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.
1 1.1 Project Introduction Traditionally, culvert head-discharge relationships are determined using empirical data published in design manu- als such as the Federal Highway Administrationâs Hydraulic Design of Highway Culverts (Normann et al., 2001), referred to in this report and in practice as HDS-5. Common culvert shapes used in practice include circular, elliptical, pipe arch, and box culverts. Historically, the determination of appropri- ate culvert size was based primarily on economic consider- ations (i.e., the appropriate size was the smallest culvert that would pass the required design discharge without exceeding the minimum allowable freeboard on the upstream side of the road crossing). More recently, however, the function of some culverts has been expanded to include aquatic and ter- restrial animal passage. Fish passage culverts typically consist of an oversized cir- cular or elliptical culvert (oversized relative to what would be required to meet the head-discharge requirements) that is partially buried in the streambed. These culverts are referred to as buried-invert or embedded culverts. The percent of invert burial can vary from zero to ~50% of the culvert height. Culvert head-discharge relationships are determined by balancing the culvert energy loss with the available differen- tial driving head across the culvert (outlet control) or by the shape of the culvert inlet and upstream driving head (inlet control). Typical culvert energy losses are associated with entrance loss, exit or outlet loss, and friction losses. Culvert design manuals, like HDS-5, provide entrance loss coefficient (ke) data (outlet control) and empirical coefficients and expo- nents for inlet control head-discharge relationships. In gen- eral, however, similar culvert hydraulic performance data are not available for buried-invert culvert geometries. The aim of this study (NCHRP Project 15-24) has been to develop a better understanding of embedded culvert hydraulics and to provide culvert design and analysis tools to engineers and to evaluate, among other things, the hydraulic characteristics of fish passage culverts. Chapter 2 reports the findings specific to the buried-invert culvert hydraulics. The hydraulics of slip-lined culverts is addressed in Chapter 3. As a result of an aging transportation infrastruc- ture, many culverts are near the end of their useful service life. In many cases, it is more economical to rehabilitate an existing or host culvert than to replace it. Inserting a smaller- diameter liner pipe inside the host culvert, a process referred to as slip lining, is a common culvert rehabilitation technique. Relative to the host culvert head-discharge characteristics, the head-discharge characteristics of the slip-lined culvert will be influenced by the reduction in cross-sectional area associated with the liner; the decrease in hydraulic rough- ness of the liner wall (slip liners are often solid-wall high density polyethylene) relative to the host culverts, which often have profiled pipe walls (e.g., corrugated metal pipe); and the geometry of the end treatment resulting from the slip-lining process. This study evaluated the head-discharge characteristics of projecting and tapered slip-lined culvert end treatments where the liner pipe projected 0.17 and 0.34 pipe diameters (D) upstream from the projecting end of the host culvert. The results of the slip-lined culvert study are presented in Chapter 3. For the case where a culvert or pipe discharges into a downstream reservoir where the reservoir velocity is essen- tially zero, the exit loss is equal to the velocity head in the pipe. When the culvert discharges into a downstream channel that has a similar alignment to the culvert, the exit loss has traditionally been estimated to be the difference in velocity head [DV2/(2g)] between the culvert and discharge channel. In the case of short, mild, or zero-slope culverts, however, sometimes the calculated exit loss represents a significant percentage of the total system energy loss. Another objective of this study was to evaluate the nature of culvert exit loss and compare experimental exit loss data with the predictive relationships for accuracy. The results of the exit loss study are presented in Chapter 4. C h a p t e r 1 Introduction
2In many wide-channel road-crossing applications, particu- larly when the channel slope is mild and/or the road embank- ment height is limited, multibarrel culverts are required. The majority of the culvert design data, such as data in HDS-5, is specific to single-barrel tests, and in general, is applied to multibarrel culvert design using the principle of superposi- tion. In other words, the discharge capacity of a single-barrel culvert is determined for a specific design head and the culvert design flow is divided by the single-barrel discharge capacity to determine the number of culverts needed. Cul- vert entrance loss (outlet control) and the inlet control head- discharge relationships are influenced by the amount of flow contraction that occurs at the culvert inlet. The supposition is that the hydraulic capacity of a single culvert barrel may change when additional barrels are installed adjacent to it due to the change in the nature of the flow contraction. Another objective of this study was to evaluate the behavior of multi- barrel culverts, relative to the single-barrel superposition assumption. The results of the multibarrel culvert study are presented in Chapter 5. A major factor affecting head-discharge relationships in culverts and open channels is the flow resistance or fric- tion loss associated with shear stresses that develop near the flow boundaries. Friction loss is typically accounted for by applying one-dimensional flow resistance relationships like the Darcy-Weisbach or Manning equations with empirically determined flow resistance coefficients (f and n, respectively). The flow resistance coefficients can vary with the relative roughness of the channel or pipe boundary, the Reynolds number (Re, discharge and viscous effects), and the cross- sectional shape of the channel flow area. When sediment is deposited in the bottom of a culvert, whether through sediment transport processes or intention- ally (embedded culvert design), the cross-sectional shape of the culvert can change relative to traditional culvert pipe shapes (e.g., circular or elliptical), which may affect the flow resistance characteristics of the culvert. It is likely to be more significant, however, that the hydraulic roughness of the wetted perimeter along the flow boundary of the flow cross- section is not constant and represents a condition referred to as âcompositeâ roughness. One-dimensional hydraulics may not be able to account for the potentially multidimensional nature of flow resistance in composite roughness channels or culverts. As part of this study, boundary roughness materials of differing hydraulic roughness were tested in a rectangular channel for both uniform roughness and composite rough- ness configurations. The experimental results were compared with 16 composite roughness predictive relationships for Manningâs n. The results of the composite roughness study are presented in Chapter 6. Historically, the culvert hydraulics experimental databases have consisted of data collected using laboratory-scale cul- verts with D ⤠6 to 8 in. To investigate potential size-scale effects in culvert hydraulics, entrance loss tests were con- ducted using 12- and 24-in. diameter circular culverts. The results of the size-scale testing are presented in Chapter 2. The data for NCHRP Project 15-24 were generated using experimental data collected in the hydraulics laboratory of the Utah Water Research Laboratory at Utah State University. A variety of different test facilities were utilized, including a 6-ft-deep by 8-ft-wide by 300-ft-long channel, a 3-ft-deep by 4-ft-wide by 48-ft-long tilting, rectangular flume, and an elevated head box and tail box between which culverts were installed for testing. The following is a list of published peer-reviewed journal and conference papers that related to this project: Allen (2012), Tullis and Anderson (2010), Tullis et al. (2008), Haderlie and Tullis (2008), Haderlie (2007), Anderson (2006), Anderson and Tullis (2006), Robinson (2005), Robinson and Tullis (2005), and Tullis et al. (2005). 1.2 Culvert Hydraulics Consistent with traditional culvert hydraulics, buried- invert culverts can operate under either inlet or outlet flow control. Under outlet control, culverts may flow full or par- tially full over a portion or the entire length of the culvert. For outlet control, the culvert discharge is determined by balancing the energy loss through the culvert with the energy available. Culvert energy losses include entrance loss, bar- rel friction loss, exit loss, and any other minor losses. The entrance loss, which is specific to the culvert inlet geometry, is typically expressed as a loss coefficient, ke, multiplied by the culvert velocity head (Equation 1-1). H k V g e e= 2 2 1 1( )- In Equation 1-1, He is the head loss associated with the culvert entrance flow condition (ft), V is the average flow veloc- ity in the culvert (ft/s), and g is the gravitational acceleration constant (ft/s2). Robinson (2005) showed experimentally that the entrance loss coefficients for a buried-invert elliptical culvert were inde- pendent of viscous or Reynolds number (Re) effects, where Re = V4Rh/n and Rh is the hydraulic radius (flow area divided by the wetted perimeter). Based on the fact that field-scale Re values are typically larger than lab-scale Re values and that the influence of viscosity tends to diminish with increasing Re, viscous or Re effects on culvert entrance loss coefficients are likely to be insignificant and were not considered in this study. Under inlet control, culvert discharge capacity is a function of the available upstream energy, the culvert inlet geometry, and a critical flow section that forms just downstream of the
3 inlet. The inlet control culvert flow capacity is typically quan- tified using empirical, quasi-dimensionless, head-discharge relationships such as those published in HDS-5. Different relationships are used for submerged (headwater above the crown of the culvert inlet) and unsubmerged (headwater below the crown of the culvert inlet) culvert inlet conditions. HDS-5 recommends Equations 1-2 and 1-3, referred to as Form 1 and 2, respectively, for unsubmerged inlet flows and Equation 1-4 for submerged inlet flow conditions. Unsubmerged Form 1 Hw D H D K K Q AD Si c u M o= +     â0 5 0 5 1 2. . ( )- Unsubmerged Form 2 Hw D K K Q AD i u M =    0 5 1 3. ( )- Submerged Hw D c K Q AD Y Si u o=     + â0 5 2 0 5 1 4 . . ( )- For Equations 1-2 through 1-4: Hwi is the headwater depth (piezometric head) measured relative to the culvert invert or streambed elevation at the inlet for buried-invert culverts (ft), D is the interior height (streambed to pipe crown) of the culvert barrel (ft), Hc is the total head at critical depth (ft), Q is the flow rate (ft3/s), A is the full cross-sectional area of the culvert barrel (ft2), So is the culvert barrel slope (ft/ft), Ku is 1.0 (Ku = 1.811 SI units), and K, M, c, and Y are all empiri- cal constants unique to a particular culvert installation. For a mitered inlet end treatment, So is multiplied by a constant of +0.7 instead of -0.5 in Equations 1-2 and 1-4. Outlet Control Testing To determine entrance loss coefficients, the culvert must be flowing under outlet control, which corresponds to subcriti- cal flow conditions in the culvert. Outlet control is achieved by installing the test culverts at a slope that is less than the critical slope (critical slope is discharge specific). For all out- let control culvert tests, the test culverts were installed in as horizontal a position as possible (i.e., zero slope) to ensure subcritical culvert flow and outlet control conditions. The culverts discharged into a tail box. A stop log assembly in the tail box was used in many cases to artificially control the tail- water depth, forcing the culvert to flow under outlet control. The entrance loss (He) for each test condition was determined as follows. The culvert entrance loss is equal to the difference between the total head in the head box and the representative, one-dimensional total head value in the culvert at the inlet. The total head in the head box was determined by measuring the piezometric head in the head box at a location where the veloc- ity head was negligible. The head box pressure tap location is shown in Figure 2-7. The total head inside the culvert inlet was determined by projecting the total head determined at pres- sure tap locations distributed along the length of the culvert invert (see Figure 2-7) back to the culvert inlet by either adding back the calculated friction loss for full-pipe flow conditions or by using gradually varied flow computational techniques for open channel culvert flow conditions. The resulting calculated upstream total head values for each of the pressure taps were averaged to give an average total head at the inlet. After the entrance loss (He) was calculated, the entrance loss coefficient was calculated using Equation 1-1 with the average culvert velocity as the representative velocity term. Using buried-invert culverts with a smooth uniform mate- rial on all flow boundaries made it possible to estimate friction losses for full-culvert flows by applying standard closed- conduit friction loss relationships and friction factors. It was also possible to calculate gradually varied flow profiles for free-surface culvert flows. Culvert entrance loss was assumed to be primarily a function of the inlet geometry of the culvert, not the roughness of the culvert material. Using smooth steel plate also facilitated accurate piezometric head measurements inside the pipe. With a smooth wall boundary, the pressure taps were oriented normal to the streamlines in the culvert. No localized turbulence regions were generated by a boundary profile as would exist with a corrugated pipe wall, for example. If streambed materials had been used for the culvert invert, it would have been difficult to account for friction loss and gradually varied flow profile variations associated with the composite hydraulic roughness flow boundary. Due to the irregular flow cross-section, four times the hydraulic radius was used as the representative pipe diam- eter in the friction relationships, as suggested by Flammer et al. (1986). The Froude number (Fr) was monitored for free-surface culvert flow conditions to verify that subcritical flow (Fr < 1.0) existed in the culvert barrel, an indicator of outlet control. The material roughness height for the steel test culverts was assumed to be 0.0018 in. (Flammer et al., 1986). Inlet Control Testing Inlet control conditions were achieved by installing the test culverts at a slope greater than the critical slope. In addition, Equations 1-2 and 1-4 suggest that culvert inlet control head- discharge relationships are slope-dependent. Based on inlet control testing of buried-invert culverts, Robinson (2005)
4concluded that for culverts with slopes ⤠3%, the effect of slope on the inlet control discharge rate was negligible. Addi- tionally, it is worth noting that the magnitude of the slope term in Equations 1-2 and 1-4 will be quite small for slopes on the order of 3%. As a result, the buried-invert culvert inlet control flow tests for this study were conducted at a uniform slope (approximately 3%). In order to ensure inlet control conditions in the test cul- verts, the Froude number corresponding to the average flow depth and velocity through the culvert was calculated to verify supercritical flow (Fr > 1.0). For inlet control flow con- ditions, the empirical constants for Equations 1-2 through 1-4 were determined for the appropriate inlet conditions (i.e., submerged or unsubmerged) as follows. For each test culvert and end treatment tested, upstream total head (Hw) and discharge (Q) data were collected for both submerged and unsubmerged inlet conditions as well as ponded and chan- nelized approach flow conditions. The total head in the head box was determined using the pressure tap in the head box for outlet control as explained previously. Note that the piezo- metric head (Hwi) in Equations 1-2 through 1-4 was replaced by the total upstream head (Hw), as Hw is the more appropri- ate head term when comparing the variation in culvert inlet performance between ponded and channelized approach flow conditions. For ponded upstream conditions, Hw and Hwi were equivalent. Once the Hw and Q data were collected for each test cul- vert, end treatment, and flow condition (i.e., submerged and unsubmerged inlet and ponded and channelized approaches), the data were plotted according to quasi-dimensionless rela- tionships corresponding to Equations 1-2 through 1-4. Regres- sion analysis was used to determine the corresponding inlet control head-discharge relation constants, K, M, c, and Y, which were unique for each culvert geometry tested. 1.3 Report Layout Summary The various topics associated with this project are dis- cussed separately in individual report chapters. The chapters are as follows: ⢠Chapter 2: Buried-Invert or Embedded Culverts ⢠Chapter 3: Slip-Lined Culverts ⢠Chapter 4: Culvert Exit Loss ⢠Chapter 5: Inlet Control Hydraulics of Multiple Circular Culverts ⢠Chapter 6: The Behavior of Hydraulic Roughness Coef- ficients in Open Channel Flow ⢠Chapter 7: Open Channel Flow Resistance: the Hy drau- lic Radius Dependence of Manningâs Equation and Manningâs n ⢠Chapter 8: Open Channel Flow Resistance: Composite Roughness A summary is presented in each chapter. Tabular experi- mental data related to Chapters 2, 3, and 5 are presented in the report appendices.
