Seismic sections extracted from acquired lines within the field and their interpretation provides the following data: 1) a depth map at top reservoir (H6 horizon), prepared with the most likely velocity map tied to the well tops and 2) A set of fault polygons representing the intersection of the interpreted fault plans with the H6 horizon, with their main dip angle and azimuth.

Data from 15 wells drilled in the northern structure were interpreted. The well log interpretation provides a complete set of composite well logs over the reservoir interval, including when available the gamma ray, neutron density, sonic and resistivity measurement [14] [15] [16] [17] [18]. The studied wells are: S-2, S-3, S-4, S-6, S-7, S-10, S-11, S-C1, S-C2, S-C3, CSE-1, CSW-1, DO-1, DO-2A, DO-A3 (Figure 3). Wells data synthesis was done using well log interpretations and correlations.

The methodology consisted in a complex analysis of data relating to top reservoir (H6 horizon) within the field of Sèmè. The Petrel Software and its numerous function modules were employed [19] [20]. The model was built using Petrel

software applications (Schlumberger) modules to assess and integrate the geological parameters. The generated model was used to display the structural architecture of H6 horizon including reservoir layers, barriers and fault. The make horizon process (Figure 4) was run using the Top H6 horizon as conformable surface, tied at each well top dataset.

The geomodeling phase is an important phase that unifies, in a 3D interpretative representation, all valuable geologic parameters. The construction of the geomodel involves successive steps, in particular: 1) the geometrical modeling based on seismic maps and fault interpretation and 2) reservoir layering based on wells correlation framework. Thus, the built top reservoir depth map was used as input. The fault pattern/polygons defined at the top of the reservoir with their corresponding fault dips were used to build up the fault sets. Well tops and correlation were used to define the vertical layering of the reservoir unit.

The seismic map used to build the geomodel is the top H6 depth map inferred from a new seismic interpretation performed in 2009. The structural configuration of the top reservoir (H6 Horizon) is characterized by a certain number of low relief independent culminations with clear orientations, either in an east-west

trend or north-northeast to south to southwest trend. Generally, structural continuity is affected by small faults of limited throws, not accurately defined by seismic. The reservoir is truncated to the east in the water bearing zone by erosional surface. Structural uncertainties are mainly related to seismic time to depth conversion as well as vertical depth inconsistencies found in few deviated wells. Seismic interpretation provided a set of fault polygon at top H6 horizon and a map of dip and angle. The fault pattern/polygons defined at the top of the reservoir with their corresponding fault dips were used to build up the fault sets. Wells correlation and Top reservoir map (well location corresponds to impact on top reservoir) gave possibility to present the structural style of the top Abeokuta formation on Sèmè field (Figure 5).

1) Two main E-W faults at North and South of the structure. These faults are limited in their Eastern part by the erosional surface between the H4 and H5 formation. The fault at the South is limited in its western part by a small horst in the central area. Another East-West fault is present at the North of well SC3, defining a closed structure in that area.

2) Different small horst and grabben systems oriented mainly N-S or SW-NE, in particular in the S4 and S10 area.

3) The accident at the south of wells DO1, S3, S6 and S4, has a direction going from East west to nearly North-South with a dip to the north. It is composed of two different faults in continuity and will have a major impact on the grid construction and flow behavior of this producing zone.

The vertical gridding was done using the top H6 horizon and the well top markers from the correlation. Well tops and correlation were used to define the vertical layering of the reservoir unit (Figure 6). Layer limits are all set conformably to the top H-6 surface: Top Abeokuta, Base A, Top B, Base B, Top C1, Top C2, Base C, Top D, Base D, Top E and Base E.

Thus, eleven explicit zonation were then built: 7 reservoir units (A, B, C1, C2, D, E) and 4 barrier units (1, 2, 3 and 4) were delineated. Each of these zones represents one reservoir layer in the grid, except the zone D that was subdivided into 2 isopach layers. The barriers were defined as explicit layers in order to easily modify their petrophysical properties during the history matching.

The different layers identified are:

Unit A: thin unit (0 - 3.2 m) not present in the south east corner (Wells Do1, S2, S3, S10, S11) with fair to good reservoir characteristics.

Barrier 1: shaly event with same extent that unit A. This level appears to have a law sealing properties, with a small break in potential between units A and B, unit A being most of the time less depleted than B (i.e. Well S4).

Unit B: good reservoir unit with an average thickness of 4.4 m, showing a global thickening from South-East to North-West.

Barrier 2: This level is most of the time corresponding to a high Gamma ray interval. Its thickness is between 1 and 2.5 m. This level corresponds to a break in pressure potential between units B and C. This break shows in some cases the unit C more depleted than B (i.e. Wells S4, SC-2, SC-3), or in some other cases the unit C less depleted than B (i.e. Wells S6, S11).

Unit C1: With a global thickness from 5 to 8 m, this level is a very good massive reservoir, produced in many wells.

Unit C2: Good massive reservoir around 5 m thick. C1 and C2 units limit was defined based on a small radioactive level, not always visible. There is no noticeable difference in pressure potential between these two units.

Barrier 3: This level corresponds to the limit between the C and D reservoir units. It is most of the time corresponding to a shaly interval (S4, S6, S3, S10), but with not so clear log response in some cases (S11). This level was cored in well S2, and is described as a 3 meter thick shaly interval, finely laminated. There is most of the time a pressure gradient break above and below this level, with the C level more depleted than the D.

Unit D: thick massive reservoir unit (around 10 m), with occasionally some shaly intervals (S3). The unit D can be subdivided into 2 isopach layers (D1 and D2).

Barrier 4: This interval is the main barrier level of the Abeokuta complex, always indicating an important pressure gradient break, between the depleted D unit above, and the non or very poorly depleted E unit bellow. It corresponds in some well with a shaly interval (S6, S7, SC-3), but has often a not clear log signature (S3, S10…). This barrier doesn’t seem to be correlated with a consistent singular sedimentological event, but most probably to a relay in multiple sealing intervals. This level was cored in well S2, and probably corresponds to a half meter shaly interval (2304 - 2304.5 core depth). Note that this well has no pressure data to prove the correlation of barrier 4, and a possible uncertainty on the core to well depth shift.

Unit E: This unit is thick (15 to 20 m) with occurrence of several shaly events. The base of this unit was mainly defined on the presence of a shaly level, not always clear. There is no pressure break visible above and below this lower limit.

In some cases, some uncertainties remain concerning the exact location of some markers, especially when no or poor pressure points are available (well S2, DO1, DO2A), or when these pressure points were taken late after end of production (CSE1, CSW1). In those cases, the well log data were mainly used with the uncertainties discussed above.

Using Petrel software applications (Schlumberger) modules, the limits and trends of the reservoir architecture and a 100 × 100 m grid was defined. Such refinement is sufficient to take into account the observable geological heterogeneities. Thus, pillar gridding process, grid skeleton and grid segmentation allow to build the 3D grid view of H6 horizon (Figure 7). Generally, structural continuity is affected by small faults of limited throws, not accurately defined by seismic.

The H6 horizon thickness is generally less than 100 m. The 3D grid view of H6 horizon is illustrated by the following geological section across the structure (Figures 8-13). The North-South orientation sections (Figures 8-10) make it possible to notice that in the direction of the sea the H6 horizon sinks following the bottom morphology of the basin.

East-West cross sections (Figures 11-13) show that, the H6 horizon has the shape of a faulted anticline less faulted and more flat in the vicinity of the southern structure.

Vertical layering of H6 horizon was previously done using gamma ray wire line of a limited number of wells (S3, S4, S7 and SC3) [7]. According to [7] only one well (SC3) located in West Sèmè North give possibility to differentiate 11 layers with a global thickness of 220 m. From other wells the number of layers vary from one (well S7) to six (well S4). The presence of horizontal barriers raised the question of the existence of several independent reservoir units, with possibility of different contact level. It was observed that for every well, the E reservoir unit is disconnected from the upper unit. On the other hand, the reservoir units A, B, C and D constituting one reservoir tank with an original OWC at 1910 m TVD remain the most likely assumption.