摘要
天然气水合物饱和度的计算通常采用阿尔奇公式、双水模型、Wood方程等方法,这些方法均基于孔隙度的求取,并需要配合岩心分析来获得公式中的有关参数,存在误差传递导致结果不正确的问题。由于天然气水合物是以固态形式赋存于地层当中,因此研究适用于含天然气水合物储层的评价模型也是解决准确评价天然气水合物储层需考虑的因素。针对沿用油气测井评价方法计算天然气水合物的孔隙度和饱和度中存在的问题,采用径向基函数作为人工神经网络,计算了我国首次采获水合物样品的神狐海域某井天然气水合物的饱和度,以其中一口井的分析数据为样本训练并建立径向基函数神经网络,有效地求出了另一口井的天然气水合物饱和度,其结果与现场孔隙水分析的饱和度基本吻合。避开了天然气水合物饱和度的模型建立及参数求取难题。
Generally Archie equation, dual water model and Wood equation are employed for computing the saturation of gas hydrates. These approaches are all based on the calculation of porosity and depend on many parameters from core analysis. This may lead to incorrect results due to the transfer of errors. Further more, gas hydrates exist in the formation with solid state which is different from conventional gas and oil reservoirs. Thus a suitable model may be one of the factors to accurate evaluation of gas hydrates.As another choice, basing on the gas hydrates drilling data of Shenhu sea area, the gas hydrates saturation is well got by the radial basis neural networks. Thus the problem for constructing the model is successfully avoided.
出处
《海洋通报》
CAS
CSCD
北大核心
2009年第4期102-106,共5页
Marine Science Bulletin
