LEVERAGING DECISION TREES TO MAP FUNCTIONAL RECOVERY SEISMIC DESIGN PROVISIONS FOR REINFORCED CONCRETE MOMENT FRAMES

Published

October 31, 2023

Author(s)

Dustin Cook

Abstract

While recent advancements have been made in performance-based earthquake engineering to stochastically quantify functional recovery times for individual buildings, most of the modern building stock is designed according to prescriptive codes and standards, such as ASCE/SEI 7 in the U.S, focused primarily on life safety goals; the minimum design characteristics required to satisfy acceptably low recovery times for buildings are currently unclear. As part of a NIST initiative to improve the functional recovery of U.S. buildings and infrastructure, this study leverages probabilistic and performance-based analysis methods to develop prescriptive recovery-based design recommendations for new buildings using decision trees to provide a robust mapping between building design characteristics and recovery outcomes. The robust yet transparent nature of decision tree models provides a novel decision support tool for subject matter experts throughout the provisions update process. More specifically, in this study I exercise a recently developed prescriptive design requirement development framework on a set of reinforced concrete moment frame buildings to propose new recovery-based importance factors, drift limits, and nonstructural components design requirement to meet target recovery-based performance objectives.

Proceedings Title

18th World Conference on Earthquake Engineering

Conference Dates

June 30-July 5, 2024

Conference Location

Milan, IT

Pub Type

Conferences

Keywords

Functional Recovery, Machine Learning, Performance-Based Earthquake Engineer, Seismic Design Provisions

Citation

Cook, D. (2023), LEVERAGING DECISION TREES TO MAP FUNCTIONAL RECOVERY SEISMIC DESIGN PROVISIONS FOR REINFORCED CONCRETE MOMENT FRAMES, 18th World Conference on Earthquake Engineering, Milan, IT (Accessed July 17, 2024)