|Projects : CUREE-Kajima Joint Research Program
PHASE VI -
INVESTIGATION OF FACTORS LEADING TO PROGRESSIVE COLLAPSE OF STRUCTURES
Category I: Evaluation and Development of Analytical Tools
Research Team: Sivaselvan Mettupalayam (Univ. of Colorado at Boulder);
Andrei Reinhorn and Gary Dargush (Univ. at Buffalo-SUNY)
TASK LIST - In the following, each of these tasks is briefly described.
Task 1 - Review existing computational models and numerical algorithms
This task will collect data about programs available for collapse analysis of structures, either static or dynamic, using micro or macro models. The task will identify capabilities and secure access to the programs for the research team. Possible programs are: ABAQUS, LS-DYNA, LARSA, RAM Perform Collapse (2004), 3DEC (2003), Merlin, ASI-Extreme Loading (2004), IDARC3D, OpenSEES, etc. In addition several pre- and post- processing packages will be identified. Among candidates: KUMO, SPIDER, PATRAN, SAP, LARSA, etc.
Task 2 - Evaluate the performance of selected computer models and algorithms
Selected computer programs will be evaluated (i) with a simple benchmark problem to identify features related to collapse and to develop a comparison table similar to Table 2 and Table 2 modified, proposed in Appendix and (ii) with a case study structure, or structures, designed according to US and Japanese standards. The task will identify the benchmark structure from the examples used to assess one of the more complex computer platforms identified in Task 1. The task will also consist of the design, analytical modeling, and response prediction (through inelastic response history analysis) of a steel frame structure with moment resisting frames in one direction and braced frames in the orthogonal direction. The size (number of stories and bays, and bay widths) will be determined through interaction between all project participants from the US and from Kajima. It is expected that a US design and a Japanese design will be evaluated. Evaluation implies analytical prediction of performance, and in particular of the collapse capacity, utilizing a representative set of ground motions from large seismic events. The study will identify the capacity limit state (Cimellaro, 2005) of the structure in the case study. This task will attempt to express the results in terms of the collapse fragility curves as defined by Cimellaro (2005).
The proposed work will be performed in conjunction with the work carried out by Professors Mahin on braced frames and Krawinkler on moment resisting frames and in conjunction with Kajima teams working on the project.
Task 3 - Identify preferred algorithms and limitations
(Analyzing case study building(s) for different scenarios and member models)
This task is dedicated to evaluating the capabilities and performances of various models and algorithms. It will consist of (a) developing a table of capabilities for all approaches considered, (b) identifying gaps in these capabilities (based on the application to the case studies), (c) enumerating possible techniques and algorithms in particular related to the handling of load redistribution, degradation, and rate of loss and final / residual capacity of systems in a global sense and establishing the most appropriate framework to address most of the gaps through additional development.
Task 4 - Develop improved models and algorithms as needed
This task is dependent on the outcomes of the previous task and cannot be fully detailed at this moment. However, based on the previous experience of the investigators, several issues will probably be identified related to: (i) limitations of computational engines versus newly developed alternatives (classic, State Space Approach, Mixed Lagrange Formulation, etc); (ii) lack of integrated packages of models and solutions for global systems; and (iii) lack of pre- and post processing capabilities with visualization. It is envisioned that some of the above issues will be addressed as follows:
1. Modeling analytically the sequence of events leading to loss of capacity to sustain gravity loads
Description: (a) Extend model of large deformations and nonlinear cyclic behavior with deterioration of strength, stiffness and slip dislocation from 2D to 3D beams, columns and connections; (b) Model sudden deterioration in solution algorithm; (c) Integrate model in platform with global interaction and complete fragmentation force and impulse redistribution
Challenges: (i) In presence of sharp gradients and discontinuities solution algorithms fail (ii) sequences of events go through unstable equilibrium states exhibiting sensitivity to initial conditions (iii) State Space Approach (SSA) seems to show good prospect of being applied in such situations.
2. Development and integration of 3D systems using the Mixed Lagrangian Formulation (MLF)
Description: (a) The approach separates the modeling of components from the global solution, thus allowing an extendible approach without changing the solution algorithm; (b) The formulation considers the coupled interaction of force / stress changes and the associated impulse. The force / impulse approach considers explicitly the sudden changes and deterioration.
Challenges: (i) Integration in large size programs (ii) stability of solution in case of non-convex systems; (iii) integration with discrete elements necessary to model fragmentation.