Use of Reservoir Simulation Parallel Computing and Optimization Techniques to Accelerate History Matching and Reservoir Management Decisions

摘要

Reservoir simulation is used in several important tasks in thernpetroleum industry. Most of such tasks require manyrnsimulation runs either because there are uncertainties inrnreservoir characterization and the model has to be calibrated orrnbecause many options have to be tested in manual orrnoptimizing procedures.rnIn the past, computer time was one of the most limitingrnvariables but nowadays computers are becoming lessrnexpensive. On the other side, people time is becoming morernvaluable so they must have tools to make the best use of theirrnminds. Despite of these facts, many important decisions inrnreservoir engineering are still strongly based on manualrnprocedures even when computers, parallel algorithms andrnautomation tools are available. Furthermore, many companiesrndo not use their network adequately.rnIn order to contribute to a change in this scenario, thisrnwork shows (1) some advantages of building tools usingrnreservoir simulation, parallel computing and optimizingrntechniques to history matching and reservoir managementrndecisions, and (2) how to make the best use of existingrnnetworks and PVM (parallel virtual machine) to acceleraternsuch processes.rnA case study using an offshore field is presented tornillustrate (1) the advantages of having a computer program tornautomate some steps of production history matching and (2)rnthe time savings obtained by parallel computing. Somernprocedures are discussed to obtain the best balance betweenrnpeople and computer time.