Well Temperature and Pressure Profiling Using Machine Learning

Abstract

The ability of wells to preserve the zonal isolation of geologic formations and avoid fluid (native or injected) transfer between those formations can be referred to as well integrity. The most often noted physical characteristics in boreholes are temperature and pressure. It is discussed how pressure well testing and temperature well testing differ and are comparable. This paper's major objective is to use machine learning to profile the well's temperature and pressure based on various variables. A good profiling of well temperature and pressure is done through the use of different elements during well testing, including thermal conductivity, formal temperature, and contact resistance. A method has been established for calculating the thermal conductivity of the formation, the beginning temperature, the skin factor, and the thermal resistance of the contact. The model used in this study therefore exhibits excellent agreement with the field collected data. The results of this study can therefore be used to more accurately compute the wellbore temperature and pressure parameters in challenging well settings.