A Well Production Prediction Method of Tight Reservoirs Based on a Hybrid Neural Network

Tight reservoirs have poor physical properties: low permeability and strong heterogeneity, which makes it difficult to predict productivity. Accurate prediction of oil well production plays a very important role in the exploration development gas reservoirs, improving accuracy has always been key issue reservoir characterization. With artificial intelligence, high-performance algorithms make reliable possible from perspective data. Due high cost large error traditional seepage theory formulas predicting production, this paper establishes horizontal productivity model based on hybrid neural network method (CNN-LSTM), solves limitations methods produces accurate predictions wells’ daily production. In order prove effectiveness model, compared with results BPNN, RBF, RNN LSTM, is concluded that CNN-LSTM are 67%, 60%, 51.3% 28% less than those four models, respectively, determination coefficient exceeds 0.95. The show can accurately reflect dynamic change law marks study as preliminary attempt application petroleum engineering, also provides new for intelligence field development.