It depends on the goal of your study/type of EOR being used, basically, you will need all the data that:

1. Describe and define your field (geology, well logs, production/injection history, completions, seismic (if available), core data.

2. Lab data ( fluid and rock props: PVT, Kr,...etc).

3. After preparing the model, you will need to simulate this model and to distribute the fluids saturation and properties in each cell of the grid.

4. run a basic model and history-match it with you actual values.

5. create scenarios for the required types of EOR and evaluate the results.

6. give reports and recommendations.

I haven't run EOR simulations myself, but I'd assume Petrel has the basic fluid properties for the drive fluid already in the simulator, so you'll just need to know the reservoir conditions (P & T) along with injection/withdrawal rates to model the EOR drive portion of the model.

Before you do that, you'll need a comprehensive geologic model - Surface grids for each key reservoir/seal unit w/ said reservoir units populated with key reservoir properties (i.e. Porosity, permiability, faults, fractures) w/in a reasonable variogram.

I believe Ekofisk is a chalk reservoir, so it might be fairly homogenous w/ high porosity & low perm. It was pretty common for those chalk reservoirs to undergo catastrophic compaction during production (pressure depletion), so pressure maintenance is critical. In such a low perm environment, I'm not sure a waterflood would work, but perhaps gas-gravity drainage or a CO2 flood would?

Depending on the EOR method, whether it is Chemical, Miscible or Thermal, you will have to choose Eclipse: Black Oil, Compositional or Thermal. The input data will be related with microscopic displacement efficiency (Kr curves) and fluids properties for both oil and displacing fluid. This data should be enter as a tables or parameter using the corresponding keyword.