Waterflooding identification of continental clastic reservoirs based on neural network

This article describes an approach based on artificial neural network to identify waterflooded zone of continental clastic reservoirs. For the logging sequence of waterflooded zone matching the characteristics of the continental oilfield, the application of artificial neural network algorithm is able to distinguish water layers, oil reservoirs and dry layers among reservoirs of waterflooded zones. The output vectors of the network represent the fluid types. Thus, better results are supposed to be obtained than traditional methods in the crossplot plate after network training. Distribution becoming nonuniform and contact between grains being loose were found after microscopic observation in the waterflooded zones. It has revealed that the waterflooded characteristics are of great significance, and it has also proved the accuracy of identification from another perspective.