Artificial Geo-Hydrological Zonation

"Groundwater inflow into underground constructions may cause hazardous working conditions and environmental problems. It is essential to localize and characterize potential flow paths and to estimate the amount of water leakages, so that appropriate sealing works can be performed. Estimating the amount of water ingress is a multi-phase process starting from qualitative and quantitative examination of rock mass. Due to association of uncertainty, vagueness and incompleteness with the monitored data set, particularly resulted from the in situ tests, accounting relevant approaches such Fuzzy Logic and Recurrent Neural Networks, are proposed in this research for learning acquisition, extraction of rules and prediction of behaviors. The computational set of artificial tools permits using the complex information entities or information granules (rock types, geologic structures, hydro-fracture potential) for data abstraction and derivation of knowledge from records (numerical or linguistic format). This is more a theoretical perspective than a set of methods or principles. As a theoretical perspective, this proposal encourages an approach to data that recognizes and exploits the information present in indexes at various levels of resolution or scales. In this sense, it encompasses the methods which provide flexibility and adaptability in the resolution at which knowledge or information is extracted and represented. The product of merging these tools is a 3D artificial-mapping of rock classification (Qi-system), geological conditions and Lugeon units that permits, as far as possible, to understand more and better the flow behavior."