Faster convergence in seismic history matching by efficient parameter searching

Seismic history matching is the process of modifying a reservoir simulation model to reproduce the observed production data in addition to information gained through time-lapse (4D) seismic. The search for good predictions requires that many models are generated particularly if there is an interaction between the properties that we change and their effect on the misfit to observed data. In this paper we introduce a method of improving search efficiency by estimating such interactions and partitioning the set of unknowns into non-interacting subspaces. We use regression analysis to determine the groups which are then searched separately but simultaneously with an adapted version of the quasi-global stochastic neighbourhood algorithm. We have applied this approach to the Schiehallion field, located on the UK Continental Shelf. The field model contains a large number of faults that affect flow at different times during production and their transmissibilities are highly uncertain. We apply history matching by modifying a model supplied by the field operator. We find that we can successfully represent the misfit function as a second order polynomial dependent on changes in fault transmissibility. Firstly, this enables us to identify the most important faults and, secondly, to modify their transmissibilities efficiently by searching sub-groups of the parameter space. Once the regression analysis has been performed we reduce the number of models required to find a good match by an order of magnitude. We also f use the regression model to quickly sample the parameter space in minutes to estimate parameter uncertainties. By using time-lapse seismic to effectively condition saturation and pressure changes in history matching we have gained a greater insight into reservoir behaviour and have been able to predict flow more accurately with an efficient inversion tool. We can now determine unswept areas and make better business decisions.