The Study of class separation in chemical composition of oil field water in Jiangling depression
using a multidimentional space non-linear mapping method combined with linear statistic analysis
is our example of applying a computerized pattern recognition technique to the classification study of oil field water.The Jiangling depression is a part of Jianghan Basin developed in Cretaceous-Tertiary.During its process of continuously sinking
having a dry or semi-dry
restricted or semi-restricted interior lake sedimentary environment
the calcium mirabilite rocks and salt rocks were developed.Faults are well developed in the depression
so the oil water relations are complicated.It was proved by the above mentioned pattern recognition method that there is a good correlation between the concentration of divalent ions(such as SO
4
2-
Ca
2+
Mg
2+
etc.)in the oil field water of Jiangling depression and oi
l-bearing conditions
and the positive correlation between SO
4
2-
and the oil-bearing conditions seems different from the traditional viewpiont of oil field water investigations.In order to interpret this result
we measured some solubility relations in the solution of simulated oil field water samples.It was noticed that in oil field water samples of Jiangling depression the Ca
2+
concentration often decreased with the increasing of SO
4
2-
concentration.It seemed to indicate that many of the sambles were almost CaSO
4
saturated.We measured the solubility of CaSO
4
in some simulated solutions and plotted a diagram of the equation log(f
±
(0))/(f
±
)-(ΣC)
1/3
to show the saturation region of CaSO
4
in oil field water based on a diffuse-lattice model of electrolyte solutions.This experiment proved that the concentration of CaSO
4
in the water associated with oil was often saturated or nearly saturated.Therefore the concentration of CaSO
4
in oil field water can be used as a good indicator for oil exploration and development in this region.Since salt and gypsum are well developed in the reservoir rocks in Jiangling depression