Spillway inadequacy remediation in the US6 November 2009
Hasan T Kocahan, Russell Wyckoff and Martin Wosnik describe the US’ experience with the Fusegate spillway control system to bring dams with undersized spillways up to current safety standards
With more than fifty applications around the world ranging from 0.9m to 9m in height, the Fusegate System has gained significant international recognition over the past few years as a reliable spillway control system. The system is typically used to increase spillway discharge capacity and/or reservoir storage capacity. Upon successful performance at Terminus dam in California, the US Army Corps of Engineers (USACE) has been considering the system for more projects. The Otter Brook and Canton dams are other examples of the ways this system can be applied to bring dams with undersized spillways to up-to-date safety standards.
The Fusegate System
The Fusegate System is a non-mechanical spillway control technology to increase reservoir storage capacity and/or to increase spillway discharge capacity.
Fusegates are free standing blocks placed side by side on a spillway sill to form a watertight barrier. The Fusegate System operates as a straight or labyrinth crested weir divided into segments or individual ‘Fusegates’, each forming one component of the weir. Each gate consists of three components; a bucket made of metal or reinforced concrete, a base, and an intake well that is connected to a chamber in the base. Each Fusegate sits on the modified spillway sill and remains in place only by gravity. Toe abutments (lugs) anchored into the sill prevent them from sliding in the downstream direction. Accumulation of seepage water in the bottom chamber is prevented by providing each chamber with two drains. The joint between adjacent Fusegates is sealed with a flat rubber gasket.
Fusegates can increase both spillway capacity and reservoir storage. For a retrofit on an existing spillway, a portion of the ogee crest is removed and provided with a flat surface. If the goal of the retrofit is only to increase spillway capacity, the crest of the Fusegates is set near the original ogee crest elevation. If the purpose is to increase storage, then the crest of the Fusegates is set above the original ogee crest elevation. For discharges up to the design flood, the Fusegate functions like an aerated labyrinth weir. Typically, the design flow is chosen to be the discharged with a return period of about 100 years or above. For discharges greater than the design flow, water begins to flow through the well and into the chamber located in the base of the gate.
When the overtopping flow reaches the selected design pool elevation, the designated Fusegate(s) will tip downstream, thus increasing the discharge capacity of the spillway. The tipping is initiated by water entering into the base chamber of the Fusegate via an intake well set at tipping design elevation. Once water enters into the base chamber, the uplift pressure rapidly increases, causing the Fusegate to rotate about the toe abutments and tumble downstream. By modifying the ballast provided for each Fusegate and the elevation of the intake well, each Fusegate can be designed to tip at a different (pre-determined) pool elevation so that an excessive flow increase does not occur downstream. Small drain holes exist at the base of the Fusegates to help drain any leakage through the seals so that uplift pressures do not occur until water elevation reaches intake wells.
The crest of the well in each Fusegate is set at a different elevation so the gates do not tilt in unison. In this manner, the increase in discharge over the spillway, as a function of reservoir elevation, can be precisely controlled. Figures 1 and 2 show the three dimensional views of a labyrinth crested Fusegate.
Constructing a new auxiliary spillway controlled by Fusegates is also another way of conducting dam safety upgrades in the case of severe spillway discharge inadequacies, where the service spillway modification is not feasible. In this scenario, the services spillway – usually equipped with mechanical gates – would discharge more frequent floods and the emergency spillway will only operate during extreme flood events.
Otter Brook application
Otter Brook Dam, located near Keene, New Hampshire, was constructed in 1958 as a flood control structure on a tributary of the Connecticut River. It is owned and operated by the USACE, New England District. The dam consists of a 40.5m high and 392.6m long rockfill embankment impounding 22.6Mm3 of water to control a basin area of 121.7km2. A pool of 6.1m depth is maintained behind the dam for recreational purposes. The remainder of the storage volume is available for floodwater containment.
The embankment constitutes a flow restriction, with resultant outlet capacity of up to 37m3/sec with the pool at the spillway crest. The water is dammed from a channel invert elevation of el. 208.2m, up to the spillway crest elevation of el. 238m and up to a top of dam elevation of el. 244.5m.
There was a 44.2m wide concrete ogee spillway with a gently sloped (1% grade) 91.4m long unlined approach channel and a more steeply sloping unlined spillway exit channel (8.2% grade, reducing after 182.9m to 4.7% grade). The discharge capacity of this spillway was 963m3/sec.
Flood routing simulations based on the updated storm data have revealed that the dam would be overtopped by 0.30 m during a PMF event as the new spillway discharge capacity requirement was to pass 1642m3/sec.
The shortlist of options considered
Initially, options to raise the dam, widen the spillway, lower the spillway, store floodwaters, remove the dam, and construct an extra spillway were reviewed. Six basic options were then selected for more detailed study:
• 1. Lower the spillway sill by 2.7m and recover the storage with the installation of 2.7m high Fusegates.
• 2. Raise the dam 0.4m and widen the spillway 15.2m.
• 3. Raise the dam 1m and widen the spillway 7m.
• 4. Add a 20.4m long new auxiliary spillway controlled either by a fuse plug or a mechanical gate.
• 5. Remove the existing spillway weir and excavate the spillway channel down 0.9m to obtain a new crest ‘weir’ capable of passing the PMF.
• 6. Widen the spillway the full 23.8m required to maintain an acceptable freeboard without raising the dam. Options 5 and 6 increased downstream flows for significant storms smaller than the PMF, and so some investigation of design and historical storms needed to be performed.
Proposed Fusegate design
The option to lower the spillway sill and install Fusegates to restore the normal pool elevation was selected as the most viable, since it limited the construction to the existing spillway weir structure and approach channel, requiring wetland mitigation in the upstream wetland only. Under this scenario, the existing concrete spillway was removed and its foundation was excavated down to el. 235.3m. A new broad crested concrete weir was then constructed with the same top elevation as the existing spillway, which formed the base for Fusegates. This ensures that the existing capacity of the reservoir and flood attenuation behaviour is not compromised. The Fusegates will topple over as the upstream water level in the lake is increased due to the extreme storm, thus increasing the capacity of the spillway channel. Figure 4 shows the three dimensional view of a straight crested Fusegate, which offered the most suitable type of Fusegate for the Otter Brook application.
The choice to excavate the spillway approach channel in order to enhance the spillway hydraulic performance, rather than raise the dam, saved the construction costs for a wall on the upstream side of the dam and avoided the need to raise the threshold of the control structure. Excavation on the environmentally sensitive east bank of the downstream channel is avoided through the use of an 80.8m long dike, which is needed to contain the floodwaters within the spillway discharge channel.
The Otter Brook application includes the installation of six 2.7m high 7.3m wide straight crested Fusegates on a newly constructed 44.2m long and 3.4m wide concrete spillway platform. The top of the platform was set at el. 235.3m on the 36.5m long portion of its right bank and at el. 236.5m on the remaining 7.3m long section on the left bank. The higher section of the platform was designed to accommodate the last Fusegate to tip for the full PMF, thereby protecting the wetland located downstream from frequent overspilling.
The approach channel was excavated in such a way that it sloped down to el. 235m at the spillway sill location with a 1% gradient. No major excavations were required downstream of the new sill. Abutment walls 0.18m wide were to be constructed on the left and right banks of the spillway to support the watertight seals that were installed between the Fusegates and the spillway abutments. A 0.4m thick pier was also constructed between the lower and elevated sections of the platform to install the watertightness seals on two adjacent Fusegates.
The Fusegates at Otter Brook are designed to tip for extremely low probability events. The first Fusegate will tip for a flood corresponding to 55% of the full PMF. Considering that the first and only time water has flown through the spillway was in 1987 with the water level 0.5m above the spillway sill, the magnitude of such a flow, which creates a 4.7m high water elevation above the sill, can be better understood.
The construction works at Otter Brook Dam began in May 2005. The works consisted of removing the existing ogee weir and spillway sill from the spillway channel; excavating the spillway channel floor upstream and downstream of the existing weir; placing a new concrete sill, a concrete pier and concrete abutment walls on both sides of the spillway channel; and installing six Fusegates on the new spillway sill. Associated with this work was also reconstruction of a wetland upstream of the Fusegates and construction of a dike on the left side of the spillway channel.
The Fusegates can be manufactured or constructed either in steel or concrete. Otter Brook Fusegates were made of pre-cast concrete by Old Castle Precast of Auburn, Maine, with stainless steel inlet wells. This selected material came out the most economical for Otter Brook with added advantages in terms of long life expectancy and extremely low maintenance costs. The Fusegates weighed 40 tons each at the time of delivery from the plant before their concrete ballasts and inlet wells were attached.
The general contractor for the construction works was George R. Cairns and Sons, Inc. from Windham, New Hampshire. The construction works were delayed for about three months due to 45.7cm of rainfall in nine consecutive days during the month of October 2005. The rain caused the swelling of every local brook, stream and river, some beyond flood storage, and many roads were washed away and houses flooded.
After cleaning up works following the October flood, the concreting of the new spillway sill began. In the meantime, the pre-cast Fusegates had been completed and were ready to be shipped to the dam site.
Before the installation of Fusegates, all concreting works related to the new spillway sill, toe abutments and the pier separating the elevated and lower section of the sill were completed except for the side abutment walls.
Once the Fusegates were delivered to the dam site, they were directly placed in their final position. Figure 6 shows the placement of Fusegates and the inlet wells.
Figure 7 shows an upstream view of Fusegates after the installation works were completed. The final stage was the excavation works and dyke construction at the left abutment.
Canton Dam application
Canton Dam is located on the North Canadian River approximately 125km northwest of Oklahoma City, Oklahoma. It is composed of an earth filled embankment, 4614.7m in length with a maximum height of 20.7m above the streambed. The top of the dam is at el. 502.3m.
The dam includes a 237.1m wide service spillway at the right abutment with 16 Tainter gates. Each gate is 12.2m wide and 7.6m high. The spillway crest elevation is at el. 491.6m.
Hydraulic studies have highlighted that the existing spillway is unable to discharge the new Probable Maximum Flood and that the dam is likely to fail during major flood events. The discharge capacity of the existing service spillway is 9600m3/sec and the new PMF studies revealed a total spillway discharge capacity requirement of 17358m3/sec.
Eight options have been studied to remedy the hydrologic deficiency:
• Raise the dam by 2.10 m (maintain the freeboard).
• Raise the dam by 0.4 (eliminate freeboard).
• Add an uncontrolled spillway.
• Additional fuse plug auxiliary spillway.
• Additional Fusegated auxiliary spillway.
• Additional gated auxiliary spillway.
• Reduction of flood control storage.
It was found that constructing an auxiliary spillway would offer the most feasible option.
The proposed plan involves the construction of an auxiliary spillway on the right abutment containing nine Fusegates at a sill elevation of el. 490.7m. The excavated material from constructing the auxiliary spillway would be spoiled below the left embankment to address the seepage and seismic issues. The rotation of the Fusegates would lower the spillway crest elevation, increasing the spillway capacity by 7762m3/sec to sufficiently discharge the PMF without overtopping the dam.
This alternative was selected as it offered a non-mechanical spillway control system with more economical feasibility on the construction and O&M costs. The projected path of the auxiliary spillway is shown in Figure 11.
Proposed Fusegate design
For Canton Dam, a total of nine Fusegates are required, each 9.75m high and 16.3m wide. The nine Fusegates will fit into the 146.3m wide auxiliary spillway channel.
The Fusegates will be accommodated on a broad-crested weir having an upstream to downstream width of 7.2m at the control section of the emergency spillway, which is at el. 489.5m. With 9.75m high Fusegates, the normal pool elevation (Fusegates crest) will be set at el. 499.2m. The approach channel will be maintained at el.488.7m and there will be a 0.5% slope in the return channel downstream of the Fusegates.
Figure 12 shows the cross section of the broad crested weir that will accommodate the Fusegates. The Fusegates will be retrofitted on the flat section of the sill denoted as L.
The Fusegates are semi-labyrinth crested and will be constructed in reinforced concrete. The inlet wells will be made of stainless steel. The watertightness will be maintained by installing EPDM seals between the Fusegates as well as between the Fusegates and the spillway sill. The photo of an example of the semi-labyrinth Fusegates is shown in Figure 13.
Intake wells are normally constructed as part of the Fusegate. Due to the shallow approach channel and wave action, the intake wells at Canton are gathered within a protective enclosure (called wet well tower) equipped with a bottom port and conduit to allow flow to enter.
Pipes embedded in the concrete sill connect individually the intake wells to the base chamber of the Fusegates. Such configuration removes the risk of premature tipping because of surface waves. A similar arrangement was used in the scope of Terminus Dam project in California.
There will be five tipping sequences for the Fusegates, where the tipping elevations range for a water level between el. 500m and el.500.4m (Maximum Water Level).
Figure 14 shows the flood routing hydrograph of the PMF through the Fusegated emergency spillway. It should be noted that the maximum flow to be simulated is 17840m3/sec.
The Fusegates are designed to tip consecutively for extremely large flood events. At Canton dam, there will be no spilling over the Fusegates crest before a flood that is in excess of 50% of the PMF. Moreover, the first Fusegate will tip for a flood corresponding to over 58% of the PMF. This probability corresponds to an inflow of 10871m3/sec.
Combined numerical and physical model study of the auxiliary spillway
USACE has decided to perform a model study of the proposed auxiliary spillway in order to optimize the structures and to ensure that the selected configuration will pass the PMF. The model study was performed by Alden Research Laboratory in Holden, Massachusetts. The physical model was constructed as a fixed-bed, undistorted model with a geometric scale ratio of 1:54, which was deemed sufficient to reduce scale effects. The PMF flow rate of 17840m3/sec prototype scales to approximately 0.9m3/sec.
The approach geometry for both spillways was first optimized by using Computational Fluid Dynamics (CFD) in order to find an approach channel configuration that can meet the PMF discharge requirements. Advantages of this combined CFD/physical model approach are that additional model runs can be made at comparatively lower costs than in the physical model, and many iterations can be conducted until a satisfactory configuration has been achieved. The selected final approach channel design would then be constructed and validated in the physical model.
The model boundaries were selected to represent the reservoir over a distance of 457m upstream of the auxiliary spillway entrance, or approximately 762m upstream of the service spillway, and 305m along the dam to the left of the service spillway. This upstream distance is required to accurately model approach flow patterns to the two spillways, in particular flow around and over the peninsula-shape outcropping (max el. 499.6m) immediately to the left of the service spillway. The downstream model boundaries were selected to represent the flood plain over a distance of 457m downstream of the auxiliary spillway exit or approximately 914m downstream of the service spillway. As a result, the physical model dimensions were 33.5m x 21.3m. The model Fusegates were cast from a resin, ‘densified’ with tin powder to obtain a conservative target density of 2.55g/cc for reinforced concrete in order to accurately simulate their geometry, mass and mass distribution and their resistance to motion.
Figure 16 shows the construction of the physical model with the existing spillway in the forefront of the photo. Figures 17 and 18 show the production of the Fusegates that will be used for testing.
The physical model was developed to ensure that (i) the PMF is safely discharged at maximum pool over both spillways and (ii) the Fusegates after tipping are properly evacuated from the return channel. It has also been instrumental to derive the basic data used to assess the stability of the Fusegates and the water feeding conditions of the wet well and of the inlet wells.
A civil contractor was selected by USACE in September 2008 to carry out the excavation of the auxiliary spillway, and detailed design of the project began in July 2009, with completion scheduled for February 2010. The concreting works, which includes a new highway bridge, the broad crested weir, the apron and the side walls will start upon completion of the excavation works in early 2011. The overall completion of the project is planned for early 2012.
For dam safety rehabilitation projects, the Fusegate System offers a wide variety of reliable and cost effective solutions. Thanks to the versatility of the system, it can be installed on service or emergency spillways, and it can also be combined with most other spillway systems either on the same or separate spillways.
Constructing an auxiliary spillway often offers an attractive solution to severe spillway discharge inadequacy problems as long as a suitable location is available. The effectiveness of such alternatives is in many cases enhanced by the use of a fusible spillway control system designed to fail automatically in case of exceptional floods in order to release some of the flood water.
The Fusegates will minimize the size of a spillway as compared to other non-mechanical spillway control systems such as ogee weirs or fixed labyrinth weirs. They will offer the same benefit of reliability and low cost of operation and maintenance as they are designed to activate only for extremely low probability flood events usually over half the PMF.
The authors are: Hasan T. Kocahan, Senior Project Engineer, hydroplus Inc, US; Russell Wycoff, Hydraulic Engineer, US Army Corps of Engineers, Tulsa District, US; and Martin Wosnik, Senior Flow Engineer, Alden Research Laboratory Inc, US
This article is based on a presentation given at the 23rd international-commission-on-large-dams (icold) Congress, held in Brazil in May 2009. For information on future ICOLD events, please visit www.icold-cigb.net