关键词: 模型, 测井, 神经网络, 地层破裂压力

Abstract: Of the mew information processing techniques, both grey mode and artificial neural network mode can be used to solve the problems of pattern- recognition and parameter calculation in fine log interpretation and reservoir description. On the basis of GM(U,N)mode and BPANN mode, this paper introduces an improved method (GMUNBPANN) to estimate parameter, and then analyzes the feasibibity of formation fracture pressure prediction by using well logging data. Taking the central gas field in H basin as an example, we have built a special grey BPANN mode to calculate the rock fracture pressure of Ahihezi andd Majiagou formation by making full use of test data and log information, and applied it successfully to 25 oil-gas wells. The result shows that this method is easy, practical and can make an accurate prediction, thus providing a new approach for the prediction of formation fracture pressure from well logging curves.

Key words: Mode, Logging neural network, Formation fracture pressure