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.
6 Introduction and Research Approach 1.1 Problem Description and Problem Statement With increased traffic, multilane highways, and complex highway interchanges, highway signs play an ever more important role in the safe operation of the nationâs transportation net- work. A detailed understanding of stresses during the service life of sign support structures is crucial for their safe and economic design. The primary load on highway signs is attributable to wind-induced mean drag. Thus, when designing economical and reliable structures to support highway signs, the first requirement is to accurately estimate mean (time-averaged) wind loads on the various types of highway signs. However, sign support structures are themselves exposed (or partially exposed) to the incoming wind. Consequently, the second requirement is to accu- rately estimate wind loads on the members of sign support structures. The drag coefficients in the current AASHTO Load and Resistance Factor Design (LRFD) Specifications for Structural Supports for Highway Signs, Luminaries, and Traffic Signals (LRFDLTS-1) (AASHTO 2015) are not always consistent with those reported in recent research studies or with those recommended by the Minimum Design Loads for Buildings and Other Structures (ASCE/SEI 7-16 2017, now ASCE/SEI 7-22 2022). This inconsistency is especially evident for complex-shaped signs, multiple signs, and dynamic message signs (DMSs). The structures supporting highway signs can be classified into three main types: bridge, cantile- ver, and butterfly (double-cantilever). Bridge-type structures are supported by vertical posts on each side of a roadway and can have long spans. The posts may contain one or two columns or poles. The horizontal structure may be a truss or a monotube. These structures are also called span-type struc- tures or sign bridges, as noted in NCHRP Report 469 (Dexter and Ricker 2002). The cantilever-type and butterfly-type structures contain a single vertical post that supports the horizontal structure (a truss or a mast arm) on which the highway signs are attached. Typically, cantilever- and butterfly-type support structures are less expensive and also more flexible than bridge-type sup- port structures. The horizontal structure extends on only one side of the post for cantilever-type structures but on both sides of the post for butterfly-type structures. Cantilever- and butterfly- type support structures typically feature shorter spans than bridge-type structures and are also subject to large-amplitude vibrations. Thus, the danger of wind-induced failure for these struc- tures is much greater when compared to bridge-type structures, as noted in several NCHRP reports (Kaczinski, Dexter, and Van Dien 1998; Dexter and Ricker 2002; Fouad et al. 2002). For truss structures, the main geometrical subtypes are 4-chord, 3-chord, and 2-chord trusses. The cross-section of a 4-chord truss is a rectangle while the cross-section of a 3-chord truss is a triangle; therefore, these structures are also commonly referred to as box and tri-chord trusses, respectively. Secondary members connect between the chords. Each 2-chord truss consists of two parallel horizontal chords with secondary members connecting the two chords in a vertical plane (also called planar trusses). The horizontal chords of the different types of truss structures generally have much larger diameters than those of the other truss members and are bolted C H A P T E R 1
Introduction and Research Approach 7 together for longer spans. Circular cross-section and L-shaped members most commonly are utilized for trusses supporting highway signs. Members with square cross-sections are some- times used for the chord members. Monotubes typically include one tube, chord, or beam, but in some cases, two horizontal tubes are employed to increase the rigidity of the sign support structure. Typically, the tubes have a constant diameter or cross-section. The aforementioned types of monotube and truss structures are now expected to support multiple signs of various sizes, aspect ratios, shapes, and weights. In particular, the goal of pro- viding drivers with up-to-date information on highway conditions has resulted in increased use of DMS cabinets. These cabinets offer a high-tech alternative to static signs and can display real- time information and advice for drivers. DMS cabinets are typically much heavier, with depths that differ from those of conventional highway static signs (Chowdhury et al. 2015). Increasing evidence suggests that monotube and truss structures supporting various large and heavy signs are subjected to much more complex wind loading than current strength design procedures commonly consider. For example, the wind loads on two side-by-side highway signs can exceed those computed for the same two signs when not adjacent to each other. Moreover, the wind loads on highway signs can increase or decrease when the signs are mounted on a sign support structure (compared to signs with no support structure). On the other hand, the pres- ence of highway signs can reduce wind loads on parts of the sign support structure shielded by the signs and can amplify wind loads over other parts of the sign support structure when compared to the corresponding wind loads estimated with no signs mounted on the support structure. Considering the risk to public safety and the considerable cost of prematurely replacing these structures, there is a demonstrated need to accurately estimate the wind loads on which the structural design of sign support structures is based. 1.2 Knowledge Gaps An important requirement in estimating wind load is knowing the correct mean value of the drag coefficient for the signs and members of their support structures. This task, in turn, requires accounting for their interactions (e.g., the generation of shielding regions and flow- acceleration regions by the signs and some of the members of the support structure). The current LRFDLTS-1 specifications do not include provisions for estimating wind loads on highway signs and their support structures in these situations. Several state departments of transportation (DOTs) report that some of the trusses supporting large DMS cabinets required frequent inspections, retrofitting, and even premature replacement of the truss because cracks developed, especially near welding areas. To avoid such problems, some states recommend that no other signs should be placed on a truss supporting a DMS cabinet. However, such measures are not necessary if designers can accurately estimate the wind loads on the support structure and on the signs supported by it and then use this information to design an appropriate structure. The drag coefficient value for DMS cabinets included in the LRFDLTS-1 specifications is extremely conservative and is not based on wind-tunnel studies or numeri- cal simulations. Recent wind-tunnel studies (Chowdhury et al. 2015) and numerical studies (Constantinescu, Bhatti, and Phares 2018) predict drag coefficients close to 1.3 for high aspect ratio DMS cabinets while the current AASHTO specifications recommend using a Cd of 1.7. For the simplest case of a static rectangular sign, the drag coefficients recommended in the current LRFDLTS-1 specifications appear to underestimate the wind loads. This conclusion is based on both the results of experimental studies (e.g., Letchford 2001) and the numerical studies discussed in an Iowa DOT report (Constantinescu, Bhatti, and Phares 2018). Meanwhile, the ASCE/SEI 7-16 (2017) recommendations for the design of aluminum sign structures seem to
8 Wind Drag Coefficients for Highway Signs and Support Structures be too conservative, specifying drag coefficient values as much as 50% higher than those in the LRFDLTS-1 specifications. Given the wide range of recommended values for drag coefficients in different standards and research studiesâand given the high probability that the LRFDLTS-1 recommendations are not sufficiently conservativeâthe recommendations for simple static rectangular signs installed above the ground need to be updated. Another knowledge gap is the unidentified reason for the relatively important differences among reported predictions of wind drag coefficients from different wind-tunnel experiments and how these predictions should be corrected for field conditions. Moreover, specific information on how the mean drag coefficient varies with the main geometri- cal parameters (sign aspect ratio, ground clearance, sign thickness, and shape) is lacking in many relevant cases. For example, LRFDLTS-1 specifications do not offer guidance for estimating wind loads for large signs with additional smaller signs attached to them or for signs of the same or differ- ent dimensions and types (static signs, DMS cabinets) that are separated by a gap. The state DOTs survey conducted as part of the current NCHRP project shows that most states use add-on signs and also place side-by-side signs on monotube and truss support structures. No procedure is cited for accurately estimating the drag coefficients in such situations given that the wind load on each sign is expected to differ from the value anticipated when each sign is at a large distance from neighboring signs. Furthermore, the presence of the sign support structure near the highway signs can alter the wind loads on the highway signs and diminish the utility of using the associated recommendations for estimating drag coefficients for isolated signs in such cases. No specific corrections are noted for the evaluation of the wind loads and associated drag coefficients when the sign is fixed in front of a relatively large bridge superstructure. The state DOTs survey indicates that large signs are installed on grade separation structures and that the forces and moments acting on the sign supports can be high. In these situations, sign designers need guidance to make rational decisions. In addition to the wind loads acting on the signs, the sign support structures are also subject to direct wind loads. For a truss support structure, this means that the drag coefficients for the chords and secondary members of the truss must be accurately determined. Estimating drag coefficients for each secondary member does not represent an easy task for trusses support- ing highway signs of different types and sizes. Because of the presence of the signs, some truss members are situated in regions where the âapproachingâ airflow velocity is significantly faster (e.g., flow-acceleration regions next to the edges of the sign) or slower (e.g., wake region behind the highway signs) compared to that of the approaching wind velocity away from the highway signs and their support structure. The chords of trusses and monotubes also experience different loads per unit length in regions close to the lateral edges of the signs and behind the signs when compared to regions situated away from the signs. Some chords and secondary members (e.g., back-face members of a 4-chord truss) are shielded primarily by other members rather than by the highway signs supported by the truss structure. No recommendations address the drag coefficient values or the wind load reduction for shielded members of truss support structures. This issue was confirmed by Ms. Bickford of Stanley Consul- tants through information obtained from engineers regularly involved in the design of highway sign support structures. The existing procedures in the literature are based on many simplifying assumptions and are not straightforward in their application to practical cases, as noted in an NCHRP project report (Fouad et al. 2002). These procedures fail to consider that the approaching air speed exceeds the incoming wind velocity for the unshielded members near the edges of the sign panel because the airflow accelerates as it passes the sign. Because the wind force is propor- tional to the square of the velocity, the underestimation of the wind load on the sign supporting structure can be significant for the portion of the member, chord, or monotube close to the edge of the sign. The literature review does not identify any method that takes into account these effects. Analysis of the designs used by state DOTs (made available via the survey) shows that many states employ trusses in which both the primary and secondary members have either a circular
Introduction and Research Approach 9 or an L-shaped cross-section. The current LRFDLTS-1 specifications provide drag coefficients only for two members (one in front of the other) for very small or very large distances between the two members. However, the current specifications do not include full information on how the drag coefficients for the two members vary as a function of the distance between the two members, cross-section of the members, aspect ratio, and flow regime (subcritical versus supercritical). 1.3 Research Objectives The following main research objectives of this study are based on the request for proposals (RFP) for NCHRP Project 15-67 and on the specific knowledge gaps identified in Section 1.2 related to accurately estimating wind loads acting on highway signs and their support structures: ⢠Objective 1. Determine accurate mean (time-averaged) drag coefficients needed to estimate wind loads for typical highway signs, including both static sign panels and dynamic message signs. Create graphs specifying the drag coefficient as a function of the sign aspect ratio and ground clearance ratio for isolated signs. Develop guidelines to estimate drag coefficients for signs with complex shapes (e.g., a large rectangular highway sign with a small sign attached at its top side) and for side-by-side signs as a function of the nondimensional gap distance between the signs. Also, establish guidelines to estimate wind-loading drag coefficients when the sign is attached to a bridge superstructure and when the highway sign is positioned next to its sign support structure. ⢠Objective 2. Develop new methods to estimate wind loads on structures supporting highway signs, using data from three-dimensional (3-D) eddy-resolving simulations of airflow past signs mounted on support structures. Ensure that methods will be suitable for consideration and adoption by AASHTO and will be accompanied by a comprehensive set of design examples. Make sure that methods apply to overhead bridge-type, cantilever-type, and butterfly-type truss structures and to overhead bridge-type and cantilever-type monotube structures. Develop recommendations on how to split the support structure into different regions (e.g., behind- the-sign, flow-acceleration, and uniform-flow regions) and indicate the values of drag coeffi- cients to use for unshielded and shielded members within these regions. ⢠Objective 3. Document all research related to the estimation of drag coefficients and wind loads on highway signs and their support structures. Propose revisions and additions to the LRFDLTS-1 specifications, working in consultation with members of both the project panel and the AASHTO T-12 Subcommittee on Structural Supports for Highway Signs, Luminaries, and Traffic Signals of the Bridges and Structures Technical Committee. Base these revisions on the new wind drag coefficients (determined as part of fulfilling Objective 1) and the new methods to estimate wind loads on structures supporting highway signs (developed as part of fulfilling Objective 2). Closely link the proposed revisions and additions to the current recommendations in the LRFDLTS-1 specifications. Clearly explain the methods so that they can readily be utilized by engineers working on the design and maintenance of transportation infrastructure components. 1.4 Research Approach Most of the current guidelines for wind design of highway structures are based on wind- tunnel studies, with some limited data from field investigations. In the past, essentially no alter- natives to wind-tunnel studies were available to estimate wind forces on structures. Advances in computational fluid dynamics (CFD) over the last two decades have provided numerical tools that allow performing accurate fully 3-D simulations of actual field conditions (e.g., high Reynolds numbers, very small blockage ratios) at a fraction of the cost of wind-tunnel studies
10 Wind Drag Coefficients for Highway Signs and Support Structures (Constantinescu, Bhatti, and Phares 2018). This research approach relies on performing 3-D numerical simulations of the turbulent airflow to estimate wind loads on highway signs and their support structures (e.g., members of trusses, monotubes). Wind-tunnel data from the literature and new data from wind-tunnel experiments performed as part of this research are used to validate the drag coefficient predictions for signs of both simple and complex shapes. Numerical simulations are crucial in generating the data needed to develop a general method for estimating wind loads on structures supporting single and multiple signs, including the chords and secondary members of truss support structures. The main data necessary to develop such a method are the wind loads on each member and on each relevant part of the member. Based on these data, corrections to the standard drag coefficient values can be proposed for each member as a function of its position with respect to the sign and the other members. Although such data can be easily obtained from 3-D numerical simulations by integrating the pressure and shear stress distributions on the surface of each member, experimental measurements will require a load cell for each member. More than that, the member must be detached from other parts of the structure to enable measurements of only the force acting on that member. These conditions are very complicated to achieve experimentally. The numerical engine used to perform the current simulations is a state-of-the-art commer- cial code STAR-CCM+ (CD-ADAPCO), which is widely used to calculate complex 3-D turbu- lent flows using the Reynolds-averaged Navier-Stokes (RANS) and Detached Eddy Simulation (DES) models for various engineering applications. The STAR-CCM+ package also contains a very powerful mesh generator, which simplifies varying the mesh density between different regions and resolving the boundary layers on solid boundaries. The governing equations are discretized and solved on unstructured Cartesian grids. The code allows the use of nested Cartesian grids with cell cutting near complex-shape boundaries. The viscous flow solver employs the Semi-Implicit Method for Pressure-Linked Equations (SIMPLE) algorithm to achieve pressure-velocity coupling for the discretized fully 3-D Navier-Stokes equations. In the SIMPLE algorithm, an intermediate velocity is obtained by solving the momentum equa- tions without the pressure gradient term. The intermediate velocity field does not satisfy the continuity equation. A pressure correction algorithm is applied to modify the pressure field so that the final velocity field satisfies the continuity equation. The convection terms in the momentum equations are discretized by using the third-order Monotonic Upstream-Centered Scheme for Conservation Laws (MUSCL) to keep numerical dissipation low. The diffusive and pressure gradient terms are discretized by using the second-order central scheme. The implicit temporal discretization is second-order accurate. The viscous solver in STAR-CCM+ is parallelized by using multiple parallel interface (MPI). The viscous flow solver displays excellent scalability when run on PC clusters with up to 128 processors for meshes containing 50 million cells. The boundary conditions in the current simulations containing highway signs, their support structures, or both are standard. No-slip (zero-velocity) boundary conditions are used at all solid boundaries (ground surface, surface of the highway sign, and members and elements of the support structure). A pressure outlet boundary condition is employed at the exit section. Simu- lations are conducted with a fully developed turbulent flow at the inlet section. For the simula- tions performed at field conditions, the vertical and lateral sizes of the computational domain are sufficiently large to make the flow blockage effects negligible, and symmetry boundary con- ditions are utilized at the lateral and top boundaries of the computational domain. When the main purpose of the simulations is validating the model, the size of the computational domain is identical to that of the corresponding wind tunnel. For such cases, no-slip boundary conditions are imposed on the lateral, top, and bottom boundaries to mimic the experimental conditions. This approach ensures that the simulation is conducted with the same flow blockage ratio as the corresponding wind-tunnel experiment. The predicted pressure and shear stress distributions
Introduction and Research Approach 11 on the highway signs and on the members and elements of the sign support structure are then integrated to calculate the wind load forces and estimate the mean drag coefficients. 1.5 Research Tasks Nine research tasks were completed to address the research objectives discussed in Section 1.3 and to develop recommendations for revision of the AASHTO LRFDLTS-1 specifications. The detailed activities associated with each of the nine tasks outlined in the RFP for NCHRP Research Project 15-67 are summarized in Table 1.1. At the suggestion of the project panel members, these research tasks are supplemented by wind-tunnel experiments, with the goal Task Activities 1. Conduct a literature review ⢠Gather literature from the United States (e.g., from NCHRP, FHWA) and abroad on wind drag coefficients and wind loads on highway signs and support structures ⢠Conduct a survey of commonly used highway signs and support structures; identify sources and availability of lab and field data of use for this study 2. Identify knowledge gaps and discuss the data required for accurate calculations of wind loads on highway signs and support structures ⢠Synthesize existing data on wind drag coefficients on various types of highway signs and identify the types of signs for which additional research needs to be conducted to estimate the drag coefficients ⢠Critically review existing methods for estimating wind loads on support structures and evaluate the adequacy of these methods; identify the limitations of these methods and discuss the main types of approximations made in these methods and their generality ⢠Identify possible types of support structures for which these methods cannot be applied 3. Propose a method and program of computational simulations to achieve project objectives ⢠Discuss why a numerically based approach, supported by experimental validation using existing validation data, is suitable to achieve the main goals of the project ⢠Develop a revised program of computational simulations to determine how the drag coefficient for highway signs varies as a function of sign geometrical parameters, shape of the sign (static and dynamic message signs, signs with complex shapes), and type of supporting structure ⢠Develop a revised program of computational simulations needed to estimate forces on the support structure (e.g., overhead bridge-type monotube structure, overhead cantilever-type monotube structure) or on the members of the overhead bridge-type or cantilever-type truss as a function of their position relative to the highway sign ⢠Use the data to propose a general and easy-to-use method to determine forces on the various types of support structures, including their members 4. Identify areas of LRFDLTS-1 that will be revised ⢠Identify types of signs for which drag coefficients will be revised (or included for the first time) in the AASHTO LRFDLTS-1 specifications; discuss the main geometrical and flow parameters affecting the drag coefficient value in the revised specifications ⢠Propose design examples to illustrate revisions, especially for cases where the drag coefficient is a function of the position of the sign relative to the support structure ⢠Identify types of support structures already mentioned in the specifications for which a new method to calculate wind loads will be proposed for inclusion in the specifications ⢠Propose design examples for each main type of overhead sign support structure to illustrate how the wind loads should be calculated as a function of the position of the sign relative to the supporting structure 5. Prepare Interim Report No. 1 ⢠Prepare draft Interim Report No. 1, including a refined work plan for Phases II and III and the rationale for the plan given findings in Tasks 1 through 4 ⢠Present Interim Report No. 1 to NCHRP panel; interact with panel on report ⢠Confirm program of numerical simulations and lab tests Table 1.1. Overview of research tasks and activities for NCHRP 15-67. (continued on next page)
12 Wind Drag Coefficients for Highway Signs and Support Structures of providing additional validation data for isolated rectangular signs, side-by-side rectangular signs, and signs that include an add-on sign. Per the task structure specified in the RFP for this NCHRP project, the research team devel- oped a three-phase research approach to complete the project: ⢠Planning: Conduct a literature review to identify knowledge gaps in the calculation of wind drag coefficients and wind loads for highway signs and sign support structures. Conduct a survey to identify the main types of highway signs and support structures. Propose a method to obtain the data needed to fill knowledge gaps (Interim Report 1). ⢠Plan Execution: Conduct research to obtain the data needed to estimate drag coefficients for highway signs and develop a method to estimate wind loads on supporting sign structures. Develop a draft of proposed LRFDLTS-1 revisions (Interim Report 2). ⢠Final Product: Prepare and submit the Final Report, including LRFDLTS-1 revisions, design examples, ballot items, and an implementation plan for the research findings (Final Report). Task Activities 6. Conduct needed research to obtain data necessary for estimating drag coefficients for highway signs and develop a method to estimate wind loads on supporting sign structures ⢠Conduct validation simulations and grid independency studies; check that computational domain size is large enough not to affect the prediction of wind loads due to flow blockage by sign and supporting structure ⢠Conduct numerical simulations to determine wind load on a highway sign and corresponding drag coefficient variation as a function of the sign geometrical parameters, shape of sign (e.g., main sign with an add-on smaller sign, side-by-side signs with a gap in between), and type of supporting structure (e.g., truss or monotube); for overhead supporting structures, note that the position of the sign relative to the structure may be a geometrical factor affecting the drag coefficient ⢠Conduct numerical simulations to study how the drag coefficients of two parallel cylinders (e.g., front and back members of a truss) vary as a function of the distance between the cylinders ⢠Conduct numerical simulations to calculate wind loads and corresponding drag coefficients on monotube support structures and on each member of truss support structures ⢠Conceptually combine findings from numerical studies to develop a new predictive method for estimating wind loads acting on different types of sign support structures; for truss structures, specify the modified drag coefficient for each member (with respect to the value for an isolated exposed identical member) as a function of the relative position of the member with respect to the sign and the other truss members 7. Develop a draft of proposed LRFDLTS-1 revisions ⢠Use data generated as part of Task 6 to develop a draft of proposed LRFDLTS-1 revisions ⢠Develop design examples, especially for the calculation of wind loads on sign support structures of different types ⢠Develop ballot items for changes and additions to LRFDLTS-1 8. Prepare Interim Report No. 2 ⢠Summarize the findings of Tasks 6 and 7 ⢠Describe the main proposed LRFDLTS-1 revisions, including design examples and ballot items ⢠Describe the proposed organization and content of the final report as well as the content of the implementation plan based on the findings and proposed methods described in Tasks 6 and 7 ⢠Present Interim Report No. 2 to NCHRP panel; interact with panel on Interim Report No. 2 9. Prepare and submit final deliverables ⢠Assemble all pertinent information from the project and write a draft Final Report ⢠Formulate the proposed LRFDLTS-1 revisions, including design examples and ballot items for proposed revisions ⢠Write the âimplementation of research findings and productsâ description of how results are implemented and how they will benefit state DOTs and other agencies; interact with panel on the draft Final Report and proposed LRFDLTS-1 revisions ⢠Submit final deliverable versions of the aforementioned documents Table 1.1. (Continued).
