Improve thin beds discrimination using the log curves deconvolution method

Tao Zhang,Chengyan Lin,Xianguo Zhang
2010-01-01
Abstract:In mature oil fields, vertical resolution of early log curves is not high enough for the characterization of thin beds and doesn't meet the demand of delicate geologic research. Aimed at this problem, the process of blind deconvolution for log curves is developed. The algorithm of this method is: form linear equations by transforming the problem of deconvolution into an extreme value problem between the expected and actual output signals, and then get the deconvolution factor by solving the equations. By this process, the influences of surrounding rocks are wiped, the energy of high frequency band in frequency spectrum is improved, the new curve is more sensitive to the variance of curve slope, low amplitude variances of curves are zoomed in and the characteristics of sedimentary surfaces are more distinct. The identification ability of processed log curves in thin beds interpretation, especially in alternating beds with single beds about 0.5 m thick, is improved after the deconvolution process. The geological rationality of deconvolution process is verified by research with cored well data.
What problem does this paper attempt to address?