Nuclear magnetic resonance (NMR)-based interpretation models are commonly calibrated using laboratory ambient core NMR measurements. For applying the core calibrated models to downhole NMR logging interpretation, the difference between the NMR responses measured at ambient and reservoir temperature needs to be evaluated. The temperature dependence of NMR relaxation time in high-quality (HQ) carbonate reservoirs has been studied, and NMR temperature dependence models were established using data analytic methods. In this paper, we extend our early studies on temperature dependence of NMR relaxation time to low-quality (LQ) carbonate formations.
For more than 95% of the LQ samples investigated, NMR relaxation time shows a positive correlation with temperature. The correlation is similar to that observed in HQ carbonate rocks but slightly less significant. Temperature-dependent correlations for predicting the geometric mean of NMR transverse relaxation time (T2,GM) from a measured temperature to any other temperature were derived from HQ to LQ carbonate rocks independently first, then a unified T2,GM correlation was derived including both the HQ and LQ carbonate reservoirs. Predicting NMR transverse relaxation time T2 distribution from one temperature to other temperatures was achieved using a dimension reduction approach involving the principal component analysis (PCA) technique. It was found that the T2 distributions at any given temperature for both HQ and LQ carbonate reservoirs can be predicted robustly from the T2 distributions at the ambient temperature by representing the T2 distributions with principal components (PCs) at the ambient temperature and then using these PCs to predict the PCs at a different temperature. The optimal number of PC components depends on the multimodality of the T2distribution.
This work extends the validity range of the data analytic methods, in particular parameter and dimension reduction methods, that quantify the temperature dependence of carbonate NMR properties. The new NMR temperature model enables the integration of NMR laboratory studies and downhole measurements for advanced petrophysical analyses in a wide range of carbonate reservoirs.