Optimization of rate of penetration using artificial intelligent techniques

Rate of Penetration (ROP) is defined as the volume of rock removed per unit area (ft) per unit time (hrs). There are several published models to predict the rate of penetration; however, most of them focus on drilling parameters such as: string revolutions per minute, weight on bit, pumping rate. Only few researchers focused on the effect of mud properties and their influence on the rate of penetration values using few or little data. The objective of this paper is to develop a new model to predict the ROP based on both the drilling parameters and mud properties using artificial intelligent techniques (ANN). Actual field measurements of more than 3333 data points of different parameters were used to build an empirical ROP model. The obtained results showed that ANN model can be used to predict the ROP with a high accuracy (correlation coefficient of 0.99 and an average absolute percentage error of 5.6%). It is very important to combine both the mechanical parameters and the drilling fluid properties to predict the ROP. The developed ANN model for estimating the ROP outperformed all of the previous available correlations.