No abstract

No abstract

Machine learning algorithms becomes popular for intelligent classification of rock images. In this paper, it selects Resnet 50 neural network model to divide the data sets based on the rock pictures taken under the white light lamp. By continuously adjusting the parameters of each layer, the intelligent classification of rocks is carries out. The training final validates accuracy reached 94.12%.

For coaled methane reservoir relative permeability curves is difficult to determine. This paper uses artificial intelligence algorithms including genetic algorithms and neural network algorithms to establish mathematical model for inversion relative permeability curves, by combination of dynamic actual production data. The results show that this method is feasible, highly practical in industry. Keywords-coaled methane reservoir neural network genetic algorithms relative permeability curves

No abstract