【摘要】 受地震资料品质、岩性、构造等诸多因素的影响,单一地震属性只能在一定程度上提供预测储层的方向,并存在多解性。地震属性融合技术用井孔资料对地震属性进行标定,建立储层含油气性与地震属性之间的关系,采取数学手段融合多种地震属性进行储层含油气性判别,避免了单一地震属性解释储层的多解性问题。BP网络具有良好的非线性拟合能力,但是易陷入局部极小值,不收敛,影响预测精度。针对该问题,采用粒子群优化其网络权值和阈值,再用BP网络对储层、非储层进行模式识别,取得较好成效。更多还原

【Abstract】 Constrained by factors such as quality of seismic data,lithology and structures,single seismic attribute can only be used to predict reservoirs to a certain extent and there are multiple possibilities.Through calibrating seismic attributes with well data,seismic attribute fusion techniques can correlate oil/gas potential with seismic attributes.For oil/gas potential prediction of reservoirs,mathematics-based multi-attribute fusion can avoid the ambiguity of single-attribute reservoir interpretation.Back propagation(BP)neural network is very good at non-linear fitting,but it is easy to get a local minimum without convergence,influencing the accuracy of prediction.To solve this problem,particle swarm is adopted first to optimize the network weight and threshold,and BP neural network is then used to differentiate reservoirs and non-reservoirs.The results are satisfactory.更多还原