Upgrading Wolf Creek dam18 March 2014
New excavation and data management systems played vital roles in building a new cut-off wall at Wolf Creek dam, in Kentucky. Report by Patrick Reynolds
Tight vertical excavation control, new foundation construction plant and an advanced data management system were vital factors that combined to enable accurate construction and early completion of a larger, and deeper, cut-off wall at the "high-risk" Wolf Creek dam, in Kentucky.
The reservoir had been kept low to reduce pressure on the leaking embankment section of the dam. But following completion of the new cut-off wall early last year, ahead of schedule, the reservoir level has been raised incrementally, and the dam's performance tested at each stage.
As Lake Cumberland returns to normal, and the local economy receives a resurgent boost, the designer and owner of the dam - US Army Corps of Engineers (USACE), Nashville District - is taking the lessons on to Center Hill dam, another of its structures suffering from seepage.
Seepage concerns at Wolf Creek
USACE has 10 dams in the Cumberland River catchment, and Wolf Creek dam in Russell County is one of its largest structures. The dam impounds Lake Cumberland, which is 162km long, the ninth largest reservoir in the US, and the largest east of the Mississippi River.
The 78.6m (258ft) high dam is a combined concrete gravity structure with a dominant, approximately 1201m (3940ft) long, earthfill embankment extending over about two-thirds of the total length. The non-zoned embankment is composed of well compacted, low plasticity alluvial clays. However, built on limestone, Wolf Creek dam has had safety problems due to leaks through the limestone bedrock below the embankment portion of the structure.
Wolf Creek was designed and built over 1938-52. USACE says concerns around seepage were identified during foundation construction due to the visible karst features. In the late 1960s, muddy flows were discovered in the tailrace and two sinkholes found near the toe of the embankment, and the seepage concerns increased. Investigations found solution channels in the limestone, and local piping and collapse of embankment fill.
Emergency grouting over 1968-70 stopped the immediate piping, and 'is generally credited with saving the dam,' says USACE, though adds that it became clear more would have to be done. In the late 1970s, a diaphragm wall - a first for USACE - was constructed down through the embankment to lock into the bedrock.
Monitoring of the concrete barrier revealed that the diaphragm wall had slowed, but not stopped, the seepage and erosion problems. USACE says that water seepage had 'found new paths under and around the wall,' and adds, 'perhaps through defects in the wall itself' as erosion of solution features continued.
Further studies were undertaken at Wolf Creek, over 2004-5, concurrently with USACE's development of its risk-informed safety programme looking at all of its dams. Conclusions were that the wall 'did not go deep enough' into the bedrock, and 'did not extend laterally far enough to intercept all major karst features.'
The failure risk was judged as highest on the USACE scale - Dam Safety Action Class I (DSAC 1), Urgent & Compelling. An independent external peer review concluded there was compelling risk of a piping failure.
Plans for a second cut-off wall
Plans were developed for a second cut-off wall - deeper, wider, and constructed immediately upstream of the first wall. The reservoir was temporarily lowered to reduce risk until completion of the wall. The water level was kept 13.1m (43ft) below the crest elevation (220.4m (723ft) asl), but sufficient to satisfy the minimum "power pool" for hydroelectric generation.
The Nashville District, part of the Great Lakes & Ohio River Division, designed the new cut-off to be constructed between the original wall and core trench of Wolf Creek embankment dam.
The new wall would extend beyond the original wall at each side - to the left, it would tie-in to the concrete dam, short of the spillway; and, to the right, it runs all the way along the embankment - its depth reducing in steps - to lock into the abutment.
In total, the new wall was designed to be approximately 1158m (3800ft) long and up to 83.8m (275ft) deep - penetrating up to 29m (95ft) into the bedrock, making it up to 22.9m (75ft) deeper than the original cut-off wall which has a variable bottom profile.
The new wall was designed with a minimum thickness of 610mm (2ft), as measured across the three types of construction employed: overlapping panels; secant piles; and, a panel-pile combination.
Construction control held the greatest challenge to consistently achieve the design thickness at the overlapping or interlocking sections. Compared to the era of the first cut-off wall, the new deep foundation project was able to call upon more advanced construction systems, and data monitoring and analyses. Both construction and IT muscle were vital to ensuring rapid and tight control of the excavation, limiting deviation of each pile and so building a barrier structure with greater integrity.
Main contractor for the works was Treviicos-Soletanche JV. Construction commenced in 2006-7 with a scheduled completion target of late 2013.
Six years of construction: stage by stage
Initial construction activities including creation of a work platform down through which the deep wall would be excavated. The platform would space for plant as well as hold a guide wall, for survey control. Preparatory works also included installing grout curtains, between which the cut-off wall would be constructed.
USACE says grouting was important for combination of reasons, including helping to: reduce risk before the wall was completed; provide site investigation data with the curtains going 15.24m (50ft) deeper than the wall; reduce construction risk should there be voids along the wall alignment; and, provide extra barrier layers, or redundancy, to the cut-off solution.
"The data showed the geology improved with depth confirming the design depths of the wall," reports USACE. "Also, no significant slurry loss occurred in the rock during wall construction."
Main construction of the cut-off wall began in 2007-8 with a run of overlapping panels excavated down through the clay embankment to sit in the top 610mm (2ft) of bedrock. The panels were 1.83m (6ft) wide by 2.8m (9'2") long, and overlapped at least 127mm (5"). Concrete was placed by tremie, working up from the base of the excavation and displacing the slurry support.
The JV contractor had proposed the panel wall as the initial construction step to minimise the time the embankment would be open and exposed to the slurry. For the closely controlled excavation job, the contractor designed and built a specific hydromill - counter-rotating cutter heads housed in a frame, locked onto and suspended by crane above the pre-built and survey-controlled guide wall. Excavation control was aided by real-time data from three biaxial inclinometers to keep tight to the vertical, and theoretical, location of each panel.
Next, the contractor sank an array of 203mm (8") dia. Pilot holes, at 890mm (35") centres, through the panel wall and deeper into the bedrock. The holes - bored using a Wassera Water hammer, which employs hydraulic pressure to drive a full face rock bit - would be guides for the main secant piling operation, but also function as exploratory holes capable of grouting work. Steering adjustments during the pilot hole excavation were made using slant face bits or a bent housing.
To complete the wall, 1197 x 1.27m (50") dia. Secant piles were bored at 890mm (35') centres to form the full cut-off barrier. The piles were bored by up to five Aker Wirth reverse circulation drilling (RCD) rigs, developed specially with Treviicos, and included a string with triple stabilised collars and a bit with stinger. Proximity to the reservoir and groundwater prevented chemical additives being used, and so the rigs relied on compressed air to create an air-water mix as a drilling fluid to support spoil removal.
While the rigs did not have their own steering capability, their bottom hole assemblies had length and weight, and onboard inclinometers, to keep the bores tightly vertical as they tracked the pilot holes. The last pile was installed in March 2013 - about nine moths early, thanks to ongoing efficiency and work process improvements despite the entire wall plus the work platform requiring for more than 229,370 cubic meters of concrete.
Lessons for rehabilitation of Center Hill
Following the works at Wolf Creek, USACE is preparing to adapt its lessons for rehabilitation of the embankments at another dam - Center Hill, near Lancaster, Tennessee, and also in the Cumberland catchment.
One key tool for Center Hill - and USACE future projects - is the IT system developed for the Wolf Creek rehabilitation. The computing muscle is described as an innovative GIS-based application that is used to manage and visualise the construction and performance data held in a relational database.
The Wolf Creek Information Management System (WCIMS) collated and managed 71 million files, including: all historical and recent geotechnical information; barrier design and construction documents; geotechnical instrumentation data; and, project management information.
But capabilities during the works were not its only strength. The IT system dramatically cut the time needed to digest and present the construction information and data to only a couple of days. Further, the task was done within a week of the last pile being bored. Therefore, the cut-off wall was accepted months early - and solely due to IT power.
In late October, Nashville District's Civil Design & Construction Branch received the Corps' 2013 Innovation of the Year Award for WCIMS. The IT plaudit followed soon after the overall rehabilitation scheme had received the 2013 Outstanding Project Award from the Deep Foundation Institute.
Patrick Reynolds is contributing editor of International Water Power & Dam Construction