The Oil and Gas industry has typically used well log averages to conceptually populate the 3D static model but in more recent times has also used probabilistic and stochastic predictions to determine the likely occurrence of reservoir properties. These methods are largely inaccurate and are rarely ever validated by blind testing. The use of log averages will never predict the precise rock property and will never predict the variation in reservoir quality. See Figure 1.
The methodology developed by iRPM determines the precise variation in rock properties and the uncertainty around that precise value. Critically, predictions are blind test against existing or new wells, which validates the property predictions.
The first step of the iRPM process is to determine the Seismic Impedance volume. Seismic Impedance volumes are determined from the traditional methods of Seismic Inversion using the Seismic band-limited or full-band trace data. The second step of the iRPM process is to determine the mathematical relationships between the 3D Seismic Impedance property and the Petrophysical interpretation results of volume of clay, reservoir porosity, permeability and hydrocarbon saturation. These Petrophysical properties are then populated into the three dimensional static model constrained and referenced to the Seismic Impedance volume. To date, static model resolution has been increased to 0.5m from the Seismic resolution of typically 2-5m. See Figure 3. This is achieved by calibrating to the higher resolution log data which is typically 0.5 feet.
The process requires at least one well log data set to calibrate and determine these relationships which are then applied across the field to each sampled Seismic trace/impedance property. In this way the extent of the hydrocarbon field and the variability of the field's properties can be determined without cognitive or conceptual biasing. The predictions can then be used to blind tested against any existing well or any exploration well yet to be drilled. Blind testing provides a true measurement of the accuracy and uncertainty of the model property prediction.
