BP神经网络在低孔渗储层水淹层识别中的应用

Application of BP Neural Network in Identification of water flooded Layer in Low Porosity and Permeability Reservoirs

针对榆树林油田低孔渗储层水淹层识别难度大,提出以BP神经网络模型为理论基础,结合研究区岩心分析、试油、以及常规测井等资料,建立油层水淹状况与测井响应值之间的对应关系,实现对水淹层的高精度解释。通过对BP神经网络模型的训练,得到满足误差条件的最佳网络。运用最佳网络对测试数据进行检验分析,最终92.9%油层水淹状况解释准确,有效解决了低孔渗储层水淹层识别难度大,精度低的问题。

Aiming at the problem that identification of water flooded layer was difficult in the low porosity and permeability reservoirs of Yushulin oilfield, taking BP neural network model as the theoretical basis, combined with core analysis of the study area, oil testing, as well as the conventional logging data, the relationship between water flooded status and logging response values was established to enhance the precision of interpretation about the water flooded layer. The best network that can satisfy error condition was got by training of the BP neural network model. Then the best network was used to test the testing data. The results show that, 92.9% of oil reservoir water flooded lay identification result is accurate, and it can effectively solve the problem of water flooded layer identification in low porosity and permeability reservoirs.