Prediction of drill penetration rate in drilling oil wells using mathematical and neurofuzzy modeling methods

Abstract

The sustainability of oil and gas well drilling operations is a vital issue from an economic and safety point of view. One of the practical methods to predict the continuation of the drilling operation is the analysis of the penetration rate of the drilling operations. Statistical and experimental studies have shown that the drilling depth, the drill weight, and the rotary table's rotation speed are the parameters that directly affect the penetration rate. However, due to their mutual influence on each other and the occurrence of the sticking-sliding phenomenon, predicting the penetration rate in the appropriate range is a complex and challenging issue. In this paper, using the field data of oil and gas wells located in the southwest of Iran, the effect of these parameters on each other was studied in the framework of statistical modeling, neural networks, and neural fuzzy systems. Using the ANFIS method and with a correlation coefficient of over 90%, a successful four-dimensional model of drill penetration rate was presented in terms of changes in drilling depth, the weight of the drill, and the rotational speed of the rotary table. By analyzing this model, it was found that the penetration rate of the drill decreases with the increase of the drilling depth, and to keep it within the acceptable range, the rotational speed and weight of the drill should be changed based on the presented policy.