Multi-parameter Fusion Intermittent Recovery Mechanism Prediction Model Based On Big Data

The most critical bottlenecks of oil recovery technology in low-permeability oilfields and conventional oilfields in the middle and later stages of development: insufficient pressure on the wells,severely insufficient fluid supply,and inability to continuously pump oil,resulting in rapid and ineffective wear of machines,rods,and pumps.The current effective method to solve the above problems is intermittent oil recovery.A reasonable intermittent oil production mechanism can increase the output of oil wells and reduce the ineffective wear of machines,rods,and pumps.Existing intermittent oil recovery mechanisms often use single parameters to estimate the supply and discharge relationship at the bottom of the well and determine the intermittent oil recovery mechanism.The feasibility of the implementation of intermittent oil recovery mechanisms depends on the accuracy of these measurement and calculation parameters,but these parameters must be obtained accurately It is very difficult,so large errors lead to poor practical application.In this paper,use big data technology,a multi-parameter prediction model of intermittent oil recovery mechanism is obtained.An intermittent oil production database was established based on the ACCESS database to facilitate the extraction and storage of oil production data for intermittent oil wells.Using the Pearson correlation analysis and curve regression analysis of SPSS software,an analysis model of the height of working fluid level and its influencing parameters was established.Based on the gray correlation method,the gray correlation degree of the working fluid level and its influencing parameters is calculated to obtain the main influencing parameters of the working fluid level;based on the support vector regression machine algorithm,data fusion is performed on the obtained main influencing parameters of the working fluid level.The height of working fluid level is predicted.Data mining was performed on the predicted working fluid level data,and the expressions of working fluid level change function in the intermittent and inter-pumping periods of the oil well were obtained;the oil recovery index was set for the purpose of maximizing the oil recovery efficiency of the oil well;based on the particle swarm search Optimal algorithm,using the oil well production index as the fitness function,to predict the best intermittent oil production mechanism of theoil well.Based on the predicted optimal intermittent oil recovery mechanism,intermittent oil recovery software was developed to facilitate the guidance of oilfield production.This study is the first to apply multiple intermittent oil production influence parameters to the setting of intermittent oil production mechanism,which can eliminate the error of single parameter inaccuracy and affect the setting of intermittent oil production mechanism.It provides a new intermittent oil production mode for low permeability wells.It has a certain guiding role in production practice.