Interpretation method of multiple array production logging based on machine learning in a horizontal well

Abstract Interpretation of production logging in oil–water two-phase flow wells is challenging, especially for horizontal wells. The complexity of the distribution of phase and velocity field of multiphase fluid has caused great problems in the measurement of traditional production logging tools, resulting in low interpretation accuracy. To address the challenges, an intelligent production profile interpretation method for oil–water two-phase flow in horizontal wells is proposed. This method uses the multiple array production suite instrument and machine learning algorithm as the core to obtain the production contribution of each layer in the oil well. First, the probability density function method is used to determine the characteristic parameters of capacitance array tool (CAT) data. Then, a support vector machine is used to identify oil–water two-phase flow patterns based on these characteristic parameters. Finally, on the basis of determining the flow pattern, an artificial neural network is used to predict the phase-specific flow rate of oil and water based on the characteristic parameters such as the spinner array tool and CAT. The validity and accuracy of this method are verified by empirical analysis of experimental data. Research shows that the method has good application prospects, and can be adjusted for stable reservoir production and development to provide strong technical support.