Development of a simulation-optimization model for multiphase systems in the subsurface: a challenge to real-world simulation-optimization

摘要

The main purpose of this paper is to demonstrate the capability of a new simulation-optimization model especially tailored to investigate the optimal management strategy of a closed coal mine in the Ruhr, Germany. This paper deals with the multiphase/multicomponent flow simulation; the optimization model (simulated annealing); the mesh generation function; the coupling of them; and the use of a parallel computer. Firstly, a mesh generation function is included in the total procedure for the modelling of complex system configurations often required when the real-world problem is dealt with. The multiphase/multicomponent flow simulator can simulate not only groundwater flow and a tracer in it but also the multiphase systems (e.g. gas-water, gas-water-NAPL system). Moreover, a parallelization strategy for the optimization procedure is proposed and implemented to overcome the enormous CPU time problem always tagged to real-world simulation-optimizations. This strategy succeeded in enhancing the efficiency of the overall procedure almost linearly by the number of the processors in a parallel computer. This model is then applied to study how to install the passive extraction wells for controlling the migration of methane continuously desorbed from coal seams inside the closed coal mine in the Ruhr, Germany. The general rule proposed as the result of the application is rather simple although it is considered very useful in many practices of coal mining operations. This paper briefly outlines the overall procedure.