| dc.description.abstract | This thesis consists of several projects related to time-lapse seismic data from the Norne field.
For the first part of this thesis, I have investigated methods for integration of the time shift data and the seismic reflection data. The time shifts were incorporated in an acoustic impedance inversion workflow, to update the monitor version of the low frequency model. This was compared directly with acoustic impedance inversion without the low frequency model update. The results indicate that the low frequency model update had a large effect, but made the interpretation of the oil water contact more challenging. The low frequency model update is therefore regarded as undesirable in this specific case. In general, the lesson learned is that the low frequency model update should be used with caution for reservoirs where both pore pressure and saturation effects are present.
A workflow for direct inversion of pore pressure and water saturation increase has been developed. The main assumption is that the time shifts are linked to reservoir pore pressure increase, and the water saturation increase is linked to the change in acoustic impedance. For the estimation of pore pressure increase, the second important assumption is that the pore pressure is below the threshold pressure for velocity sensitivity at the time of the base survey. By using a published pressure dependent velocity model for the Norne field, a method for estimating the pore pressure at the time of the monitor survey is suggested. For the estimation of the water saturation change, a simple rock physics modeling workflow based on well log data is used. The proposed inversion scheme takes the uncertainty caused by rock variability, fluid properties and the non-linear relationship between the water saturation and the acoustic impedance into account.
For the second part of this thesis, the interpretation of the results from a time lapse imaging method, using seismic data from the Norne field is presented. This was the first real 3D dataset test for this method. The interpretation of the results indicates that the application of the method was not successful in this specific case. | en |
