神经网络储层预测技术在HG地区的应用

Application of neural networks reservoir prediction technique to HG area

利用人工神经网络储层预测技术，进行隐蔽油气藏预测，有成功的，也有失败的。预测中，如何正确进行时窗有效参数选择、样本建立、样本检验以及作图参数选择，将直接关系到储层预测的成败。我们利用人工神经网络储层预测技术，在HG地区进行隐蔽油气藏预测时，针对隐蔽油气藏的特点，探索出了一套在精细的储层标定及层位解释后，沿层进行时窗选择、样本建立、有效参数提取、样本检验及作图参数选择的工作方法，并取得了可喜的地质效果。

We have seen some successful and some failure examples for predicting the hidden reservoir with artificial neural networks reservoir prediction technique. In prediction, how to pick the window, to select the available parameter, to build the pattern, to cheek the pattern and to choose the drawing parameter, are related directly to success or failure of reservior prediction. In this paper. when we predicted the hidden reservoir in the HG area with artificial neural networks reservoir prediction technology according to the speciality of hidden reservoir, we explored a method in picking the window, selecting the available parameter, building the pattern, checking the pattern and choosing the drawing parameter along with the layer after reservoir identificated and interpretated. A satisfactory result has been achieved.