Artificial Neural Network Approach to Predictive Modeling of Cuttings Transport Ratio in an Inclined and Horizontal Wells

Abstract

In recent times several drilling problems related to poor hole cleaning have been reported.  Problem such as a high torque and drag, poor hydraulics and pipe sticking have resulted into an increase in operational costs. To therefore achieve and successful drilling operation in horizontal and inclined wells; a good cutting transport ratio and higher hole cleaning efficiency are required in this is largely dependent on factors such as are fluid properties and rheology, size and density of cutting, pipe eccentricity, flow rate, rate of penetration, drill pipe rotation, average fluid velocity, and hole geometry and inclination. In this study, a new empirical correlation was established for the prediction of cuttings transport ratio from wellbore to the surface. The artificial neural network (ANN) coupled with the back-propagation (BPNN) learning algorithm technique was employed to develop the empirical correlation based on 405 experimental data. The input variables are made up of the  mud density and mud rheological properties  hole inclination angle, flow rate (GPM), drill pipe rotary speed (RPM) and size and density of cutting. The result obtained showed that the developed correlation using ANN technique could be applied in predicting the cuttings transport ratio with a high accuracy for average absolute errors (AAE) less than 0.5%.  Moreover, the correlation coefficients (R) was higher than 0.9 for both number of neurons but six number of neurons gave the higher correlation coefficient especially in testing stage as compared with three neurons. The application of this method is cost effective with zero use of auxiliary equipment, which will undoubtedly help in accurate estimation of hole cleaning operation; an important function in drilling operation.