神经网络方法在电缆地层测试数据反演中的应用

利用电缆地层测试资料正确反演地层参数是油田勘探中的一项重要任务。将神经网络方法引入到电缆地层测试数据的反演中,可以在考虑管线存储效应和井筒表皮效应对测量产生影响的情况下,同时获得地层的渗透率和井筒的表皮系数。该方法利用地层测试早期和中期的数据,不仅可以缩短地层测试器的测量时间,而且可以充分利用测试过程中的动态数据。验证结果表明,该反演方法是可行的。

Application of neural network method to wireline formation test data inversion

Formation parameter inversion from wireline formation test data plays an important role in oilfield exploration. The neural network method was applied to the wireline formation test data inversion. The model can work out formation permeability and dimensionless skin coefficient accurately considering storage effect and skin effects. The model can not only reduce test time ,but also take fully advantage of dynamic information by early and intermediate period data. It is testified that the neural network model is feasible.