HIGH PERFORMANCE COMPOSITIONAL 3D RESERVOIR SIMULATION IN CONJUNCTION WITH UNSTRUCTURED GRIDS
Ivens da Costa Menezes Lima1; Anthônio Nunes Moreira Netto1; João Henrique Bessa Gomes1; Edilson Pimentel Drumond Filho1; Joaquim Bento Cavalcante Neto1; Kamy Sepehrnoori2; Francisco Marcondes1
1 Universidade Federal do Ceará; 2 The University of Texas at Austin
doi:10.20906/CPS/CILAMCE2017-0555
Resumo
Large-scale simulation projects involving hundreds of thousands of gridblocks require application of parallelization techniques in order to achieve practical computational times. There are essentially three ways of parallelizing a simulator by using distributed memory, shared memory, and a combination of the two mentioned approaches. This work is based on the first approach, where the domain is divided among the processors involved and each process is responsible only for its portion of the computations. This approach has two main advantages; it reduces the required memory per process and allows the simulations to be carried out using clusters with large number of processors. In this work, we used open source libraries to partition the computational domain, manage the grid information between the processors, and solve the linear system of equations generated from the discretization of partial differential equation modeling fluid flow in the reservoir. ParMetis (Parallel Graph Partitioning and Fill-reducing Matrix Ordering) is used to partition the computational domains, FMDB (Flexible Distributed Mesh Database) is responsible to manage the grid information between the processors, and PETSc (Portable, Extensible Toolkit for Scientific Computation) solves the linear systems of equations. The numerical approach is based on the Element based Finite Volume Method (EbFVM) in conjunction with unstructured grids. In fact, the main challenge was to manage the grid, fluids, and reservoir data set in such a way that the communications between the processors were reduced. We used an in-house compositional, multicomponent/multiphase simulator called UTCOMP, which was developed at The University of Texas at Austin, in order to perform this implementation. We show that the EbFVM adapts well for modeling reservoirs with complex geometries and also we present its performance in a parallel mode. The results are presented in terms of oil and gas production curves, speedup curves and CPU times for various case studies.
Palavras-chave: Parallelization; EbFVM; 3D unstructured grids; UTCOMP simulator