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.
101Â Â Conclusions 4.1 Applicability of Results to Improving Current Practice The overarching goal of the project was to aid directly in the safe and economical design of highway sign support structures. The project results are of direct practical use to civil engi- neers who are engaged in the design of highway traffic signs and their support structures. The research results (e.g., methodologies to estimate wind loads) were developed with the end user in mind. The main target audiences are (1) staff members at federal and state government agen- cies, who are in charge of designing support structures for highway signs; and (2) members of the AASHTO T-12 Technical Committee on Structural Supports for Highway Signs, Luminaires, and Traffic Signals, which is responsible for updates to the AASHTO specifications used for the design of highway sign support structures. Additional audiences also will benefit from this research, such as diverse engineering consultancies and construction organizations. These audiences are based not only in the United States but also internationally given that design guidelines developed under the auspices of NCHRP projects are readily adopted in other countries. Inclusion of the main research findingsâwind drag coefficients for highway signs and their support structuresâin the LRFDLTS-1 specifications will ensure that the main results of NCHRP Project 15-67 have a significant impact on current practice. The LRFDLTS-1 specifications account for only a subset of the relevant geometrical and flow parameters that affect the drag coefficient for commonly used highway traffic signs. Recent research suggests that some of the recommended values of the drag coefficient underestimate the wind loads for thin rectangular static signsâin part because the drag coefficient for thin signs is only a function of the sign aspect ratio in the LRFDLTS-1 specifications. Thus, the effect of the ground clearance distance is ignored. The specifications recommend a rather large drag coefficient value for dynamic message signs. This value does not appear to be based on wind-tunnel experiments or on any other accurate method to estimate wind loads. The LRFDLTS-1 specifications do not make any direct recom- mendations on estimating the drag coefficient for signs with complex shapes that are commonly used on U.S. highways or for multiple signs placed next to each other. Applying the findings of this research (new values of drag coefficients for highway signs) in practice should be straightforward. The main challenge was to include the new wind drag coefficients in a document that is widely available to concerned audiences. One of the main requirements of the RFP for NCHRP Project 15-67 was the formulation of revisions to LRFDLTS-1 specifications. Once incorporated into the LRFDLTS-1 specifications, the new drag coefficient values will serve as the primary reference for estimating wind loads on highway signs. C H A P T E R 4
102 Wind Drag Coefficients for Highway Signs and Support Structures The other main research task undertaken as part of NCHRP Project 15-67 was to propose methods to estimate wind drag coefficients on the members of truss structures and on the tubes and chords of monotube and truss structures supporting highway signs. Although methods were proposed in the past and are being used by some state DOTs, the LRFDLTS-1 specifications have not adopted any of them. This decision may be attributable to insufficient research sup- porting those methods and to the procedures themselves, which can be confusing and difficult to apply. This situation also poses the main challenge for the new methods proposed as part of this project. In this regard, it is essential to achieve clarity in the formulation of the design methods and to develop a comprehensive set of design examples covering the main types of sign support structures. The design examples in Appendix B cover the main types of configurations that are of inter- est for practical applications. These design examples could form a companion document to the revised LRFDLTS-1 specifications and would be accessible from a website maintained by NCHRP. Although the objective of NCHRP Project 15-67 was to estimate wind loads on highway signs and their support structures to enhance strength design, the drag coefficients and the method for calculating the wind loads can also be used to estimate equivalent (along the wind direction) static wind loads for fatigue design. Appendix B includes two such examples (in Sec- tions B1.4 and B5.4) where the equivalent static natural wind gust pressure and the equivalent static natural truck gust pressure are estimated by employing the procedure recommended in Section 11.7.1 of the LRFDLTS-1 specifications, which is based on using drag coefficients esti- mated under steady wind conditions but with a different wind velocity than the one used for strength design. This report suggests using the new method to estimate the drag coefficients for highway signs when making such fatigue design calculations. Once these methods are adopted in the LRFDLTS-1 specifications, the applicability of these methods to improving current practice should be immediate because these methods are formu- lated in a way that is easy to understand and can be applied to a wide range of support structures. The research team worked closely together with members of both the project panel and the AASHTO T-12 Technical Committee on Structural Supports for Highway Signs, Luminaires, and Traffic Signals to ensure that the final product met these standards. 4.2 Conclusions and Recommendations The research conducted as part of this project accomplished its main objectives by develop- ing new general methods to estimate wind loads on highway signs and their support structures that can be incorporated into the LRFDLTS-1 specifications. These methods not only apply to multiple highway signs placed on overhead bridge-type and cantilever-type truss structures and on overhead bridge-type and cantilever-type monotube structures but also account for the inter- actions and shielding effects between the signs and the support structures. Thus, the new methods fill an important research gap that arises because current LRFDLTS-1 specifications take into account just a subset of the relevant geometrical and flow parameters that affect the drag coef- ficient for commonly used highway traffic signs. Moreover, the specifications offer only very simplified rules to estimate wind loads on sign support structures. If the LRFDLTS-1 specifica- tions incorporate more detailed and more accurate information on wind drag coefficients for highway signs and more accurate methods to estimate wind loads on sign support structures, the structural (strength) design of these structures should improve. As a result, the service life of the highway sign support structures used throughout the United States will be extended. The current research project demonstrated that the drag coefficient for static rectangular signs varies significantly not only with the sign aspect ratio b/h but also with the ratio h/(h + hg) between the sign height and the sum of the ground clearance distance and the sign height.
Conclusions 103 The last effect is not accounted for in the current LRFDLTS-1 specifications. When b/h > 3, the drag coefficient for the static sign sharply increases with higher values of h/(h + hg). This result also explains in part why reported drag coefficients for static signs as a function of b/h from different wind-tunnel studies are not always consistent. The current research project also showed that the drag coefficient value recommended in the LRFDLTS-1 specifications for dynamic message signs is not justified, especially because the recommended value is independent of b/h and h/(h + hg). Consistent also with the recent wind- tunnel study by Chowdhury et al. (2015), the current research confirmed that increasing the sign thickness results in a slight decrease of the drag coefficient for the sign. The current research project showed that an add-on sign has the effect of slightly increasing the drag coefficient for the full sign. Wind-tunnel experiments conducted as part of this project demonstrated that the numerical model accurately predicted drag coefficients for signs that include an add-on sign. The current research project also found that the presence of the sup- port structure behind the sign increases the drag coefficient for the sign and that the wind loads acting on side-by-side signs are a function of the nondimensional gap distance between the two signs. For a certain range of the nondimensional gap distance, the wind loads acting on side- by-side signs can be up to 30% higher compared to the corresponding wind loads acting on the signs that are not in the vicinity of another sign. This important result was also confirmed by new wind-tunnel experiments conducted as part of this project. These findings are significant considering that the current LRFDLTS-1 specifications do not provide any direct recommenda- tions on estimating the drag coefficient for signs with complex shapes that are commonly used on U.S. highways or for multiple signs placed next to each other. The project findings led to a new method for estimating the drag coefficient on a highway sign placed on a monotube or truss structure that supports one or multiple signs. Once the drag coefficient for an equivalent isolated sign is estimated as a function of b/h and h/(h + hg), several modification factors are applied to account for the effects produced by an add-on sign, the proximity of another sign, and the sign support structure. The current project research suggests that the effect of sign thickness can be neglected. The current project research also found that highway signs placed on a grade separation struc- ture are subject to significantly larger wind loads when compared to signs with no bridge struc- ture. Moreover, depending on the bridge design (e.g., presence of a barrier rail or a separation rail), the vertical distribution of the wind loads acting on the sign can be very nonuniform. This variation is especially large for signs of large height with a top edge situated above the top of the rail. This information is critical for correctly designing the supports used to affix the sign on the bridge. As part of the current research, a method was developed to divide the surface of the sign into several horizontal subzones and to prescribe the drag coefficients for each subzone as a function of the wind direction, sign aspect ratio, and relative position of the top of the sign with respect to the top of the barrier rail or separation rail. As a result, the total wind load acting on the sign can be estimated more accurately with this new method. The current project research also showed that drag coefficients for circular cross-section and L-shaped members are a function of not only the Reynolds number but also the aspect ratio, defined as the ratio between the member length and the member width. For aspect ratios that are representative of chords and secondary members in monotubes and trusses, the drag coeffi- cients may be significantly lower than those assumed for infinitely long members at the same Reynolds number. A physically based estimation of wind loads on a structure supporting one or multiple signs should consider the shielding of the structure induced by the signs as well as the increase in wind load associated with the acceleration of the approaching wind as it passes the signs (i.e., the
104 Wind Drag Coefficients for Highway Signs and Support Structures amplification zone). Simulation results confirmed the presence of such amplification zones and enabled estimation of the increase in wind loads for the members (or parts of members) situated in the acceleration zones (which form close to the edges of the signs) and the length of these zones. A similar phenomenon is present in each gap region between side-by-side signs when multiple signs are placed on the support structure. Simulation results also allowed (1) estimating the reduction in the wind loads acting on chords and secondary members that were partially shielded by other members of the sign support structure and (2) quantifying the amplification or reduction of these wind loads with respect to the ones acting on an identical isolated member. The wind loads acting on the secondary members and the parts of the chords situated directly behind the signs were found to be negligible. These data supported proposing a general method to estimate wind loads on monotube and truss structures supporting one or multiple signs. The idea is to first calculate a drag coefficient for each member as a function of the Reynolds number, cross-sectional shape, and aspect ratio (assuming that the member is situated far from any other members and thus free of shielding effects). Rules are proposed that indicate how to assign each part of a chord and each secondary member to the appropriate flow region (i.e., behind-the-sign, flow-acceleration, gap, or uniform-flow region). Then, a modification factor is introduced for truss chords to account for interactions among chords and for shielding effects attributable to the possible presence of another chord in front of the chord in question. A second modification factor is introduced for both chords and secondary members to account for possible amplifica- tion of the drag coefficient inside the flow-acceleration regions and over parts of each gap region and to consider possible shielding effects by other secondary members. The proposed method also allows estimations of wind loads on gusset plates in truss struc- tures, using a similar approach in which the drag coefficient of an isolated gusset plate fully exposed to the incoming wind is modified by a modification factor accounting for shielding effects. Extensions of the method are proposed to approximately estimate both wind loads on vertical columns and wind loads applied transverse to the plane of sign support structures so that a full analysis of the structure can be undertaken under relevant loads (e.g., normal, transverse, or a combination of both). The design examples included in this report indicate that, for some cases, the wind loads esti- mated using the new methods can be significantly higher than those obtained by applying the rules in the current LRFDLTS-1 specifications. For example, some of the highway signs were found to be subject to wind loads as much as 40% higher than the ones estimated based on the current LRFDLTS-1 specifications. However, the assumption should not automatically be made that new methods always predict larger wind loads when compared to the existing procedures in the current LRFDLTS-1 specifications. For example, the wind loads acting on the truss structures in two design examples in Appendix B (Design Examples 2 and 3) were as much as 20% lower when estimated using the new methods. Although the calculations in the proposed methodology are more involved, the main advantage lies in more accurate wind load estimations and the fact that the new method is directly related to the physics of the airflow as it interacts with obstacles in the form of highway signs and members of sign support structures. 4.3 Suggested Research The proposed methods to determine drag coefficients on highway signs and sign support structures based on high-resolution CFD simulations are very general and can be extended to other types of signs, luminaires and support structures. The suggestions for future research in this section would benefit from using the research approach developed as part of the current study to estimate wind loads for other relevant types of structures that are covered by AASHTO LRFDLTS-1 specification recommendations or are of interest to other manuals and design guidance documents used by state DOTs.
Conclusions 105 Although the current research project includes recommendations for determining wind loads on vertical columns of circular cross-section used to support monotubes, additional research is needed on how to more accurately calculate the drag coefficients for vertical columns as a func- tion of their cross-section (e.g., circular, hexagonal, dodecagonal, octagonal, square), Reynolds number, and especially aspect ratio. The expectation is that such a study will result in lowering the recommended values of the drag coefficients for vertical columns given that the current AASHTO recommendations are based on values obtained for infinitely long cylinders of various cross-sections. Additional research is also recommended to refine procedures for estimating wind loads on signs and their support structures for transverse loads. Using the approach developed in the current research project, more accurate recommendations for estimating shielding effects can be developed. In addition, research is recommended to estimate wind loads on both sides (e.g., front and back sides with respect to the incoming wind direction) of structures spanning diagonally across roadway intersections that support multiple traffic signs and signals. More accurate recommendations would be beneficial for estimating both wind loads on the signs situated on the back side of the support structure and wind loads on the parts of the support structure in front of those signs. The proposed methods can be extended to determine drag coefficients for traffic signals, temporary traffic sign supports used in work zones for traffic control, flat elements of non- rectangular shape that connect truss chords on the vertical columns, luminaire supports, and vertical poles of various cross-sectional shapes and high-mast lighting towers. Of particular interest is using an approach similar to the one in the current project to estimate wind loads on horizontal trusses and develop a general procedure to estimate wind loads on trussed towers. Another recommendation is that more research should be conducted to identify the types of loads that should be considered in the design of the different sign support structures (e.g., normal, transverse, a linear combination of normal and transverse loads) given that various specifications include different recommendations.
