Tracking Tertiary delta sands (Urdaneta West, Lake Maracaibo, Venezuela)

Abstract

This paper illustrates a novel approach (based on seismic facies classification and 3D visualization) for tracking deltaic sand systems in the Urdaneta West producing field, Lake Maracaibo, Venezuela. As part of a major field review (volume evaluation and well planning), the following oil-bearing reservoirs were studied:• Oligocene sands (braided deltaic systems) • Eocene sands (lower deltaic systems) Among the geologic challenges associated with developing these reservoirs, the thinness of the sand layers and the lack of amplitude response of heavy oil-bearing sands were key. Consequently, a dedicated geophysical workflow was designed to address these issues. It comprises the following steps:• Improving seismic resolution by applying high-frequency imaging (HFI) processing ( Figure 1) • Implementing neural-network seismic facies classification techniques (loop-level trace-shape analysis for the Oligocene reservoirs and volume classification for the Eocene sands) to evaluate nonamplitude-supported plays • Calibration of seismic reservoir indicators This workflow integrated all available seismic attributes (including poststack attributes and impedance data), improved the seismic interpretation, and provided the means to predict sands (in a nonamplitude-supported play).The geophysical results have been compared with the geologic data (cores and logs) and have confirmed the depositional model established by the geologists. Geologic challenges and technical solutions. The UrdanetaWest field is operated by Shell Venezuela on behalf of Petróleos de Venezuela SA (Pdvsa). Even though this field has been producing for nearly 40 years (with more than 100 well penetrations over the study area), exploitation and development of Oligocene/Eocene oil-bearing sands remain complex due to the following geologic challenges:• Thin sand layers (~10 ft on average) with limited lateral extent (fluvial and braided-delta deposits) • Pore fluid characteristics (12-14°API oil with a low gasoil ratio of 50 scf/bbl) resulting in a nonamplitude-supported play. Amplitude/impedance variations are not indicative of hydrocarbon presence, but can be used as lithology indicators • Poor continuity of seismic events within the Eocene reservoirs resulting in complex seismic interpretation • Density of faults (particularly in the Eocene reservoirs) affecting event correlation and possibly reservoir connectivityTo minimize these risks, Shell Venezuela conducted a major field review performed by a multidisciplinary project team. The main objectives of this project were:• To increase seismic resolution using HFI processing (resolution was enhanced from 100 to 30 ft, but individual sand layers still could not be seismically characterized) • To implement multiattribute seismic facies classification SEPTEMBER 2004 THE LEADING EDGE 909 Figure 1. Comparison between original volume (top) and reprocessed HFI volume (bottom) on SW-NE seismic line (reverse SEG convention; trough = increased acoustic impedance). Inset is structure map of Oligocene unconformity with...

Evolution of seismic interpretation during the last three decades

“…In its original implementation, waveform classification was totally unsupervised. Subsequent implementations allowed the interpreter to introduce and modify specific waveforms, such as average waveforms representing good and bad wells (Poupon, 2004). SOM can be applied to any suite of attributes, providing a clustering technique that can be an excellent means of characterizing geomorphology (Strecher and Uden 2002;Coleou et al, 2003;Roy et al, 2011).…”

A knowledge-integration framework for interpreting seismic facies

“…Interpreters use automated seismic facies classification to identify important stratigraphic or reservoir characteristics from the seismic data, usually without well data to guide the classification. Examples are the mapping of thin clastic reservoirs interbedded between coal and shale layers (Chandra et al, 2003), the delineation of channel systems (Poupon et al, 2004;Cao et al, 2005), the prediction of thin-bed reservoirs (Xie et al, 2004), and the identification of lithofacies geometry within carbonate buildups (Farzadi, 2006). More recent studies include the interpretation of palaeokarst geobodies and sedimentary patterns of a carbonate turbidite (Farzadi and Hesthammer, 2007), the mapping of lithologic changes in a tidal channel and pinnacle reefs (Marroquín et al, 2009a), and the identification of hydraulic fracturing on a shale formation (Roy and Marfurt, 2011).…”