Artificial neural network permeability modeling of soil blended with fly ash

• Main Authors: Dungca, Jonathan R., Galupino, Joenel G.
• Format: text
• Published: Animo Repository 2017
• Subjects: Soil permeability—Testing; Waste products; Fly ash; Neural networks (Computer science); Civil Engineering
• Online Access: https://animorepository.dlsu.edu.ph/faculty_research/1915
• Institution: De La Salle University

The determination of the permeability properties of soil is important in designing civil engineering projects where the flow of water through soil is a concern. ASTM D2434 Standard Test Method for Permeability of Granular Soils (Constant Head & Falling Head) is being followed to determine the vertical permeability, while for horizontal permeability, there are none. In this study, tests such as Atterberg limit, relative density tests, and particle size analyses are done to determine the index properties of soil blended with fly ash. Subsequently, microscopic characterizations tests, elemental composition tests and permeability tests are done to determine the chemical and physical properties of the soil mixes. A new permeability set-up was used in determining the horizontal permeability soil mixes. Data were extracted during the experiment and a relationship between the properties of soil and the permeability was established. An artificial neural network model was used to predict the coefficient of permeability when the percentage of fly ash is available. © Int. J. of GEOMATE. All rights reserved.