New NMR Porosity Correction Algorithm for Steady-State Buildup Effect in Unconventional Reservoirs

Abstract

Abstract Nuclear magnetic resonance (NMR) logging for subsurface exploration is gaining acceptance as an everyday measurement for conventional and unconventional reservoirs, as well as new energy applications such as carbon storage and geothermal formation evaluation. Out of many NMR sensor designs, ones with a saddle-point magnetic field profile have gained broad acceptance because of their short echo spacing and many operational advantages. Despite bringing advantages, the saddle-point design can lead to systematic overestimation of porosity in combination with T1-T2 logging, which has been especially observed in unconventional reservoirs. We have developed a multiphysics tool model—a digital twin—that quantifies the recently discovered steady-state buildup (SSB) of the NMR signal in a multivalued gradient magnetic field. This effect arises mostly from off-resonance signals. The model incorporates the static and dynamic fields of the NMR sensors, associated physics, and the transmitter and receiver electronics. It is validated by laboratory experiments. This SSB effect has been found to be the key cause of the observed porosity overestimation, which is most significant in unconventional and carbonate reservoirs. The application of the digital twin has allowed us to improve the tool forward modeling accuracy by calculating the correction factors for each combination of T1, T2, pulse sequence, and environment. A new kernel-corrected NMR inversion has been developed. The novelty of the inversion includes not only a forward modeling correction, but also post-inversion correction for the bias caused by regularization and an additional procedure for the reduction of the artifacts particularly evident at short relaxation time constants. The overall improved inversion delivers accurate, higher-definition T1-T2 maps that were not attainable before. Validation of the new inversion has been performed on field datasets. Results demonstrate reconciliation of kernel-corrected NMR porosity with independent porosity measurements in unconventional and conventional formations. High-definition maps enable better segregation of different fluid types in unconventional reservoirs and delineation of producible intervals with increased clarity and confidence.