ELRA地区致密砂岩储层非均质性强,孔渗关系复杂,产能预测较困难。文中利用神经网络非线性拟合的独特优势,建立储层每米产油量和测井参数关系,预测致密砂岩储层产能。经新探井试油数据检验,该方法的产能预测符合率大于83.3%,满足生产需求。 Tight sandstone reservoirs in ELRA are highly heterogeneous and have complicated pore-permeability relationships, making it difficult to predict production capacity. In this paper, the unique advantage of neural network nonlinear fitting is used to establish the relationship between oil production per meter and logging parameters and predict the production of tight sandstone reservoirs. The new exploration well test oil data test, the method of predicting the production capacity of more than 83.3%, to meet production needs.