In this study, we present the result obtained from incorporating well log information into a conjugate gradient optimization scheme in Full Waveform Inversion (FWI). We test this approach on synthetic datasets generated using the three sonic logs from Hussar. Using formation tops to guide the interpolation, the sonic logs are interpolated to form the 2D velocity model used in this study. The initial velocity model for the inversion is a linear v(z) velocity model. We adopt the conjugate gradient algorithm as described by Magnus R. Hestenes and Eduard Stiefel. Our results show that combining well information with conjugate gradient directions in FWI can save computational time, as well as getting a good inverted model after a few iterations. The inverted model shows encouraging results and this proves that the algorithm works well and can resolve thin beds in the model.
View full article as PDF (0.73 Mb)