deflections during construction. The team used the results of the analysis to determine threshold levels for an extensive system of inclinometers and surface monitoring points around the site to confirm the cofferdam’s performance. Dewatering Design. The dewatering Cut-off wall excavation sequence passing the 200 sieve), partially placed underwater and partially placed in the dry to the level of the cut-off’s working plat- form. Once the cut-off wall was installed, the dike was built up to its final elevation with an impervious clay cap. The completed dike was over 30 m (100 ft) at its highest. The river bottom foundation for the marine dike was another concern, in parti- cular, settlement due to the compressibility of the natural soils and slope stability of the dike because of weak founding soils and sloping bedrock into the river. The team conducted a bathymetric survey as well as a penetration test into the river bottom silt. The test consisted of a 1.8 m (6 ft) diameter by 3 m (10 ft) steel pipe filled with rock was lowered on the silt and the penetration measured. If it penetrated more than 15.2 cm (6 in), additional foundation rock was required. This testing was performed on a 30 m (100 ft) x 30 m (100 ft) grid. The compaction of the “dumped” sand was a primary concern. Trench stability calculations by MRCE indicated that a minimum phi angle of 30-degrees was 4.4 kg/m (70 pcf) to maintain trench stability. Obviously, trench failures in the marine dike in the Ohio River would have been catastrophic to the project. The project team opted for a program of vibroflotation to improve the hydraulically- placed sands, with CPT testing used to confirm the improvement. Landside Excavation. The landside required along with slurry units weights of 3 excavation was analyzed using both limit equilibrium slope stability analysis and finite element analysis using Plaxis. The latter was used to evaluate potential 84 • DEEP FOUNDATIONS • NOV/DEC 2013 system design was a combined effort between the JV, Moretrench and MRCE. Moretrench made recommendations on the well sizes and details, and MRCE used MODFLOW to analyze the performance. The analysis indicated that the worst case water infiltration (exclusive of major voids specification for ensuring contact with the underlying rock. The final specification indicated that when the hydromill was touching rock, the QC people would gather at the desanding unit and mutually agree when “fresh” chips of rock were being produced, at which point they would call bottom (of course we had our borings as an initial guide). For the clamshell work, the mutually conceived specification said that when the 20 ton clamshell was no longer bringing up decomposed or fractured rock, then bottom was achieved. A rigorous QA/QC program insured permeability. The in the cut-off wall) would be 11,830 L/min (2,600 gpm), with an expected flow of about 1820 L/min (400 gpm). The dewatering system was sized for 22,750 L/min (5,000 gpm) to insure sufficient capacity if problems arose. All of the stake holders were focused on producing a final product that exceeded expectations and did everything possible to meet and exceed the requirements. The end result was that the pumping rates averaged 683 L/min (150 gpm) to draw down the water levels required for construction. The entire system was presented to the BOC who expressed a high degree of comfort with the team, the design, and in the end, the results. Cut-off Wall. Design issues for the cut- off wall centered on its continuity, seal to the underlying rock, permeability and trench stability. Continuity was defined as a minimum 254 mm (10 in) section at any given point in the nominal 800 mm (31.5 in) wall. MRCE and the JV developed a program included confirmatory on-site strength and permeability testing using 2,300 samples, of which 890 were from the batch plant and 1,410 from various elevations in the trench. Samples were tested on site at the JV’s QC lab. Trench stability calculations were carried out as an infinite 2-D trench, where longer open trenches were used (landside), and as a 3-D analysis where shorter panel lengths were planned due to small differential head and weaker soils (marine dike). The engineers used limit equili- brium to compute stability of the trench with assumed active wedge failure. The (72 pcf) to 1,281 kg/m (80 pcf). The team used the results in construction through tables specifying the required unit weight of slurry for a given river stage to maintain the required factors of safety. was varied, from an assumed 1,153 kg/m 3 assumed slurry unit weight in the trench 3
