A new fish passage design developed specifically for high-head hydroelectric plants uses controlled decompression to minimize fish injury during passage while also being less expensive than the traditional downstream passage alternatives.
By Ryan Greif, Kai Steimle, Richard Brown and Dan Gessler
Migratory and anadromous fish stocks are declining, making fish passage at hydropower facilities a critical component of the Federal Energy Regulatory Commission licensing and relicensing processes. Downstream fish passage is a two-phase problem. First, it is difficult for migrants to find a downstream route in a reservoir. Second, once migrants are directed, it is challenging to provide safe passage conditions over dam structures.
This second phase is particularly difficult at high-head dams. Traditional solutions for transporting bypassed fish involve large, expensive structures typically made of concrete, such as fish ladders, chutes or fish locks, or trap-and-haul operations. Reservoir elevation fluctuations make intakes challenging to design. Once collected, juvenile fish require low gradient conditions, which translate into large structures at high-head dams. Alternatively, hauling operations, which use barges or transport trucks to move fish downstream, are well-suited to high-head dams and reservoir fluctuations, but they often create considerable delay for migrants, stressful transport conditions and high ongoing operational costs.
The proposed 10-MW Applegate Dam Project is a run-of-river retrofit project at Applegate Dam on the Applegate River in Oregon that illustrates the challenges of providing cost-effective downstream passage. Applegate Dam is 242 feet high and is operated by the U.S. Army Corps of Engineers for flood control, fish and wildlife enhancement, water quality control, recreation and irrigation storage. The hydroelectric project, being considered for development by AG Hydro, would generate about 28.3 GWh annually. This relatively small size limits the budget for construction and operation of the project in comparison to larger, peaking hydroelectric plants.
The Oregon Department of Fish and Wildlife has targeted Applegate Lake for reintroduction of anadromous steelhead and Southern Oregon/Northern California Coast (SONCC) coho salmon, which are listed as threatened under the Endangered Species Act. Typical pool elevations vary by about 98 feet over the course of a year due to seasonal operations. These conditions make traditional downstream fish passage cost-prohibitive. For example, a bypass chute would be more than 3.5 miles long to maintain the maximum 1.3% slope needed to achieve the velocity range of 6 to 12 feet per second recommended by the National Oceanic and Atmospheric Administration.1 In addition, the variable pool elevations experienced would make design of the entrance difficult. A trap-and-haul operation would accommodate the variable pool but would be prohibitively expensive for an original project of this size.
The most cost-effective solution for the Applegate project was an in-conduit Eicher screen that would divert downstream migrants around the turbines. However, this high-head dam requires an innovative approach to manage bypass flow velocities and fish decompression rates to minimize barotrauma. Mead & Hunt developed the decompression raceway concept for AG Hydro using computational fluid dynamics (CFD) modeling to refine the configuration and fish testing to identify protective decompression rates. This new solution to high-head downstream fish passage could be applied across a wide range of site conditions and target species.
Decompression raceway concept
The essential processes and components of the decompression raceway design (e.g. gradual and uniform velocity changes and in-conduit diversion screens) are adapted from tested and commonly-used passage methods and criteria. However, by recombining these elements and adding a controlled decompression process, the authors have created a system that can be applied at high-head hydro facilities to provide fish passage at a lower cost than traditional methods. (Configurations for the reservoir intake, turbine water diversion, river return, counting station, and other appurtenant facilities vary depending on site conditions and project goals and will not be discussed in this article.)
Figure 1 on page 22 shows the main components of the decompression raceway. Its operation can be summarized in the four-step sequence for an installation with two or more decompression raceways arranged for continuous passage:
- Filling Period: A backwash spray system cleans the screens, then the raceway is filled with pressurized water from the penstock. This water fills the standpipe that connects the raceway to atmospheric air and maintains the pressure within the raceway (10 minutes +/-);
- Fishing Period: Diversion water containing fish is routed into the raceway. Diversion flow is diffused through wedge-wire screen sections on the upper half of the raceway. Flow velocity directs fish toward a holding area behind a bar rack sized for predator screening. A valve behind a flat wedge-wire screen maintains a minimum attraction velocity in the holding area (15+ minutes);
- Decompression Period: After a preset time interval, diversion flows are re-routed to another raceway, and flow through the first raceway stops. The raceway is decompressed to atmospheric pressure by gradually draining the standpipe (2 minutes +/-); and
- Draining Period: Once atmospheric pressure is reached, a fast-opening valve drains the water and fish into a tank or channel for transport to an evaluation station or river return (2 minutes +/-).
In a typical downstream passage application, fish are transported from the reservoir to the decompression raceway (near the downstream toe of the dam) through a pipe with an in-conduit diversion screen. As the fish move down in elevation, the water pressure increases and fish lose neutral buoyancy as the air inside the swim bladder compresses. (If fish have access to an air pocket while under pressure, they may gulp air to regain neutral buoyancy.) At the end of the pipe, the fish flow into the decompression raceway and gradually decelerate as the water is diffused through wedge-wire screens. When the fish reach the holding area, water velocity is less than 1 foot per second (fps).
A wedge-wire screen at the back of the holding area prevents fish from escaping, while a flow-through velocity of about 0.4 fps is maintained. Fish accumulate in the holding area for a set time, at which point a valve seals the inlet and the inflow is diverted to a second decompression raceway. The raceway is isolated from the pressurized inflow pipe and the pressure is maintained with a water-filled standpipe connected to the raceway behind a screen. As soon as inflow stops, the standpipe is drained to gradually return the contents of the decompression raceway to atmospheric pressure. The drainage rate is controlled with a valve to limit the rate of decompression so the fish have time to expel excess air from the swim bladder via their pneumatic ducts. A flush valve is then opened to discharge the fish and remaining water into a pipe for return to the river. Lastly, a water spray backwash system is used to clean the screens and fill the decompression raceway and standpipe for the next cycle.
The objectives for designing a decompression raceway should include:
- Returning fish from a pressurized environment to atmospheric pressure at a controlled rate, minimizing barotrauma;
- Controlling water flow through the raceway to provide uniform deceleration, prevent screen impingement, and minimize flow separation and eddy formation;
- Sizing holding areas to allow ample swimming space and oxygen supply;
- Providing appropriate lighting or surveillance consistent with project-specific goals;
- Providing debris management systems as needed for the site; and
- Maintaining a relatively quick (less than 30 minutes) cycle time to minimize stress and delay.
The decompression raceway system returns fish to atmospheric pressure by gradually draining a standpipe, but a hydro-pneumatic pressure tank could be used. Uniform deceleration is achieved through the use of wedge-wire screens and discharge valves located at the top of the decompression raceway. Each screen diverts a portion of the flow until velocity is reduced to about 0.4 fps. According to NOAA's National Marine Fisheries Service (NMFS, now NOAA Fisheries), the velocity component that is perpendicular to the wedge-wire diffuser screen face at any point should be less than 0.4 fps to prevent fish impingement.1 Flow separation and eddy patterns can be minimized by incorporating gradual expansions and contractions where needed. The holding area size is determined by the peak number of juveniles to be held during any one fishing period. Oxygen supply is not typically a critical issue, as the flow-through velocity ensures a fresh water supply during the fishing period, and the decompression and draining periods will not typically exceed 10 minutes total.
Sight glass and/or video surveillance equipment can be included in the design to allow for monitoring and admit ambient light, which is often important to allow migrating juvenile salmonids to quickly adjust to downstream river conditions upon release, to avoid predators. Bar racks or other barriers can be installed in conjunction with flap gates or multiple outlets to exclude larger predators and debris from the holding area. It is possible to achieve cycle times of about 30 minutes because decompression, draining and filling can be completed quickly. The corresponding total travel time of fish from upstream reservoir to the river would be 10 to 30 minutes, depending on when fish enter the raceway and the configuration of the intake and river return facilities.
The decompression raceway design achieves substantial cost savings by:
- Eliminating the need for large concrete structures (fish locks or ladders);
- Minimizing the footprint area and associated acquisition/easement costs required for high-head passage;
- Reducing operating costs through automation while ensuring safe passage under a range of operating conditions;
- Using commercially-available products to maximize supplier competition and reduce capital costs;
- Keeping maintenance low with the addition of a dual-purpose backflush cycle using a pressurized water spray system; and
- Allowing periodic collection (single unit) or continuous flow-through (two or more units) to achieve site-specific goals.
CFD model tests
Alden Research Laboratory performed CFD model testing to verify the hydraulic performance of the decompression raceways for the Applegate project. Feedback from early model results was used to guide the design process and resulted in increased confidence in the final design. The objectives of the CFD modeling efforts were:
- Configure the decompression raceways for uniform deceleration;
- Configure the raceways to prevent flow separation or eddy formation; and
- Characterize screen approach velocities to evaluate compliance with NOAA Fisheries fish passage criteria developed to minimize the potential for fish impingement against screens.
The modeling objectives are interrelated. Gradual, uniform deceleration prevents flow separation and eddies without introducing guide vanes or flow splitters, which could create a strike hazard for fish entrained in the flowing water.
Flow patterns were modeled using FLUENT software by ANSYS, a three-dimensional CFD model. The fluid solvers solve the Reynolds Averaged Navier-Stokes equations and use a body-fitted computational grid. Model simulations were conducted under steady state conditions with an inflow rate of 31.4 cubic feet per second (corresponding to a maximum velocity of 10 fps in the upstream 24-inch-diameter pipe) and outflows of 25 cfs through the diffuser screens and 6.4 cfs through the holding area outlet. The CFD model domain included a 60-degree horizontal pipe bend immediately upstream of the decompression raceway, which represented actual site conditions for Applegate.
Figure 2 is a color contour profile of velocity magnitude along the central axis. Figure 3 on page 28 is an isometric streamline plot with zero-mass particles color-coded for velocity. All figures represent an inflow rate of 31.4 cfs, with 25 cfs passing through the diffuser and 6.4 cfs through the holding area screens. CFD model results for the final design show relatively uniform deceleration through the raceway, averaging 0.1 fps/foot. Note the regular spacing of the velocity color contours in Figure 2 and flow-through velocity in the holding area of less than 0.4 fps. Results from earlier design stages indicated flow separation at abrupt pipe expansions, so sidewall expansion angles of 7 degrees were used for final design.
The NOAA Fisheries criterion for juvenile fish screen approach velocity is 0.4 fps or less for screens with active debris removal provisions. NOAA Fisheries defines approach velocity as the vector component of velocity that is perpendicular to and upstream of the screen, measured as close as possible to the boundary layer turbulence generated by the screen face, and computed by dividing the maximum screened flow by the effective screen area. The effective screen area is defined as the total submerged screen area, excluding major structural members, but including the screen face material, projected onto a plane.1
Assuming a uniform flow distribution through the diffuser screens, the approach velocity is about 0.2 fps for the 31.4 cfs maximum design flow. However, CFD modeling revealed some areas where flow was not uniform. Diffuser screen approach velocities were approximated by the vertical velocity component at the screen face and ranged from -1 fps (reverse flow) to 1 fps, with higher velocities generally observed near the entrances to the 24-inch pipe risers on top of the raceway. The vertical component of velocity was within the recommended maximum for most of the screen area. Reverse flow occurred primarily in the first two of eight screen panels and may have been caused by non-uniform entrance conditions resulting from the 60-degree horizontal bend immediately upstream. Design modifications that may help create more uniform flow conditions at the diffuser screens include providing a 20-foot-long straight pipe section upstream of the raceway and/or increasing the number of diffuser screen panels.
Overall, the hydraulic characteristics of the diffuser screens should prevent fish impingement because the high sweeping velocity will move fish quickly to the holding area.
The results of CFD model tests verified the expected hydraulic performance of the decompression raceways, and the computed velocities demonstrated a uniform deceleration with no flow separation and minimal potential for impingement of fish against screens.
Hyperbaric pressure tests
Barotrauma injuries can occur when fish are rapidly decompressed, with injuries increasing as the ratio of pressure change (acclimation pressure/nadir pressure) increases.2 (Nadir pressure is the lowest pressure fish are exposed to.) Common injuries associated with barotrauma include swim bladder ruptures, exophthalmia (eye-pop), emboli in the gills and fins, and internal hemorrhaging. However, salmonids regulate gas inside the swim bladder through the pneumatic duct. This duct is connected to the esophagus, allowing gas to be gulped or expelled through the mouth. We hypothesized that fish decompressed from high pressures to surface pressure would not suffer from barotrauma injuries if they were decompressed at a rate that allowed the fish to expel gas from the swim bladder.
Juvenile salmonids migrating downstream at the Applegate project will undergo rapid pressurization as they travel through the intake tower and outlet conduit and then through a 24-inch-diameter bypass pipe to the decompression raceways. They will be held at high pressure (up to 90 pounds per square inch absolute) for 0 to 30 minutes (depending on when they enter the raceway), then be gradually decompressed to atmospheric pressure (14.7 psia) before release into a gravity-flow pipe to return to the river.
Dr. Richard Brown of Pacific Northwest National Laboratory, operated by Battelle, led hyper/hypobaric pressure tests to identify the time required to decompress juvenile salmonids from 90 psia to surface pressure (a ratio of pressure change of 6.1) without barotrauma injury,3 assuming the salmonids could become neutrally buoyant at these high pressures. Hatchery-reared juvenile chinook salmon and juvenile steelhead were used for testing. Two scenarios were tested to investigate the relationship between initial acclimation depth and injury rates. Table 1 on page 32 provides data on the test specimens.
The first scenario (see Figure 4) simulated the worst-case scenario, which entails fish acclimated to 74 psia for 16 to 24 hours. Scenario 1 simulated fish that remain under high pressure in the bypass system for an extended amount of time and were allowed to become acclimated (i.e. allowed to become neutrally buoyant by providing an air bubble in the chamber for pressure regulation). Following acclimation, these fish were decompressed from 74 psia to 12.1 psia over 1.7 minutes.
In the second scenario (see Figure 5), fish were acclimated to a pressure of 19.9 psia for 16 to 24 hours. This period simulates fish acclimated to 22 feet of water depth (24.2 psia) before entering the bypass system. As in the first scenario, fish were decompressed from 74 psia to 12.1 psia over 1.7 minutes.
Damage occurs when the gas inside the fish expands, based on the ratio of acclimation pressure to exposure pressure.4 In Scenario 1, a fish would encounter a ratio of pressure change of 6.1 (90/14.7). In Scenario 2, fish would enter the bypass system at a pressure of 24.2 psia. After pressurizing to 90 psia for 6.7 minutes, fish would be reduced to 14.7 psia. The ratio of acclimation pressure to exposure in this scenario would be 1.6 (24.2/14.7). These ratios of pressure change represent the amount that the volume of the swim bladder will increase upon decompression. For example, a ratio of pressure change of 6.1 will lead to the swim bladder volume becoming 6.1 times larger upon decompression than it was at the acclimation depth.
The depth of acclimation before exposure to rapid decompression is an important variable; the deeper the acclimation, the greater the chance of injuries or mortality.2 Therefore, in Scenario 2, fish were acclimated to a pressure-equivalent depth that is likely the deepest at which most of the fish will be able to achieve neutral buoyancy. The freshwater depth of 22 feet is the median value for the maximum depth at which juvenile chinook salmon can attain neutral buoyancy.5
In total, 70 chinook salmon and 132 steelhead were tested with the Scenario 1 pressure simulation. None of the chinook salmon were injured. After 60 steelhead were tested, one was found to have a ruptured swim bladder. The scenario was altered by increasing the decompression time from 1.7 to 3 minutes to allow the fish more time to expel gas from the swim bladder. With the increased decompression time, 72 additional steelhead were tested and none exhibited injury.
A total of 70 chinook salmon and 72 steelhead were tested with the Scenario 2 pressure simulation. No fish were injured.
The low occurrence of injury (one fish in Scenario 1) was likely due to the fact that pressures were decreased slowly enough that fish could expel gas from the swim bladder as the volume expanded. Video indicated the injured fish was the only one that did not expel gas as decompression occurred. The video also shows that when the swim bladder ruptured, gas was expelled through the vent and mouth.
During this research, fish were decompressed from 74 psia to 12.1 psia, which multiplied the volume of air inside the swim bladder by 6.1 times. Steelhead may have been more susceptible to injury because they were able to become neutrally buoyant at high pressures (74 psia, the pressure present at 47 feet of freshwater) while none of the chinook salmon could. Thus, the steelhead had a higher relative volume of air in their swim bladders before decompression. The reason for the differences between the two species is unknown. However, it is unlikely that either species would be expected to be neutrally buoyant at these pressures under natural conditions because the mean maximum depth for juvenile chinook salmon is 22 feet.5 Nevertheless, even under the worst-case scenario tested, there was no injury to fish if decompression occurred over 3 minutes, which was slow enough to enable the fish to expel gas from their swim bladders.
Providing safe downstream fish passage over high-head dams is difficult. Traditional solutions for transporting bypassed fish can be expensive and can create considerable delay and stressful conditions for migrants. Reservoir elevation fluctuations make structural approaches challenging to design. The decompression raceway developed for the Applegate project represents an innovative approach that is more cost-effective than traditional downstream passage methods at high-head dams.
CFD model tests were used to verify the expected hydraulic performance of the decompression raceways, and the computed velocities demonstrate a uniform deceleration with no flow separation and minimal potential for impingement of fish against screens. Pressure tests demonstrated minimal potential for injury (no injuries for 142 fish tested) related to barotrauma for the simulated conditions based on application at the Applegate project.
The authors thank Larry Swenson with NOAA Fisheries for his keen interest and helpful collaboration in adapting the application of the design concept for the Applegate project to meet the Northwest Regional Office's performance goals.
The decompression raceway system described is patent-pending, and all associated rights have been assigned to Mead & Hunt; the company intends that this system be made available for the benefit of fisheries everywhere.
1Anadromous Salmonid Passage Facility Design, National Marine Fisheries Service, Northwest Region, Portland, Ore., 2008.
2Brown, R.S., et al, "Quantifying Mortal Injury of Juvenile Chinook Salmon Exposed to Simulated Hydro-turbine Passage," Transactions of the American Fisheries Society, Volume 141, No. 1, 2012, pages 147-157, doi: 10.1080/00028487.2011.650274.
3Brown, R.S., and B.D. Pflugrath, "Applegate Dam Pressure Investigations: Final Report," prepared for Symbiotics, Portland, Ore., 2011.
4Brown, R.S., et al, " Assessment of Barotrauma from Rapid Decompression of Depth-acclimated Juvenile Chinook Salmon Bearing Radiotelemetry Transmitters," Transactions of the American Fisheries Society, Volume 138, 2009, pages 1285-1301, doi: 10.1577/T08-122.1.
5Pflugrath, B.D., R.S. Brown, and T.J. Carlson, "Maximum Acclimation Depth of Juvenile Chinook Salmon: Implications for Survival during Hydroturbine Passage," Transactions of the American Fisheries Society, Volume 141, No. 2, 2012, pages 520-525, doi: 10.1080/00028487.2012.670187.
Ryan Greif is a water resources engineer with Mead & Hunt. Kai Steimle, formerly with project developer AG Hydro, is now an aquatic ecologist with Newhalen Associates LLC. Richard Brown is a senior research scientist with Pacific Northwest National Laboratory. Dan Gessler is vice president director of Colorado operations at Alden Research Laboratory.