Modeling Sub-Seismic Reservoir Discontinuities: Integrating Production Data With Seismic Data and Geological Analogues

The "Z" Field, located in Block A-18 of the Northern Malay Basin, has been producing gas from complex, multiplestacked sandstone reservoirs for approximately four years. Most reservoirs are located between around 3600 and 4300 ft TVDSS, and generally exhibit Class III AVO response, with better-quality zones exhibiting low acoustic impedance resulting in a high amplitude seismic expression. Consequently, the success rate for targeted production wells is high, with 100% of seismic anomalies drilled at these levels encountering commercial gas deposits. Well "X" targeted one such amplitude anomaly and consistently produces gas at a rate of more than 20 MMscf/d. Two years later, another well, "Y”, was drilled approximately 1 km away. Both wells lie within the same seismic anomaly, share similar electric log profiles and were originally thought to lie within the same sand body. However, reservoir pressure in well "Y" was found to be around 160 psi greater than in well "X", and CO2 concentrations were 7-8% higher, indicating the reservoir was compartmentalized. Paleoenvironment is believed to be estuarine, with most sands interpreted as tidal bars. Analogue studies show that such deposits are frequently heterolithic, containing thin shale beds that can act as baffles or barriers. The zone of interest was subjected to a chemostratigraphic study, which revealed that the provenance of the sand differed between the two wells. This difference in provenance indicates that the reservoir is likely to be divided into at least two genetically distinct sand bodies. After closer examination, several narrow zones of high acoustic impedance were identified in seismic attribute extractions. These zones were defined as permeability barriers in the static geomodel, which was subsequently used as the basis for the dynamic simulation model. The dynamic model was then history-matched to production data, and the position and extent of the modeled shale barriers were adjusted until an optimal history-match was achieved. Complex, compartmentalized reservoirs are never easy to model. However, by incorporating multiple geological, geophysical and reservoir engineering data sets, geologically consistent models of static and dynamic reservoir behavior can be achieved.