Close 

Dam Safety: Adding a Spillway Protects Integrity of Jackson Bluff Dam

With the fuse plug at its 12-MW Jackson Bluff project being similar in design to the one that failed at Silver Lake, the city of Tallahassee sought an innovative way to provide adequate flood control capacity while maintaining the integrity of nearby ecosystems and protecting people and property downstream.

By Rob McGarrah, Yiying Xiong and Kim Hansen

A potential failure mode analysis session held at the 12-MW Jackson Bluff Hydroelectric Project in Tallahassee, Fla., revealed that soil conditions at the fuse plug area were similar to that at the Silver Lake project. The session was held in July 2003, and the unexpected failure of the Silver Lake fuse plug had occurred just two months earlier. The Silver Lake fuse plug eroded significantly deeper than its design due to the unfavorable foundation material properties. This raised concerns about the functionality of fuse plug structures founded on sandy soils.

Geotechnical investigations at the fuse plug area at Jackson Bluff, performed by Ardaman & Associates of Orlando, Fla., confirmed that the condition of the foundation materials was not favorable for controlled fuse plug release. An inflow design flood study was performed in 2004 and investigators concluded that the hydraulic capacity provided by the fuse plug was necessary to meet the IDF requirements; therefore, abandoning the fuse plug was not a viable option. The city of Tallahassee, owner of the Jackson Bluff project, evaluated several alternatives to armor the existing fuse plug area and create a non- erodible sill to prevent extensive scouring.

However, evaluations showed that these alternatives could not satisfy the hydraulic requirements and would cause significant environmental disturbances downstream. The design eventually selected incorporates a conventional concrete, ogee-shaped crest spillway and downstream slope armoring located next to the existing gated spillway with a standard Bureau of Reclamation Type III stilling basin and sheet pile cutoff walls to enclose the spillway and stilling basin. The sheet pile cutoff walls act as seepage and scouring cutoffs to reduce uplift under the spillway slab and undercutting at the toe of the stilling basin. The existing fuse plug and emergency spillways were replaced, with a remote dike at the same crest level as the main embankment. Project construction commenced in September 2010 and was completed in August 2011.

Background

The Jackson Bluff project is one of only two hydroelectric facilities in Florida. It is located on the Ochlocknee River, about 66 miles upstream from its mouth at the Gulf of Mexico and about 20 miles west-southwest of Tallahassee. The reservoir upstream of the dam, Lake Talquin, has a surface area of about 10,200 acres and a gross volume of about 150,000 acre-feet at the normal water surface elevation of 68.5 feet mean sea level. Lake Talquin is relatively shallow, long and narrow, with a total length of more than 10 miles. Normal head at the dam is about 40 feet. Normal operation of the facility is to maintain the lake level at 68.5 feet with 1-foot bandwidth in either direction. The Jackson Bluff project is a run-of-river facility.

Main structures of the dam consist of a powerhouse with three generating units; a 199-foot-long gated spillway; a 3,600-foot-long earth embankment with a crest elevation of 77 feet; a 950-foot-long emergency spillway with a crest elevation of 72.3 feet; and a 562-foot-long fuse plug spillway with a crest elevation of 74.3 feet. The emergency spillway is a broad-crested grass-lined weir. The fuse plug spillway was designed to deploy at overtopping down to its concrete sill elevation of 68.3 feet.

The total project discharge capacity at the maximum pool level of 77 feet is 148,000 cubic feet per second, which is equal to half of the probable maximum flood of the project. The accepted IDF of the project is half of the PMF.

Downstream conditions

Downstream from Jackson Bluff Dam, the Ochlocknee River travels through four counties before entering into the Gulf of Mexico. A camping area, a mobile home park, and a road with 45 homes are about a half-mile downstream from the dam. These residential developments are on relatively low ground. A flood stage elevation was established for this area for flood warning and evacuation purposes.

More developments exist further downstream along the river. Nearly 200 residential structures are along both sides of the river within 30 miles downstream of the dam.

This view of the Jackson Bluff Dam shows the project's powerhouse (right), gated spillway and embankment. Jackson Bluff, owned by the city of Tallahasse, Fla., is one of just two hydroelectric projects in Florida.
This view of the Jackson Bluff Dam shows the project's powerhouse (right), gated spillway and embankment. Jackson Bluff, owned by the city of Tallahasse, Fla., is one of just two hydroelectric projects in Florida.

PFMA session

In June 2003, a PFMA session was held as part of the Part 12 dam safety inspection at Jackson Bluff Dam. Two months before the session, a dam failure event occurred at Silver Lake Dam in Michigan. This event resulted in larger-than-expected erosion of the fuse plug spillway, which was founded on sandy soil. In light of the Silver Lake event, several potential failure modes of the Jackson Bluff Dam fuse plug and emergency spillway, which were founded on similar sandy materials, were discussed at the PFMA session. The discussion identified several scenarios, including a deep breach of the fuse plug and erosion and breaching of the emergency spillway prior to fuse plug activation. The emergency spillway crest was 2 feet lower than the fuse plug and could be overtopped at a flow slightly above the 100-year flood level. Overtopping of the emergency spillway could cause erosion of the soil material and even failure of the spillway. Participants in the PFMA session recommended that an updated IDF study be performed to determine the hydraulic requirements of the project and if the fuse plug and emergency spillway could be abandoned.

IDF study

Although the likelihood of fuse plug activation was determined to be small, the potential consequence of a malfunction could be significant because of the residential and commercial developments within a half-mile downstream from the dam. Shortly after the PFMA session, the city of Tallahassee requested that Mead & Hunt Inc. of Rancho Cucamonga, Calif., perform an updated IDF study. An IDF study is a dam failure analysis that determines the requirement of the total project capacity for a Federal Energy Regulatory Commission-regulated dam. This analysis simulates hypothetical dam failure scenarios and determines the resulting downstream flooding impacts. The original IDF study for Jackson Bluff was conducted by R.W. Beck in 1983 and determined the IDF for the project to be one-half of the PMF.

Mead & Hunt developed a hydraulic model using the BOSS DAMBRK one-dimensional dynamic flood-routing program, developed by BOSS International. This program computes an outflow hydrograph from the dam (with or without a dam failure) and uses the dynamic wave method to route the outflow hydrograph downstream. The model determines water surface elevations, flow rates, time to peak flow, and flow and stage hydrographs at different locations along the river.

Several meetings and workshops were conducted between the city of Tallahassee, Mead & Hunt and FERC during the evaluation process to reach consensus on the breach parameter selection. The most likely and most severe breach location was determined to occur at the 3,600-foot-long earthen embankment. FERC guidelines recommend an average breach width of three to five times the breach width for earthen embankments. Based on this recommendation, the embankment breach in the dam failure analysis was assumed to develop down to the foundation of the embankment, resulting in a breach depth of 40 feet, and the average breach width was five times the breach depth.

However, another breach scenario was also considered.

In 1957, two sections of the earthen embankment were washed out during a flood due to head-cutting caused by seepage flow through the embankment. The washed-out sections were 600 feet wide and 30 feet deep on average. Although this historical embankment failure was not caused by overtopping, and the embankment was modified and repaired using modern techniques and materials to avoid a recurrence, FERC requested additional dam failure analyses be performed using the historical failure dimensions.

The dam failure analysis using the first set of breach parameters (based on FERC guidelines) suggested that a dam failure at a flood above the current IDF would not impose significant incremental hazards to downstream areas, and, therefore, the current total project capacity was adequate. The dam failure analysis using the second set of breach parameters (based on the 1957 dam failure) suggested that a dam failure at any flood up to the PMF would cause significant incremental hazards downstream. Based on these results, the project did not have adequate spillway capacity and would have to double its spillway capacity to meet the capacity requirement. However, doubling the spillway capacity was not practical or economical.

The USGS gage record showed that a flow slightly above the five-year flood would cause the river stage at the residential area half a mile downstream to rise to the flood stage, and some houses in this area would be flooded. The five-year flood is equivalent to only 12% of the total project capacity; therefore, an increase to the total spillway capacity would not reduce downstream flooding.

FERC approved an approach that included improving Tallahassee's existing flood warning system rather than increasing the spillway capacity.

Evaluating the fuse plug and emergency spillway area

Based on results from the updated IDF study, the capacity of the fuse plug and emergency spillway were needed to meet the spillway capacity requirements for the dam. Therefore, deleting the hydraulic capacity of the fuse plug and emergency spillway was not an option.

Limited information was available for the existing geotechnical condition at the dam. Evaluation of this information indicated the foundation material was composed of layers of sand and clay, and was susceptible to erosion. Several potential remediation measures for the fuse plug and emergency spillway were proposed in the evaluation report, including improving the condition of the existing fuse plug and emergency spillway to ensure proper functioning of these two structures; modifying the downstream area to avoid head-cutting; constructing a new gated spillway to replace the fuse plug and emergency spillways; and revising the emergency action plan to include the additional homes that may be affected by the emergency spillway failure.

The evaluation report also recommended a subsurface exploration at the fuse plug and emergency spillway area for further evaluation of the modification alternatives and a topographic survey for detailed engineering design. The potential remediation concepts were discussed at a meeting with FERC.

Geotechnical investigation

The subsurface exploration and topographical survey of the project were performed in August 2006. Five test borings were taken to a depth of 50 and 70 feet below the ground surface to identify the soils within the existing fuse plug and emergency spillway area. In general, these soils consist of inter-mixed layers of silty sand, fine sand, clay, silt, and clay-infused silt. Limestone rock or very stiff-to-hard calcareous clay silt was encountered between the approximate elevations of 20 feet and 35 feet.

After obtaining the field geotechnical and topographic data, the remediation alternatives were further evaluated for feasibility.

Examining the options

Armoring the existing fuse plug and emergency spillways was initially considered to be the most straightforward and potentially most economical approach. Because of the erodibility of the foundation materials, downstream erosion and head-cutting needed to be addressed. Head-cutting is an erosion pattern that begins at a critical section downstream of the spillway and progresses upstream into the reservoir. Revising the EAP alone to passively react to a potential failure was not considered to be a viable option. Several other structural modification methods were also considered:

- Re-grading to flatten the downstream slope to allow use of a grass-lined channel;

- Lining the downstream slopes with concrete, articulated blocks, soil cement, and riprap;

- Constructing a single concrete core wall that would be stable if erosion occurred; and

- Installing stepped lines of diaphragm walls.

Because of the loose characteristics of the foundation materials, the downstream core wall or stepped diaphragm walls would have to be anchored deep into the foundation, resulting in high construction cost. The large size of the downstream slope area also made re-grading and lining uneconomical. Further, passing flows through the fuse plug and emergency spillway area would result in significant environmental impacts to the immediate downstream area. This alternative was eventually abandoned.

Duckbill spillway to replace fuse plug and emergency spillway

A duckbill spillway, to accommodate the tight physical space, was considered to replace the fuse plug and emergency spillway. A duckbill spillway is a wrapped side-channel spillway. The concept was to build a duckbill spillway at the fuse plug area with the same hydraulic capacity as the combination of the fuse plug and emergency spillway and place a roller-compacted-concrete sill to prevent deeper erosion into the foundation materials. Hydraulic computations showed that a substantial concrete exit channel would have to be constructed from the duckbill spillway to the main river, resulting in significant environmental impacts to the area immediately downstream. Thus, this alternative was also abandoned.

RCC auxiliary spillway

Modification options at the fuse plug area were ruled out due to high construction cost, construction difficulties and significant potential environment impacts. Several other alternatives were evaluated that involved abandoning the fuse plug and emergency spillway and constructing a new spillway next to the existing gated spillway. A new gated spillway was not considered to be a viable option due to the requirement of increasing spillway capacity at relatively low reservoir elevations, which would worsen the downstream flooding.

An RCC auxiliary spillway was evaluated. RCC spillways offer the advantage of being placed directly over the earthen dam, and the downstream steps can act as energy dissipaters. Therefore, they have become a popular option for emergency spillways over the past two decades.

Because the RCC spillways were typically used only for emergency spillways and had rarely been activated, the theories of this type of spillway were not very mature. Several approaches were taken to analyze the energy dissipation effects of the spillway downstream steps. The results turned out to be less than ideal. Consequently, this alternative was also abandoned.

Conventional ogee-shaped crest spillway with stilling basin

The alternative determined to be most viable was a conventional ogee-shaped crest spillway with a Bureau of Reclamation Type III stilling basin adjacent to the existing gated spillway. This configuration was a well-studied model and has been tested at many dams around the world. It provided a structurally stable spillway and eliminated the uncertainty regarding energy dissipation. In this concept, the fuse plug and emergency spillway would be replaced by a remote dike by building the existing crest to the same elevation as the earthen embankment. This alternative would result in the fewest environmental impacts to the downstream area.

Workers are shown during the stilling basin construction phase of the Jackson Bluff project renovation.
Workers are shown during the stilling basin construction phase of the Jackson Bluff project renovation.

Ogee crest auxiliary spillway design concepts and components

The selected design alternative included a remote dike to enclose the fuse plug and emergency spillway and an 825-foot-long ogee crest overflow concrete spillway adjacent to the existing gated spillway with Reclamation's Type III stilling basin for energy dissipation. The ogee crest was set at an elevation of 69.5 feet, 1 foot above the normal pool elevation. A stone apron was designed in front of the concrete ogee structure to reduce erosion of the embankment materials due to the high approach flow velocity.

An upstream sheetpile cutoff wall was designed to extend the entire 825-foot length beneath the ogee structure into the foundation down to the elevation of an existing sheetpile core wall. The upstream cutoff wall was designed to reduce seepage flow through the embankment material. A downstream sheetpile cutoff wall was also designed to extend the entire 825-foot length at the downstream edge of the stilling basin. The downstream cutoff wall was designed to prevent potential downstream erosion from developing further upstream beneath the stilling basin and auxiliary spillway. On both sides of the auxiliary spillway, the spillway structures are isolated from the rest of the embankment by sheetpile training walls. Reclamation's Design of Small Dams was used as the primary reference for design of the key components.

Advantages

The advantages of this design included:

- Reduced impact to the downstream forest by passing the flows exiting the paved spillway through a natural creek into the main river;

- Reduced downstream environmental impacts by balancing the structure size and hydraulic capacity;

- High discharge coefficient of the ogee crest structure, resulting in a shortened paved spillway crest length;

- A closer configuration of both flow passage structures (gated spillway and overflow auxiliary spillway), allowing for better coordination of flow distribution during flood events;

- Relatively consistent foundation materials at the existing engineered embankment, reducing the complexity of erosion protection design;

- Shortened hydraulic jump and reduced downstream scouring potential by using the Type III stilling basin; and

- Greatly reduced seepage and erosion potential by using sheetpile cutoff walls along with the under-drain system beneath the spillway structures.

Hydraulics

The new auxiliary spillway had a total discharge capacity of 69,900 cfs, which replaced the hydraulic capacity through the fuse plug and emergency spillway. The spillway would be activated at a flood slightly less than the 50-year flood at the project site. The updated spillway rating curve with the new auxiliary spillway was incorporated into the BOSS DAMBRK model, which was developed for the IDF study to determine the hydraulic impact of the new auxiliary spillway to downstream areas. The model results showed the new auxiliary spillway would not increase the flooding impact to the downstream area under the IDF condition.

Gate operation protocol

During the permitting process, concerns were raised about minor incremental increases in the downstream impacts as a result of the new spillway being activated under interim flood conditions. To address this, a modified gate operating protocol was developed that would result in no changes to downstream conditions as compared to the existing facility impacts. This protocol would have the facility operators match the post-modification downstream flows with the pre-modification downstream flows.

Geotechnical analysis

Subsurface exploration was conducted at the proposed auxiliary spillway location in 2007 and geotechnical analyses were performed using the field data. The slope stability of the proposed auxiliary spillway was analyzed using the UTEXAS4 computer program developed by Shinoak Software of Austin, Texas, for the pre-, during, and post-construction conditions under various flooding conditions. The factors of safety under the analyzed conditions were determined to be adequate.

In addition, a seepage analysis was performed using the 2D finite-element program SEEP2D to simulate various conditions of the embankment before, during, and after construction. The total seepage flow rate through the entire 825-foot-long auxiliary spillway after construction was determined to be 0.3 cfs under normal pool level condition and 1.2 cfs under the IDF condition. These results were considered to be insignificant and would not cause erosion of the foundation materials.

Lessons learned

The Jackson Bluff auxiliary spillway project evolved from a potential failure mode identified at a PFMA session and went through eight years of development. Although this failure mode was initially identified to have a low probability, the potential impact to the downstream area could be significant.

During the process of the hydraulic analysis, conceptual design, geotechnical investigation, final engineering design, and construction, Tallahassee officials worked closely with FERC and Mead & Hunt to develop the best option for the project.

During the investigation and design process, many factors had to be considered. Dam safety, impacts to the upstream and downstream residents, impacts to the surrounding environment, economics, constructability and public relationships were all part of determining the most viable option.

Although this project brings no benefit to the hydroelectric generation of the project, it greatly improves the safety of the dam structures and best protects the people living downstream of the project.

Rob McGarrah is general manager of the city of Tallahassee's electric system. Yiying Xiong, P.E., was project manager and Kim Hansen, P.E., was senior project manager with Mead & Hunt Inc. Both are now with MWH.

More HR Current Issue Articles
More HR Archives Issue Articles

To access this Article, go to:
http://www.hydroworld.com/content/hydro/en/articles/hr/print/volume-31/issue-05/article/dam-safety-adding-a-spillway-protects-integrity-of-jackson-bluff-dam.html