The 3D analysis and estimation of transient seepage in earth dams through PLAXIS 3D software: neural network

• Published in: Environmental earth sciences 2019-09, Vol.78 (18), p.1-7
• Main Authors: Bayat, Mandana, Eslamian, Saeid, Shams, Gholamreza, Hajiannia, Alborz
• Format: Article
• Language: English
• Subjects: Algorithms; Coefficients; Computer programs; Dam engineering; Dam failure; Dams; Downstream effects; Earth; Earth dams; Elasticity; Erosion; Mathematical analysis; Neural networks; Permeability; Piping; Seepage; Software; Soil; Soil permeability; Specific gravity; Upstream
• ISSN: 1866-6280; 1866-6299
• DOI: 10.1007/s12665-019-8405-y

The issue of seepage and its management is essential in earth dams which if exceeded from a certain stage the destruction of dam is probable. Despite the studies run through consulting engineers, the hydraulic behaviors of both body and foundation cannot be predicted. To understand the function of such structures, the geotechnical software PLAXIS 3D based on FEM is applied to determine the transient seepage analysis on the subject earth dam for 1 year. The outputs in downstream drain are compared with the actual obtained results from seepage at the same points with 15-day intervals for 1 year as an experimental sample. The obtained data here thorough PLAXIS 3D in addition to available actual seepage, permeability coefficients in x, y, and z directions, transient head upstream, soil-specific gravity for both the wet and dry conditions, Poisson coefficient, c, ɸ, and elasticity modulus material are trained through a neural network algorithm. Here, an equation is yield through which the seepage is calculated in transient head upstream at any point of the subject earth dam. This equation has a great accuracy of (P value > 0.05), and is applied in calculating the actual seepage of earth dams in critical situations such as earth dam failure due to piping operation or dam body erosion.