Evaluation of training techniques of artificial neural networks for geothermometric studies of geothermal systems
L. Díaz-González, C.A. Hidalgo-Dávila, E. Santoyo and J. Hermosillo-Valadez
-
A multivariate statistical analysis with artificial neural networks for geothermometric studies of geothermal systems has been carried out.
Na, K, Mg, Ca y Li compositions of geothermal fluids were evaluated for determining their contributions in the estimation of deep equilibrium temperatures of geothermal wells.
log(Na/K) showed the highest contribution ranging from 69% to 75%.
The results obtained in this geochemometrical study will enable in the future to develop a multicomponent geothermometer for the exploration and exploitation of geothermal systems.
|