arup

In the aftermath of previous moderate to large earthquakes (e.g. Loma Prieta 1989, Northridge 1994, and Christchurch 2011), nearby university campuses (e.g. Stanford, California State University, and University of Canterbury) have experienced widespread building damage, disruption of lecture and research, and reduction of undergraduate and graduate attendance. Other potential consequences of an earthquake could include life-safety issues and a loss of university reputation. While the value of ensuring campus-wide resilience in the face of earthquakes is apparent, the strategy to achieve measurable improved performance can be difficult to develop and quantify. This paper presents a methodology for estimating disruption and recovery of space on a university campus based on the REDi downtime assessment methodology (founded upon FEMA P-58). This is used to predict downtime for individual buildings due to building repairs and delays to the initiation of repairs (i.e. impeding factors). Hundreds of earthquake scenarios were considered to impact an example campus, and the related results are presented. Particular attention is paid to metrics of primary interest to university campuses, including restoration of different occupancy space over time and displacement of student populations.

Recent earthquakes have demonstrated that the vertical component of ground motion can be damaging to building structural and nonstructural components. Recorded vertical accelerations from these recent earthquakes significantly exceed the vertical spectral demands used for seismic design prescribed in building codes in the United States. The focus of this paper is to determine, through non-linear time history analyses, the expected behavior of a code-compliant 3-story steel moment frame structure under vertical accelerations consistent with the design level earthquake intensity. This study differs from prior research because all components, including gravity elements, are explicitly modeled to capture potential non-linear behavior. Prior studies compared elastic demand-capacity ratios for some structural elements, which may not provide a meaningful measure of damage due to the high frequency nature of vertical response. Insights from the analysis are used to recommend future research needs.

Over the past decade, the earthquake engineering profession has made significant progress in developing models and tools for predicting loss of functionality and downtime in buildings, most notably the seismic performance assessment methodology for buildings (i.e., FEMA P-58) developed by the Federal Emergency Management Agency. While FEMA P-58 represents an important step-changes in practice, its methodology for predicting downtime contains significant simplifications and assumptions that impact the accuracy of its results. Downtime is arguably the most difficult performance metric to estimate, but it is also one of the most important, providing insights into how long critical buildings might be rendered unusable, which in turn can be used to predict restoration times for business operations or important community services. This paper provides an overview of the state of the art in predicting earthquake-induced downtime in buildings. It describes the REDi downtime assessment methodology, a sophisticated process for predicting downtime in individual buildings that builds on FEMA P-58, and summarizes several improvements that have been made to the methodology since its initial publication in 2013. It also discusses important remaining challenges in predicting downtime in buildings.

Earthquake-induced loss of functionality and downtime in individual buildings can cause major economic and societal impacts, ranging from displacement of individual families and businesses to disruption of emergency medical services and government functions. The ability to predict recovery times for buildings after an earthquake is a critical component in understanding and mitigating the seismic risks facing a community. The seismic performance assessment methodology proposed by FEMA P-58 created a framework that translated structural performance into metrics that would enable stakeholders to make risk-informed decisions. Upon that framework, the REDi downtime assessment methodology was created to convert repair times from a FEMA P-58 assessment into building downtime through realistic labor allocation, consideration of delays to initiation of repairs, and calculation of intermediate goals along the path to full recovery. Since its initial development in 2013, the REDi downtime assessment methodology has been implemented in several projects. From these project and research applications, further improvements have been made to the previously published methodology. These improvements, key takeaways, and remaining challenges are discussed.

This paper presents a series of validation exercises for seismic site response analysis using the recordings from three well-instrumented, urban lakebed seismograph stations in Mexico City (CAO, SCT and TXS2). The dynamic behavior of Mexico City Clay has long been a topic of great interest among geo-seismic researchers due to the lack of consensus as to the cause of the unusually high site amplifications and long durations of ground motion measured in the lakebed. While some researchers have postulated that trapped waves in the basin of Mexico City are the cause, others have attempted to disprove this theory through validation of one-dimensional site response analysis with mixed degrees of success. The current study concludes that the cause lies both in the response of the deep basin and the unique properties of Mexico City Clay (including low damping and shear wave velocity). This study demonstrates that nonlinear site response analysis with bi-directional excitation using LS-DYNA results in comparable site amplifications and surface acceleration spectra to the recorded data, provided that the base input motion appropriately accounts for the deep basin (either directly through downhole array measurements or via GMPEs specifically developed for Mexico City “rock” or hard layer). The site response analysis proposed herein explicitly incorporates the unusual properties of Mexico City Clay in the time-domain nonlinear approach (low damping, strain rate effects, cyclic degradation). This study also indicates that the surface response is highly sensitive to the assumed shear wave velocities, requiring consideration of modifications to in situ geophysical measurements due to phenomena such as the thixotropy of the clays.

As analytical models become more numerous and complex, automation becomes essential to allow time for critical thought and innovation. This companion paper details a suite of tools that have been developed to enable advanced nonlinear site response analyses to be carried out in an automated fashion. These automation tools were built within various programming environments and have the ability to handle pre and post-processing tasks as well as interface directly with the finite element analysis program LS-DYNA. Pre-processing begins with a program written in Excel Visual Basic for Applications, which generates data to be fed to the processer (LS-DYNA) with a JavaScript API. After running the model, JavaScript is again employed to prepare the output data for the post-processor, which was written in Matlab. The functionality and development of each rapid automation tool is presented and the resulting time savings are discussed. Altogether, these rapid automation tools have decreased the required worker hours for pre-and post-processing of site response analyses from over two days to under half an hour per soil column.

The REDi Rating System (v1.0) was published in late 2013 as a guideline for implementing “resilience-based earthquake design” to achieve “beyond code” resilience objectives in the design of new construction. Since its publication, the REDi guidelines have been incorporated into several RFP’s and have been adopted for the design of high-profile projects including the 181 Fremont Tower in San Francisco (designed by Arup) and the Long Beach Civic Center (designed by NYA and SOM). In addition, the REDi downtime methodology (which enables calculation of the time required to achieve re-occupancy and functionality) have been utilized for existing building portfolio risk analysis for the University of British Columbia, Vancouver (UBC) to provide a significantly more detailed view of the risk relative to standard methods (i.e. HAZUS) and the REDi guidelines and framework will inform future mitigation strategies on campus. In this article, the designers will share their experiences and lessons learned from incorporating the REDi framework in their projects, including communication with their respective clients, influence on their design approach, the adoption of specific design and/or planning enhancements, and the overall design and peer review process.