Skip to contentUnited States Department of Transportation - Federal Highway Administration FHWA Home
Research Home   |   Hydraulics Home
Report
This report is an archived publication and may contain dated technical, contact, and link information
Publication Number: FHWA-RD-02-078
Date: November 2003

Bottomless Culvert Scour Study: Phase I Laboratory Report

2. EXPERIMENTAL APPROACH

TEST FACILITIES AND INSTRUMENTATION

The experiments were conducted in the FHWA Hydraulics Laboratory located at the Turner-Fairbank Highway Research Center (TFHRC) in McLean, VA. Test facilities and instrumentation used during the experiments are described in this section.

Figure 1. View of the flume in the Hydraulics Laboratory. Photo. This picture shows an overhead view of the Hydraulics Lab at the Turner-Fairbank Highway Research Center in McLean, Virginia. The flume is on the right side of the picture. It is elevated above the lab floor, and there is a walkway next to the flume that is accessible via a set of stairs in the foreground.

Figure 1. View of the flume in the Hydraulics Laboratory.

Hydraulic Flume

The experiments were completed in a 21.34-meter- (m) long by 1.83-m-wide (70-foot- (ft) long by 6-ft-wide) rectangular flume with a 2.4-m-long by 1.83-m-wide (8-ft-long by 6-ft-wide) recessed section to allow for scour hole formation. A 9.14-m (30-ft) flow development section from the head box to the transition section consisted of a plywood floor constructed 0.1 m (4 inches) above the stainless steel flume bottom. The plywood floor was coated with a layer of epoxy paint and sand to approximate the roughness of the sand bed channel in the recessed section. The walls of the flume are made of a smooth glass. The flume was set at a constant slope of 0.04 percent and the depth of flow was controlled with an adjustable tailgate located at the downstream end of the flume. Flow was supplied by a 0.3-cubic meter/second (m3/s) (10-cubic foot/second (ft3/s)) pumping system. The discharge was measured with an electromagnetic flow meter.

Electromagnetic Velocity Meter Operation

A 13-millimeter (mm) spherical electromagnetic velocity sensor (Marsh-McBirney 523) was used to measure equivalent two-directional mean velocities in a plane parallel to the flume bed.

A fluctuating magnetic field was produced in the fluid surrounding the spherical sensor that was orthogonal to the plane of four carbon-tipped electrodes. As a conductive fluid passed around the sensor, an electric potential was produced proportional to the product of the fluid velocity component tangent to the surface of the sphere and normal to the magnetic field and the magnetic field strength. The electrodes located at four locations on the sensor detected the voltage potential created by the flowing water. The voltage potential produced was proportional to the velocity of the fluid flowing in the plane of the electrodes. Two orthogonal velocity components in the plane of the electrodes were measured. Detailed information on the meter operation is available in the technical manual, Instruction Manual Model 511 Electromagnetic Water Current Meter, provided by the probe manufacturer, Marsh-McBirney Inc.

Post-Processing and Data Analysis

Post-processing and data analysis were performed using the LabVIEWTM graphical programming technique for building applications such as testing and measurement, data acquisition, instrument control, data logging, measurement analysis, and report generation. LabVIEW programs are called virtual instruments (VI's) because their appearance and operation imitate physical instruments such as oscilloscopes and multimeters. Every VI uses functions that manipulate input from the user interface or other sources and displays that information or moves it to other files or other computers.

A VI contains the following three components:

  • Front panel: Serves as the user interface (figure 2).
  • Block diagram: Contains the graphical source code that defines the functionality of the VI (figure 3).
  • Icon and connector pane: Identifies the VI so that you can use one VI in another VI. A VI within another VI is called a subVI. A subVI corresponds to a subroutine in text-based programming languages.
Figure 2. Example of a front panel. Computer screen capture. This screen capture is of the front panel of a lab view program. It serves as the user interface. The top bar reads GKY2 method scour data Ken Young (metric new).VI. This example screen displays a scatterplot graph with an X axis from 0.000 to 175.000 and a Y axis from 1.0 to 12.0, where X and Y may be any measurements or calculations made by the computer. Plot 1 points are concentrated on the lower left side of the graph, with a few scattered points moving to the upper right side of the graph. The regression line begins at coordinates 0.000, 2.6, curves up steeply at first, then levels out to end at coordinates 160.000, 9.0. This plot is shown to demonstrate the graphing capabilities of the software

Figure 2. Example of a front panel.

Figure 3. Example of a block diagram. Computer screen capture. This screen capture is of the block diagram screen of a lab view program. The purpose of this screen capture is to show the visual programming environment of LabView, in which different aspects of the program take the form of boxes interconnected by lines. The top bar reads GKY method scour data Ken Young (metric new).VI Diagram. The screen displays graphical source code that defines the functionality of the virtual instrument. The left side of the screen contains a box that reads "Path for Data." Above this box and slightly to the right is a box that displays the number of runs. Both these boxes lead to a shaded box that reads scour VW. This box has five lines coming out of it. To the right of this box but not connected to it by a line are nine connected boxes, all with 0.10 inside them. One box is connected to the top box in this series; it reads 0. Under these boxes is a box that displays the formula Array Y max over Y0. Under this box and slightly to the left is a box that displays the formula Array A over Y0 top hat 2. Under this box and slightly to the right is a box that displays Array Y max measured. To the right of the series of nine connected boxes is a box that reads, "Number of coefficients." Under this box is a box that displays the formula Y max over Y0 times y2 over Y0 data. Under this box is a box that reads, "Y-Y." Next to this box is a box that reads, "Y, M-Y, C." Four boxes are under this box. In order of position, top to bottom, they read, "model description," "MSE," "RSQ," and another "RSQ." Above the "Y, M-Y, C" box are five text boxes. In order of position, top to bottom, they read, "MSE," "RSQ (fit)," "coefficients," "RSQ," and "MSE." There are a series of eight boxes to the right of the screen; each contains a grid of six cells and is connected to boxes that are numbered sequentially, one through eight. Under these boxes is a box with the formula Y max over y2 times fit minus A over Y0 top hat 2.

Figure 3. Example of a block diagram.

The regression analysis was done using the nonlinear Levenberg-Marquardt fit algorithm to determine the set of coefficients that minimize the chi-square quantity:

Equation 1. Chi square equals the summation of I equals zero to N minus 1 times the square of the following quotient: Y subscript I minus F of X subscript I and A subscript 1 through A subscript M, all divided by theta subscript I.     (1)

In this equation, xi and yi are the input data points, and f (xi; a1,...,aM) is the nonlinear function, where a1,...,aM are coefficients. If the measurement errors are independent and normally distributed with a constant standard deviation σi = σ, this is also the least-square estimation.

MODEL BOTTOMLESS CULVERT SHAPES

Three bottomless culvert shapes were constructed and tested: (1) a rectangular model with a width of 0.61 m (2 ft) and a height of 0.46 m (1.5 ft) (figure 4), (2) a CONSPAN model with a width of 0.61 m and a height of 0.45 m (1.46 ft) (figure 6), and (3) a CONTECH model with a width of 0.61 m and a height of 0.42 m (1.36 ft) (figure 8). All three models were evaluated with 45-degree wingwalls (figures 5 and 7) and without wingwalls. The models were constructed of Plexiglas®. Marine plywood was used for the vertical face of the models and for the wingwalls. The models were mounted in the centerline of the flume.

Figure 4. Rectangular model, vertical face. Photo. This picture shows a rectangular model of a bottomless culvert shape. The rectangle is 24 inches wide and 18 inches high, and it is surrounded by sediment that is lowest at the edges of the rectangle and peaks at its center. The sediment's highest point reaches 8 inches high, or 10 inches from the top of the rectangle.

Figure 4. Rectangular model, vertical face.

Figure 5. Rectangular model with wingwalls. Photo. This picture shows the rectangular model in figure 4 with wingwalls attached. The wingwalls are attached to the rectangle at a 45-degree angle, and sediment can be seen peaking in the middle of the rectangle, 14 inches from its top.

Figure 5. Rectangular model with wingwalls.

Figure 6. Conspan model. Photo. This picture shows a Conspan model of a bottomless culvert shape. A space is cutout of the solid portion; the space is 24 inches wide and 17.6 inches high. Sediment is highest at the ends of the solid portion and lowest in the cutout.

1 inch = 25.4 mm

Figure 6. CONSPAN model.

Figure 7. Conspan model with wingwalls. Photo. This picture shows the model in figure 6 with wingwalls attached. The sediment is distributed fairly evenly across the bottom of the shape.

Figure 7. CONSPAN model with wingwalls.

Figure 8. Contech model. Photo. This picture shows a Contech model of a bottomless culvert shape. A space is cutout of the solid portion; the space is 24 inches wide and 16.4 inches high. Sediment is highest at the ends of the solid portion and lowest in the cutout.

1 inch = 25.4 mm

Figure 8. CONTECH model.

EXPERIMENTAL PARAMETERS

Steady flow experiments were conducted for approach flow depths of 0.106 m, 0.212 m, and 0.304 m (0.35 ft, 0.7 ft, and 1 ft) and approach velocities ranging from 0.091 to 0.304 m/s (0.3 to 1 ft/s). The discharges to obtain the approach flow conditions varied from approximately 0.019 to 0.14 m3/s (0.7 to 5 ft3/s). The particle size (D50) was varied from 1.2 to 3.0 mm (0.004 to 0.01 ft) for the scour experiments.

Riprap experiments were conducted for uniform particle sizes of 9.5 mm, 12.7 mm, 20.3 mm, and 25 mm (0.375 inch, 0.5 inch, 0.8 inch, and 1 inch). The velocity was increased incrementally until discernible areas of particles were dislodged, which was considered to define the failure condition for that particle size. Because of time constraints, riprap experiments (figure 9) were conducted for the rectangular model with vertical headwalls only. Vertical headwalls were considered a worst-case condition and wingwalls should reduce the riprap size determined from these experiments.

Figure 9. Rectangular model from the scour protection task. Photo. This picture shows an overhead view of a rectangular model with vertical headwalls only. There are large particles of uniform size piled against the headwalls and extending into the rectangular cutout.

Figure 9. Rectangular model from the scour protection task.

Previous | Table of Contents | Next


The Federal Highway Administration (FHWA) is a part of the U.S. Department of Transportation and is headquartered in Washington, D.C., with field offices across the United States. is a major agency of the U.S. Department of Transportation (DOT).
The Federal Highway Administration (FHWA) is a part of the U.S. Department of Transportation and is headquartered in Washington, D.C., with field offices across the United States. is a major agency of the U.S. Department of Transportation (DOT). The hydraulics and hydrology research program at the TFHRC Federal Highway Administration's (FHWA) R&T Web site portal, which provides access to or information about the Agency’s R&T program, projects, partnerships, publications, and results.
FHWA
United States Department of Transportation - Federal Highway Administration