Physical Model Studies of the GCID
Pumping Plant Fish Screen
Structure Alternatives

Progress Report No. 1

1:30 Scale Model Investigations
Alternative D

Return to abstract.


Table of Contents


Tables

  1. GCID screening option D-3, 2-dimensional simulation results (n = 0.025 for the river channel)
  2. GCID screening option D-3, 2-dimensional simulation results--simulated riffle (n = 0.025 for the river channel)
  3. Initial testing flow simulation set points
  4. Expanded testing program for the final screen concept design

Figures

  1. General location map of GCID pumping plant and existing fish screen facilities
  2. Conceptual layout: plan view of proposed A alternative
  3. Conceptual layout: plan view of proposed D alternative
  4. Sacramento River main channel invert elevations at Montgomery Island for the years 1991 and 1995
  5. Photograph of the D alternative physical model as constructed in the laboratory
  6. Head loss coefficient versus Reynolds number relationship for an angled, vertical wedge-wire, screen configuration
  7. RMA2 numerical simulation of a simplified riffle in the main river channel along Montgomery Island
  8. Photograph of the ADV setup for velocity measurements along the screen
  9. Conceptual layout of the D alternative screen configuration and associated modifications
  10. Original D alternative screen configuration test results, flow simulation 7
  11. Opposite bank guidewall modification test results, flow simulation 1
  12. Opposite bank guidewall modification test results, flow simulation 7
  13. 4 screen orientation modification test results, flow simulation 1
  14. 4 screen orientation modification test results, flow simulation 7
  15. Pumping plant forebay guidewall modification test results, flow simulation
  16. Reduced screen length modification test results, flow simulation 1
  17. Reduced screen length modification test results, flow simulation 7
  18. Bypass channel entrance modification test results, flow simulation 1
  19. Final test results, flow simulation 1
  20. Final test results, flow simulation 7
  21. Final test results, flow simulation 12
  22. Final test results, flow simulation 13
  23. Final test results, flow simulation 14
  24. Final test results, flow simulation 9
  25. Final test results, flow simulation 10
  26. Final test results, flow simulation 11
  27. Plan view of the D alternative fish screen showing oxbow channel flow velocity vectors in the vicinity of the screen
  28. Repeatability verification results, flow simulation 7
  29. Repeatability verification results


Purpose

This report presents the results of the D alternative, positive barrier fish screen physical model investigations for GCID (Glenn-Colusa Irrigation District). The study was performed to evaluate, improve, and document the viability of the concept as a means to protect the fishery resource.


Application

The information included in this report is provided to the GCID TAG (Technical Advisory Group) to assist in the evaluation of proposed screen alternatives and to provide design data for the selected alternative.


Introduction

The GCID Pumping Plant is located in north-central California, about 100 miles north of Sacramento, on an oxbow of the Sacramento River. Figure 1 is a general location map. The pumping plant exports water from the Sacramento River to the west side of the Sacramento River Valley for irrigation purposes.

In 1972, a rotary drum screen facility was constructed to provide fish protection from pumping plant entrainment. The facility originally consisted of 40 drum screens 8 ft wide and 17 ft in diameter. In 1970, the Sacramento River experienced the largest flooding since the construction of Shasta Dam. The result was a meander cutoff downstream from Montgomery Island, which caused a decrease in river length of almost 1-1/2 miles. This meander cutoff has caused a drop in water surface elevations of about 3 ft at the north end of Montgomery Island. These changes occurred over several years as the river stabilized. Lower water surface elevations resulted in lower than desired water depths in front of the drum screens. As a result, through-screen velocities exceeded resource agency fish screening criteria during high diversions. In 1991, the NMFS (National Marine Fisheries Service) filed an injunction against the irrigation district to restrict pumping during the peak winter-run chinook salmon downstream migration period.

The district initiated an aggressive program in conjunction with resource agencies to identify options for both short- and long-term resolutions of the screening problem. To improve interim screen performance, flat panel wedge wire screens were placed in front of the drum screens in 1993. In 1995, the drum screens were removed from service.

Pursuit of a long-term solution has generated a number of screening alternatives, which have, in turn, been subjected to detailed evaluation. In 1994, HDR Engineering, Inc., prepared a draft feasibility report which reviewed eight alternatives for replacement or modification of the existing screen facilities. Since then, these alternatives have been reduced to two.

The two remaining alternatives, labeled "A" and "D" are shown as figures 2 and 3, respectively.

Both of these alternatives are to be investigated under this study. Alternative A consists of a new screen facility located just upstream from the existing facility. The A screen concept is a four-bay-multiple-V structure with bypass and evaluation facilities. Screen alternative D consists of modifying the existing screen facilities by increasing the length of the flat panel screen structure. The proposed screen is about 1,000 ft long, extending about 500 ft upstream from the existing structure.

Both of the previously described alternatives will initially be evaluated and optimized using a 1:30 scale physical model. Upon completion of these investigations, one or both alternatives will be modeled at a smaller scale to provide design and operation data for the prototype facility. A report series will be generated for documentation of the physical modeling of the screen alternatives. This report covers the 1:30 scale model investigations of the D alternative and constitutes the first report in the series.


Objectives of the Model Studies

Prior to designing a fish screening facility, the objectives and operational constraints of the facility must be established. This procedure requires identifying applicable State and Federal resource agency fish screening criteria and objectives specific to the site. This process was conducted through the GCID screen replacement TAG. The following organizations participate on the TAG:

In conjunction with these organizations, several consultants also participate as members of the TAG. These consultants provide biological, engineering, and legal expertise.

Through this process, the following major objectives were identified for the D alternative screen concept.

Not present in the objectives for the 1:30 model are evaluations of operating criteria, intermediate screen bypasses, and screen baffling. These topics were not included for the following reasons:


Conclusions

The results of these investigations demonstrate that the D alternative is a viable design. The study results show the long flat plate screen concept can be designed to meet the listed objectives. A series of modifications to the screen design was identified and instituted through the model study to improve performance. These screen modifications and the final design are shown on figure 9. A brief summary of screen performance versus objectives for the final D alternative screen configuration is given below.


Similitude

The physical model of the D alternative must resemble the prototype geometrically and kinematically to predict prototype performance under specified operating conditions (Bureau of Reclamation, 1986). Geometric similarity is achieved with the ratios of all prototype to model geometric parameters being equal. Kinematic similarity is achieved with the ratios of all prototype to model velocities being equal. Froude law similitude is employed to establish the kinematic relationship between model and prototype. This similitude is based on maintaining model and prototype Froude numbers, which are equal in all cases. The required geometric and kinematic ratios for this 1:30 Froude scale model are as follows:

Geometric

Lr = Lp/Lm = 30
Ar = (Lr)2 = 900
Vr = (Lr)3 = 2,700

where:
Lp = prototype characteristic length
Lm = model characteristic length
Lr= length ratio
Ar = area ratio
Vr = volume ratio

Kinematic

tr = (Lr)1/2 = 5.48
vr = (Lr)1/2 = 5.48
ar = 1
Qr = (Lr)5/2 = 4,930

where:
tr = time ratio
vr = velocity ratio
ar = acceleration ratio
Qr = discharge ratio


Physical Model

The fish screen model was constructed at the Bureau of Reclamation WRRL (Water Resources Research Laboratory) in Denver, Colorado. The 1:30 scale model covered about 3,000 ft of the oxbow channel, including the D alternative screen structure, the pumping plant, and part of the downstream bypass channel. The scale was chosen to achieve the study objectives and yield efficiency of model operation. Froude number similitude criteria were used to establish kinematic similarity between model and prototype. Figure 5 is a photograph of the river model for the D alternative as constructed in the laboratory.

Modeling of the screen under this investigation merits some important considerations. The prototype screen is sized such that it consists of 0.071-in wedge wire on 0.164-in centers, representing a 3/32-inch slot opening, which yields an open area of about 55 percent. The size of the prototype screen prevents modeling this detail at a 1:30 scale. However, for the modeling purposes of this application, representing only the resistance characteristics of the screen is important. The resistance characteristics of the prototype, which are defined by the head loss versus discharge relationship, can be adequately modeled, provided the Re (Reynolds number) of the through-screen flow regimen is sufficiently high. Reynolds number is a non-dimensional ratio of inertial forces to viscous forces, expressed as:

Reynolds number equation

Previous work performed by Yeh et al. (1988) in this area has indicated that for Re greater than or equal to 250, the screen head loss coefficient is not significantly sensitive to large changes in approach velocity. An evaluation similar to Yeh et al. (1988) was conducted for the GCID model in a WRRL flume. However, to evaluate model scale effects, through-screen velocity was used rather than approach velocity. Figure 6 illustrates head loss versus Re relationship for a screen angled 10o to the flow. These results show that a minimum Reynolds number of 80 (based on the through-screen velocity) is adequate for representing the prototype screen resistance. Therefore, to adequately model the prototype screen requires similarity of screen porosity and a through-screen Re of greater than about 80. This condition was achieved in the model by using 3/16-in perforated plate having a 56-percent open area to model the prototype screen. Model through-screen velocity for a 0.33-ft/s approach velocity gives an Re of about 120.


Numerical Model

The river system hydraulics near GCID were estimated using the hydrodynamic model RMA2 (RMA2 is marketed under the name Boss FastTabs by Boss International), which is a two-dimensional, depth-averaged, finite element model developed by the U.S. Army Corps of Engineers. Numerical modeling was performed under contract by Ayres Associates. Much of the model development had previously been conducted by Ayres (Resource Consultants and Engineers, 1994) as part of an effort to study options for a gradient restoration structure across from Montgomery Island.

Numerical flow simulations were conducted to provide hydraulic data on river flow splits around Montgomery Island and determine estimated water surface elevations within the oxbow channel. These data were needed to establish entrance and exit boundary conditions for operation of the physical model. A total of 15 flow scenarios were run for the D alternative screening concept. Of these scenarios , 13 flow combinations were identified to establish the system (river and pumping plant) hydraulics assuming no gradient restoration structure in the main river channel. Table 1 lists the flow combinations modeled and the major hydraulic data derived for each. These simulations were conducted using 1991 main river and bank topography data. The main channel data are considered to represent recent low river gradient conditions at Montgomery Island. The oxbow channel was modeled as a trapezoidal channel, 2:1 side slopes, with a 145-ft-wide bottom at elevation 128.0. At about 200 ft upstream from the screen structure, the oxbow invert elevation was lowered to elevation 127.0. Simulations 1 to 11 were each repeated using three values of channel rugosity corresponding to Mannings n values of 0.02, 0.025, and 0.03. These roughness values cover the expected range of channel conditions and, therefore, give the likely range of hydraulic parameters.

Table 1. - GCID screening option D-3--2-dimensional simulation results (n = 0.025 for the river channel).
Run
No.
Qriver Qpump QHamilton City Manning "n"
(Bypass
channel)
Qintake Qbypass Water Surface Elevation (ft)
North
Gage
South
Gage
GCID
Screens
(Input) (Output)
1 7,000 2,000 5,000 0.003
0.025
0.02
2,534
2,578
2,629
534
578
629
135.9
135.9
135.9
134.5
134.5
134.5
134.5
134.5
134.5
2 7,000 2,500 4,500 0.025
0.02
2,914
2,945
414
445
135.8
135.8
134.4
134.4
134.9
134.9
3 8,000 2,400 5,600 0.03
0.025
0.02
2,928
2,970
3,018
528
570
618
136.2
136.2
136.2
134.8
134.8
134.8
135.5
135.5
135.4
4 8,000 2,850 5,150 0.025
0.02
3,262
3,292
412
442
136.1
136.0
134.6
134.6
135.1
135.0
5 9,000 2,750 6,250 0.03
0.025
0.02
3,285
3,325
3,375
535
575
624
136.5
136.5
136.5
135.0
135.0
135.0
135.7
135.6
135.6
6 9,000 3,000 6,000 0.025
0.02
3,480
3,520
480
520
136.4
136.4
134.9
134.9
135.4
135.4
7 10,000 3,000 7,000 0.03
0.025
0.02
3,570
3,615
3,665
570
615
665
136.8
136.8
136.8
135.2
135.2
135.2
135.9
135.9
135.9
8 12,000 3,000 9,000 0.03
0.025
0.02
3,800
3,872
3,948
800
872
948
137.5
137.5
137.4
135.8
135.8
135.8
136.7
136.7
136.6
9 20,000 3,000 17,000 0.03
0.025
0.02
4,652
4,793
4,950
1,652
1,793
1,950
139.9
139.9
139.9
138.2
138.2
138.2
139.5
139.4
139.4
10 40,000 3,000 37,000 0.03
0.025
0.02
7,230
7,404
7,572
4,240
4,415
4,584
144.5
144.5
144.5
142.8
142.8
142.8
144.2
144.2
144.2
11 60,000 1,000 59,000 0.03
0.025
0.02
9,060
9,255
9,427
8,088
8,275
8,440
148.5
148.5
148.5
146.7
146.7
146.7
148.3
148.3
148.2
12 8,000 300 7,700 0.025
0.02
1,364
1,474
1,064
1,174
137.0
136.9
135.4
135.4
136.8
136.8
13 5,000 1,000 4,000 0.025
0.02
1,611
1,670
611
670
135.5
135.4
134.2
134.2
135.2
135.1

To assess the impact of the riffle aggradation identified in the 1995 main channel survey on the system hydraulics, two additional simulations were conducted. The riffle was modeled as a broad-crested weir placed at the location of the natural riffle. The simplified riffle was depicted as a rock structure with a 20-ft-wide (stream-wise direction) crest and a 1:100 downstream slope. The riffle was superimposed on the 1991 river topography as shown on figure 7. Weir crest elevations of 133.0 and 134.0 were run for the condition of 7,000-ft3/s river (north gauge) and 3,000-ft3/s pumping at GCID. Table 2 gives the major hydraulic parameters with the simplified riffle in the main channel. The numerical simulations of the riffle were conducted to provide limited data indicating main channel aggradation impacts on operation. The flow conditions given in table 2 were not modeled in the physical model.

Table 2. - GCID screening option D-3--2-dimensional simulation results--simulated riffle
(n = 0.025 for the river channel) (GMF = gradient maintenance facility).
Run
No.
Qriver Qpump QHamilton City Manning "n"
(Bypass
channel)
Riffle
Crest
Elevation
Qintake Qbypass Water Surface Elevation (ft)
Up-
stream
from
GMF
Down-
stream
from
GMF
North
Gage
South
Gage
GCID
Screens
(Input) (Output)
14 7,000 3,000 4,000 0.025 134 3,612 612 136.2 134.3 136.3 134.2 135.2
15 7,000 3,000 4,000 0.025 133 3,420 420 135.8 134.3 135.9 134.2 134.8


Test Setup

Water is supplied to the model from a 250,000-gallon sump via the laboratory pumping system. Discharge delivered to the model is measured using the laboratory venturi meters. The system uses a flow controller to maintain desired flow rate. Model tailwater elevations are maintained using stoplogs at the downstream end of the bypass channel. Model water surface elevations are monitored using point gages set at specific locations (i.e., intake channel, screen structure forebay, bypass channel entrance). The pumping plant was simulated using three separate pump and manifold systems in the model. Pump intakes 1 and 2, 3 to 8, and 9 and 10 were manifolded to separate pumps. Pumped discharges were measured using a Controlotron ultrasonic flowmeter for pumps 3 through 8 and paddle wheel type flowmeters for pumps 1 and 2 and pumps 9 and 10. The bypass discharge was measured using a 12.5o v-notch weir. Model velocities were measured using an ADV (acoustic Doppler velocimeter).


Testing

Testing under this phase of the hydraulic model study has been consistent with achieving the required objectives. Both dye and confetti tests were performed for flow visualization purposes to determine general flow patterns associated with this alternative. Velocity measurements were conducted to quantify near-screen hydraulic conditions. The results of these flow visualization and velocity measurement tests lead to modifications that will improve performance of the D alternative screen.

Flow Visualization

Flow visualization tests were conducted to evaluate the upstream transition from the channel to the screen structure, the opposite bank guidewall orientation, and the downstream transition from the screen structure to the bypass channel. These tests employed both confetti and dye to establish surface and sub-surface flow patterns, respectively. Tests were documented using video taping and photographs.

Velocity Measurements

All velocity measurements were acquired using a Sontek ADV (acoustic Doppler velocimeter). Figure 8 is a photograph of the ADV setup used for acquiring velocity measurements along the screen for this investigation. The ADV instrument can measure local velocities in water to a resolution of 0.001 ft/s and has a maximum sampling rate of 200 Hz and an accuracy of ±0.5 percent of the measured value. For the model study, each velocity reported was an average of about 750 samples obtained at 25 Hz.

The relatively large measurement sample size was selected to reduce the uncertainty associated with measuring the normal velocity component in the presence of a strong sweeping velocity field. Measurement uncertainty is proportional to the inverse square root of the number of the sample size, or simply, the greater the sample size the lower the uncertainty of the estimate of the population mean.

Velocity data measured in the study are reported as the mean value of the data sample. The mean value was determined as:

mean value equation

The average uncertainty of each measurement can be characterized by the standard deviation, which is defined as:

standard deviation equation

The standard deviation represents the average uncertainty of the separate measurements of u1,...,un. The uncertainty of the mean or best estimate of velocity is the standard deviation of the mean or probable error. The value of u can be considered more reliable than any one measurement considered separately because it is comprised of all n measurements of u. The uncertainty in any set of n measurements is defined as:

standard deviation of the mean equation

Velocity measurements were acquired along the screen structure for two baseline flow simulations prior to and after each successive modification to the model. Baseline flow simulations consisted of pumping plant discharges of 2,000 ft3/s and 3,000 ft3/s, both with bypass discharges of 500 ft3/s. Minimum water surface elevations and corresponding river flows were estimated for these pumping conditions using available numerical data from the gradient restoration feasibility study (Resource Consultants and Engineers, 1994). Estimated values used for the physical model are given in table 3. Better estimates of the river flows for these conditions were obtained following completion of RMA2 numerical modeling. These values are shown in parentheses below the estimated values in table 3. For consistency in comparison of modifications, the river values given as the upper values in table 3 were carried through the model study. The simulation number given in table 3 corresponds to the sequencing of numerical simulations of table 1.

Table 3. - Initial testing flow simulation set points.
Simulation
No.
Qriver
(ft3/s)
Qpumping
(ft3/s)
Qintake
(ft3/s)
Qbypass
(ft3/s)
w.s. el.screens
(ft)
1 7,000
(7,500)
2,000 2,500 500 135.5
7 10,000
(12,000)
3,000 3,500 500 136.7

Velocities were measured at the centerline of each 40-ft-wide bay for the new screen structure and at the centerline of every fifth bay along the existing structure. Point velocities were measured at the 0.6 depth, thus representing the approximate vertical average velocity.

Evaluation of flow visualization and velocity data guided modifications tested in the model. This approach resulted in tests of:

Each of these modifications was developed based upon the results of previous tests. Again, screen velocity measurements were used to identify possible causes of poor screen performance. Figure 9 is a conceptual layout identifying the associated modifications.

The final screen concept configuration was tested under a wide range of flow conditions. These hydraulic conditions tested are included as table 4.


Results

The primary result of the testing is the realization of improved screen performance for the D alternative screen. This improved performance is demonstrated by the increase in screen effectiveness for the upstream 300 ft of screen area, the elimination of eddy zones on both sides of the channel transition to screen forebay, and the establishment of near-uniform screen velocity distributions under non-baffled conditions.

The major results of the model study are presented as x-y velocity plots for each configuration tested. The dependent variable is given as the measurement location along the screen structure and the independent variable represents the magnitude of velocity at each measurement location. Sweeping and normal components of velocity are plotted for each test (figs. 10 to 29). Velocity data for each test in tabular form are included in the appendix. The tabular data also provide measurement sample size, SDEV (standard deviation), and SDOM (standard deviation of the mean) (normal velocity component only).

Table 4. - Expanded testing program for the final screen concept design.
Simulation No.
(reference table 1)
Qriver
(ft3/s)
Qpumping
(ft3/s)
Qintake
(ft3/s)
Qbypass
(ft3/s)
w.s. el.screens
(ft)
1 7,000 2,000 2,500 500 135.5
7 10,000 3,000 3,500 500 136.7
9 20,000 3,000 5,029 2,029 139.6
10 40,000 3,000 8,090 5,090 144.7
11 60,000 1,000 10,350 9,350 148.7
12 4,000 500 1,100 600 135.1
13 5,000 1,000 1,650 650 135.3
14 8,000 300 1,400 1,100 136.9
Test Results

The original D alternative screen configuration was tested under flow simulation 7. Figure 10 represents the results of test No. 1. The normal and sweeping velocity components are shown by open circles and solid circles, respectively, for all data plots. As shown, the normal component screen velocity distribution is non-uniform. Negative sweeping and normal component velocities existed along the first upstream bay. Dye tests indicated that this condition was a result of a large eddy zone generated by the upstream channel transition to the screen structure. Flow visualization tests also showed approach flow separated from the opposite bank at the upstream end of the screen structure and impinged largely on the upstream one-third of the screen. As a result, a large eddy zone existed along the opposite bank guidewall.

The opposite bank guidewall was extended into the channel and shaped to turn the approach flow and align it with the screen structure. The guidewall was shaped until dye traces indicated approach flow remained attached along its full length. The reshaped guidewall provided good uniformity of approach channel flow along the screen. Dye injected into the oxbow channel upstream from the screen at three points across the channel tracked nearly parallel along the screen length. Near-bank flow entered the screen within the first quarter of the screen length, mid-channel flow entered the screen over the middle half of the screen, and opposite-bank flow moved parallel to the opposite bank entering the screen along the downstream one-quarter of its length.

The modified opposite bank guidewall was then tested under flow simulations 1 and 7. Figures 11 and 12 represent the results of these tests. Approach and sweeping velocities improved because of the guidewall changes. However, poor flow conditions persisted near the upstream transition to the screen.

To improve flow conditions at the screen's upstream end, the leading 300 ft of screen structure was angled 4o into the approach flow (fig. 9). This modification improved the alignment of the approach channel and screen. Figures 13 and 14 show the effects of this modification. The screen realignment eliminated the eddy in front of the first bay. However, screen approach velocities on the first two screen bays exceeded allowable criteria. The high velocities were caused by the close proximity of the upstream bend in the oxbow channel. Flow leaving the channel bend approached the upstream end of the screen before completing the turn. This condition caused the angle of attack on the screen to be significantly larger near the upstream end of the screen.

Two modifications were tested to further improve screen approach velocities. First, the pumping plant forebay guidewall was moved closer to the screen structure, thus reducing the forebay area, particularly at the upstream end of the screen structure (see fig. 9). Figure 15 shows the resulting screen velocity distribution for flow simulation 1. Reducing the forebay area improved the overall uniformity of approach flow along the screen but fell short of achieving the uniformity of approach velocity needed at the upstream end of the screen. The testing clearly showed the screen had to be shortened or moved downstream to avoid the direct influence of the channel bend. To test this assumption, the screen length was reduced by 150 ft (fig. 9). This reduction resulted in a screen length of 1,003 ft and a screen area of about 9,100 ft2 at a water surface elevation of 136.4. Figures 16 and 17 show the improvement in the screen velocity distribution obtained.

The final modification to the D alternative screen tested in the 1:30 scale model consisted of changing the bypass channel entrance geometry. This effort was undertaken to increase sweeping velocity on the downstream most screen bays. A submerged berm was placed along the opposite bank guidewall near the entrance to the bypass channel. The berm was designed to reduce the channel area and provide a smooth transition to the bypass channel. The berm tested increased near-screen sweeping velocities by about 30 percent at the downstream end of the screen (fig. 18). Additional efforts in this area were not considered warranted for the objectives of the 1:30 model. Final geometry of the bypass intake will depend on the final screen length chosen for the design.

Final Concept Testing

Upon completion of the initial modifications, tests were conducted to document screen performance for a wide range of river and pumping flow combinations. The flow combinations tested are listed in table 4. Figures 19 through 26 show the results of these tests. Of special note are reverse flow conditions that occur near the downstream end of the screen during low pumping (figs. 21 through 23) or high river flow conditions (figs. 24 through 26). Reverse flow is indicated on the figures by negative values of the normal velocity component. Reverse flow conditions occur when flow in excess of pumping demand moves through the upstream portion of the screen. This condition is accentuated by the curvature of the oxbow channel. Figure 27 shows velocity vectors measured at several cross sections along the oxbow channel. Flow is directed into the near bank as it moves around the bend upstream from the screen structure. The angle at which flow approaches the screen, and therefore, flow through the screen, is greatest at the upstream end. Flow combinations that result in large bypass flows will likely result in some reverse flow at the lower end of the screen structure. Dye was injected in the regions of reverse flow to determine if the condition created eddies or slack water conditions in front of the screen that might favor predators. The reverse flow through the screen was found to merge smoothly with flow entering the bypass channel moving continuously downstream.

Prior to completion of model testing, a final test was conducted to verify repeatability of the data. The final configuration was again tested under flow simulation 7 conditions. Figure 28 represents the results of this test. These results were then compared with the results obtained for test No. 11 shown on figure 20. Figure 29 represents the comparison plot of these two tests. The results show a satisfactory agreement of data.


Bibliography

Bureau of Reclamation, Hydraulic Laboratory Techniques, U.S. Department of the Interior, 1980, reprinted 1986.

Resource Consultants & Engineers (RCE), Inc., Riverbed Gradient Restoration, Sacramento River Mile 206, California, Advanced Data and Topography for the Design Memorandum , U.S. Army Corps of Engineers, December 1992.

Resource Consultants and Engineers (RCE), Inc., Riverbed Gradient Restoration Structures for the Sacramento River at the Glenn-Colusa Irrigation District (GCID) Intake, California 2-Dimensional Modeling of a Natural Riffle and Gradient Restoration Facility , U.S. Army Corps of Engineers Contract Number DACW05-90-C-0168, 1994.

Yeh, Harry H., Shrestha, Mandira, Free Surface Flow Through a Screen , University of Washington, Department of Civil Engineering, 1988.


Return to abstract.