Use this URL to cite or link to this record in EThOS:
Title: Towards a practical quantitative framework for repeated time-lapse seismic data interpretation
Author: da Motta Pires, Paulo Roberto
ISNI:       0000 0004 8499 497X
Awarding Body: Imperial College London
Current Institution: Imperial College London
Date of Award: 2019
Availability of Full Text:
Access from EThOS:
Access from Institution:
Time-lapse seismic data, also referred to as 4D seismic, occupy a central position in reservoir management. This research pursues a practical, expedite framework to generate quantitative information even from noisy time-lapse seismic datasets, provided that they support consistent qualitative interpretations. The proposed method takes advantage of repeated time-lapse seismic data, which, however in a discrete fashion, can potentially capture the fluid flow dynamics in the reservoir. We reconstruct the continuous expansion of time-lapse seismic anomalies (visual delimitation of bright-spot regions) by performing the geometrical interpolation of their boundaries. Our method relies solely on the shapes of and on the extensions covered by these boundaries, minimising the importance of the magnitude of the time-lapse seismic signal. This interpolation employs the Fast-Marching Method and trajectory (pathline) tracing to estimate the anomaly boundary arrival times in the reservoir. Combined with the classic theory of immiscible displacement, these times generate quantitative information, namely water saturation maps, without resorting to rock-physics modelling or to time-consuming reservoir modelling workflows. Although we target real applications, we validate our method with a synthetic dataset, which realistically represents a waterflooded sandstone reservoir. The estimated water saturations reveal close agreement with their actual levels. These results render our method at least as accurate as well adjusted reservoir simulation models, though requiring significantly less data and only a fraction of the time spent in conventional history matching. We also discuss potential gains of our method if supported by additional reservoir information and if integrated to reservoir modelling routines.
Supervisor: King, Peter Sponsor: Petroleo Brasileiro, S.A.
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral