The Research On Water Saturation Evaluation Model Of Sweep-frequency Dielectric Logging

With the development of oilfield exploration,there are more and more low-resistivity reservoirs and the formation water salinity is varied.Conventional logging data combined with Archie formula is not accurate enough to evaluate complex reservoirs.Because of the apparent difference in the permittivity of oil and water,dielectric logging has shown good applications in evaluating complex reservoirs.And in recent year,a new dielectric tool named ADT that introduced by foreign oil service company has worked well in oil field in our country.At present,the response mechanism of sweep-frequency dielectric logging is not perfect and the research on dielectric saturation model in China is still in its infancy.In this paper,after fully investigating the research status of dielectric logging,26 pure natural cores,including 20 pure sandstone and 6 carbonate rock samples,were measured by dielectric sweep frequency in the laboratory.Through the experimental data,the influence of saturation,salinity and porosity on the dielectric constant is analyzed qualitatively.At the same time,the experimental data of four frequency points are extracted.These data are used to further analyze the relationship between different physical parameters and core dielectric constant.The experimental results show that the relative permittivity of rock is positively correlated with each physical parameter.And with the increase of water saturation,porosity and salinity,the dispersion increases.Through the analysis of the data of the dry rock sample,the skeleton dielectric constant of the pure sandstone rock sample is about 5.99,while that of the carbonate rock sample is about 7.The dispersion model of pore water was obtained by using simulated annealing algorithm method.A sweep dielectric fluid saturation evaluation model was established,which combined with experimental data and effective dielectric model.The cross validation of experimental data and actual logging data proves that the evaluation effect of the model is good.