Research On Wells And Layers Selection And Tapping Potential Of Water Flooding At Extra High Water Cut Period

Fracturing is one of the important means for the development of low permeability reservoir, the better the effect of oil exploitation makes this measure has been used until now. However, with the continuous development of petroleum reservoir, the increase of water content leads to the poor effect of fracturing measures, the implementation of petroleum reservoir fracturing measures need to be more targeted and the fracturing effect is predictable. In this paper, a lot of research work has been done on the optimization of fracturing well layer and block productivity prediction in the six block of apricot.Based on the analysis of the production data and geological data and a lot of literature research on the fractured well in the six area of apricot, the index system of fracturing influence factors is preliminarily established, and each of the selected indicators were analyzed, combined with the theory of correlation analysis between the fracturing effect and influence of various factors. Using the multiple correlation coefficient method to determine the main control factors of fracturing effect, and the results of grey correlation degree analysis are used to extract the sensitive factors of fracturing effect, which is an important foundation for the subsequent work. The well layer selection is mainly based on fuzzy comprehensive evaluation, and because of the many factors affecting the well layer, the data processing workload is big, it is very necessary to compile the software which can deal with the data in batch process. According to the principle of fuzzy comprehensive evaluation, the database software for the selection of well and layer for fracturing is worked out, and the well level evaluation is completed, and the hierarchical evaluation chart is plotted to determine the final evaluation level range. For evaluation of fracturing wells, select FI=0.42, 0.55 two data points as the demarcation point of evaluation index, evaluation of reservoir level, for evaluation of fracturing layer, select FI=0.45, 0.57 two data points as the demarcation point of evaluation index, evaluation of reservoir level to select an ideal fracturing layer. Comprehensive consideration of the influence factors of fracturing layer after the evaluation of fracturing layer and reference to the actual experience in the field, and the ideal layer screening parameter range of the well layer fracturing measures is determined finally. Block capacity prediction is based on the factor analysis. Through factor analysis eliminate correlation between variables is caused by the information overlap, extraction common factor at the same time, to achieve the purpose of dimensionality reduction, convenient for the calculation of subsequent. Support vector machine(SVM) data mining on the existing data, combined with contour map repeatedly mediate the key parameters, the relative error of the 27 groups is less than 10%, the average relative error is 12.73%. The average relative error of BP neural network prediction is 15.94%, and the relative error of multiple linear regression is 28.57%. SVM has a strong and stable learning ability. Support vector machine(SVM) data mining based on existing data, combined with contour map repeatedly to mediate the key parameters, determine the optimal parameter combination to get the optimal support vector machine measures to increase oil content prediction model, and the prediction model is analyzed, and the applicability of the prediction model is judged. The grey prediction model is established for the production of the fractured well layer, and the relevant parameters are optimized to obtain the optimal forecasting model.Through the error analysis of the above capacity prediction method to judge the practicality of productivity prediction model.