Report No.: CCEER-13-19

Title: Development of Generalized Fragility Functions for Seismic Induced Content Disruption

Authors: Guzman Pujols, Jean C., Ryan, K.L

Date: December  2013

Performing Organization:
Department of Civil Engineering/258
University of Nevada, Reno
Reno, NV 89557


This study reports on the evaluation of the response of building contents subjected to a series of 1-D shaking experiments in a linear and nonlinear structure, with the goal of developing simplified seismic fragility functions to be used for content disruption. The experiments were conducted as a payload project in collaboration with the NEESR Grand Challenge project at the University of Nevada, Reno (UNR). The test bed structure was a full scale two-by-one bay, two-story steel frame, spanning 3 biaxial shake tables at the NEES Equipment Site at UNR. The room contents were staged in a small space enclosed by dry wall partitions with an open doorway on the second story of the test bed structure. The contents consisted of several items representing an office setup, including a desk with a computer, a desk chair, and two bookcases with typical content. In addition, a storage cart with several kitchen items was included in the room to observe the performance between different types of furnishing supports and its effect on the content response. For the last test series of the program, another room with similar contents on the south bay, second story of the test bed structure was included in the experiments.
As a supplementary study, the same evaluation technique is applied to another series of experiments that were conducted at E-Defense in collaboration with the NEES TIPS project (Tools to Facilitate Widespread Use of Isolation and Protective Systems). In the experiments, a full-scale, five-story, steel moment frame building was subjected to 2-D and 3-D motions using the E-Defense shake table and tested with two isolation systems and fixed at the base. Building contents and equipment typically found in hospitals and offices were staged in two enclosed rooms on the fourth and fifth floor of the building, respectively. The rooms were also enclosed by dry wall partitions with an open doorway. For the hospital setup, the contents included a medical storage cart with different supplies, a table, a hospital bed with bedside cart, an IV pole, an instrument stand and an exam light. For the office setup, contents included two desks, a computer, three desk chairs, a copy machine, and two bookcases with typical content. Cameras were mounted in two corners of the room, as well as two locations on the floor at the doorway openings.

For both experiments, the content response is evaluated collectively, by defining qualitative categorical ratings of overall content disruption based on observance of specific behaviors, such as sliding, toppling, rolling or falling. Then, these DRs are correlated to EDPs such as peak floor acceleration (PFA) and peak floor velocity (PFV). The best EDPs to predict the observed disruption (based on correlation between EDP and DR) are selected from among several evaluated. The primary objective of this evaluation is to develop simplified fragility functions to predict the statistical occurrence of the DR based on the observed EDP. The simplified fragility curves can be broadly applied for performance evaluation of building contents from a design or a loss estimation perspective. The final objective is to use the supplementary study as an independent data set to validate findings from the first study. In addition, the sensitivity of the fragility functions to vertical excitation is also evaluated.

Two alternative implementations to develop the simplified fragility functions (used in both studies) were considered. Each method was found to have limitations for the derivation of the fragility parameters, and as a result, the fragility functions were developed by a hybrid approach that incorporated components of both methods. The results from the NEESR-GC, using PFA as the EDP in the fragility development, could not be validated by the results from the E-Defense dataset, due to the different sources of variability in the E-Defense experiments. Fragility functions for the collective seismic response of unanchored building contents were found to be best represented with horizontal velocity demand parameters. The final fragility curves presented as a function of PFV can be applied broadly to protect building contents.