Stochastic Modelling of Thin Mud Drapes inside Point Bar Reservoirs with ALLUVSIM-GANSim
Abstract: Modelling fine-scale mud drapes inside point bars is critical to model fluid flow in complex sedimentary environments. These features have been modelled using deterministic or geostatistical modelling tools (e.g., object-, event-, and pixel-based). However, this is a non-trivial task due to the need to preserve geological realism (e.g., connectivity within channels), while being able to condition the generated models to point data (e.g., well-log data). Generative Adversarial Networks (GAN) has shown promise in large-scale geological facies modelling scenarios (e.g., braided river and carbonate reservoirs), but their potential for small-scale mud drapes is still largely unexplored. Here, we propose a geo-modelling workflow for fast modelling of conditional mud drapes which may exist in real world scenarios. Initially, improved ALLUVSIM produces realistic unconditional models of mud drapes along the accretionary surfaces, serving as GAN training data. GANSim is then employed for achieving geomodelling conditioned to secondary data. Thirdly, temporal pressure conditioning via Monte Carlo Markov Chain validates the quality of GAN-generated mud drapes. In synthetic data tests, the pre-trained generator exhibits ability for mud-drapes-feature extraction and conditioning diverse information in producing multiple realizations. An application example in a modern meandering river verifies method effectiveness and practicability. This workflow provides technical support for predicting mud drapes inside underground point bar reservoirs in the future.
Some references about PgGAN and GANSim
- Karras, T., Aila, T., Laine, S., & Lehtinen, J. (2017). Progressive growing of GANs for improved quality, stability, and variation. ArXiv Preprint, arXiv: 1710.10196.
- Song, S., Mukerji, T., & Hou, J. (2021a). Geological facies modeling based on progressive growing of generative adversarial networks (GANs). Computational Geosciences, 25(3), 1251-1273. https://doi.org/10.1007/s10596-021-10059-w
- Song, S., Mukerji, T., & Hou, J. (2021b). GANSim: conditional facies simulation using an improved progressive growing of generative adversarial networks (GANs). Mathematical Geosciences, 53(7), 1413-1444. https://doi.org/10.1007/s11004-021-09934-0
- Song, S., Mukerji, T., & Hou, J. (2022). Bridging the gap between geophysics and geology with generative adversarial networks. IEEE Transactions on Geoscience and Remote Sensing, 60, 1-11. https://doi.org/10.1109/tgrs.2021.3066975