Petroleum Reservoir Permeability Prediction based on Principal Component Analysis and Support Vector Regression

By studying the well-logging technology and basic characteristics of the petroleum reservoir together with the well-logging parameters which have impact on reservoir permeability, a reservoir permeability predicting method is introduced, based on Principal Component Analysis(PCA) and Support Vector Regression(SVR) according to the traditional predicting methods of reservoir permeability in this paper. PCA is used to reduce the dimensionality of well-logging parameters through which the optimizing relevant parameters of permeability are chosen. Then, the optimized relevant well-logging parameters are imported into SVR in order to predict reservoir permeability. The experimental results show that the extracted parameters through PCA have a good relevance with the reservoir permeability, and SVR has also a higher accuracy in predicting permeability. It has been shown the strengths and practical values of PCA and SVR for the prediction of petroleum reservoir permeability.