An efficient parallel-computing method for modeling nonisothermal multiphase flow and multicomponent transport in porous and fractured media

This paper presents an efficient massively parallel scheme for modeling large-scale multiphase flow, multicomponent transport and heat transfer in porous and fractured reservoirs. The new scheme is implemented into a parallel version of the TOUGH2 code and the numerical performance is tested on a Cray T3E-900 (a distributed-memory parallel computer with 692 processors) and IBM RS/6000 SP (a distributed-memory parallel computer with 3328 processors). The efficiency and robustness of the parallel-computing algorithm are demonstrated by completing three over-million-cell simulations using site-specific data for site characterization. The first application is the development of a three-dimensional unsaturated zone numerical model simulating flow of moisture in the unsaturated zone of Yucca Mountain, Nevada; the second problem is the modeling of flow of moisture, gas, and heat at the same site. The third application is the study of flow-focusing phenomena through fractures for the same site. Simulation results show that the parallel-computing technique enhances modeling capability by several orders of magnitude in speedup of computing times for large-scale modeling studies.