神经网络技术与地震参数隔层预测

Non-permeable-bed prediction using neural network technique and seismic parameters

1. 西安工程学院信电系,西安 71O054;2. 西安石油学院石油工程系,西安 710065;3. 辽河油田研究院,盘锦 124000

Engineering University, Xi' an 710054

论述了油田区块开发前期和中期利用三维地震保幅叠偏资料以及钻井地质和测井资料, 采用神经网络技术进行储层间的隔(夹)层预测方法及实际应用问题。通过实例分析, 认为该方法技术的精度和可靠性较高, 其结果为油田区块开发方案的制定提供了较可靠的依据。

In the early and middle stage of oilfield development, the neural network technique is usedfor the prediction of nonpermeable beds beween reservoirs, on the basis of 3-D amplitude-preserving mi-gration seismic data along with gedogic and well logging data. This technique is expounded in the paper.A red example analyrsis shows that the technique is of high precision and rdiability. The resu1t obtainedby the technique provides basis for designing the oilfield devdopment scheme.