Productivity Predicting Model Of Tight Oil Reservoir Based On Flow Unit Classification

The tight sand oil reservoir is an important part of unconventional hydrocarbon resources,which has been paid more attention.The poor physical property and its complicated seepage mechanism make it hard to predict the productivity.Based on the analysis of Yanchang formation in Ordos Basin,the paper forms a productivity prediction system of tight sand oil reservoir,which can be used in the reservoir project design,by the hydraulic flow unit theory and engineering mathematical methods.Firstly,the R35 pore throat radius is selected to classify the flow unit,by the use of mercury injection experiment results,and the cored intervals were classified into four classes.PERM?POR?SH and AC well log data,which have strong correlation with R35,are selected as flow unit division parameters used in the uncored intervals.A BP neural network model was constructed and used for learning and training data from cored intervals.The model is used for hydraulic flow unit prediction when it reached the preset accuracy.So the classification can be extended to the uncored intervals,and a geological flow unit model of reservoir can be obtained.Then,in order to predict the productivity of tight reservoir accurately,the key control factors of each flow unit are obtained by grey analysis method,based on the geology and stimulations factors analysis.A fractured vertical well productivity prediction model considering threshold pressure gradient and stress sensitivity is constructed,and then developed into four,combined with the key control factors and the threshold pressure gradient equation.The calculation results show that the models are reliable and applicable.At last,combined with the prediction model and recognition of geological characteristics,favorable areas of productivity of each formation are obtained.And they can be useful for the production adjustment.So the model is meaningful for the production and development.