Artificial Neural Network Model for Predicting the Equivalent Circulating Density from Drilling Parameters

The equivalent circulating density (ECD) is considered a critical parameter during the drilling operation. The ECD measurement should have a high degree of accuracy not to cause well control problems. The practical way for measuring the ECD is costly and the other alternative way for ECD estimation from the mathematical methods provides low accuracy. The main goal of this study is to employ machine learning techniques for predicting ECD from only the surface drilling parameters without any downhole measurements. The study provides ECD predictive model by using an artificial neural network (ANN). The data used was collected during drilling horizontal section that covered a wide range for ECD and drilling parameters (3,570 points). The model was trained, tested, and optimized to provide high accuracy prediction for ECD. The results showed an overall strong ECD prediction with a correlation coefficient (R) greater than 0.99 and an average absolute percentage error (AAPE) less than 0.24%. Furthermore, the model was validated with an unseen data set and proved the high accuracy performance level (R of 0.98 and AAPE of 0.3%) that enhances the model application in the practical drilling operations that will save extra cost and time.