Scientific Methodology

Why does TransGen use SGR-based methods?

In "Basic project" mode TransGen uses Shale Gouge Ratio and/or Clay Smear Potential as the measure(s) of fault seal potential of the fault-rock which is then used to calculate fault-rock permeability and ultimately faulted transmissibilities and transmissibility multipliers. In "Flexible project" mode, the choice of the measure(s) of fault seal potential is user-controlled with the option to include up to five user-defined measures of fault seal potential in the calculations (e.g. SGR, CSP, SSF, etc, etc).

However,  SGR-based methods work!  We have been using the method since 1994 and from our own experience it has been shown to correctly predict fault seal in hundreds of cases world-wide.  It has become the industry standard method of choice for fault seal prediction in exploration and production.  It is a well understood and robust technique which predicts capilliary seal and critical pore-throat radii.  

Field studies show the internal structures of fault zones are known to be extremely variable over short distances.  Even if faults are sampled directly from a core from the reservoir they may have little predictive value at grid-block or reservoir scales:  

Observations at too fine a scale get bogged down in complexity and do not have a predictive value.  An appropriate scale for observation and prediction is at the grid block scale.   This is an appropriate scale for the SGR method.

What  is wrong with choosing a single-value for the transmissibility multiplier for each fault?

Fault rocks are a mixture of the material incorporated into the fault from their walls and in composition represent an average of the wall rock compositions.  Their properties at the grid-block scale should vary less than the wall rocks.

Transmissibility multipliers are used to incorporate the fault-properties (thickness and fault-rock permeability) by modifying the transmissibility at each cell-cell connection.  This transmissibility of a connection that needs to be modified varies as a function of the length and permeability of the juxtaposed cells.  In many models, both of these vary.  Even given constant dimensions, connections between high permeability cells will have higher transmissibilites than lower permeability cells.  Even in order to incorporate a constant fault property,  the transmissibility multiplier should be lower for high tranmissibility connections than low transmissibility connections.

Choosing a constant transmissibility multiplier for a whole fault in a variable-permeability sequence produces geologically unrealistic fault effects.    A separate transmissibility multiplier needs to be generated for each cell-cell connection, even if the fault properties are constant.

How can TransGen predict the small-scale variations in fault-rock permeability?

Luckily TransGen does not need to predict the small-scale variations in fault zone permeability that are seen in core measurements.   Fault permeability on the grid block scale  is likely to be more of an average and to be much less variable.

For a reservoir simulation, what matters is the upscaled  fault-rock permeability at the grid-block scale and upscaling any core based fault-data to this scale is difficult.  The grid-block fault-permeabilities will be dominated by highest permeability part of the fault-permeability distribution.  Estimating the population distribution from only a small number of samples is itself difficult, but further complicated by factors such as microfracturing and sample damage.  The most critical measurements are those which are most likely to be erroneous.