A Newly Developed Drilling Rate Model Optimizes Drilling Efficiency Using Artificial Intelligence

Abstract

Abstract Rate of penetration is defined as the depth of the drilled rock per time unit while drilling. Alternatively, drilling rate or rate of penetration (ROP) is the speed of drilling. It is usually stated as drilled footage per hour (ft/hr). ROP has a major impact on the cost optimization and accelerating delivery of wells. ROP is a critical factor in calculating cost per foot in each well. ROP optimization ensures efficient mud solid equipment by controlling the mud control system and its compatibility with drilling cuttings. It is also valuable to optimize the ROP without jeopardizing the well drilling performance order to avoid the inducing drilling problems. The objective of the introduced paper is to present a new rate of penetration model that combine the mechanical parameters, bit and rig hydraulics, and drilling fluid rheological properties. The model has been validated in the field, and compared with offset wells. The results has shown improvement of well drilling performance and safe optimization. Many of ROP models were developed before such as ROP model based on drilling fluid parameters, surface drilling parameters, or machine learning models or on bit design modifications models. Noticeably, those models have not been validated, and not feasible for drilling operation. The new ROP model combines several factors, has been validated, and designed for drilling operation. Mechanical parameters, rheological properties, bit & rig hydraulics were collected for the same field, hole section, and mud type to ensure the compatibility of the developed model with applications of ROP optimization for the given field. The relationships between ROP with mechanical parameters, drilling fluid rheological properties and bit and rig hydraulics were analyzed and evaluated to conclude the best and strongest relationships. The optimization was used and implemented by including valuable and novel technique that enabled the developed model to perform significant improvement in comparison with drilled offset wells. The result showed the validation and comparison values of measured ROP was 95 % of regression which is a high value of squared regression of R2. By comparing the optimized ROP in real time with offset wells’ ROP, 50% ROP improvement was recorded. The developed model can tell the user the optimized drilling parameters, proper drilling fluid properties, enhanced rig & bit hydraulics that can lead for comparable and proper optimization without jeopardizing the well drilling performance. The developed model can be easily link to real time operations center (RTOC) to ensure real time monitoring model and ROP analyzer system. The new model can be compiled with other artificial intelligent tools in real time.