Predicting Formation Fracture Pressure Based on BP Neural Network
Presently, the theoretical models of calculating formation fracture pressure are generally used, but there are few statistical models that can obviously reflect the formation fracture pressure and the rock mechanics parameters. In this paper, taking into account the relation of formation fracture pressure and rock mechanics parameters, the four parameters which have a close connection to formation fracture pressure, such as Young’s modulus, bulk modulus, poison’s ratio and depth are selected. Thus based on the BP neural network, the statistical models of formation fracture pressure are established to calculate fracture pressure of carbonate formation. The research shows that BP neural network model is complex, and its model is difficult to set up, but the error of estimate’s formation fracture pressure is little, the precision is high.
KeywordsFormation fracture pressure BP neural network Rock mechanics parameter
Unable to display preview. Download preview PDF.
- 1.Stephen, R.: Prediction of Fracture Pressure for Wildcat Wells. JPT 34(4), 863–872 (1982)Google Scholar
- 2.Draou, A., Osisanya, S.: New Methods for Estimating of Formation Pressures and Fracture Gradients from Well Logs. SPE 63263 (2000)Google Scholar
- 3.Huang, R.: A Model for Predicting Formation Fracture Pressure. East China Petroleum Institute Journal 4(4), 335–346 (1984)Google Scholar
- 4.Zhou, N., Yang, Z.: Overview on Pressure Prediction of Formation Fracture. Chongqing Institute of Technology (Natural Science) 13(1), 36–39 (2011)Google Scholar
- 5.Qi, B., Xie, G., Zhang, S., et al.: Application of Fracture Pressure Log Interpretation Methods in the LG area. Nature Gas Industry 29(10), 38–41 (2009)Google Scholar
- 6.Li, M., Lian, Z., Chen, S., et al.: Rock Mechanical Parametric Experiments and the Research of Formation Fracture Pressure Prediction. Oil Drilling & Production 31(5), 15–18 (2009)Google Scholar