A METHODOLOGY FOR INCORPORATING UNCERTAINTIES AND MULTIPLE OBJECTIVES IN THE COMPUTER-AIDED DESIGN OF SEISMIC-RESISTANT STEEL STRUCTURES (EARTHQUAKES, OPTIMIZATION, PROBABILITY)

This dissertation focuses on the development and preliminary testing of a methodology for the probabilistic limit states design of seismic-resistant steel structures. Emphasis is placed on the formulation of a mechanism which allows a designer to include the effects of uncertainties and multiple design objectives in an optimization-based design process.; Sources of uncertainty in the seismic design environment are identified and described. The concept of GOOD, BAD and HIGH, LOW preference pairs is proposed as a mechanism for allowing a designer to impart aspects of his or her engineering judgement and intuitive knowledge to the design process. Scaling procedures for combining the aforementioned effects, and the statistics of frame response into a single design entity called designer dissatisfaction are given.; The design method is demonstrated via the design of a three story, single bay, moment resistant steel frame using a computer-aided design system called DELIGHT.STRUCT. Linear and non-linear time history analyses are built into the design procedure itself rather than serving as a check at the end of the design process. The frame's performance is assessed on the basis of its statistical response to gravity loads alone, gravity loads plus a family of moderate earthquakes and finally gravity loads combined with an ensemble of rare severe earthquake ground motions. Design objectives include the frame volume, minimum story drifts and maximum hysteretically dissipated energy. The Phase I-II-III Method of Feasible Directions is used to solve the constrained optimization problem. Examples are presented to demonstrate the proposed method for a selection of single design objective and multiple design objective problems having various design parameter layouts.