Drill-bit Vibration as a Safe and Viable Replacement for Radioisotope-Based Well Logging Tools for Measuring Density, Porosity and Clay Mineral/Organic Matter Content

Sealed radioisotope sources containing radioactive materials have been used for decades in oil and gas well logging tools, primarily to determine the density, porosity, and hydrogen content, and, in some instances, elemental compositions, of the various rock types and their contained fluids. Two types of sealed radioactive sources are the most used: gamma-ray used by the formation density log and neutron source used by the neutron density and porosity logs. Both tools produce continuous, isotropic, non-interruptible radiation posing radiological security risks. Thus, it is essential to explore replacement technologies to determine subsurface properties without using radioisotope sources. The adoption of non-radioisotopic alternatives in the oil and gas industry has seen less success because the proposed alternatives did not prove substantial added value in terms of accuracy, cost, and applicability in severe drilling environmental conditions. Validation of these alternatives is a slow process since it requires running the new tool along with the existing tools in the same wells. Teverra proposed a new technology to address these shortcomings and answer the need for a nonradioactive logging tool. The technology leverages drill bit vibration data to accurately reproduce the density, porosity, and clay content logs. Since drill bit vibration data are widely available, the technology can be readily validated against the historical vibration and logs collected in existing wells. Phase I study established a proof of concept, through developing a machine learning model that uses vibration data measured while drilling, along with the electronic drilling recorder (EDR) data recorded at the surface, to generate the density, porosity, and clay content logs. The prediction accuracy obtained is promising and can be substantially improved in the proposed Phase II study using advanced signal processing and machine learning techniques. In combination with the pure data-driven approach, we envision developing a physicsbased model relating drilling dynamics data to the formation physical properties to further validate the proposed technology and ensure consistency between theoretical and empirical findings. The proposed logging technology based has a potential to obtain substantially higher resolution logsin real time, overcoming the typical problem of lagging behind the bit of the Logging-While-Drilling (LWD) data, improving well geosteering and formation characterization. This new technology has huge economic and technological benefits, while minimizing adverse environmental impact. In the upcoming Phase II of the project, we will conduct comprehensive numerical modelling to simulate the drilling process and study physical dependencies on the generated synthetic vibration data to steer the developments. Finally, we will develop and commercialize a software platform hosting this technology to provide real-time modeling services.