FTCN: A Reservoir Parameter Prediction Method Based on a Fusional Temporal Convolutional Network

Predicting reservoir parameters accurately is of great significance in petroleum exploration and development. In this paper, we propose a parameter prediction method named fusional temporal convolutional network (FTCN). Specifically, first analyze the relationship between logging curves parameters. Then, build design fusion module to improve results curve inflection points, which integrates characteristics shallow convolution layer deep network. Finally, conduct experiments on real datasets. The indicate that compared with baseline method, mean square errors FTCN are reduced by 0.23, 0.24 0.25 predicting porosity, permeability, water saturation, respectively, shows our more consistent actual geological conditions. Our innovation new introduce model innovatively. main contribution can well predict even when there changes formation properties. research work provide reference for analysis, conducive interpreters’ efforts rock strata identify oil gas resources.