Inverting for the petro-elastic model via seismic history matching

Optimisation of well performance and identification of unswept regions is a key goal in reservoir management. This may be realised by quantitatively integrating time-lapse (4D) seismic with reservoir modelling and simulation to improve the knowledge of a reservoir's current state and possible future behaviour. To address this, we have developed an automated method of seismic history matching where simulation models are converted into predicted acoustic impedances, using a petro-elastic transform and suitable rescaling, and compared to equivalent observed data together with more conventional dynamic well data. Our method is applied to the Schiehallion UKCS field. We update the reservoir permeability and the petro-elastic transform parameters simultaneously to improve the match to observed seismic and dynamic well data. We find that shale requires a separate pressure response to improve the history match.