Browsing by Subject "Mineral joint inversion"
Now showing 1 - 1 of 1
- Results Per Page
- Sort Options
Item Combining spectroscopy elemental well logs, inverted nuclear logs, and high-resolution borehole images for improved mineral and petrophysical interpretations(2021-05-10) Eghbali, Ali; Torres-Verdín, CarlosEstimation of petrophysical and geomechanical properties of complex hydrocarbon reservoirs relies on accurate knowledge of the mineral composition of rocks. Laboratory technics such as X-Ray Diffraction (XRD), Quantitative Evaluation of Minerals using Scanning Electron Microscopy (QEMSCAN), and thin-section analysis commonly used for mineral identification and quantification are performed on a limited number of test plugs and investigate only a small volume of each core plug. For continuous well-log-scaled mineral identification, petrophysicists have relied on downhole gamma-ray spectroscopy logging tools. However, the vertical resolution of these well logs is relatively low, and the measurements are spatially averaged across laminated or heterogeneous rocks. Furthermore, standard mineral interpretations, which are obtained from commonly used element vs. mineral concentration empirical correlations, are shown to be inadequate in complex reservoir rocks. To overcome these technical difficulties, I implement a numerical interpretation technique that uses high-resolution borehole images, spectroscopy, and inverted standard nuclear logs that have mitigated shoulder-bed and tool effects, to identify the minerals that compose the rocks and estimate mineral concentrations and petrophysical properties. High-resolution images such as downhole micro-resistivity are used to detect the boundaries of petrophysical layers within the depth interval of interest. Standard logging tool responses are inverted into piecewise constant earth model properties using a Markov Chain Monte Carlo algorithm for each layer to correct for tool spatial averaging. I use existing but limited point-by-point surface mineral data from rock-sample XRD and/or QEMSCAN to build a database of mineral priors. Using stochastic inversion methods and the mineral database priors, I solve a nonlinear optimization problem that simultaneously determines mineral concentrations, porosity, and water saturation when fluid-sensitive well logs are available. Results obtained from the latter interpretation workflow are validated using synthetic field cases with core data.