Reliability-based design optimization of a vehicular live load model. (1st August 2018)

• Record Type: Journal Article
• Authors: Kamjoo, Vahid; Eamon, Christopher D.
• Is Part Of: Engineering structures. Volume 168(2018)
• Page Start: 799
• Page End: 808
• Publication Date: 2018-08-01
• Subjects: Reliability-based design optimization -- Bridges -- Steel -- Prestressed concrete; Structural engineering -- Periodicals; Structural analysis (Engineering) -- Periodicals; Construction, Technique de la -- Périodiques; Génie parasismique -- Périodiques; Pression du vent -- Périodiques; Earthquake engineering; Structural engineering; Wind-pressure; Periodicals; 624.105
• DOI: 10.1016/j.engstruct.2018.05.033
• Languages: English
• ISSNs: 0141-0296

Abstract

Highlights: Current bridge live load models result in large discrepancies in reliability. Reliability-based design optimization is used to develop a new live load model. Significant improvement was found using the optimized model. Abstract: Typically, RBDO is used to select physical characteristics of a structure in order to optimize performance under uncertainty, where design variables are often taken as geometric or material parameters defining the structure. In this study, the effectiveness of using RBDO to develop a design load model is explored, which is to be used to design a wide range of bridge girders subjected to location-specific traffic loads. The objective is to minimize variation in reliability index among different bridge girders designed using the load model. Design variables are taken as the number of axles, axle spacing, and axle weights of the design vehicle representing the live load model. A probabilistic constraint is imposed, limiting the minimum reliability index of each girder design. Random variables considered are girder resistance in moment and shear, and bridge dead load and vehicular live load components. A single loop procedure is implemented by using a non-iterative, modified reliability method, the final results of which are verified with direct Monte Carlo Simulation. The optimization problem is solved with a genetic algorithm. It was found that procedure employed could be used to develop a design load model that resulted in substantial improvement over the design models in current use, where the optimized models significantly reduced the range and coefficient of variation of reliability index among the bridge cases considered.