Title :
Application of Genetic Neural Network to Water-Flooded Zone Identification
Author :
Tang, Hong ; Liu, Hong-qi ; Jin, Song
Abstract :
It is very important for oilfield development to identify water flooded zone effectively and evaluate the flooding degree or grade quantitatively because of high water cut for most oilfields in China. Automatic identification of water flooded zones is realized by applying neural networks. Firstly, a standard database of fluid logging facies and water flooded grade standard were built up based on the statistic data from the key wells, And then samples were trained by genetic neural network algorithm(GA). Compared with the simple BP neural network, genetic neural network algorithm is more robust and easily convergent. The results of practical application in certain oilfield show that a trained genetic neural network proved to be powerful and effective.
Keywords :
Educational institutions; Floods; Genetic algorithms; Genetics; Permeability; Reservoirs; genetic algorithm; logging interpretation; neural network; water-flooded zone;
Conference_Titel :
Computational and Information Sciences (ICCIS), 2011 International Conference on
Conference_Location :
Chengdu, China
Print_ISBN :
978-1-4577-1540-2
DOI :
10.1109/ICCIS.2011.90