The shear wave velocity information is very important for the precise prediction of low permeability reservoirs in offshore Paleogene fan delta.However,due to the limita-tions of cost and acquisition technology,effective acquisition of shear wave velocity informa-tion has become one of the urgent problems to be solved in the fine characterization of low permeability reservoirs.Meanwhile,due to the diversity of low permeability reservoirs and the complexity of sedimentary environment,lead to deficiencies in predicting small samples shear wave velocity based on the data-driven empirical formula method and the physical-law-driven petrophysical model method.Therefore,the research on machine learning algo-rithm based on Gaussian process regression is carried out.The algorithm has the advantages of small training data demand,high prediction accuracy and uncertainty evaluation of results.Taking the log data curve of Paleogene low permeability area in Bohai M oilfield as the ap-plication object,the results show that the machine learning algorithm based on Gaussian pro-cess regression can quickly predict the elastic wave velocity,and can realize the quantitative analysis of the uncertainty of the prediction results.
关键词
低渗储层/高斯过程回归/小样本机器学习/弹性波速度预测/储层分类
Key words
low permeability reservoir/Gaussian process regression/small sample machine learning/prediction of elastic wave velocity/reservoir classification
维普
万方数据