to-length ratio), slenderness ratio (height-to-thickness ratio), pre-compression load, and reinforcement ratio. This application elucidates the intricate failure mechanisms of masonry walls under combined IP and OOP loadings and quantifies the consequent reductions in IP capacity due to the presence of OOP loading. Such insights highlight the criticality of considering IP-OOP interactions in ensuring safe and reliable masonry wall design and evaluation.
  The second part of this thesis aims at addressing a practical need for uncertainty analysis of masonry walls as required for reliability-based calibration of design codes and probabilistic performance-based design for masonry buildings. In this context, statistics estimation and reliability analysis are pivotal aspects. At the same time, the second part of this thesis will address a fundamental need in the field of uncertainty quantification: developing efficient algorithms/estimators for mean, variance, and failure probability when quantities of interest involved are obtained using computationally expensive high-fidelity models, such as those developed in the first part. In the second part, two novel statistics estimators (for mean and variance, respectively) and an innovative reliability analysis algorithm are proposed. The proposed estimators/algorithms leverage limited expensive high-fidelity models (e.g., finite element models) with a large number of cheap low-fidelity models (e.g., design-code models), via a multi-fidelity (MF) approach. This not only enhances computational efficiency but also ensures the accuracy of estimations of mean, variance, and probability of failure in reliability analysis. Case studies on masonry walls under IP or OOP loadings demonstrate the adaptability and practicality of these methods, marking a significant improvement over several existing approaches. Aside from the uncertainty analysis, this research also opens new possibilities for integrating high-fidelity computational models into design-code model enhancements in the masonry community.

• Subjects / Keywords

  • Structural Engineering
  • Masonry Walls
  • Computational Modeling
  • Uncertainty Quantification

• Graduation date

  Spring 2024

• Type of Item

  Thesis

• Degree

  Doctor of Philosophy

• DOI

  https://doi.org/10.7939/r3-wq4j-hf34

• License

  This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.