– Campus Golm, Haus 9, Raum 1.10
Parameter Estimation from Noisy Observations of a Stochastic Heat Equation
Gregor Pasemann (HU)
Femke Vossepoel (Delft University of Technology, Netherlands)
Fluid extraction from a subsurface hydrocarbon reservoir results in compaction of the reservoir, in particular cases leading to subsidence and induced seismicity. Coupled flow-geomechanical models simulate the processes responsible for this and help to understand and mitigate the consequences. The objective of the presented work is to estimate the governing parameters of these models through assimilation of observations of surface deformation using ensemble-based assimilation methods.
Sensitivity studies evaluate the influence of the reservoir geology to the compaction of hydrocarbon producing reservoirs with a particular interest on the role of heterogeneities and the presence of faults. Starting from a simplified configuration, the complexity of the model is increased in a number of synthetic data assimilation experiments. In the more complex and non-linear simulations of fault behaviour, the use of ensemble-based particle methods is especially attractive.
The resulting data assimilation system can be used to assess the value of various sources of information and help to effectively balance the uncertain parameters in the subsurface. Ultimately, this not only provides the most likely description of the compaction-related phenomena, but also helps to better understand, predict and manage the surface effects.