Project Title/ID Number Development and Validation of Performance Models— 4122002
Start/End Dates 10/1/02—9/30/03
Project Leader Greg Deierlein (Stanford/Faculty)
Team Members Rohit Kaul (Stanford/Grad Student), Arash Altoontash (Stanford/Grad Student)
Project goals and objectives

High fidelity simulations are essential to the PEER PBEE methodology to accurately model seismic response, damage and collapse of buildings and bridges. Central to the effort of thrust area 4 is the development of OpenSees (Open System for Earthquake Engineering Simulation) - a versatile computing framework that integrates analysis models for seismic hazard, ground motions, soil response, soil-foundation structure interaction, and structural response. The focus of this project (4122002) is to develop and validate generalized hinge models for simulating strength and stiffness degradation in reinforced-concrete structures. Emphasis is on spring assemblies, generalized hinge elements, and cyclic hysteretic models for beam-columns and their connections that are sufficiently robust to capture highly nonlinear response at collapse, yet practical for application to large systems.

Role of this project in supporting PEER’s vision

We are developing, implementing, and validating structural analysis models for OpenSees in collaboration with other researchers in thrust area 4 – primarily the team led by Fenves at Berkeley. We are also collaborating with Lowes of the Univ. of Washington who is developing detailed meso-scale models and damage indices for beam-column joints. The generalized hinge models we are developing complement more computationally demanding fiber-type models that Filippou has developed. We are making use of research results from other PEER projects including:

  1. tests on non-ductile beam-columns and beam-column joints by Moehle and Lehman
  2. tests of bridge piers by Mahin, Xiao, Ashford, and Pardoen
  3. sensitivity formulations by Der Kiureghian and Conte
  4. the column performance database assembled by Eberhard
We are also collaborating with researchers in Japan and Taiwan who are conducting tests of RC and composite steel-concrete subassemblies.
Methodology employed

Our model formulations are based on principles of concentrated stress-resultant plasticity models, which are fairly well established for hardening behavior. We are extending these basic hardening models to include strength and stiffness degradation under reverse cyclic loading effects, including the interaction of axial-flexural-shear effects and large deformations. We developing and implementing the models in the native C++ modular object-oriented framework of OpenSees. Calibration and validation studies are making use of existing test data from other PEER projects and from prior research by the PI and others on reinforced concrete and composite steel-concrete structures.

Brief description of past year’s accomplishments and more detail on expected Year 6 accomplishments

This project builds on a previous PEER project, in which we implemented a beam-column element subassembly with rotational and shear springs for large-deformation response. The prior implementation included a super-element assembler to combine springs and several nonlinear uniaxial material models. The material models were developed in collaboration with a project directed by Krawinkler on the performance of stiffness and strength degrading structures. Over the course of implementing these elements, we worked with Fenves and the OpenSees development team to resolve bugs and limitations that we discovered in OpenSees.

Since October 2001 (Year 5), we realized the following accomplishments:

  1. Theoretical development and implementation of a new 2D P-M yield surface beam-column element has been completed. The implementation utilizes object-oriented modular programming to allow convenient extension of the model to include different yield surface function, hardening/softening parameters, damage indices, and hysteretic response variables. Element was published on the OpenSees website (CVS release version) in 1st quarter 2003. Work is currently underway to extend this element to model P-M-V behavior.
  2. Implementation of a 2D beam-column joint model to account for joint panel deformations and member-based bond-slip response. Inelastic behavior is included through single or parallel/series material uniaxial material models. Transformation equations handle large deformation behavior, and element includes routines for sensitivity analyses. Element was published on the OpenSees website (CVS release version) in 1st quarter 2003. Work is currently underway to extend the joint model for 3D behavior with large deformations/rotations.
  3. Investigation of spurious internal forces arising from alternative definitions of viscous damping. This study was initiated by questions from Krawinkler’s team (thrust area 3), who encountered unexpected problems associated with the definition of damping coefficients in OpenSees. Our study discovered the source of the problem and reaffirms the practice of specifying either discrete damping coefficients or mass/stiffness proportional damping using the updated tangent stiffness.

Figure 1.
Larger View

Figure 2.
Larger View

  1. Preliminary calibration studies of the beam-column and joint elements have been completed.
  2. PEER report on beam-column and joint element development, implementation, and validation. (1st draft complete, May 2003; final publication expected July 2003).
  3. Preliminary implementation of new damage class has been implemented.
  4. In 1st quarter 2003, we released a new version (V.2) of the OpenSees TCL editor, which is available for download in the “tools” link on the OpenSees webpage.
Other similar work being conducted within and outside PEER and how this project differs

What most distinguishes this project from other research on inelastic analysis is as follows:

  1. it takes advantage of state-of-the-art computational technologies available in the object-oriented OpenSees framework
  2. the elements capture rapid stiffness/strength degradation that occurs in concrete frames as they approach their collapse limit-state
  3. integration with other models for structural and geotechnical components and probabilistic performance assessment tools, and
  4. the emphasis on calculating damage parameters to evaluate the component performance in the context of the PEER PBEE methodology.
Plans for Year 7 if this project is expected to be continued
None – the project should be completed by Dec. 2003.
Describe any instances where you are aware that your results have been used in industry
The models developed in this project have been implemented in OpenSees and are being used to simulate the testbed buildings. In time, we expect these types of models will find their way into engineering practice, either through the implementations in OpenSees or third-party implementations in commercial software.
Expected milestones
  • Feb. 2003 - Formal CVS release of 2D beam-column and joint models in OpenSees.
  • Sept. 2003 - Release of P-M-V beam column and 3D models.
  • July 2003 – Interim PEER report on 2D P-M beam-column and joint models
  • Dec. 2003 – Final PEER report on 2D P-M-V beam-column model
  • New 2D beam-column models, 2D and 3D joint models, and material models in OpenSees
  • On-line user and developer documentation
  • PEER Project Reports (based on 2 PhD theses)
  • Journal Publications on element formulations and sensitivity analyses