Meta-algorithm assisted interval inversion for petrophysical properties prediction

Abstract

This research proposed integration between the GSS method and the interval inversion for regularizing the prediction algorithm of borehole logging data. The proposed algorithm has been tested using synthetic borehole logging data set contaminated with 5 % Gaussian noise. The GSS-based inversion scheme showed a smoother data distance convergence. The GSS-based inversion could decrease the conditional number of the sensitivity matrix. Furthermore, the algorithm has been applied to real data of a tight sand reservoir. The results from the Damped Least Square (DLSQ) scheme have been compared to those of the GSS-based scheme. The data distance of the DLSQ showed a rough convergence data distance line, while it showed a smooth convergence of the data distance line in the case of the GSS-based inversion. The convergence of the data distance of both schemes has been stabilized at 9.5%.