Bo Li

Cascading hazards challenge conventional performance assessment because the effects of a subsequent hazard depend on the residual state left by the primary event. Conventional single-hazard or independently combined multihazard assessments may misrepresent cascading hazard risk, as they often treat hazards as isolated or additive events and neglect the residual system state created by the primary hazard. This study develops a state-dependent framework for cascading hazard risk analysis within a performance-based engineering context. The proposed framework propagates primary-hazard effects through three connected components: vulnerability amplification, consequence amplification, and residual exposure updating. Vulnerability amplification is represented by a reduction in the median capacity of secondary-hazard fragility functions due to residual damage caused by the primary hazard. Consequence amplification is represented by increased casualty rates and repair-cost ratios under degraded post-event conditions. Residual exposure is updated to evaluate cumulative consequences on the remaining population and asset value and to avoid double-counting. A numerical illustration is conducted for a coastal school building subjected to an earthquake–tsunami sequence. Results show that, relative to an uncoupled assessment, the coupled analysis increases predicted total casualties and direct economic losses by approximately 54.93% and 31.67%, respectively. These findings demonstrate that residual damage and degraded post-event conditions can materially alter cascading hazard consequences. The proposed framework can support risk-informed mitigation, emergency planning, and performance-based design decisions for infrastructure exposed to cascading hazards.

In the wake of the Fukushima nuclear disaster, assessing the seismic safety of nuclear facilities, especially the secondary system, has become crucial. Floor response spectra (FRS) are keys to this assessment, typically generated through time history analyses. However, prior research shows that different time histories matching the same target spectrum can produce highly variable FRS estimates. This study examines the impact of spectrum-matched time histories on biased FRS estimates using machine learning methods. We generated 1750 spectrum-matched time histories based on five distinct target response spectra and applied them to eight typical steel moment-resisting frame buildings to compute their FRS. Machine learning techniques were then used to analyze the sensitivity of FRS estimates to modifications in ground-motion parameters due to spectral matching. The results indicate that changes in spectral accelerations are the primary factor influencing biased FRS estimates, followed by the damping scaling factor index for damping ratios of 7%, 2%, 10%, and 1%. These findings will aid in generating more accurate spectrum-matched time histories, reducing the risk of biased FRS estimates and improving seismic safety assessments for nuclear facilities.

Floor response spectra (FRS) are critical for the seismic analysis of the nuclear secondary system as per American Society of Civil Engineers Standard 4-16. Despite the use of spectrum-matched time histories, significant variations in FRS estimates can lead to inaccuracies in seismic analysis and system design. Our companion article explored the influences of ground-motion parameters modification due to spectral matching on FRS accuracy using machine learning, revealing that changes in spectral acceleration and damping scaling factor index IDSF for damping ratios ξ=7%, ξ=2%, and ξ=10%, significantly contributed to biased FRS estimates. Building on these insights, we propose enhanced spectrum-matched time histories that include additional IDSF values for different damping ratios. Using a nuclear service building as a case study, we demonstrate that these enhanced time histories significantly improve FRS accuracy, benefiting seismic safety practices in the nuclear industry.

Strong-motion records often show unrealistic displacements with drift, making them unsuitable for direct use in analyzing structural responses. To remove the drift, low-cut filtering is commonly used. However, the conventional zero-padding scheme to extend the accelerogram before filtering leads to unwanted effects, such as filter transients and an artificially extended duration. Furthermore, stripping off the added portion afterward may cause compatibility issues, resulting in inaccurate velocity and displacement time series derived from the acceleration data. To overcome these limitations, we introduce a new approach for extending and filtering velocity data. In this approach, the velocity trace is extended linearly at the end to double its length, followed by mirroring the data around the starting point. An adjustment function is then applied to ensure that the filtered velocity trace starts with zero after the extensions are removed. This novel technique effectively minimizes distortions in the resulting traces, producing compatible traces while preserving the original data duration. The advantages of this approach are illustrated through two case studies, showcasing its superiority over the zero-padding scheme and its performance compared to the Pacific Earthquake Engineering Research Center data. Moreover, a quantitative assessment using an extensive dataset of ground motions from Japan, encompassing a broad spectrum of magnitudes (Mj 2.3–7.2, in which Mj denotes Japan Meteorological Agency magnitude), epicentral distances (0–1500 km), and site conditions, validates the robustness of the proposed method. Our method completes the drift correction of ground motion in one shot, eliminating the need for subsequent baseline correction required by current approaches to yield consistent acceleration, velocity, and displacement traces.

Seismic risk analysis has received significant attention from both academia and industry in recent years. When conducting seismic risk analyses, selecting hazard-consistent ground motions is crucial. Traditional methods for selecting hazard-consistent ground motions often have limitations, addressing only a few periods and using discrete intensity levels, which can lead to inaccurate risk estimates. This study introduces a novel methodology that overcomes these limitations by using equivalent earthquakes derived from special seismic hazard disaggregation. This approach generates a scenario earthquake set to simulate target response spectra consistent with the seismic hazards of a target site through importance sampling. A selection scheme is also proposed to identify hazard-consistent ground motions from a database. Furthermore, a new method for performing probabilistic seismic demand analysis is presented. A numerical example demonstrates that the selected ground motions maintain strong consistency with the site's seismic hazard curves across a wide range of periods. Unlike traditional methods, this approach does not depend on the fundamental period of specific buildings, making the selected ground motions applicable for accurate seismic risk analysis of multiple structures at the same site.