®®®® SIIA Público

Título del libro: Water-Rock Interaction
Título del capítulo: Determination of flowing pressure gradients in producing geothermal wells by using artificial neural networks

Autores UNAM:
OCTAVIO GARCIA VALLADARES; EDGAR ROLANDO SANTOYO GUTIERREZ;
Autores externos:

Idioma:
Inglés
Año de publicación:
2010
Resumen:

A predictive model based on an application of the artificial neural networks (ANN) for obtained the flowing pressure gradients in geothermal wells was applied. The ANN model uses geometrical and physical data commonly measured in producing wells. The prediction of pressure gradients was successfully achieved using the following input data: wellbore geometry (i.e., wellbore depth, inclination angle of pipe, and wellbore diameter), mass flow rate and bottomhole temperature measurements collected from a world database of geothermal wellbores. For the ANN, several computational architectures based on the Levenberg-Marquardt optimization algorithm, the hyperbolic tangent sigmoid transfer-function, and the linear transfer-function were effectively used. The best fitting training data set was obtained with an ANN architecture of 15 neurons in the hidden layer, which made possible to predict the flowing pressure gradients with a satisfactory efficiency (R2 = 0.9910). The results provided by the ANN model between measured and simulated data were in good agreement. © 2010 Taylor & Francis Group, London.


Entidades citadas de la UNAM: