Studies on Probability-Based Design for Residential Construction.

摘要

Housing construction in the United States is based largely on accepted practices rather than engineering principles. And while the performance of residential structures has been generally acceptable, the economic losses associated with recent natural disasters has prompted a reexamination of the basis for the design of these structures. Unfortunately, such re-examinations are often done with limited, anecdotal information. In addition, engineering specifications, which apply to a wide variety of buildings, rarely consider the unique experience of residential construction and establish minimum performance expectations (i.e., target reliability) based on past design practice for 'engineered' structures. The objectives of this study were to: (1) quantify the 'experience' of conventional residential construction in terms of modern concepts of reliability as used in the development of engineering specifications; (2) lay the groundwork necessary to improve probability-based design procedures as applied to residential (wood-frame) structures; (3) identify ranges of reliability that may be associated with acceptable performance such that minimum or target reliabilities may be contemplated for efficient residential building design to eliminate unacceptably low performance while maintaining a historically consistent perspective on housing value (i.e., safety vs. affordability); and (4) investigate the efficacy of a multi-story live load reduction factor to account for some degree of time independence between maximum loads realized on any one story level of a home.