The spatial variability of relative permeability curves has not attracted much attention yet. This paper addresses this issue,
and extends the self-calibration technique for the generation of absolute and relative permeabilities conditioned not only
to permeability data but also to saturation and pressure data.
The paper starts with a sensitivity analysis presenting a synthetic example where the spatial variability of relative permeability
is relevant, then proceeds with a derivation of the algorithm used to condition a realization of relative permeabilities to
pressure and saturation data (both steady state and transient) and concludes with the demonstration of the technique with
one synthetic example.
The paper shows that the spatial variability of relative permeabilities is important in reservoir characterization. It also
demonstrates how the self-calibrating simulation method can be used to generate realizations of spatially variable relative
permeability curve parameters which are consistent with measured values of the state of the reservoir.