An integrated prediction and optimization method for drilling rate of penetration

Abstract

The rate of penetration (ROP) is a crucial indicator in drilling engineering and has been the focus of interest for decades. Previous researches established many models in terms of rock physical and mechanical properties, drilling technology and equipment capabilities, drilling fluid rheological properties, and drilling engineering parameters. However, most of the above models are based on engineering experience and logical reasoning. In this paper, we investigate drilling modeling and optimization based on machine learning and intelligent optimization algorithms. The ROP modeling is achieved using a BP neural network model, which is further applied in the optimization of weight on bit (WOB) and revolution rate (RPM). The results demonstrated that the optimized operational parameters, as recommended by WOB and RPM, can increase the drilling efficiency by 12.3%. This methodology can be further applied in actual field operations as suggested parameters to the drilling efficiency further.