Machine Learning Applications in the Petroleum Industry

Abstract

Machine learning tools have been used in different engineering-related aspects and the energy sector is no exception. This chapter focuses on summarizing the applications of machine learning tools in different areas of the upstream petroleum industry, that is, exploration and production of hydrocarbons, viz., oil and gas. This chapter begins by introducing commonly used machine learning tools and their applications in the energy sector. This is followed by a brief review on the applications of these tools in the petroleum industry such as for exploration of hydrocarbons, wellbore drilling, rock and fluid properties estimation, hydrocarbons production optimization, forecasting, and production maintenance through applications of waterflooding. Data preparation process which is very critical for successful utilization of machine learning tools is also presented including commonly used preprocessing techniques. In the last sections of this chapter, application of machine learning tools in three areas of upstream petroleum industry is discussed. These areas are: prediction of the borehole drillability, rock petrophysical properties of porosity and permeability, and hydrocarbon production forecasting and optimization. Different machine learning models employed in these three areas are summarized and their performance in arriving at the solution is reviewed