Arresting failure propagation in buildings through collapse isolation

Building collapses cause significant economic losses and fatalities. From 2000 to 2019, disasters, largely attributed to building collapses, resulted in approximately US$2.97 trillion in economic losses and 1.23 million deaths. These collapses often stem from the propagation of local initial failures triggered by events such as earthquakes, floods, storms, landslides, explosions, vehicle impacts, or design errors. The increasing frequency and intensity of extreme events highlight the urgent need for robust building designs insensitive to initial damage. Current robustness design approaches aim to completely prevent collapse initiation by providing extensive connectivity within the structural system, enabling load redistribution after local failures. However, this strategy is ineffective and potentially counterproductive with larger initial failures, as it can lead to cascading failures and total collapse. Past research has primarily focused on preventing collapse initiation, neglecting the crucial aspect of preventing collapse propagation. This study presents an innovative approach to address this gap, focusing on controlled failure propagation to limit the extent of collapse after large initial failures.

Existing literature extensively covers the prevention of collapse initiation following local failures, often focusing on enhanced connectivity to facilitate load redistribution. The Ronan Point tower collapse in 1968, caused by a lack of connectivity, spurred significant research in this area. While completely preventing collapse is ideal, the reality of unforeseeable incidents necessitates strategies to mitigate the impact of unavoidable failures. Studies have examined various failure scenarios induced by different extreme events, including earthquakes, floods, and impacts. The limitations of relying solely on increased connectivity for robustness in the face of large-scale initial failures are well-documented. This research builds upon this existing literature by proposing a paradigm shift, focusing on controlled collapse propagation rather than solely on preventing collapse initiation.

This research introduces a novel design approach called hierarchy-based collapse isolation. This approach distinguishes between small initial failures (where load redistribution prevents collapse initiation) and large initial failures (where partial collapse is inevitable). The design aims to (1) arrest uncontrolled collapse propagation from large initial failures and (2) ensure alternative load paths (ALPs) to prevent collapse initiation after small failures. This is achieved by establishing a specific failure hierarchy, prioritizing the failure of connections over columns, thereby limiting the propagation of collapse. The study focused on framed building structures, utilizing computational simulations validated by a full-scale experimental test. A 15m × 12m precast reinforced concrete building with two floors was designed and constructed. Two designs were compared: Design H (hierarchy-based collapse isolation) and Design C (conventional design complying with building codes). Design H incorporated optimized partial-strength connections and enhanced columns. High-fidelity collapse simulations using the applied element method (AEM) were performed to investigate the behavior of both designs under small and large initial failures. The simulations were validated against experimental data from a full-scale collapse test. The full-scale test involved two phases: phase 1 (quasi-static removal of two penultimate-edge columns) and phase 2 (dynamic removal of a corner column). The building was instrumented with various sensors to monitor its structural response. The computational simulations considered various parameters, including material properties (concrete and steel) obtained from laboratory testing, and utilized realistic models for concrete behavior under compression and tension. The AEM accurately represented large displacements, cyclic loading, fracture, contact, and collisions. The study involved multiple iterations between numerical simulations and design adjustments to determine optimal design parameters.

Simulations predicted that both Design H and Design C could prevent collapse initiation after small initial failures exceeding building code requirements, demonstrating the efficacy of ALPs. However, under large initial failures, Design H effectively isolated the collapse to the directly affected region. In contrast, Design C experienced total collapse due to uncontrolled failure propagation. The full-scale experimental test confirmed these predictions. Phase 1 demonstrated the prevention of collapse initiation after the removal of two penultimate-edge columns. Load redistribution through beams, columns, and slabs prevented collapse. The analysis of the structural response showed that the flexural response of the beams and the contribution of the floor slabs were key to preventing collapse. The reinforcement detailing played a crucial role in enabling this load redistribution. Phase 2, involving the removal of the corner column, demonstrated the effectiveness of collapse isolation. The collapse was arrested along predefined segment borders due to connection failures occurring before column failures. The rupture of continuous reinforcement bars halted the transmission of forces to the columns. The debonding of slab reinforcements further contributed to the segmentation of the structure. The upright part of the building experienced significant dynamic horizontal forces but remained standing, highlighting the protective effect of the proposed design. While the upright part sustained damage, safe evacuation would have been possible, saving lives that would have been lost with a conventional design.

The findings demonstrate that the extensive connectivity employed in conventional robustness design can be detrimental during large initial failures, leading to catastrophic collapse. The hierarchy-based collapse isolation design successfully mitigates this risk by controlling the failure hierarchy, ensuring connection failures precede column failures. The full-scale experimental validation confirms the efficacy and cost-effectiveness of the proposed approach, without compromising the ability to prevent collapse initiation after small failures. This last line of defense against major building collapses offers a significant advancement in structural engineering. The achieved segmentation limits the extent of damage, potentially facilitating safer evacuations and rescue operations. The successful implementation and validation of hierarchy-based collapse isolation provide a robust framework for future building design, enhancing resilience against extreme events.

This research presents a novel hierarchy-based collapse isolation design, validated through full-scale testing, that effectively arrests collapse propagation after large initial failures without impairing the prevention of collapse initiation after small failures. This approach offers a crucial last line of defense against catastrophic building collapses. Future research should explore the application of this approach to diverse building types, materials, and failure scenarios. Further investigation into optimizing connection design and failure criteria could lead to more refined and cost-effective implementations.

The study focused on a specific precast reinforced concrete building configuration. The generalizability of the findings to other building types and materials requires further investigation. The full-scale test involved a specific sequence of column removals; other failure scenarios should be considered. While the AEM simulations provided valuable insights, the accuracy of the model depends on the accuracy of the input parameters and assumptions, potentially impacting the simulation outcomes.