Dam and Lock Construction
Monongahela River, USA: As part of the project to build a new navigation lock at Point Marion, Pennsylvania, the US Army Corps of Engineers constructed a cofferdam that incorporated the land wall of the existing lock.
A similar project, undertaken by the TVA in 1961, ended in catastrophic failure when a large portion of the land wall of Wheeler Lock slid ten meters into the dewatered excavation, killing two workers. River traffic had to be suspended until the lock was reconstructed. The root cause of the failure was an undetected seam of clay in the foundation rock, but no stabilization measures had been taken and no instrumentation had been installed. The use of rock anchors and instrumentation was still in its infancy at the time.
To ensure the safety of the Point Marion cofferdam (shown in yellow), the design team reviewed more than 150 boring logs and tested soil and rock samples for shear strength. Then they determined what stabilization measures would be required and developed an instrumentation plan that could verify the effectiveness of those measures.
The instrumentation had to be operational prior to any excavation in the coffered area. More than forty inclinometers were installed in the land wall and other parts of the cofferdam. The inclinometers extended into more competent rock below the existing foundation, so that measurements could reveal sliding of the wall as well as rotation. Standpipe piezometers were installed to monitor saturation levels in the granular-fill cells and to monitor uplift pressures in the foundation of all the cofferdam elements.
Stabilization measures were necessary because the land wall of the existing lock had a relatively narrow base and had been built on weak sedimentary rock with a low shear strength. Excavation activity would occur just three meters away and would extend nearly four meters below the foundation of the wall.
A carefully sequenced plan of construction called for the installation of 139 vertical anchors in the land wall to prevent overturning and an additional 335 inclined anchors to prevent horizontal movement.
The land wall contained a culvert (see drawing above) to transport water to and
from the lock chamber, so the upper row of inclined anchors had to be positioned
to pass beneath the culvert. This required excavation of about ten meters of soil
immediately behind the wall. As excavation was underway, inclinometers detected
rapid movements in the foundation that were approaching the shear strains estimated
to mobilize peak strength.
Piezometers, too, were valuable in detecting a potential stability problem. The excessive saturation level measured in one of the granular-fill cells revealed that river water was leaking through the sheet pile interlocks. The problem was quickly remedied by cutting additional weep holes on the landward side of the cell and sealing the leak on the riverward side. The uplift piezometers showed no excess uplift pressures, verifying that the foundation was successfully sealed by pregrouting during the installation of rock anchors.
In summary, the instrumentation program at Point Marion Lock was of critical importance to the safety and success of the project. The instruments not only verified design assumptions, but also detected developing problems early so that corrective measures could be taken. Begun in 1991 and completed in 1994, the Point Marion Lock project achieved the goals of its designers: the economical replacement of the original lock with minimal disruption of river traffic.
Thanks very much to Brian Greene and Andrew Schaffer of the US Army Corps of Engineers, Pittsburgh District, for providing this story. A detailed account of the project was written by Brian Greene, John Gerlach, and Andrew Schaffer and published in the Bulletin of the Association of Engineering Geologists, Vol XXX, No 3, 1993 , pp 265-279. The same article also provides a list of papers written on other aspects of the project.