Assessment of Available Collapse Simulation Methods for Use in Performance-Based Seismic Engineering Project

Summary

While it is a key aspect of structural analyses that support current Performance-Based Seismic Engineering (PBSE), "structural collapse" is not clearly defined or directly simulated in the PBSE methodology. "Collapse" is now either assessed indirectly based on lateral displacement ("drift") analysis results that exceed engineering judgment-based predefined limits, or is implied when the numerical analysis solution algorithm being used fails to converge. The ability to accurately simulate structural collapse is critical for implementing PBSE. Accurate determination of collapse will enable designers to identify prudent levels of displacement demand that remain away from the collapse limit state.

This project:

1. Identifies and assesses critical issues concerning nonlinear mechanics of structures at or near sidesway collapse, to improve analytical accuracy and consistency
2. Verifies the accuracy of practitioner-oriented structural analysis software packages for simulating sidesway collapse
3. Develops guidelines on the levels of sophistication necessary to accurately simulate large displacement (near-collapse) response

As a major output, the project will produce calibrated measurement science methodology and guideline publications for improving simulation and assessment procedures that evaluate seismic sidesway collapse of structures using practitioner-oriented nonlinear dynamic analysis software. This will impact how practitioners perform nonlinear dynamic structural analysis, design, and assessment, while advancing PBSE design philosophy through implementation by practicing structural engineers.

Description

Objective - This project supports improved Performance-Based Seismic Engineering (PBSE) via evaluating the capabilities of commercial structural analysis software to simulate structural sidesway collapse accurately. This project:

1. Identifies and assesses critical issues concerning nonlinear mechanics of structures at collapse or near-collapse, to improve analytical accuracy and consistency;
2. Verifies the accuracy of practitioner-oriented structural analysis software packages for simulating sidesway collapse; and,
3. Develops guidelines on the levels of sophistication necessary to accurately simulate large displacement (near-collapse) response.

What is the new technical idea? NIST GCR 09-917-2 (NIST 2009) identified improving analytical models and demand assessment for buildings near collapse from seismic loading as the second highest priority research area in support of full implementation of Performance-Based Seismic Engineering (PBSE). NIST GCR 13-917-23 (NIST 2013) identified improving analytical models and simulation capabilities for buildings in near-collapse seismic loading as a high priority research topic. Commonly used practitioner-oriented collapse assessment analysis tools do not directly simulate collapse but instead require monitoring of other demands (e.g., drift) that are indirectly associated with collapse, such as the analytical results reported in FEMA P695 (FEMA 2009). In spite of this imprecision, the ASCE/SEI 7 Standard (ASCE 2010) uses the probability of structural collapse as the key parameter to define seismic risk and corresponding minimum design loads.

This project posits the new technical idea of assessing collapse behavior based on combining basic nonlinear structural collapse mechanics theory with sophisticated research-oriented nonlinear finite element modeling and simulation. Comparisons of these collapse mechanisms with maximum drifts computed using (less sophisticated) practitioner-oriented software packages will be performed to verify current collapse prediction capabilities in those packages.

The knowledge gained will then be applied to develop a rational, rigorous, and comprehensive methodology to measure structural performance based on the corresponding limits of acceptance criteria currently used in the ASCE/SEI 41 standard (ASCE 2013) that were developed in FEMA 273 (FEMA 1997) over 18 years ago and are currently being investigated by a NIST extramurally funded project. Once this methodology is developed, practitioners will be able to perform detailed ASCE/SEI 41 nonlinear dynamic sidesway assessment on both global and local structural responses using practitioner-oriented software packages with confidence.

What is the research plan? The first phase of this project involves theoretical investigation of basic structural collapse mechanics to identify the fundamental issues that must be included in an analytical model to capture sidesway collapse behavior. Key elements of this work are identifying the means to simulate the behavior analytically and verifying the numerical algorithms and associated accuracies obtained from current practitioner-oriented software packages. Simple structural moment frames and frame components are used in the study, keying on sidesway mechanism formation. Of specific interest are:

1. The need to incorporate large displacement theory into widely used nonlinear geometric stiffness models for practitioner-oriented software modeling, up to structural collapse
2. The level of sophistication required in modeling material nonlinearity for accurate simulation of inelastic dynamic behavior at structural collapse
3. The computation of plastic rotations in nonlinear structural response at yielding locations, including nonlinear dynamic response up to structural collapse

The second phase of the research includes developing LS-DYNA finite element models to simulate collapse of structures subjected to seismic events. Archetypical steel moment frame structures that have been designed and modeled in a previous Engineering Laboratory (EL) project (Harris and Speicher 2015) are used in this phase. The goal is to capture as accurately as possible the sidesway collapse mechanism that is often experienced in flexible moment frames modeled using lower-end, practitioner-oriented software packages. Results are then used to calibrate the computer models using practitioner-oriented software packages such as SAP2000, PERFORM-3D, and OpenSees. Guidelines will then be developed for using these practitioner-oriented software packages to simulate sidesway structural collapse accurately.

Once the mechanism of collapse in the modeling and simulation process is accurately captured using practitioner-oriented software packages, the third phase of the research revisits geometric nonlinearity in practitioner-oriented software packages and investigates how different small-displacement formulations can impact the actual large displacement response of framed structures. This work again focuses on moment-resisting structural steel systems, which have been studied in other EL projects. Using the existing building models allows more project time to be spent on investigating collapse assessment of these same buildings. Significant effort is spent on developing computer models of framed structures using various platforms and performing simulations using a large number of ground motions, and results will be documented in several journal papers.