The Optimization Of Fracturing Wells And Layers And The Study On Productivity Prediction Model In XingLiu District

The fracturingeffect for thetransformation oflow permeability oil and gas reservoir and the fieldsin thelater stage of high water cutis very obvious. With thecontinuous development of oil and gas fields,reservoir waterrate increased year by year and the fracturingeffect is worse year by year. In order to make sure the good effect and economic benefit that the fracturing transformation can bring, it is essential important for the fracturing whether it will succeed effectively that how to select the candidate wells and layers which has fracturingpotential and predict oil increased accurate. Based on the investigation of the large number of literaturesin home and abroad about the fracturing wells and layers selection andafter the measures productivity prediction model, combined with the data from the actual block Xingliu block in Daqing oilfield has beenfractured to study above problems.Based on the analyze of fractured wells production condition and the geological conditions in Xingliudistrict, carry out the single factor analysis and multivariate factors analysis to influencing effect of fracturing wells and layers.Multiple factors analysis mainly use complex correlation coefficient method, according to the results of correlation coefficient, removing high correlation influence factors and determining the main factors affecting the fracturing effect.At the same time make the statistical analysis on fracturing wells and layers data in Xingliu district, preliminarily determine the main affecting factors scope which the fracturingeffect is better, optimization comprehensive use fuzzy comprehensive evaluation method to optimize the ideal fracturing wells and layers. Using the principal component analysis method to establish the principal component models of fracturing wells and layers, through regressingthe curve between comprehensive principal component value with fracturing increased oil production to get the equation between increased oil productions with main influence factors for the prediction. Using cluster analysis to divide thefracturing wells and layers into 5 categories and analysis these five kinds of clustering results for discriminant analysisverification, the qualified validation results shows that the new candidate fracturing wells can be classifiedaccording to the result of discrimination analysis.According to the classification results of clustering analysis, combining the target block has been fractured wells data to make the multiple linear regression and multivariate nonlinear regression and make the prediction model of incremental oil production of fracturing.Meanwhile by BP neural network training on fracturing wells and layers data to get BP neural network prediction model of fracturing oil increment. According to the error analysis results, compared the three methods, with the minimum error as the standard to determine the productivity model of fractured wells and layers.