Guidelines for Analysis Methods and Construction Engineering of Curved and Skewed Steel Girder Bridges (2012)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

140 4.1 Summary Based on the results of the research conducted on this project, the following conclusions may be drawn: •• Conventional 1D line-girder and 2D-grid methods of analysis are capable of predicting accurate construction responses in many situations; however, there are definite bridge geometries where significant reductions in accuracy can be expected. This research has provided a scoring method engineers can utilize as an aid to gage the accuracy of these simplified tools. Several examples are provided illustrating how the scoring system can be applied most effectively. •• The research identified a number of critical shortcomings in commonly used conventional methods and, in each case, provided mechanistic evaluations of the reasons for the shortcomings and recommended improved procedures that remove or alleviate these flaws. For I-girder bridges, the key critical flaws identified were: 1. The common dramatic underestimation of I-girder torsional stiffnesses by using solely the St. Venant torsional stiffness GJ/L in 2D-grid analysis methods. This flaw was addressed by the development and use of an equivalent St. Venant torsion constant that approximates the increase in the girder torsional stiffnesses due to the restraint of warping. 2. The common usage of equivalent beam elements for cross-frames that are unable to capture the physical load-deformation characteristics of these components. This problem was addressed by developing a procedure to obtain relatively accurate equivalent beam properties using a Timoshenko beam approximation rather than a Euler-Bernoulli equivalent beam element. In addition, “exact” equivalent beam elements were developed for a complete range of practical I-girder bridge cross-frame types. 3. The lack of any direct method of evaluating flange lateral bending stresses due to skew in I-girder bridges. The NCHRP Project 12-79 research developed an approximate procedure that works directly with the more accurate values of the cross-frame forces obtained using the above two improvements. 4. The lack of consideration of locked-in force effects associated with SDLF and TDLF detailing of the cross-frames. These locked-in forces are due to the lack of fit between the cross- frames and the girders in the initially fabricated (cambered and plumb) girder geometry. This issue was addressed by recommending a streamlined procedure for calculating these effects using cross-frame element initial strains, initial stresses, or initial (fixed-end) forces. Thorough case study examples were presented to provide practical guidance for when the influence of the above locked-in effects should be considered in design. Several areas of important improvements were also identified for tub-girder bridge analysis: 1. A method was developed for simplified estimation of the internal torques due to skew in tub girders, C H A P T E R 4 Conclusions and Recommendations

Conclusions and Recommendations 141 2. The impact of skew on box-girder cross-section distortion was directly evaluated and it was shown that the distortion associated with skew effects is typically minor, and 3. A method of accounting for a localized spike or “saw-tooth” in the longitudinal average normal stress distribution in the top flanges of tub girders, caused by the interaction with diagonals in the top flange lateral bracing (TFLB) system, was developed. •• Lastly, this research developed a guidelines document providing recommendations on the level of construction analysis, plan detail, and submittals for curved and skewed steel I- and tub-girder bridges. These guidelines were developed in a specification and commentary format suitable for direct incorporation into other specifications or guideline documents. 4.2 Recommendations for Implementation The recommendations for implementation of the results of this research are aimed at evolutionary improvements to the current state of practice for steel girder bridge engineering. These recommendations are primarily focused on the following items, which relate directly to the original scope of the research: 1. Improvements to Conventional Analysis Methods: Specifically, improvements to the modeling of I-girder torsional stiffness, the modeling of cross-frames’ overall stiffness in 2D-grid analyses, the calculation of flange lateral bending stresses from 2D-grid analyses, calculation of fit-up forces due to cross-frame detailing, and simplified analysis improvements for tub-girder bridges. 2. Definition of Erection Engineering Tasks: Specifically, a detailed list of recommended items to investigate as part of the erection engineering effort, with commentary, building on existing engineering guidelines as currently published by AASHTO. 3. Recommendations for Appropriate Level of Analysis Refinement: Specifically, a set of simple tables providing “letter grade” assessments of the anticipated accuracy of various analysis methods (1D and conventional 2D-grid vs. 3D benchmark solutions) corresponding to the framing and geometry of a given bridge. The first of these recommendations takes the form of explicit definition of the suggested improvements. The implementation of most of these recommendations would have to be undertaken voluntarily by the structural engineering software industry, but it is hoped that market pressures would encourage implementation. The implementation of the remainder of these improvements would be through education of the design community. It is recommended that the second and third recommendations would take the form of a guidelines document, published by AASHTO in the form of a guide specification that could be adopted, referenced, or excerpted by the various state DOTs. Various modifications to the AASHTO (2010 and 2010b) Specifications could be provided to make the Specifications consistent with the detailed guidelines. 4.2.1 Improvements to Conventional Analysis Methods This research produced a number of recommendations for improvements to conventional analysis methods. These recommendations are detailed in Section 3.2 of this report. A summary of these recommendations with implementation strategies is provided in Table 4-1. 4.2.2 Definition of Erection Engineering Tasks Currently, there is no nationally recognized guideline addressing erection engineering and erection plans for curved and/or skewed steel girder bridges. The closest nationally recognized

142 Guidelines for Analysis Methods and Construction Engineering of Curved and Skewed Steel Girder Bridges Improvement Report Section Description Implementation Strategy Improved I-Girder Torsion Model for 2D-Grid Analysis 3.2.2 Current 2D-grid methods typically neglect warping stiffness, which is a key parameter in the torsional stiffness of an I-girder. Methods for improving the torsional model of an I-girder are provided. Provide specific methodologies (presented in this report) to the bridge software industry and encourage their implementation in commercial bridge design software. Provide education (through this report and through associated presentations/publications) on this topic to the bridge engineering community. Encourage implementation in commercial bridge design software. Improved Equivalent Beam Cross-Frame Models 3.2.3 Current 2D-grid methods use simplified models of cross-frame stiffness that mispredict cross-frame load-deformation characteristics. A method for improving the modeling of cross-frame stiffness is provided. Improved Calculation of Girder Flange Lateral Bending Stresses 3.2.4 Current 2D-grid methods use a simplified approach for calculation of I-girder flange lateral bending stresses and do not provide any direct calculation of flange lateral bending stresses due to skew effects. An improved method is provided. Calculation of Locked-in Forces due to Cross- Frame Detailing 3.2.5 Currently, bridge engineers do not typically include locked-in forces in bridge design. Guidance on proper evaluation of lack-of-fit forces is provided. Provide specific methodologies (presented in this report) to the bridge software industry and encourage their implementation in commercial bridge design software. Provide education (through this report and through associated presentations/publications) on this topic to the bridge engineering community. Encourage implementation in commercial bridge design software. Simplified Analysis Improvements for Tub-Girder Bridges 3.2.6 2D-grid analysis is not capable of directly predicting all responses in tub girders, particularly with regard to internal framing responses. Improvements for simplified analysis methods are provided. Estimation of Fit- up Forces 3.3.5 Currently, bridge engineers do not typically evaluate fit-up forces in a consistently correct manner. Guidance on proper evaluation of fit-up forces is provided. Provide education (through this report and through associated presentations/publications) on this topic to the bridge engineering community. Software implementation of the recommended improvements from Sections 3.2.2 through 3.2.5 will permit the estimation of fit-up forces using simplified 2D-grid methods in I-girder bridges. Implementation of Section 3.2.5 to 3D FEA methods is essential for comprehensive evaluation of fit-up forces using these methods. The improvements recommended in Section 3.2.6 are not directly related to the evaluation of fit-up forces in tub-girder bridges. I- Table 4-1. Analysis improvements and recommendations for implementation.

Conclusions and Recommendations 143 guideline is the AASHTO/NSBA Steel Bridge Collaboration Guide Specification S10.1 – 2007, Steel Bridge Erection Guide Specification. This report addresses this lack of guidance in Appendix B, Recommendations for Construction Plan Details and Level of Construction Analysis. This appendix provides specific guidelines and commentary on recommendations for construction plan details and recommendations for methods of structural analysis and calculations. These guidelines are comprehensive and address all aspects of erection engineering plans and calculations. An owner-agency could adopt the guidelines as a complete specification, could reference the guidelines in their erection specifica- tions, or could adopt all or portions of the guidelines in their specifications. 4.2.3 Recommendations for Appropriate Level of Analysis Refinement Section 3.1 of this report outlines simplified equations to check for (and prevent) large second-order amplification in I-girder bridges. When tub girders are fabricated with proper internal cross-frames to restrain their cross-section distortion as well as a proper top flange lateral bracing (TFLB) system, second-order amplification of the overall deformations is practically nonexistent. Simplified rules are provided for identifying cases where overall overturning stability and potential uplift at bearing locations is more likely in both I-girder and tub-girder bridges. A basic scoring table is provided for assessing the anticipated accuracy of 1D line-girder and conventional 2D-grid methods as a function of the framing and geometry of a given bridge. With the implementation of the recommended improvements to the conventional 2D-grid methods, the accuracy of a 2D-grid analysis is improved to the extent that comparable solutions to 3D FEA are obtained for the assessment of gravity-load responses during construction as long as: •• There are at least two I-girders connected together, and •• They are connected by enough cross-frames such that the connectivity index I R n m C cf = +( ) 15000 1 is less than 20 (IC ≤ 20). 4.2.4 Recommendations for Specific Revisions to AASHTO Documents The research accomplished by NCHRP Project 12-79 covers both design engineering and erection engineering, as well as detailing, fabrication, and erection of steel girder bridges. As a result, there are some areas where more than one option exists for specific implementation of the recommendations of this project in the form of revisions to AASHTO documents. In some cases, a single recommendation for revisions to AASHTO documents is provided below; in other cases, when appropriate, a second recommendation is listed as an option. Also, as previously mentioned in Section 4.2.1, the success of the recommendations resulting from this research related to improvements to conventional analysis methods are critically dependent upon implementation in commercial software. Specific updates to provisions in the AASHTO LRFD Bridge Design Specifications are important to provide the endorsement and authority of AASHTO behind these recommendations, while these provisions must be written in a way to maintain freedom for software providers and engineers to use any legitimate method of analysis that provides sufficient accuracy for a given design. Detailed presentation of the procedures in AASHTO guidelines documents is critical for end users to understand the methods and how to use them.

144 Guidelines for Analysis Methods and Construction Engineering of Curved and Skewed Steel Girder Bridges Thus, the four primary options for implementation of specific revisions to AASHTO documents are: 1. Revisions to the AASHTO LRFD Bridge Design Specifications. 2. Revisions to the AASHTO LRFD Bridge Construction Specifications. 3. Revisions to appropriate AASHTO/NSBA Steel Bridge Collaboration Guideline or Guide Specification documents. 4. Separate publication and dissemination to the bridge industry. A key advantage to implementation in an AASHTO/NSBA standard is that doing so will offer thorough vetting of practice recommendations with broad representation, including owners, design engineers, engineers who perform erection calculations and analysis, fabricators, and contractors. The specific recommendations are listed in Table 4-2, with primary (and if appropriate, secondary) implementation suggestions. 4.3 Further Research Needs The NCHRP Project 12-79 research has provided a relatively comprehensive assessment and synthesis of the adequacy of simplified 1D and 2D analysis methods for prediction of the constructability and of the constructed geometry of curved and/or skewed steel girder bridges. A guidelines document has been developed based on this research, providing recommendations on the level of construction analysis, plan detail, and submittals for curved and skewed steel girder bridges. Nevertheless, there are a number of related areas that merit further study: •• Fit-up Practices—A focused, comprehensive investigation of the impact of various decisions and procedures on the fit-up of steel girder bridges during erection would be very fruitful. A fit-up decision is made on every steel bridge project and usually, due to lack of other direction, the decision is made by the fabricator. The decision impacts constructability of the bridge members during erection, loads in the steel bridge system, and the final bridge geometry. This practice has been customary from the earliest days of steel bridge construction, but there has been little study of actual implications of the decision. This investigation should address the various impacts on fit-up forces, locked-in stresses, and final constructed geometry. The collective knowledge of fit-up issues in the steel bridge industry today (2012) is based almost entirely on qualitative experience. Partial knowledge of each aspect of the issues is typically highly compartmentalized: steel detailers, fabricators, and erectors have knowledge and preferences on detailing practices, designers have knowledge and preferences on how to perform structural analysis for final conditions, and owners have knowledge and preferences regarding the final geometry of bridges. A comprehensive knowledge of all aspects of these issues by all parties is lacking. Further- more, all parties only have limited understanding regarding the possible implications of detailing methods on the structural behavior. This research should involve more than just the application and exercise of sophisticated analytical tools; the analytical assessments will be most useful if they are coupled with high-resolution, high-quality field measurements. The emphasis should be placed on I-girder bridges with NLF, SDLF, and TDLF detailing, but some assessment of the specific causes of fit-up issues in tub-girder bridges also would be a valuable contribution. •• Early Concrete Deck Stiffness and Strength—More extensive coupled field and analytical evaluation of the effects of early concrete deck stiffness and strength gains, including the influence of staged concrete deck placement would be very valuable. Prior research addressing this consideration shows generally that significant early stiffness and strength gains can exist. However, the studies have been limited to only a few bridges and a few parameters of the (continued on page 154)

(continued on next page) Table 4-2. Recommendations for specific revisions to AASHTO documents.

Table 4-2. (Continued).

(continued on next page)

Table 4-2. (Continued).

(continued on next page)

Table 4-2. (Continued).

(continued on next page)

Table 4-2. (Continued).

154 Guidelines for Analysis Methods and Construction Engineering of Curved and Skewed Steel Girder Bridges concrete mix design and methods of construction. A more comprehensive understanding of the actual early-age behavior during and after placement of concrete decks is needed if engineers are to take optimum advantage of the early strength gains. Furthermore, as noted in the NCHRP Project 12-79 research, there is no such thing as a conservative displacement prediction. Sufficient measurements and corresponding analytical predictions are needed to allow the calculation of confidence limits for the predicted displacements during and after the concrete deck placement. •• Innovative Framing Arrangements—Further studies of innovative framing arrangements to mitigate nuisance stiffness effects in skewed girder bridge construction would be useful. For example, as detailed in the Task 8 report of the NCHRP Project 12-79 research, further research should be conducted to investigate the use of skewed intermediate cross-frames at angles larger than 20°, combined with a split-pipe connection detail to mitigate the prob- lems of connecting to the girders at a sharp skew angle. As demonstrated in the NCHRP Project 12-79 research, nuisance stiffness effects are mitigated best by making the intermedi- ate cross-frames parallel to the bearing lines in parallel-skew bridges and by “fanning” the intermediate cross-frames between the skewed bearing lines in bridges with non-parallel skew. The behavior of straight and curved bridges with these arrangements should be inves- tigated in more detail, including the consideration of other impacts that these cross-frame arrangements may have on the design behavior and construction. •• Tub Girders with Pratt TFLB Systems—The impact of Pratt top flange lateral bracing (TFLB) system layouts on the internal forces in tub-girder bridges needs to be better understood. The NCHRP Project 12-79 research showed that a conventional 2D-grid analysis, coupled with commonly used tub-girder bridge component force equations, has particular difficulty in predicting the response of these types of bridges. Further improvements to 2D-grid analysis methods may be possible to make these methods viable for the design of tub-girder bridges with Pratt TFLB systems. •• Live-Load Effects—Lastly, the emphasis of the NCHRP Project 12-79 research was on analysis for construction engineering of steel girder bridges. Parallel studies should be conducted to evaluate the accuracy of simplified methods of analysis for the prediction of the live-load response of bridges. Of concern is the tedious nature and limited accuracy of traditional load distribution factor calculations for horizontally curved and/or skewed girder bridges as a function of the complexity of the bridge geometry. Engineers need to better understand the limits of their analysis calculations regarding the live-load response of steel girder bridge systems.