| Abstract Submission No. | ABS-4.1-0043 |
| Title of Abstract | Machine learning aided estimation of mineralogical elements from downhole NMR data for characterisation of gas hydrate reservoir |
| Authors | Amrita Singh*, Maheswar Ojha* |
| Organisation | CSIR National Geophysical Research Institute, Hyderabad |
| Address | 6 CHUNILAL ROY LANE B.C ROAD BURDWAN, West Bengal, India Pincode: 713101 E-mail: amrita.29oct@gmail.com |
| Country | India |
| Presentation | Oral |
| Abstract | Downhole nuclear magnetic resonance (NMR) data are useful for detecting and quantifying mineralogical constituents in sediments by analysing grain size distribution. However, because of the variety of hydrate occurrence patterns within the subsurface lithology, its application to gas hydrate reservoir classification remains difficult. The NMR instrument is unable to detect solid gas hydrates in pores and has a low amplitude in the NMR T2 distribution, which can lead to inaccurate subsurface lithology estimation. Therefore, based on grain size, we assess the precise amount of different lithologies in terms of clay, silt, and sand and propose a technique to recover NMR T2 distributions signals without gas hydrates from measured NMR signals (with gas hydrate). The site NGHP-02-05 was selected from Area C of the Indian National Gas Hydrate Program's second expedition (NGHP-02) in the KG offshore Basin. We estimate the volumetric percentages of sediment constituents using the supervised probabilistic automatic relevance determination (probabilistic-ARD) technique and employ the cumulative trapezoidal method to preserve NMR T2 distribution signals free of gas hydrates. When comparing medium (silt) to large (sand) grains to fine (clay) grains, the results reveal a considerable difference in the amplitude distribution, suggesting that the gas is mainly distributed in larger pores. The findings also indicate that silt and sand predominate over clay in the lithology. |
| Keywords | Krishna-Godavari Basin; NMR T2 distribution; lithology quantification; gas hydrate. |