Improved Seismic Inversion Using Dual-Sensor Streamer Technology

Dhelie, Per Eivind ¹; Reiser, Cyrille ²
(1) MultiClient, Petroleum Geo-Services, Lysaker, Norway. (2) Reservoir, Petroleum Geo-Services, Weybridge, United Kingdom.

Building a 3D reservoir model is a complex task which involves all the E&P disciplines. The reservoir geophysics part of it plays a key role for the seismic attribute extraction. Obtaining an accurate 3D model has historically been difficult, as conventional seismic data has a relatively narrow band width. Due to lack of low frequency information in conventional seismic data, both well log information and seismic velocity information are used as a-priori information to complement the missing low-frequency information. The low frequencies are crucial in order to obtain the correct absolute elastic properties. Therefore, the main challenge would be to have available seismic data with a much broader bandwidth. This would allow the derivation of reliable reservoir properties away from the well.

New dual-sensor streamer technology uses additional velocity sensors that combined with pressure sensors, delivers a broader spectrum data uncompromised by the receiver side notch. This allows the seismic streamer to be towed much deeper (> 15m) than conventional streamers thereby capturing more of the low frequency seismic signal. The broader bandwidth combined with less noise allows the dual-sensor data to use more of the recorded low frequency information in the inversion process rather than relying on well log and a-priori information. In short, more of the 3D reservoir model is based on actual seismic measurements rather than a-priori single location well log information extrapolated and flooded into the model.

This paper presents several case histories where inversion results are compared using both conventional and dual-sensor seismic data. Comparisons of these reveal that the low frequency information improves the reliability of the inversion process and the derivation of the reservoir properties. The reservoirs are better delineated (clearer top and base) which will have a significant impact when building the reservoir model. Thin reservoirs and geobodies can be interpreted with higher accuracy. There is also improved separation between different sand and shale bodies, as well as higher statistical coherency in reservoir cross plot analysis. Comparison of Vp/Vs ratio (result of the pre-stack inversion) shows far less dispersion when results are derived from inverted dual-sensor data compared to conventional seismic data. In total this means that the elastic attribute estimation is more accurate and reservoir properties can be predicted with higher certainty.

AAPG Search and Discovery Article #90135©2011 AAPG International Conference and Exhibition, Milan, Italy, 23-26 October 2011.