Rapid post-earthquake safety assessment system for buildings using sparse acceleration measurements

Tsuchimoto, Koji

Permalink

https://hdl.handle.net/2142/113973 Copy

Description

Title

Rapid post-earthquake safety assessment system for buildings using sparse acceleration measurements

Author(s)

Tsuchimoto, Koji

Issue Date

2021-12-01

Director of Research (if dissertation) or Advisor (if thesis)

Spencer, Jr., Billie F.

Doctoral Committee Chair(s)

Spencer, Jr., Billie F.

Committee Member(s)

• Bergman, Lawrence A.
• Fahnestock, Larry A.
• Wada, Akira
• Work, Daniel

Department of Study

Civil & Environmental Eng

Discipline

Civil Engineering

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• Rapid safety assessment
• post-earthquake
• sparse acceleration measurements
• damage index
• maximum inter-story drift angle
• safety classification
• damage-sensitive features
• convolutional neural network
• parametric model updating
• nonparametric system identification
• experimental validation
• cost-benefit analysis

Abstract

After a major seismic event, building safety should be assessed by qualified experts prior to reoccupation and resumption of operations. Such inspections are generally performed by teams of two or more experts and are time consuming, labor intensive, subjective in nature, and potentially put the lives of inspectors in danger. Structural Health Monitoring (SHM) provides the potential to accelerate the required amount of evaluation. The goal of the research proposed herein is to develop a cost-effective approach for rapid safety assessment of buildings after seismic events by using sparse acceleration measurements. First, the problem definition of the rapid safety assessment system is discussed, and an appropriate performance measure is developed. Herein, the maximum inter-story drift angle (IDA) is considered as a reliable damage index (DI) to classify the safety of buildings after seismic events occur. A damage sensitive feature (DSF) is then defined by a comparison between linear and nonlinear responses that can be used to assess the condition of buildings. Condition assessment using a convolutional neural network (CNN) is employed to uncover the complex relationship between DSFs and the damage index. Issues associated with full-scale implementation are also addressed. First, validation of the proposed approach is conducted using a numerical model of a five-story steel building. A three-dimensional (3D) nonlinear analysis model is created and subjected to random ground motion. Because the use of the 3D analysis model is too computationally expensive to generate the required training data, a simplified nonlinear model is developed for this purpose, along with a corresponding linear model. The training data for the CNN incorporates uncertainties in both the analysis model and ground motion. While the numerical results confirmed the potential of the proposed approach, experimental validation on large-scale structures is required for field implementation. Moreover, an extension to the assessment of high-rise buildings, such as those used for residences in many modern cities, is desired. To this end, data for a 1/3-scale, 18-story experimental steel building that was tested on the shaking table at E-Defense in Japan is employed. The importance of online model updating of the linear analysis model to calculate corresponding DSFs during the operation is also discussed. Experimental results confirm the efficacy of the proposed approach for rapid post-earthquake safety assessment for high-rise buildings. Finally, a cost-benefit analysis with respect to the number of sensors used is presented. This research demonstrates the efficacy of the proposed approach for rapid post-earthquake safety assessment of buildings, using sparse acceleration measurements.

Graduation Semester

2021-12

Type of Resource

Thesis

Permalink

http://hdl.handle.net/2142/113973

Copyright and License Information

Copyright 2021 Koji Tsuchimoto

Owning Collections