U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590

Skip to content
Facebook iconYouTube iconTwitter iconFlickr iconLinkedInInstagram

Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations

This report is an archived publication and may contain dated technical, contact, and link information
Publication Number: FHWA-HRT-06-138
Date: October 2006

Effects of Inlet Geometry on Hydraulic Performance of Box Culverts


The theory outlined in HDS-5, as described in this chapter, is the basis for analyzing data from this study.


Empirical regression coefficients developed by fitting experimental data to semitheoretical equations are listed in HDS-5 for each inlet shape to relate headwater to discharge intensity for a range of flows.

Analysis of inlet-control data depends upon whether or not the inlet is submerged. If the inlet is not submerged, headwater can be expressed as a simple energy balance between the critical section and the upstream section, which are almost adjacent to one another. Alternatively, headwater can be analyzed as flow over a weir. In the latter instance, discharge is determined in part by the headwater raised to a power, M, which typically has a factor of two-thirds, or 0.667. If the inlet is submerged, flow through the inlet is conceptually like flow through an orifice, albeit an irregular shaped orifice. Orifice flow depends in part on the effective area of the orifice flow and the change in pressure. Since headwater depths are usually measured from the invert of a culvert, a more general expression for the orifice flow includes the distance from the invert.

Figure 12 illustrates an unsubmerged inlet control condition. Experimental data from this type of condition have been analyzed two ways. First, a simple energy balance can be made between sections 1 and 2 illustrated in figure 12. The critical specific energy Hc can be computed for any culvert shape, although it is tedious computation for some shapes. Data is regressed to determine regression coefficients K and M for computing the entrance loss, He. HDS-5 describes this analysis as the form 1 analysis for unsubmerged inlet control flow (figure 13). Most of the circular culvert inlets listed in HDS-5 are based on the form 1 analysis. The other way to analyze this type of data is to treat it as flow over a weir as illustrated in the form 2 equation (figure 14). Then the regression coefficients essentially absorb the critical specific energy control. All of the rectangular box culverts in HDS-5 are based on the form 2 equation, and in all cases, the exponent M was 0.667. Circular culverts and elliptical culverts with tapered throats are also based on the form 2 equation, but the exponent M varies a little for these inlets.

Figure 12. Diagram. Typical inlet control flow condition.

View Alternate Text

Figure 13. Equation. Unsubmerged form 1, inlet control.

View Alternate Text

Figure 14. Equation. Unsubmerged form 2, inlet control.

View Alternate Text

For inlet control under submerged conditions, regression constants c and Y are determined using the equation in figure 15.

Figure 15. Equation. Submerged form, inlet control.

View Alternate Text
Where for figures 13, 14, and 15:
HWiis headwater depth above inlet control section invert.
Dis interior height of culvert barrel.
Hcis specific head at critical depth (dc + Vc2/2g).
Qis discharge.
Ais full cross-sectional area of culvert barrel.
Sis culvert barrel slope.
K, M, c, Yare regression constants.
Kuis 1.811 for SI and 1.0 for English units.


Outlet control is the general case that covers everything except the special case of free surface flow passing through critical flow near the inlet. Outlet control flow occurs anytime the barrel flows full throughout, as illustrated in figure 16, regardless of the slope of the culvert.

Figure 16. Diagram. Outlet control for full flow condition.

View Alternate Text

If the culvert is installed on a mild slope (0.7 percent), outlet control will occur for both a submerged or unsubmerged inlet and a submerged or unsubmerged outlet. If the culvert is installed on a steep slope (3 percent), outlet control will occur if the tailwater is sufficiently high to submerge the critical depth at the upstream end of the culvert or to cause full barrel flow throughout. If a culvert is installed on a steep slope and only a portion of the barrel is submerged, a hydraulic jump will occur in the barrel but the headwater will still be inlet controlled. The controlling relationship for outlet control is an energy balance between sections 1 and 2, as illustrated in figure16, but now the two sections are separated by all the losses and elevation changes that occur between the tailwater channel and the approach channel. The outlet control head-discharge relationship, in which head relates to the total energy, is shown in the equation in figure 17.

Figure 17. Equation. Head-discharge relationship, outlet control.

View Alternate Text
HWois headwater depth above the outlet invert.
TWis tailwater depth above the outlet invert.
Vuis approach velocity.
Vdis downstream velocity.
HLis total energy loss.

The total energy losses include the entrance, friction, and exit losses, as shown in figure 18.

Figure 18. Equation. Total energy losses.

View Alternate Text
HLe is entrance loss.
HLfis friction loss.
HLo is exit loss.

The entrance and exit losses are commonly expressed as a fraction of the barrel velocity head, as shown in figures 19 and 20.

Figure 19. Equation. Entrance loss.

View Alternate Text

Figure 20. Equation. Exit loss.

View Alternate Text
Keis an entrance loss coefficient.
Kois an exit loss coefficient, usually assumed to be 1.0 for design purposes.
V2/2gis velocity head inside the culvert barrel.
Vd2/2gis velocity head in the downstream channel near the outlet, often neglected in design.

Since this study was focused on the effects of inlet geometry, the primary emphasis was on the entrance loss coefficient for outlet control, but exit losses were measured to determine if multiple barrels or skewed inlets affected the exit loss coefficient. It would be reasonable to express the entrance loss in terms of a difference in velocity heads, especially for wide culverts with low contraction ratios, but in this study the data were not analyzed that manner.

Previous | Contents | Next

Federal Highway Administration | 1200 New Jersey Avenue, SE | Washington, DC 20590 | 202-366-4000
Turner-Fairbank Highway Research Center | 6300 Georgetown Pike | McLean, VA | 22101