Improved Correlations for Heavy-Oil Viscosity Prediction with NMR

Abstract

Delineation of the viscosity in heavy-oil reservoirs is viable for heavy-oil exploration and production activities. Low-field nuclear magnetic resonance (NMR) relaxometry has proven to be a powerful tool for a downhole assessment of heavy-oil viscosity. This prediction is derived from the sensitive relationship between viscosity- and NMR-measured properties, such as spin-spin (T2) and spin-lattice (T1) relaxation. In this paper, we demonstrate the versatility of NMR relaxometry by presenting three different approaches for obtaining viscosity correlations that incorporate T1, T2, and hydrogen-index (HI) data as a function of temperature and five inter-echo spacings (TE) to account for the signal loss of fast-relaxing components in heavy oil. The results of this work show that NMR-relaxometry measurements correlate very well with heavy-oil viscosities and that each methodology demonstrates reasonable success with heavy-oil predictions. In addition, these relationships demonstrate substantial correlation improvement, especially in comparison with other published relationships that are also tested. Lastly, insight regarding heavy-oil-sample storage time and emulsion is briefly discussed to emphasize the significance of fluid-sample handling within the scheme of heavy-oil viscosity prediction and core/fluid-analysis protocols. Viscosities predicted with a correlation derived from aged and emulsified samples were observed to be within a 17% error in comparison with those predicted with a correlation developed from recently collected heavy oils. From a well-logging point of view, the uncertainty introduced by using aged samples for the interpretation of well-logging data is not considered excessive and may be correctable. These results give new and valuable insight into the considerations that are necessary for sufficient core and fluid analysis.