Browsing by Author "Elyasi, Niloofar"

Filter results by typing the first few letters
Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Advancing Structural Engineering Through Data-Driven Methodologies: Seismic Vulnerability Assessment and Backbone Curve Determination
    (University of Waterloo, 2025-04-11) Elyasi, Niloofar
    Structural engineering has traditionally relied on analytical and experimental methods to ensure the safety of structures. These methods, while effective, often require significant resources, time, and expertise, limiting their applicability across diverse contexts. Meanwhile, vast amounts of data collected from surveys, experimental studies, and seismic events remain largely underutilized, providing a unique opportunity to develop advanced data-driven methodologies. This thesis aims to harness the potential of the available data repositories to address critical challenges in structural engineering, with a focus on seismic vulnerability assessment and backbone curve determination. Through the use of machine learning (ML), this thesis introduces innovative methodologies at both the system and component levels. A rapid visual screening (RVS) framework is developed to quickly assess the seismic vulnerability of low-rise reinforced concrete (RC) buildings. By incorporating ML models, this framework outperforms traditional evaluation methods with higher accuracy and broader applicability. Using post-earthquake survey datasets from a variety of seismic events, it proposes a region-independent tool, eliminating reliance on subjective judgments and region-specific calibrations. For backbone curve determination, used for analyzing the seismic behavior of RC columns, this thesis introduces a novel ML-based methodology. By employing experimental datasets and advanced regression techniques, it offers a practical and efficient alternative to the conventional methods. This approach not only predicts backbone curve parameters with high accuracy but also ensures accessibility for broader applications, especially in resource-limited environments. In summary, this thesis bridges system-level and component-level challenges, underscoring the potential of data-driven approaches in structural engineering. By providing a foundation for integrating innovative approaches into the field, this research advances both academic insights and practical applications. These contributions respond to the demand for efficient and reliable solutions, supporting safer structures and more effective resource management in modern structural engineering practices.