Estimation of the thermal conductivity of a heterogeneous CH4-hydrate bearing sample based on particle swarm optimization

Abstract Knowledge of the thermophysical properties of methane hydrate-bearing sediments (MHBS) plays a critical role in the evaluation of fluid production potential from gas hydrate accumulations in geologic media. The process of estimating the values of key flow and transport parameters related to MHBS is of great importance. In this paper, we developed an integrated optimization workflow by integrating a global optimization algorithm, specifically particle swarm optimization (PSO) with the TOUGH+Hydrate v1.5 hydrate simulator. The optimization technique was employed to estimate the composite thermal conductivity of MHBS (λMHBS) through history-matching the simulated results with the experimental observation over time automatically in a process of stepwise heating an aqueous-rich heterogeneous MH-bearing core sample (SH = 0.40 and SA = 0.57) at P = 7.0 MPa. Two different composite models were tested and compared for their performance in predicting the temperature response. A sensitivity study was conducted to provide guidance on the selection of PSO parameters to accelerate the convergence speed. The optimal λMHBS obtained from the Bejan linear model is 2.16 W/mK with an absolute average deviation (AAD) of 2.66%, which is superior than the Somerton nonlinear model. Based on the results from the sensitivity study, the coupling of PSO algorithm with T+H v1.5 was proved to be robust and reliable. It is found that a particle size over 20, a maximum velocity coefficient around 0.5 and a well-defined parameter range guided by literature improves the efficiency in searching the optimal solution, achieving a global best cost with fewer iterations. The optimization procedure developed in this paper can be adopted in estimation of other groups of parameters related to MH-bearing systems (i.e. thermal, flow, geomechanical properties and the kinetic rate parameters of MH reactions) and in the optimization of production strategies from MH reservoirs (e.g. well positioning, pressure drawdown level and rate, etc.) to improve the energy efficiency of the production process.