Research On Artificial Intelligence Interpretation Methods For Oil-water Two-phase Flow Production Profile Logging In High Water Cut Condition

Production profile logging is an important method for monitoring the dynamics of oil and gas wells,which can reflect the production status and reservoir characteristics of oil and water wells.However,the oil field enters into high water-bearing stage,the oilwater two-phase flow presents complexity and diversity,and the traditional output profile logging interpretation method often has large errors and limitations.By analyzing the current status of the output profile logging interpretation method,an oil-water two-phase output profile logging interpretation method based on artificial intelligence algorithm is proposed,aiming to improve the accuracy and efficiency of the output profile logging interpretation.The flow pattern is an important factor to determine whether the output profile logging interpretation is accurate or not.First of all,according to the numerical simulation of oil-water two-phase flow and oil-water two-phase production logging simulation test experiment observation to study the effect of flow rate and water content on oil-water two-phase flow pattern,and clarify the characteristics of high water-content oil-water two-phase flow.Experimental data from production wells are collected to provide a data base for flow pattern identification using artificial intelligence algorithms and output profile logging interpretation.The results of the artificial intelligence algorithm flow pattern identification and the experimental results basically agree,and a new idea is proposed for the oil-water two-phase flow pattern identification.Finally,a new method for output profile log interpretation is investigated.The output profile logging data are processed using artificial intelligence algorithms(including BP neural network,GA-BP neural network,and conditional generation adversarial network).These algorithms can automatically learn the characteristics and laws of the data,divide the training and testing sets with the production well experimental and logging data as samples,input the output profile logging interpretation influence factors,and output the predicted fractional phase flow rate.The prediction results show that the proposed artificial intelligence algorithm-based output profile logging interpretation methods can significantly improve the efficiency and accuracy of fractional phase flow rate prediction.Among them,the conditional generation adversarial network has the highest prediction accuracy.Using flow type classification as a guide,sample data from production well experiments and logging data under different flow types were used to predict the fractional phase flow rate,and the predicted flow rate of sample data not classified by flow type was compared and analyzed,and the results showed that the flow rate accuracy of the new artificial intelligence output profile interpretation method based on flow type identification was higher,which further improved the accuracy of the output profile logging interpretation.It is concluded that the artificial intelligence algorithm based on the high water-bearing oil-water two-phase output profile logging interpretation method has better applicability...