Resumen

SANTILLAN, David; FRAILE-ARDANUY, Jesús  y  TOLEDO, Miguel Ángel. Prediction of Gauge Readings of Filtration in Arch Dams using Artificial Neural Networks. Tecnol. cienc. agua [online]. 2014, vol.5, n.3, pp.81-96. ISSN 2007-2422.

Artificial neural networks are mathematical structures inspired by the brain of live beings which can generate relatively simple non-linear numerical calibration models. The present work models the flow of water filtered through the rocky base of a pilot arch dam using a multi-layer perceptron neural network. Seepage through a rock mass is difficult to model because it is impossible to obtain a detailed characterization of the medium through which it passes and because of the complexity of the process. The final result is a model composed of three hidden neurons grouped in a layer, using as input variables the water level in the reservoir and their three velocities from prior periods. The structure of the neural network is determined considering the influence of each of the input variables on the output variables. This is based on an extensive set of possible input variables extracted from analytical or conceptual models of the physical phenomenon to be modeled.

Palabras llave : Arch dams; seepage; artificial neural networks; dam monitoring.

        · resumen en Español     · texto en Español     · Español ( pdf )