Reliability-based design of corrugated web steel girders in shear as per AASHTO LRFD
Introduction
AASHTO adopted the concept of Load and Resistance Factor Design for bridges and published its first Bridge Design Specifications in 1994 [1]. In the framework of LRFD, designs are generally economical and meet a pre-set safety level, which is represented by a target reliability index. In the development of resistance factors for AASHTO LRFD, the specifications considered typical construction materials, including timber, reinforced concrete, prestressed concrete, and structural steel. For structural steel, the specification addressed girders, trusses, arches, cable-stayed, and suspension.
Around the world, the most common type of steel bridges is composed of a cast-in-place concrete slab that is supported on multiple steel girders oriented along the bridge centerline. Such a structural form is widely used for short and medium spans because it can accommodate different alignments and geometric constraints, including curvature and skewness. The concrete slab and steel girders can act together as a composite system through shear connectors to support the entire load on the bridge. The steel girders, which can have I- or box cross-section, carry the imposed load by the internal actions of shear and bending moment (with the help of the slab, if composite action is present). The system is economical and long lasting for the vast majority of bridges, although stability concerns limit the stresses, especially during construction.
Structural design codes and specifications include rules and guidelines that specify minimum acceptable level of safety for constructed facilities. They aim at protecting public health, safety and the general welfare in relation to construction. Bridge specifications become law of a particular jurisdiction when formally enacted by the appropriate authority, such as the department of public work. Nowadays, the majority of the bridges around the world are designed following the Load and Resistance Factor Design (LRFD) philosophy. Such an approach considers the variability of loads and material properties in structural design in order to ensure uniform safety. Reliability calibration procedures for the AASHTO LRFD Bridge Design Specifications are well documented in the available literature [2,3]. New research on the statistical parameters of loads and material properties often results in proposed modifications to existing load and resistance factors [4,5]. Development of new material that is not adequately covered in the current specifications, such as fiber reinforced polymer concrete, requires the development of resistance factors to meet the target reliability [6]. Also, the move from working stress design to LRFD-based approach in the field of geotechnical engineering has prompted research in the area [[7], [8], [9], [10]]. Adapting the AASHTO Bridge Design Specification into a country that have different traffic characteristics and construction practice than the US may require revisiting the load and resistance factors [11].
With the development of LRFD, reliability-based design is now considered the standard in designing new bridge structures. In this context, load and resistance data are treated as uncertain quantities characterized by appropriate probabilistic models. In the framework of LRFD, load effects on structural members are scaled-up by a load factor, and resistance is scaled down by a resistance factor. In this way, it is expected that the structural member would achieve the reliability index that is consistent with the design standard.
To determine the appropriate load and resistance factors, a calibration process is needed to achieve the target reliability. This process is based on a series of reliability analyses assuming a large database of structural members and statistical data on both load and resistance models. For bridges in the US, Nowak [2,3] presented a calibration procedure to develop load and resistance factors for LRFD bridge code. In this work, the load and resistance were treated as random variables and described through their bias factors, coefficients of variation, and probability distributions. The calibration process assumed a target reliability index of 3.5, which was selected based on the performance of existing bridges that were designed based the AASHTO standard specifications to ensure consistent and uniform safety margin for all structures. More recently, Nowak and Iatsko [4] revised the load and resistance factors for the AASHTO LRFD bridge design specifications due to the fact that the existing specifications used statistical parameters dating back to the 1970s and 1980s.
From the above review, it becomes clear that probabilistic-based approaches are becoming the de-facto tool to design structural members and there is always a necessity to update design codes and standards by re-calibrating load and resistance factors in light of the availability of new statistical data, development of new materials, and advancement of reliability methods. Additionally, there is a growing interest in using new structural forms, such as corrugated web steel girders (CWSGs) for the construction of bridges [12]. For example, Abbas [13] investigated the behavior of girders with corrugated web made of high-performance steel. The popularity of CWSGs as a viable alternative to welded plate I-girders has led to its adoption in many bridge projects; examples can be found in [14,15]. One notable example is the demonstration bridge built by the Pennsylvania Department of Transportation [16]. This bridge is made of four equally spaced girders with corrugated web formed using high-performance steel plates (HPS-485 W). Large-scale tests on the demonstration bridge enabled the development of new shear and flexural strength criteria, but no reliability-based studies were conducted.
The present paper aims at addressing a deficient research area concerning CWSGs; namely, the development of reliability-based resistance factors for shear design. To do so, first there is a need to evaluate the current shear resistance factor of welded plate girders as per the AASHTO LRFD Bridge Design Specifications [17], and then calibrate the resistance factor for CWSGs in shear as per the current edition of the specifications. Moreover, sensitivity analyses are carried out to investigate the most influential design parameters on the reliability index of CWSGs.
Section snippets
Problem statement and research objectives
No matter how good their performance is, the widespread use of CWSGs in bridge construction necessitates the backing of mainstream design codes and specifications, which is lacking at the moment. Design codes offer design guidelines and load and resistance factors to ensure a uniform design of structural components to meet a certain safety level. In terms of the modern load and resistance factor design (LRFD) philosophy, this means that all designed components shall achieve a target reliability
Representative bridge models
The first step in any resistance factor calibration process requires the preparation of representative bridge models to perform reliability analysis. In the present study, three bridge cross-sections were selected as shown in Fig. 1. The first bridge cross-section (Fig. 1(a)) consists of 210 mm thick concrete slab on seven girders spaced at S= 2 m, the second (Fig. 1(b)) consists of 230 mm thick concrete slab on five girders spaced at S= 3 m, and the third (Fig. 1(c)) consists of 250 mm thick
Statistical parameters of resistance and load models
In this section, the statistical parameters for load and resistance models are presented. These data are the basic required information to carry out a reliability analysis. Generally, to carry out a standard structural reliability analysis and in order to accurately estimate the reliability index, the basic requirements are the mean, coefficient of variation, as well as the type of distribution for each random variable. The random variables in the context of structural reliability include the
Design equation and target reliability
Modern structural design codes are based on the concept of LRFD, in which a structural member is designed to have a capacity that is greater than the maximum load effect during the expected life of the member, that is:in which, ϕ is the resistance factor that is often equal or less than the unity; Rn is the nominal resistance; Qi is the nominal effect of load component i, γi is the load factor for Qi, which is often greater than unity.
Keeping up with the same form of Eq. (14), as per
Limit state function and reliability analysis
In the calibration process, the design equation (Eq. (15)) serves as the basis to develop the limit state function, G:in which G is a random variable that represents the safety margin; R is the random variable representing the resistance; Q is the random variable representing the load effects. Considering the components of load effects represented in Eq. (15), the random variable Q used for the Strength limit state function is given by:
Reliability analysis is carried
Results and discussion
The results of reliability analyses performed for designs based on the current AASHTO LRFD bridge design specifications [17] are presented first by considering welded steel plate girders and then by addressing corrugated web steel girders. Results from the first set of analyses are compared to previously published work, while the results of the second set are used to calibrate the resistance factor for corrugated web steel girders in shear based on the obtained target reliability index.
Sensitivity analysis
The reliability indices computed in Section 7.2 of this study depend on several factors that include the design parameters (i.e., geometry and material properties of the corrugated web: {tw, hw, b, b/c, θ, Fy, E}), girder spacing, S, span length, L, in addition to the average daily truck traffic, ADTT. To quantify the relative contribution of each design parameter to the reliability index, we employ the one-at-a-time type of reliability-based sensitivity analysis. This is the simplest and the
Summary and conclusions
A series of reliability analyses were performed to calibrate the resistance factor for corrugated web steel girders in shear following the current AASHTO LRFD bridge design specifications. The reliability analyses employed the Importance Sampling method. The study included the computation of the reliability indices for welded plate girders in shear considering the current load and resistance factors. It was found that the current resistance factor, ϕv = 1.0 doesn't always ensure a design that
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This research was supported by the SCMASS research group of the University of Sharjah and the American University of Sharjah, United Arab Emirates. The authors are grateful for these supports.
References (40)
- et al.
Reliability evaluation of axially loaded steel members design criteria in AASHTO LRFD bridge design code
Reliab. Eng. Syst. Saf.
(2013) - et al.
Calibration of resistance factors for reliability based design of externally-bonded FRP composites
Compos. Part B Eng.
(2008) - et al.
Lateral-torsional buckling strength and behaviour of high-strength steel corrugated web girders for bridge construction
Thin-Walled Struct.
(2018) - et al.
Normalized shear strength of trapezoidal corrugated steel webs: improved modeling and uncertainty propagation
Thin-Walled Struct.
(2019) - et al.
Shear strength of trapezoidal corrugated steel webs
J. Constr. Steel Res.
(2011) - et al.
Shear strength and design of trapezoidally corrugated steel webs
J. Constr. Steel Res.
(2009) - et al.
Shear strength of trapezoidal corrugated steel webs
J. Constr. Steel Res.
(2013) AASHTO LRFD Bridge Design Specifications
(1994)Calibration of LRFD bridge code
J. Struct. Eng.
(1995)- et al.
Calibration of LRFD Bridge Design Code
(1999)