14 Nov 2024

Application of Deep Learning for Segmenting Seepages in Levee Systems

Abstract: Seepage is a typical hydraulic factor that can initiate the breaching process in a levee system. If not identified and treated on time, seepages can be a severe problem for levees, weakening the levee structure and eventually leading to collapse. Therefore, it is essential always to be vigilant with regular monitoring procedures to identify seepages throughout these levee systems and perform adequate repairs to limit potential threats from unforeseen levee failures. This paper introduces a fully convolutional neural network to identify and segment seepage from the image in levee systems. To the best of our knowledge, this is the first work in this domain. Applying deep learning techniques for semantic segmentation tasks in real-world scenarios has its own challenges, especially the difficulty for models to effectively learn from complex backgrounds while focusing on simpler objects of interest. This challenge is particularly evident in the task of detecting seepages in levee systems, where the fault is relatively simple compared to the complex and varied background. We addressed this problem by introducing negative images and a controlled transfer learning approach for semantic segmentation for accurate seepage segmentation in levee systems.

15 Oct 2024

Deployable Resilient Installation Water Purification and Treatment System (DRIPS): Relief Well Biofouling Treatment of Dams and Levees

Abstract: The US Army Corps of Engineers (USACE) conducts regular inspections and maintenance of relief wells to ensure their proper functionality and to identify early signs of malfunction or potential failure. Expenses associated with labor, materials, and transportation are the primary cost drivers of relief-well maintenance. To minimize labor hours and materials, a treatment approach intended to improve logistics and reduce material costs during relief-well treatment was developed and tested. This approach employed external UVC, mechanical brush treatments, and chlorinated-gas-infused water to produce liquid sodium hypochlorite (NaClO). Preliminary bench-scale testing with chlorine, oxalic acid, and UVC informed the selection of field testing methods and optimal amendment concentrations. Field demonstrations were conducted annually over three years. During the demonstrations, the system underwent continuous optimization to enhance its efficiency. Different locations in Mississippi (Grenada Dam, Eagle Lake, and Magna Vista) were selected for testing. Both new and traditional treatment approaches yielded adequate results, achieving microbial reduction at 96% to 100%. The development and refinement of this system demonstrated that relief wells can be treated within a comparable timeframe and with similar efficiency while utilizing fewer purchased chemicals and materials.

29 Sep 2021

Backward Erosion Testing: Magnolia Levee

Abstract: Using a confined flume device, an experimental study investigated the critical horizontal gradient of soils obtained from a site identified as potentially vulnerable to backward erosion piping (BEP). Tests were conducted on glacial outwash material obtained from a sand and gravel quarry in the vicinity of Magnolia Levee in the community of Magnolia, OH. The two bulk samples collected from the quarry had similar grain-size distributions, grain roundness, and depositional environments as the foundation materials beneath the levee. Samples were prepared at various densities and subjected to gradual increases of flow in a wooden flume with an acrylic top until BEP was observed. The critical average horizontal gradient ranged from 0.21 to 0.30 for a bulk sample with a coefficient of uniformity of 1.6, while tests conducted on a bulk sample with a coefficient of uniformity of 2.5 yielded critical average horizontal gradients of 0.31 to 0.36. The critical average gradients measured during these tests compared favorably to values in the literature after applying adjustments according to Schmertmann’s method.

28 Sep 2021

Backward Erosion Progression Rates from Small-Scale Flume Tests

Abstract: Backward erosion piping (BEP) is an internal erosion mechanism by which erosion channels progress upstream, typically through cohesionless or highly erodible foundation materials of dams and levees. As one of the primary causes of embankment failures, usually during high pool events, the probability of BEP-induced failure is commonly evaluated by the U.S. Army Corps of Engineers for existing dams and levees. In current practice, BEP failure probability is quantitatively assessed assuming steady state conditions with qualitative adjustments for temporal aspects of the process. In cases with short-term hydraulic loads, the progression rate of the erosion pipe may control the failure probability such that more quantitative treatment of the temporal development of erosion is necessary to arrive at meaningful probabilities of failure. This report builds upon the current state of the practice by investigating BEP progression rates through a series of laboratory experiments. BEP progression rates were measured for nine uniform sands in a series of 55 small-scale flume tests. Results indicate that the pipe progression rates are proportional to the seepage velocity and can be predicted using equations recently proposed in the literature.

24 Sep 2020

Study of Sand Boil Development at Kaskaskia Island, IL, Middle Mississippi River Valley

Abstract: Mississippi River flooding in 2013 and 2016 caused severe underseepage and development of several medium to large high-energy sand boils behind the landside levee toe at Kaskaskia Island, IL. This levee system is located between St. Louis and Cape Girardeau, MO, and is part of the Kaskaskia Island Drainage and Levee District on the Middle Mississippi River. Flooding on the Mississippi River in 2013 and 2016 was below the design flowline for this levee. This report documents a case history study into the causes of seepage, piping, and sand boil development at a levee reach at Kaskaskia. Site-specific geotechnical data were collected and evaluated to determine the causes for poor performance at the studied levee reach locations. Data collected involved design documents, geologic and geotechnical borings, closely spaced cone-penetrometer tests (CPTs), electrical resistivity surveys, laboratory soil testing of sand boil ejecta, CPT samples from targeted stratigraphic horizons in the subsurface, and both piezometer and river-stage data. These data indicate sand boils present within this levee reach involved a chronic seepage condition that became progressively worse through time. This condition was directly related to the underlying site geology, namely the top stratum thickness and the depositional environment in this levee reach.