4th International Conference on Earthquake Engineering Taipei, Taiwan October 12-13, 2006 Paper No. 138 SELF-LEARNING SIMULATION FOR MODELING OF BEAM-COLUMN CONNECTIONS IN STEEL FRAMES Gun Jin Yun 1 Jamshid Ghaboussi 2 and Amr S. Elnashai 3 ABSTRACT The inelastic response of beam-column connections significantly affects the seismic performances of steel frames. Accurate models of beam-column connections are important both in understanding the seismic behavior and in design. An effective approach towards developing accurate models for connections is self-learning simulation, which is presented in this paper. The latter simulation method has the unique advantage in that it can extract material models from global structural response. Self- learning simulation is based on auto-progressive algorithms that employ the principles of equilibrium and compatibility, and the self-organizing capability of artificial neural network material models. Steel beam-column connections are represented by point plastic hinges that are modeled by a new neural network representation for modeling of cyclic behavior of connections. The neural network based connection model has significant advantages over conventional models in that it can handle complex behavior due to local buckling and tearing of components. Moreover, its implementation is more efficient than conventional connection models since it does not need interaction equation and plastic potential. In this paper, the performance of the neural network based connection model is verified with experimental data and the proposed self-learning simulation methodology is demonstrated with a simulated structural testing of a two-story one bay frame with semi-rigid connections. Keywords: Nonlinear Finite Element Analysis, Neural Network, Steel Beam-Column Connection, Inelastic Hysteretic Model, Self-Learning Simulation INTRODUCTION The behavior of connections significantly affects global responses of building structures under earthquakes because they dissipate a large amount of energy due to hysteretic damping. Therefore, accurate modeling of the real connection behavior under cyclic loading is essential for the seismic performance assessment. Even though retrofitted and/or new seismic design can protect connections from brittle failures, they, to a certain extent, are still vulnerable to earthquake shaking due to unexpected 3D loading scenario, variations in weld details and quality, geometric imperfection and residual stresses in heat affected zones, etc. Many mathematical models have been developed for modeling the cyclic behavior of connection regions with various degrees of complexity. Even though simplified models such as bilinear or tri-linear model with kinematic hardening representation have 1 Graduate Research Assistant, Department of Civil and Environmental Engineering, University of Illinois at Urbana- Champaign, Urbana, Illinois, 61801, U.S.A. Email: gyun@uiuc.edu , Tel:1-217-333-6957 2 Professor Emeritus, Department of Civil and Environmental Engineering, University of Illinois at Urbana-Champaign, Urbana, Illinois, 61801, U.S.A Email: jghabous@uiuc.edu , Tel:1-217-333-6959 FAX:1-217-265-8039 3 Bill and Elaine Hall Endowed Professor of Civil Engineering, Director of Mid-America Earthquake Center, Department of Civil and Environmental Engineering University of Illinois at Urbana-Champaign, Urbana, Illinois, 61801, U.S.A. Email: aelnash@uiuc.edu , Tel:1-217-265-5497, FAX:1-217-265-8040