隐蔽油气圈闭场源信息靶中灰色BP网络标定法

将油气圈闭激发的场源信息视为灰色信息 ,通过综合关联滤波方法对提取的信息进行灰色优化处理 ,以突出隐蔽油气圈闭异常 ,实现隐蔽油气圈闭范围、形态、靶中的定位预测 .提出了一种耦合灰色系统和神经网络的灰色 BP网络方法 ,实现灰色 BP网络滚动训练预测靶位聚焦 .用该法进行隐蔽油气圈闭靶中标定 ,可有效克服“矛盾”样本带来的不适定性问题 ,并提高油气圈闭识别率和定位预测的智能性、合理性与准确性程度 .在 TK测区提取的 3个圈闭上进行了“靶中”定位预测 ,并对测区已出高产油气流的沙四井油气圈闭进行靶中标定 ,验证效果良好

Positioning the Center of a Subtle Oil-Gas Trap by Grey BP Neural Network

It is put forward to combine gray system with BP neural network for positioning the center of a subtle oil gas trap. The field source formation emitted by a oil gas trap is taken as gray information, and the information is optimized by comprehensive relational filtering method in order to reveal the anomaly of the subtle oil gas trap, to predict its range and shape, and to position its center. To use the method for positioning the center of a subtle oil gas trap can effectively overcome the instability caused by "contradictory samples", increase the recognition capability of subtle oil gas traps, and enhance the intelligent degree, rationality and accuracy of positioning the centers of them. Three oil gas traps are recognized in Tuoku area, their centers are also positioned. The predicted center of No.1 oil gas trap is very close to the position of Sha 4 well in the area, from which high yield oil gas is developed. This proves the reliability of the method.