Options
Simultaneous Estimation of Multiple Hydrate Core Characteristics from a Production Time-Series Using Coupled ANN–GA Methodology
Journal
Natural Resources Research
ISSN
15207439
Date Issued
2022-06-01
Author(s)
Singh, Raghvendra Pratap
Vishal, Vikram
Abstract
Experimental retrieval of hydrate-core properties is limited due to scarcity of samples, difficult operating conditions, and the requirement of several empirical studies, which make core characterization an expensive and highly time-consuming process. In a first attempt, this study employed a coupled ANN–GA framework to estimate multiple parameters simultaneously from a production time-series obtained from thermally stimulated depressurization-based gas recovery. Five intrinsic hydrate-core properties (i.e., porosity, permeability, saturation of hydrate and aqueous phases, and effective permeability constant) and one process-dependent parameter (heat transfer coefficient) were retrieved in this study. A representative case study and sensitivity analysis were performed first to highlight the physical aspects of thermally stimulated depressurization and therein involved the key parameters. Next, the details of ANN development, validation, and inverse methodology were presented and discussed. The study concludes that ANN is a robust surrogate for simulating thermally stimulated gas production from hydrate cores. In addition, ANN–GA coupled methodology can simultaneously retrieve multiple hydrate-core characteristics from a production time-series with reasonable accuracy. All five intrinsic properties were retrieved with the highest accuracy, with below 5% error. However, the retrieved value of the heat transfer coefficient was less accurate than others, primarily it being a process-dependent parameter.
Subjects