The invention discloses a
polymer-flooding reservoir yield prediction method and
system based on a double attention mechanism and a storage medium, relates to the field of oilfield development, and aims to solve the problems that a bidirectional dependency relationship and local features of long-time-
series data are difficult to capture at the same time and feature
weight distribution and time-series information focusing effects are insufficient in an existing method. According to the technical key points, the method comprises the following steps: S1, collecting
field development data of the
polymer flooding reservoir and screening optimization characteristics; s2, a CNN-BiLSTM-DAM fusion model is constructed, and the function implementation process of the model comprises the steps that local features of input data are extracted based on a convolutional layer, a dynamic weight is distributed to the local features of each
time step through an attention mechanism layer, forward and backward dependency relationships of weighted
time sequence feature data are captured through a bidirectional loop structure of a BiLSTM layer, and a dynamic weight is distributed to the local features of each
time step; performing
weight adjustment on the
time sequence features output by the BiLSTM layer through the attention mechanism layer, and capturing
global time dependence through the BiLSTM layer; and S3, the optimized features are input into the trained CNN-BiLSTM-DAM fusion model, and the yield of the
polymer flooding reservoir is predicted.