Hydraulic Loss Coefficients for Culverts (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.

5 2.1 Summary Concerns about roadway crossings for fish, debris, and terrestrial animals have promoted the development of alter- native designs for culverts that are larger than traditional culverts, use buried-invert (embedded) circular or ellipti- cal barrel shapes or bottomless arches, and often span the existing bank-full channel or feature a simulated streambed. Matching the culvert streambed (material composition and slope) with the adjacent upstream and downstream channel reaches allows the culvert streambed to aggregate and erode at natural rates, greatly reducing the potential for artificial fish passage barriers to form, such as perched outlets, as often occurs with traditional culverts. In current practice, entrance loss coefficients and inlet control head-discharge relationships for buried-invert cul- verts designed for fish passage applications are either ignored or approximated using traditional culvert design data due to a lack of data specific to these alternative culvert geometries. In this study, experimental methods were used to determine entrance loss coefficients and inlet control head-discharge relationships for circular culverts with 20%, 40%, and 50% invert burial depths and an elliptical culvert with a 50% invert burial depth. In general, the entrance loss coefficients for buried- invert culverts were higher than entrance loss coefficients for traditional culverts of the same cross-sectional shape without invert burial. The influence of approach flow con- ditions (ponded or channelized) on entrance loss coef- ficients and inlet control head-discharge relationships is also reported. This chapter outlines the experimental methods used to determine entrance loss coefficients and inlet control head-discharge regression constants relative to these alternative culvert geometries and presents the data relevant to the hydraulic design and evaluation of these culverts. 2.2 Introduction Due to increased concern about the environmental impact of traditional culvert designs, more environmentally sensi- tive culvert designs are now being implemented in the field. Traditionally, the smallest culvert capable of passing a design flood was installed in order to minimize costs. Issues associ- ated primarily with debris and fish passage through culverts, however, have promoted the implementation of larger culverts and alternative culvert barrel geometries. These larger culverts, which in some cases span the entire streambed width, are typically installed so that the pipe invert is located below the natural streambed grade. Sub- strate is placed inside the culvert up to the level of the natural streambed grade and arranged in such a way as to simulate a streambed throughout the culvert. Additionally, the culverts are generally sized and the substrate placed so that little or no discontinuity exists between the simulated streambed in the culvert and the adjacent upstream and downstream reaches. These culverts are commonly referred to as buried-invert or embedded culverts. Similar fish passage culvert environments are also created using bottomless culverts, such as pipe arches. The results of this study are applicable to both buried-invert and bottomless culverts. Examples of buried-invert culverts with a simulated streambed are shown in Figure 2-1. One advantage of buried-invert fish passage culverts over traditional culvert designs is the elimination of high flow velocities in the culvert at shallow flow depths due to the increase in culvert flow area. Another advantage of buried- invert culverts is that the simulated streambed can aggre- gate and erode streambed materials similarly to the natural streambed, thus maintaining a more natural system. Many aspects of buried-invert fish passage culvert designs are primarily influenced by fish passage considerations, such as the composition of the simulated streambed. Structural capacity, hydraulic requirements, and public safety are also C h a p t e r 2 Buried-Invert or Embedded Culverts

6important to consider in decisions regarding culvert size, pipe material, and culvert end treatments for buried-invert culvert designs. A review of publications related to the hydraulics of buried-invert culverts for fish passage produced a significant amount of information; very little information was found regarding buried-invert culvert hydraulics (i.e., entrance loss coefficients and inlet control head-discharge relationships). Two documents summarized a variety of design proce- dures for buried-invert culverts (Bates et al., 2003 and Jordan and Carlson, 1987). Because the design of buried-invert cul- verts for fish passage is a relatively new process, Bates et al. (2003) recommended that current design methods be imple- mented conservatively until the hydraulics of buried-invert culverts are more completely understood. Bates et al. (2003) also suggested that it is imperative that the discharge capacity of buried-invert culverts be evaluated. No procedure or data for doing so, however, were discussed in the document. Jor- dan and Carlson (1987) produced a discharge coefficient for a buried-invert culvert with a square-edged vertical headwall end treatment. Their design procedure was not consistent with current FHWA’s culvert design method, published in a report entitled Hydraulic Design of Highway Culverts (Nor- mann et al., 2001), referred to here and in practice as HDS-5. Jordan and Carlson (1987) suggested that further research is necessary to determine the discharge coefficients of other frequently used buried-invert culvert inlet geometries. In current practice, there is some uncertainty in deter- mining the head-discharge relationships for buried-invert culverts due to a lack of hydraulic data specific to buried- invert culvert geometries (e.g., entrance loss coefficients, veri- fied friction loss predictive methods for composite hydraulic roughness culvert flow, and inlet flow control head-discharge relationships). At present, head-discharge relationships for buried-invert culverts are either not determined or are approximated using traditional culvert design data such as the data presented in HDS-5. The objective of this part of NCHRP Project 15-24 was specifically to investigate the hydraulics of buried-invert culverts, including the experi- mental determination of buried-invert entrance loss coef- ficients and inlet control head-discharge relationships for a variety of traditional culvert end treatments. Due to the wide range of design possibilities for buried-invert culverts, a vari- ety of buried-invert culvert end treatments were evaluated to determine their influence on buried-culvert inlet control head-discharge relationships or outlet control energy loss characteristics. 2.3 Research Objectives The objectives of this research included determining the entrance loss coefficient, ke, and the inlet control head- discharge relationships for circular culverts with invert burial depths of 20%, 40%, and 50% and an elliptical culvert with a 50% invert burial depth. All buried-invert culverts were tested with four different end treatments: (1) thin-wall pro- jecting, (2) mitered flush to 1.5:1.0 (horizontal to vertical) fill slope, (3) square-edged inlet with vertical headwall, and (4) 45° beveled entrance with vertical headwall, the bevel extending 1 in. vertically for every 24 inches of horizontal culvert span. The four end treatments tested are illustrated in Figures 2-2 and 2-3. Each end treatment was tested with two different approach flow conditions—ponded and channelized. The ponded approach represented a reservoir condition with negligible velocities everywhere except near the culvert inlet. Significantly higher approach velocities developed when two parallel guide Figure 2-1. Examples of buried-invert culverts for fish passage.

7 walls were installed, one on each side of the culvert inlet, cre- ating the channelized approach with a ratio of channel width to culvert horizontal span of 2. The wing walls were approxi- mately four times the culvert span in length. Figure 2-4 illus- trates ponded and channelized approach conditions. To make an accurate comparison between channelized and ponded approach flow conditions, it was determined that the total upstream head (Hw) would be a more appropriate term for quantifying inlet control head-discharge relationships than the piezometric head (Hwi). Consequently, for both approach flow conditions, the upstream head was measured in a loca- tion with negligible velocity head (i.e., Hwi = Hw). By doing so, the upstream head term in Equations 1-2 through 1-4, Hwi, was replaced by the total head term, Hw. For the channelized approach flow conditions, some energy loss occurs between the reservoir and the culvert inlet due to flow contraction at the channel inlet and friction loss. Hw at the culvert inlet was approximated by subtracting the friction loss (Manning’s n = 0.009 assumed) and a contraction loss at the channel entrance (entrance loss coefficient = 0.36 assumed) from the total head measured at the reservoir pressure tap. As the data from this lab-scale culvert study will likely be applied to larger, field-scale buried-invert culverts, another research objective was to gain some understanding regarding the dependence of entrance loss coefficients on culvert diam- eter (size-scale effects). Entrance loss testing was conducted using 12-in. [inside diameter (I.D.) = 11.75 in.] and 24-in. diameter (I.D. = 23.45 in.) traditional circular culverts with square-edged inlet with headwall end treatments as shown in Figure 2-5. The entrance loss coefficients were determined for each culvert tested with submerged and unsubmerged ponded inlet conditions. 2.4 Experimental Method A detailed discussion of the testing procedures associ- ated with the determination of ke for outlet control and the empirical constants associated with the head-discharge relationships (Equations 1-2 through 1-4) for inlet control is presented in Chapter 1. Commercially available circular PVC pipe with the dimensions shown in Table 2-1 were used for the entrance loss coefficient size-scale testing. The buried- invert culverts were fabricated using smooth steel plate for the culvert wall and the flat invert that represented the simu- lated streambed. The circular culverts with 20%, 40%, and 50% invert burial depths had an inside diameter or hori- zontal dimension (Dh) of approximately 18 inches with the vertical rise (D) dimension varying with burial depth. The dimensions of the elliptical culvert were Dh = 25 inches and a vertical span of 17 inches; D = 8.5 inches with the 50% invert burial geometry. The aspect ratio of the elliptical culvert was based on dimensions of commercially available culverts. Figure 2-2. Buried-invert culvert inlet end treatments using a circular culvert with a 50% invert burial.

8Figure 2-3. Buried-invert culvert end treatments evaluated: (A) thin-wall projecting, (B) mitered flush to 1.5:1 (horizontal to vertical) fill slope, (C) square-edged inlet with vertical headwall, and (D) 45° beveled inlet with vertical headwall. (Ponded) (Channelized) Figure 2-4. Culvert approach flow test conditions. (A) (B) (C) (D)

9 12-in. circular culvert 24-in. circular culvert Figure 2-5. Overview of entrance loss size-scale testing. Dimension trevluc lacitpillE strevluc ralucriC 20% buried invert 40% buried invert 50% buried invert 50% buried invert D (in) 14.57 11.02 9.05 8.66 Dh (in) 18.11 18.11 18.11 25.20 20 20 20 20L (ft) Table 2-1. Test culvert geometries. Each buried-invert culvert had a wall thickness of 0.125 in., was approximately 20 ft long, and was supported continu- ously along the flat invert by two parallel, 4-in. square steel box beams that ran the length of the culvert. The culverts were also supported with vertical columns at regular intervals to minimize culvert deflection (maintain a constant slope) during testing. A schematic of the cross-sectional geometry for each buried-invert test culvert is provided in Figure 2-6. The buried-invert culvert cross-sections illustrated in Figure 2-6 were uniform over the entire length of each test culvert. Water was supplied to the test culverts via a head box, which measured 24 ft long by 22 ft wide by 5 ft deep. Water was supplied to the head box via 20-in. and/or 8-in. supply lines, both containing calibrated orifice flow meters. A sche- matic of the culvert test facility is shown in Figure 2-7, and a photo overview is provided in Figure 2-8. The location of the

10 Figure 2-6. Buried-invert test culvert cross-sectional geometries. Figure 2-7. Culvert test facility.

11 pressure tap that was used to measure Hw in the upstream reservoir is identified in Figure 2-7. The entrance loss coefficient, ke, and the inlet control regres- sion constants, K, M, c, and Y, were determined for both sub- merged and unsubmerged inlet conditions. The entrance loss coefficient data are presented as a function of (Hw/D), where Hw represents total upstream head and D is the buried-invert culvert vertical height (streambed to pipe crown). 2.5 Experimental Results Outlet Control The results of the size-scale effects testing for the circular culverts with square-edged inlets and a vertical head wall are shown in Figure 2-9, where the entrance loss coefficient, ke, is plotted as a function of Hw/D. At Hw/D values greater than 1.0, the 24-in. diameter circular culvert yielded an average entrance loss coefficient of 0.515, while the 12-in. diameter Figure 2-8. Culvert test setup. circular culvert produced an average entrance loss coefficient of 0.511. The average value of experimental uncertainty for the size-scale entrance loss testing was approximately 1.5%. The fact that the 12- and 24-in. diameter tests produced essentially the same entrance loss coefficient and that the experimental loss coefficient was consistent with the typical published ke value for square-edged inlets with headwall end treatment (ke = 0.5) suggests that there were no biases in the entrance loss coefficient experimental data associated with culvert size. Based on this result, it is also assumed that the buried-invert culvert hydraulic performance data from this study may be applied to larger field-scale culverts. The entrance loss coefficient data for the circular culvert with 20%, 40%, and 50% invert burial and the elliptical culvert with 50% invert burial are shown in Figures 2-10 through 2-13. The entrance loss coefficient, ke, is plotted as a function of Hw/D and classified by end treatment and approach flow condition (ponded or channelized). For each of the four end treatments tested [thin-wall projecting, mitered flush to 1.5:1.0 (horizontal to vertical) fill slope, square-edged inlet with vertical headwall, and 45° beveled inlet with vertical headwall], the entrance loss coefficient varied significantly with Hw/D for Hw/D values less than 1.0 to 1.5. Above that range, ke remained relatively constant. For all end treatments tested, channelized approach flow had no appreciable effect on the entrance loss coefficients, with the exception of the thin-wall projecting end treatment. For the thin-wall pro- jecting inlet, the channelized approach condition was more efficient (smaller ke value) than the ponded approach condi- tion. The increase in efficiency for the thin-wall projecting channelized approach flow condition was likely due in part to a decrease in the amount of flow contraction at the inlet relative to the ponded condition. Figure 2-9. Size-scale effect entrance loss coefficients for traditional circular culverts (square edged with vertical headwall and ponded approach flow).

12 Figure 2-10. Circular culvert, 20% buried invert, entrance loss coefficient data. Figure 2-11. Circular culvert, 40% buried invert, entrance loss coefficient data. Table 2-2 summarizes the average ke values for each end treat- ment for the buried-invert circular and elliptical test culverts. As a comparison, Table 2-3 presents the entrance loss coeffi- cient values reported in HDS-5 for traditional circular culverts with end treatments similar to those evaluated in this study. The entrance loss coefficients for all the buried-invert culvert geometries and various end treatments were larger (more head loss) than the entrance loss coefficients for traditional circu- lar culverts with the same or similar end treatments. The ke values for the circular culvert with 20%, 40%, and 50% invert burial depths and the elliptical culvert with 50% invert burial depth were relatively uniform for a given end treatment. Con- sequently, a single representative or average ke value is pre- sented in Table 2-4 for each end treatment tested. The average experimental uncertainty of 1.8% associated with the data and the even larger uncertainty associated with predicting culvert velocities in the field are such that the scatter in the ke data rela- tive to the average values associated with the various buried- invert culvert barrel geometries is considered negligible. Inlet Control Inlet control head-discharge relationships were created for unsubmerged and submerged inlet conditions (i.e., 0.3 <

13 Figure 2-12. Circular culvert, 50% buried invert, entrance loss coefficient data. Figure 2-13. Elliptical culvert, 50% buried invert, entrance loss coefficient data. Average ke (± extreme % deviation from average) Culvert end treatment 20% 40% 50% 50% lacitpille ralucric ralucric ralucric Thin-wall projecting (ponded) 1.01 0.97 0.96 1.11 (±3.5%) (±3.2%) (±6.6%) (±9.8%) Thin-wall projecting (channelized) 0.91 0.89 0.93 0.96 (±3.1%) (±4.4%) (±4.6%) (±9.1%) Mitered to 1.5H:1.0V w/ vertical headwall 0.83 0.89 0.88 0.91 (±4.9%) (±3.1%) (±4.1%) (±4.6%) Square-edged inlet w/ vertical headwall 0.55 0.54 0.54 0.60 (±9.7%) (±8.0%) (±5.6%) (±7.2%) 45° beveled inlet w/ vertical headwall 0.32 0.34 0.32 0.32 ( ±11%) (±5.7%) (±4.3%) (±14.1%) Table 2-2. Average buried-invert culvert entrance loss coefficients.

14 Circular culvert inlet end treatment ke 9.0 epols llif morf gnitcejorP 7.0 epols llif ot deretiM 5.0 llawdaeh htiw degde-erauqS 45° beveled edge with headwall 0.2 Table 2-3. Circular culvert entrance loss coefficients from HDS-5 (Normann et al., 2001). Culvert inlet end treatment ke 00.1 )ralucric( gnitcejorP 01.1 )lacitpille( gnitcejorP 09.0 )ralucric( epols llif ot deretiM Square-edged with headwall (circular) 0.55 Square-edged with headwall (elliptical) 0.60 45° beveled edge with headwall 0.35 Table 2-4. Recommended buried-invert culvert entrance loss coefficients. Figure 2-14. Circular culvert, 20% buried invert, inlet control Form 2 (Equation 1-3) data. Hw/D < 5.0) for each of the four end treatments for the circular culverts with 20%, 40%, and 50% invert burial depths and the elliptical culvert with 50% invert burial depth. Figures 2-14 through 2-17 present inlet control quasi-dimensionless rela- tionships in Equation 1-3 (Form 2) for the four culverts tested and the various end treatments, approach flow conditions, and inlet submergence conditions. Regression of the unsubmerged data in Figures 2-14 through 2-17 produced the empirical constants K, M, c, and Y associated with Equations 1-2 through 1-4. The inlet con- trol regression constants generated for buried-invert culverts are shown in Table 2-5 and are classified by culvert geom- etry, inlet end treatment, approach flow condition, and inlet submergence condition. For comparative purposes, Table 2-6 includes inlet control regression constants for traditional circular culverts with end treatments similar to those of the buried-invert culverts tested in the current study. These regression constants for Equations 1-2, 1-3, and 1-4 are published in HDS-5 (Normann et al., 2001). Experimental uncertainty for the generation of inlet control head-discharge relationships for this study was < 1%. Channelized approach flow had a significant effect on the head-discharge relationship for the thin-wall projecting end treatment. The channelized approach flow had a higher discharge for a given Hw than the ponded condition. Con- sequently, separate regression constants were derived for

15 each approach flow condition for the thin-wall projecting end treatment. Regression constants for the square-edged inlet with vertical headwall, 45° beveled inlet with vertical headwall, and mitered flush to 1.5 horizontal to 1.0 vertical fill slope end treatments were developed by combining both the ponded and channelized data since channelization had a minimal effect on the inlet efficiency. This suggests that the inlet geometry has a much greater effect on the inlet efficiency than the reduction of flow contraction due to channelization for the square-edged, 45° beveled, and mitered inlet end treat- ments. Tabular support data for the Chapter 2 experimental results are included in Appendices A and B. 2.6 Conclusions Prior to this study, entrance loss coefficients and inlet con- trol head-discharge relationships for buried-invert culverts were either ignored or approximated using traditional circu- lar culvert data because of the lack of data specific to buried- invert culvert geometries. The entrance loss coefficient data and Figure 2-15. Circular culvert, 40% buried invert, inlet control Form 2 (Equation 1-3) data. Figure 2-16. Circular culvert, 50% buried invert, inlet control Form 2 (Equation 1-3) data.

16 Figure 2-17. Elliptical culvert, 50% buried invert, inlet control Form 2 (Equation 1-3) data. Culvert type Unsubmerged Submerged Form 1 Form 2 K M K M c Y 20% Buried invert circular Projecting end, ponded 0.0860 0.58 0.4293 0.64 0.0303 0.58 Projecting end, channelized 0.0737 0.45 0.4175 0.62 0.0250 0.63 Square headwall 0.0566 0.44 0.4001 0.63 0.0198 0.69 45o Beveled end 0.0292 0.57 0.3869 0.63 0.0161 0.73 Mitered end, 1.5H:1.0V 0.0431 0.58 0.4002 0.63 0.0235 0.61 40% Buried invert circular Projecting end, ponded 0.0840 0.76 0.4706 0.69 0.0453 0.69 Projecting end, channelized 0.0927 0.59 0.4789 0.66 0.0441 0.52 Square headwall 0.0490 0.71 0.4354 0.68 0.0332 0.67 45o Beveled end 0.0358 0.62 0.4223 0.67 0.0245 0.75 Mitered end, 1.5H:1.0V 0.0317 0.77 0.4185 0.68 0.0363 0.65 50% Buried invert circular Projecting end, ponded 0.1057 0.69 0.4955 0.71 0.0606 0.54 Projecting end, channelized 0.1055 0.59 0.4955 0.69 0.0570 0.48 Square headwall 0.0595 0.59 0.0595 0.59 0.0402 0.65 45o Beveled end 0.0464 0.46 0.4364 0.69 0.0324 0.67 Mitered end, 1.5H:1.0V 0.0351 0.59 0.4419 0.68 0.0504 0.44 50% Buried invert elliptical Projecting end, ponded 0.1231 0.51 0.5261 0.65 0.0643 0.50 Projecting end, channelized 0.0928 0.54 0.4937 0.67 0.0649 0.12 Square headwall 0.0819 0.45 0.4867 0.66 0.0431 0.61 45o Beveled end 0.0551 0.52 0.4663 0.63 0.0318 0.68 Mitered to slope 0.0599 0.60 0.482 0.67 0.0541 0.50 Table 2-5. Buried-invert culvert inlet control regression constants.

17 the inlet control head-discharge design curves determined in this study will assist in evaluating the hydraulic capacity of buried-invert culverts. Based on the experimental results of the entrance loss coefficient data and inlet control head-discharge data for the 20%, 40%, and 50% buried-invert circular culverts and the 50% buried-invert elliptical culvert with four different end treatments, the following conclusions are made: 1. Buried-invert culvert entrance loss coefficients, ke, are higher than entrance loss coefficients for traditional cir- cular culverts with the same or similar end treatments. Buried-invert circular and elliptical culverts with the 45° beveled with vertical headwall end treatment produced ke values 65% higher, on average, than the traditional circular culverts with the same end treatment. The buried-invert circular culvert with the thin-walled projecting end treat- ment produced ke values, on average, that were 9% larger than the traditional circular culvert with the same end treatment. Traditional culvert (invert not buried) ke data for this study were based on values published in HDS-5. 2. For square-edged with headwall and thin-wall projecting inlets, the shape of the culvert (20%, 40%, and 50% bur- ied-invert circular or 50% buried-invert elliptical) had no significant effect on ke. The elliptical buried-invert culvert with mitered flush to 1.5:1.0 (horizontal to vertical) fill slope and the 45° beveled inlet with vertical headwall end treatments produced larger ke values than for the circular buried-invert culverts. 3. Under unsubmerged inlet conditions, ke for buried-invert culverts varies significantly with Hw/D. This may be due in part to variations in flow contraction at the culvert inlet with Hw/D. For a submerged inlet, ke is relatively inde- pendent of Hw/D (ke = constant) and is higher than the unsubmerged inlet values. The submerged ke values are listed in Table 2-4 and are recommended for use in design, as they constitute a more conservative value. 4. With the exception of the thin-wall projecting end treat- ments, a channelized approach flow with a channel to culvert width ratio of 2 had no significant effect on ke or on the inlet control head-discharge relationships for buried-invert circular and elliptical culverts when using total upstream head rather than piezometric head. The channelized approach was slightly more efficient than the ponded approach for the thin-wall projecting inlet for both inlet and outlet flow control. For all buried-invert culvert geometries and end treatments tested, the Form 2 equation (Equation 1-3) matched the experimental data more closely than the Form 1 equation (Equation 1-2) for unsubmerged inlet control flow conditions. Test Culvert & End Treatment degrembuS degrembusnU Form 1 Form 2 K M K M c Y HDS-5 Circular CMP 45.0 3550.0 – – 05.1 0430.0 gnitcejorP Mitered to slope 0.021 1.33 – – 0.0463 0.75 Square headwall 0.0078 2.0 – – 0.0379 0.69 HDS-5 ralucriC Beveled ring, 45° bevels 0.0018 2.50 – – 0.0300 0.74 Smooth tapered inlet throat – – 0.534 0.555 0.0196 0.90 Rough tapered inlet throat – – 0.519 0.640 0.0210 0.90 Table 2-6. HDS-5 traditional circular culvert inlet control regression constants (Normann et al., 2001).