Supplemental Guidance on the Application of FHWA’s Traffic Noise Model (TNM) (2014)

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.

3 C H A P T E R 2 2.1 Introduction When modeling noise levels at receptors located adjacent to an elevated roadway on a structure (bridge or viaduct), FHWA TNM is capable of predicting the noise generated by vehi­ cles traveling on the highway structure, taking into account direct noise paths and diffracted noise influences of any noise­ blocking features (parapets, noise barriers, etc.). However, FHWA TNM Version 2.5 does not enable the direct model­ ing of noise reflected off of barriers or retaining walls on the opposite side of the roadway nor noise reflected off of the underside of the structure itself. While Version 3.0 of FHWA TNM will be capable of modeling reflected noise, the treat­ ment of such reflections will be limited to vertical or nearly vertical surfaces such as far­side barriers and retaining walls. Reflections will not be applicable to horizontal surfaces such as the underside of bridges and viaducts. In addition to the noise that can be reflected off of the under­ side of structures, vibrations of a structure can be created by vehicles traveling on a bridge, and these vibrations result in noise being radiated from/by the bridge superstructure. Vehi­ cles traveling over bridge expansion joints also create noise that can travel to adjacent receivers located above and below the elevation of the structure roadway as well as to receivers directly underneath the structure. In the majority of instances, the structure­related noise conditions described above occur simultaneously, and their individual noise level contributions cannot be segregated. Therefore, the NCHRP Project 25­34 research team has eval­ uated several candidate modeling techniques and has consid­ ered the conditions identified above both individually and in combination. Based on these evaluations, several best model­ ing practices for the development of adjustments to the basic FHWA TNM predictions have been developed to account for these structure­reflected and structure­radiated noise condi­ tions that cannot be modeled directly. It is envisioned that such practices will be applied similarly to the basic (modeled) noise levels generated by either FHWA TNM Version 2.5 or Version 3.0. This chapter provides a summary of the evaluations per­ formed and the resultant suggested best modeling practices for determining adjustments to FHWA TNM values to account for structure­related noise contributions. A more detailed dis­ cussion, including the development of the suggested practices, is included in Appendix A (available on the NCHRP Proj­ ect 25­34 web page at http://apps.trb.org/cmsfeed/TRBNet ProjectDisplay.asp?ProjectID=2986). 2.2 Modeling Techniques Evaluated Candidate modeling techniques evaluated for modeling structure­reflected noise primarily focused on addressing noise reflected from the underside of bridge structures to adjacent receivers located below (lower in elevation) elevated portions of a highway. Sound paths to adjacent receivers located level with or above a highway are generally influenced by reflections off of the pavement, which are accounted for in FHWA TNM. The evaluations of noise reflections off of roadway features such as safety barriers, median barriers, and retaining walls are discussed in Chapter 5. For above­ road receivers, structure­radiated noise is not believed to be a major issue since it is masked by direct path vehicle noise pre­ dicted by FHWA TNM and tire­pavement interaction varia­ tions (not a part of this research project). However, in certain situations, receivers located above the roadway are influenced by noise from expansion joints. The techniques developed to address expansion­joint noise for receivers located below the elevation of the roadway were also evaluated for application to receivers elevated above the roadway. 2.2.1 Best Modeling Practices #1A and #1B Best Modeling Practice #1A, an image­source technique that constructs an image of any receptor influenced by noise Structure-Reflected Noise and Expansion-Joint Noise

4reflecting off of the undersides of structures, was evaluated and tested for several projects. A complete description of this technique can be found in a 2002 paper by Reiter and Bowlby.1 Additional discussion of this technique is contained in Appendix A. This technique starts with FHWA TNM skew section views to help identify which sections of roadways and which vehicle types are involved in the reflections that may reach individual receptors. Figure 1 from the 2002 Reiter and Bowlby paper is an example. For any receptor affected by noise reflections, its associated “reflection-contributing” sources are then modeled at that receptor’s image location. For each affected receptor, its noise level from reflected sources is then added to the noise level generated in the base FHWA TNM run to obtain its total noise level. This technique was utilized as a screening process by NCHRP Project 25-34 research team members during the noise evaluation of a Tennessee Department of Transpor- tation (TDOT) project that involved the proposed wid- ening of Interstate 40 (I-40) and a proposed four-level interchange reconstruction in Nashville. The reflecting structure was a ramp with relatively low traffic volumes, so structure-radiated and joint noise were not concerns. Since the time that the screening analysis was completed, the project has been constructed. The team tested and evalu- ated this technique from the 2002 paper using Best Modeling Practice #1B, which involved limited simultaneous measure- ments at three sites that were identified as being affected by reflected noise. For comparison, simultaneous measurements were also taken at three additional sites where traffic char- acteristics were similar, but where the elevated ramp struc- ture is not present and where reflected noise is not an issue. The reflected noise component estimated by Best Modeling 1 Reiter, D. D. and W. Bowlby, “Assessing Noise Reflections off the Underside of Elevated Bridge Structures: Procedures Using the FHWA Traffic Noise Model,” Transportation Research Record: Journal of the Transportation Research Board, No. 1792, Transportation Research Board of the National Academies, Washington, D.C., 2002, pp. 50–56. 2 Reiter, D. D. and W. Bowlby, “Assessing Noise Reflections off the Underside of Elevated Bridge Structures: Procedures Using the FHWA Traffic Noise Model,” Transportation Research Record: Journal of the Transportation Research Board, No. 1792, Transportation Research Board of the National Academies, Washington, D.C., 2002, Figure 4, p. 53. Figure 1. TNM skew section showing roadways contributing to reflected sound.2

5 Practice #1A was compared to the estimated reflective noise component derived from Best Modeling Practice #1B. Details related to this comparative analysis are contained in Appen­ dix A. While Best Modeling Practice #1B indicated a wider variation of the reflective noise values (3 to 8 dB range) than the 4 to 5 dB range estimated via Best Modeling Practice #1A, the average value was approximately 5 dB using both prac­ tices. This indicated that either practice (image­source mod­ eling or comparative measurements) appears to represent a viable Best Modeling Practice for estimating the contribution of structure­reflected noise. Best Modeling Practice #1A was also used by the research team in its noise analysis of the widening and reconstruc­ tion of the Interstate 95 (I­95) Section Girard Avenue Inter­ change (GIR) project in Philadelphia, Pennsylvania, and by the Washington State Department of Transportation (WSDOT) to evaluate the effects of noise reflections on com­ munities adjacent to the two­level bridge carrying Interstate 5 over the Ship Canal in Seattle. The processes, applications, and limitations associated with Best Modeling Practices #1A and #1B are described below. 2.2.1.1 Best Modeling Practice #1A: FHWA TNM Modeling of Reflected Noise by Developing Image Receptors The process of Best Modeling Practice #1A is the following: 1. Model direct highway noise contributions from all road­ ways using FHWA TNM. 2. Use the technique described in Reiter and Bowlby 2002 (see footnote 1) to estimate adjustments due to reflections off of the underside of structures. 3. Apply adjustments to obtain structure­noise­adjusted predicted noise level. The applications and limitations of Best Modeling Practice #1A are the following: 1. Since Best Modeling Practice #1A is solely based on noise modeling, it can be applied to any type of highway project, i.e., construction on new location or reconstruction of an existing highway. 2. Use requires detailed geometric and traffic information. 3. Use does not account for the variation of reflected noise associated with different types of superstructures, i.e., spread box beams, adjacent box beams, segmental bridges, steel I­beams, steel deck pans, etc. 4. Best Modeling Practice #1A deals only with structure­ reflected noise and does not account for any other structure­ related noise. 2.2.1.2 Best Modeling Practice #1B: Comparing Noise Measurements at a Site Containing Reflections with a Site without Reflections The process of Best Modeling Practice #1B is the following: 1. Model direct highway noise contributions from all road­ ways using FHWA TNM. Model for each traffic condi­ tion at all receivers associated with each measurement period. 2. Conduct multiple sets (minimum of three) of noise mea­ surements at selected setback locations where reflected noise is believed to be a contributing factor. 3. Conduct multiple sets (minimum of three) of simultane­ ous measurements at locations with similar setbacks that have similar traffic and topographic features, but where reflections from the underside of a structure are not a con­ tributing factor. 4. For each measurement setback distance, calculate the difference between the values for Items 2 and 3, above. This is the reflected noise adjustment factor. 5. For each measurement setback distance, apply the reflected noise adjustment factor (Item 4) to the FHWA TNM noise level from Item 1 to obtain the structure­noise­adjusted predicted noise level. The applications and limitations of Best Modeling Practice #1B are the following: 1. Use requires detailed geometric and traffic information. 2. Use inherently accounts for the type of superstructure. 3. Use requires exclusion of extraneous noise sources. 4. Use requires sufficient equipment and personnel to per­ form simultaneous measurements and to collect simul­ taneous traffic data, which are required to normalize the measured levels to one set of traffic conditions. 5. Use requires finding a location without reflections that has similar traffic and topography for comparison with the reflective location. 2.2.2 Best Modeling Practice #2: Using Noise Measurement Data to Develop Combined Structure-Related Predicted Noise Levels Development of a best modeling practice that relies on noise measurements to establish adjustment factors associated with structure­related noise to apply to basic FHWA TNM values involved a multistep approach. This multistep approach was a refinement of an approach used by NCHRP Project 25­34 research team members during a 2011 noise analysis of an adjacent section of I­95.

6For the multistep approach, the research team initially conducted noise measurements directly underneath a span of the I-95 viaduct at Schiller Street in Philadelphia where other highway noise sources do not exist. Three sets of simul- taneous measurements were taken underneath the viaduct at three positions: • Site 1: Within 5 ft of the bottom of the deck near an expan- sion joint. • Site 2: Within 5 ft of the bottom of the deck at a point mid- way between expansion joints. • Site 3: At 5 ft above the ground at a location midway between Positions 1 and 2. These measurements were performed using American National Standards Institute (ANSI) Type I noise meters and compatible microphone cables. Commonly available and relatively inexpensive equipment (connected pieces of half- inch electrical conduit costing less than $25.00 supported by speaker stands) was used to position the microphones at locations close to the underneath of the viaduct superstruc- ture. Results of the measurements are included in Table 1. The measurements show very little difference in noise lev- els at positions underneath the structure, illustrating that, at this location, ground level noise levels resulted from a com- bination of joint and deck noise, with neither of these noise sources predominating. In addition, there was little difference between noise levels measured just below the deck and those measured 5 ft above the ground. Based on these observations, it was assumed that a measurement taken at a point below the outside of the viaduct (drip edge location) would represent the combined noise level due to deck and joint noise at that location. To estimate the combined contribution of deck and joint noise at points at various distances (setback locations) from the structure, drop-off equations associated with vari- ous drop-off rates were developed. For the purpose of establishing an initial reference distance for calculating structure-related noise at setback locations, it was assumed that the noise emanates from the underside of the deck at the centerline of the structure, midway between the drip edges. In establishing structure-related noise levels at setback locations, the location of drip edge noise was assumed to be midway between the bottom of the bridge deck and the ground. The input parameters illustrated in Figure 2 were used to determine the reference distance (Dref) and the dis- tance from the assumed midpoint source of structure-related noise (S) to the drip edge location (Aref) using the following procedure: • Measure the height of the structure from the underside of the deck to the ground (h). Divide the distance by two (h/2) to calculate the midpoint between the ground and the underside of the deck. This is designated the drip edge midpoint (Aref). • Measure the width of the structure from drip edge to drip edge (w). Divide the distance by two (w/2) to calculate the midpoint or centerline of the structure (Mw). The under- side of the deck at Mw is the assumed location of the source of the structure-related noise (S). • To calculate Dref, the distance from the source of the deck noise (S) to the drip edge midpoint Aref, the formula ( ) ( )= +2 22 2D w href was employed. Figure 3 illustrates the relationship between Dref and the location of the analysis points at various setback distances from the drip edge. The height, width, and measured noise level at the drip edge of the structure are entered into the Structure-Related-Noise Calculation Worksheet (see Table 2). The spreadsheet calculates Aref, Mw (or S), and Dref. Setback distances from the drip edge of the structure are included in Table 2 for standard distances of 25, 50, 100, 200, and 400 ft. A blank row (Axxx) is provided for inserting an additional setback distance if desired. The spreadsheet also calculates the structure-related noise at the analysis points based on the three drop-off rates of 3.0, 4.5, and 6.0 dB per double distance (dB/DD) using the following formula: )(= − 10 10L L Log D DA DE AP refx Date Beginning Time of Measurement Measured Noise Level Leq in dB(A) Position 1: Near Joint, within 5 ft of Bottom of Deck Position 2: Away From Joint, within 5 ft of Bottom of Deck Position 3: 5 ft Above Ground between Positions 1 and 2 Leq Leq Leq 4/15/2013 3:47 pm 63.6 63.2 63.1 4:08 pm 64.4 64.1 64.2 4:24 pm No Data 64.2 64.2 Table 1. Structure-radiated and expansion-joint noise under I-95.

7 A detailed discussion of the development of this worksheet- based methodology can be found in Appendix A. In testing the appropriateness of this methodology at setback locations related to the following projects, the research team found the practice to yield reasonable results: • I-95 projects in Philadelphia, Pennsylvania: Section AFC (nine sets of measurements at Schiller Street) and five Sec- tion GIR measurement sets taken at Eyre Street, Sergeant Street, Susquehanna Avenue (two sets of measurements), and Cambria Street. • Pennsylvania Turnpike Susquehanna River Bridge project in Highspire, Pennsylvania: (four sets of measurements). • Indiana Department of Transportation (DOT) project: (two sets of measurements). • Arkansas I-40: (two sets of measurements). where LDE = Equivalent sound level (Leq) noise measurement in dB(A) taken at 5 ft above ground under structure drip edge. LAx = Calculated structure-related-noise level at an analy- sis point Ax, located x feet from the drip edge. DAP = Distance from point S to the analysis point Ax. DRef = Distance from point S to Point ARef. The value of “10” in the formula represents a drop-off rate of 3 dB per doubling of distance (dB/DD). For the 4.5 dB/ DD calculation, this value is 15 in the formula. For a 6 dB/DD drop-off rate, the value in the formula is 20. 3 Source: Environmental Acoustics Figure 2. Input parameters.3

8Detailed information associated with the testing process is included in Appendix A. The processes, applications, and limitations associated with Best Modeling Practice #2 are given below. 2.2.2.1 Best Modeling Practice #2: Process, Applications, and Limitations The process of Best Modeling Practice #2 is the following: 1. Model direct highway noise contributions from all road- ways using FHWA TNM. Model under a variety of free- flow traffic conditions at all receivers associated with each measurement period. 2. Conduct multiple (minimum of three) sets of noise mea- surements at the drip edge ground level location and at a minimum of two setback distances for the purpose of vali- dating the FHWA TNM runs and determining the extent of structure-related noise contributions. If third-octave band measurements were conducted, review frequency graphs for setback locations to help verify the limits of structure-related-noise contributions. See Appendix A for frequency graphs associated with one-third-octave band measurements for the tested projects. 3. Apply the adjustments from the Table 2 worksheet to lev- els at setback locations to determine total modeled noise levels at each setback location. 4. If expansion-joint noise is the predominant source of structure-related noise, assume that the noise emanates from the joint above the measurement point rather than at the midpoint of the structure and adjust the worksheet Dref value to be the distance from the drip edge micro- phone to the bottom of the structure’s deck. 5. Apply the values from the Table 2 worksheet to FHWA TNM predicted levels for the proposed project using the drop-off rates that best correlate with the measured levels. The applications and limitations of Best Modeling Practice #2 are the following: 1. Use requires detailed geometric and traffic information. 2. Use inherently accounts for the type of superstructure. 3. Use requires exclusion of extraneous noise sources. 4. Use requires sufficient equipment and personnel to per- form simultaneous measurements and to collect simulta- neous traffic data. 5. It does not account for any reflected noise from other sources of highway noise that affect setback locations unless such reflected noise reaches the ground level drip edge location. 6. Since this best modeling practice was developed based on actual existing conditions and tested against these con- ditions, it is likely to be most applicable to projects that involve reconstruction and/or widening of existing high- ways as opposed to highways on new locations. In any case, measurements should be taken at structures that resemble the structure type and configuration planned for the pro- posed highway improvement project. 7. While measurements conducted during the development of the Table 2 worksheet did not indicate substantial varia- tion of expansion-joint and/or deck noise levels due to the variety of observed traffic conditions, users may want to 4 Source: Environmental Acoustics Figure 3. The relationship between Dref and the location of the analysis points.4

9 Input Data: 27kced fo edisrednu ot dnuorg morf ,erutcurts fo thgieH :h A ref : Center point between ground and underside of structure (h/2). 132 13.5 Mw: Midpoint of structure (w/2). The underside of the deck at this point is the assumed source of structural noise (S). 67 66.0 Set-back Calculations: Aref 06766.0 A25 6.462952 A50 6.3671105 A100 0.26761001 A200 0.06762002 A400 6.75764004 AXXX 0.6676 Aref 06766.0 A25 9.362952 A50 4.2671105 A100 1.06761001 A200 0.75762002 A400 4.35764004 AXXX 0.6676 Aref 06766.0 A25 3.362952 A50 2.1671105 A100 1.85761001 A200 0.45762002 A400 2.94764004 AXXX 0.6676 PennDOT I-95 at Schiller Street 4/16/2013 11:11am Northbound Side at 25 feet and 50 feet Analysis Point Distance from Drip Edge (ft.) Distance from S to Analysis Point (ft.) Measured Noise Level at Drip Edge Leq in dB(A) Calculated Noise Level, Drop-off Rate = 3.0 dB/DD w: Width of structure Dref: Reference distance - from S to Aref Measured Noise Level at Drip Edge, dB(A) 66 Analysis Point Distance from Drip Edge (ft.) Distance from S to Analysis Point (ft.) Measured Noise Level at Drip Edge Leq in dB(A) Calculated Noise Level, Drop-off Rate = 6.0 dB/DD Measured Noise Level at Drip Edge Leq in dB(A) Calculated Noise Level, Drop-off Rate = 4.5 dB/DD Analysis Point Distance from Drip Edge (ft.) Distance from S to Analysis Point (ft.) Table 2. Structure-related-noise calculation worksheet. test this methodology under different traffic conditions as well as test the characteristics of their project’s specific structure type, employing techniques used by the research team in developing the worksheet drop­off methodology (see Appendix A). 2.2.3 Best Modeling Practice #3: Using a Combination of Best Modeling Practices As illustrated in the testing of the various candidate modeling techniques, several of the projects evaluated were affected by contributions of structure­reflected noise addi­ tional to deck­radiated and/or expansion­joint noise. This required the incorporation of Best Modeling Practices #1A and #2. If appropriate, Best Modeling Practice #1B could also be employed in such a situation. In addition, there may be situations where two different practices may be considered for application. For example, WSDOT used Best Modeling Practice #1A to adjust FHWA TNM modeled noise levels to account for structure­reflected noise on the two­level Ship Canal Bridge in Seattle, Wash­ ington. At the same time, the research team applied Best Modeling Practice #2 to several selected setback receptors to

10 determine this practice’s potential to account for structure­ related noise for such a project. This comparison indicated that these two best modeling practices produced similar val­ ues for the selected receptors. The selection of the appropri­ ate methodology for a project such as this would most likely depend upon whether the structure­related noise is associated with reflections, deck­radiated noise, expansion­joint noise, or some combination of these sources. Where structure­reflected noise is the predominant source, Best Modeling Practice #1A is probably most appropriate, whereas Best Modeling Prac­ tice #2 could be considered where all sources are present, but where sources of deck and/or joint noise predominate.