Abstract Acoustic impedance inversion results are strongly dependent on the initial model. Conventional initial model based on well-log interpolation exhibits significant uncertainties in sparsely drilled areas. Meanwhile, conventional interpolation algorithms, constrained by stationarity assumptions, fail to characterize complex geological structures. Therefore, constructing an inversion initial model with more reliable geological information becomes essential for improving inversion accuracy. Based on geological interpretation, we first derive subsurface density parameters from well logs. Subsequently, structure-constrained full waveform inversion (FWI) is implemented using the variable-density acoustic wave equation to obtain a high-resolution velocity model. This improved velocity model helps construct an initial acoustic impedance model with enhanced lateral resolution for subsequent inversion. Furthermore, to address the ill-posedness of conventional L1 regularized inversion, we adopt a nonconvex L1-2 regularization constraint to achieve accurate reconstruction of subsurface structures. The practical data application results show that the proposed method effectively improves the accuracy of acoustic impedance inversion in tectonically complicated areas, providing new insights for complex reservoir prediction while exhibiting applicability and potential for widespread implementation.
Chen et al. (Thu,) studied this question.