Date of Award
8-2018
Document Type
Thesis
Degree Name
M.S. in Engineering Science
Department
Civil Engineering
First Advisor
Elizabeth Ervin
Second Advisor
Yacoub Najjar
Third Advisor
Christine Surbeck
Relational Format
Dissertation/thesis
Abstract
Many structural collapses, all over the world, are related to inspection errors. The U.S. I- 35 Bridge in Minneapolis and the Canterbury Television building in New Zealand are a few examples. U.S. infrastructure conditions are poor, scoring a D+ according to ASCE report 2017. When the structures are inspected, the cost is high and service is lost. Inspectors use their trained but subjective judgment, which can lead to errors. Improved inspection techniques are required to reduce the cost, time, and subjectivity.
SHETM (Structural Health EvaluationTM) is an in-house software that has the ability to employ data to improve damage detection techniques. Prior work shows that the experimental data has too much uncertainty to identify the best detection algorithm, so this work uses finite element output for damage detection in a steel frame model. Commercial packages SAP2000 and ABAQUS are employed for cases with sequential beam removal for the algorithms of MAC (Modal Assurance Criterion), COMAC (Coordinates Modal Assurance Criterion), Modal Flexibility, Modal Curvature, and Modal Strain Energy indicators. Two cases are compared for different mode sets, and the finite element output reveals the best indicators for various reinforcing braces. Throughout all cases, the Flexibility Absolute Difference provided the most trusted results. This study serves as the patch between experimental data and finite element output by providing advice on algorithm use. Addition of any experiment’s uncertainty can now be better evaluated by separating algorithm noise versus data noise.
Recommended Citation
Nguyen, Ai D., "Using Finite Element Output to Detect Damage in a Steel Frame Model" (2018). Electronic Theses and Dissertations. 2782.
https://egrove.olemiss.edu/etd/2782
Accessibility Status
Searchable text