10 October 2023 in Geophysics
Elastic parameters such as P-wave and S-wave velocity and density are of great significance for subsurface quantitative interpretation and reservoir prediction. Current prestack amplitude-variation-with-angle (AVA) inversion methods have been widely used in the industry to obtain subsurface elastic parameters. Conventional AVA inversion methods are theoretically based on a linearized amplitude-variation-with-offset model that formulates the relationship between prestack seismic reflection coefficients and the subsurface model of elastic parameters, which is called physical model-driven inversion. However, due to the linearized physical model, it is difficult to obtain high-accuracy and high-resolution inversion results using model-driven inversion when the incident angles are greater than 30°. In recent years, several neural network-based prestack AVA inversion methods, known as data-driven inversion, have been developed to address this issue. However, these methods typically require a large amount of labeled data for training networks, and there is no physical mechanism in the network training process. To solve this problem, a joint data-driven and physics-driven inversion of prestack AVA elastic parameters is proposed. Under the framework of semisupervised learning, a 2D convolutional neural network and a recurrent neural network are used to establish the mapping between several adjacent prestack AVA gathers and 1D elastic parameters in the time domain. The full Zoeppritz equation suitable for large incidence angles is used as a physical constraint for training the neural network. The loss functions of the network are constructed by combing well-log data and prestack AVA seismic data. This approach enables us to use fewer labeled data for training the network and increases the physical interpretability of the training process. To further improve the inversion results, the inverse distance weighted correlation coefficient of seismic data is applied to the loss function of seismic data and well-log data. Synthetic and field data examples demonstrate that the joint data-driven and physics-driven prestack AVA elastic parameter inversion improves accuracy and resolution compared with conventional model-driven inversion.
9 January 2025 in Geophysics
Prestack inversion of elastic and anisotropic parameters in horizontally transversely isotropic (HTI) media is a critical technique for characterizing high-angle, directionally aligned fractured reservoirs. Traditional model-driven inversion methods rely on approximate reflection coefficient equations for anisotropic media, but their accuracy and robustness are often limited by model assumptions and data noise. In contrast, data-driven inversion methods based on deep learning suffer from limited generalization and poor physical interpretability, particularly in scenarios with scarce labeled data. We develop a method within a semi-supervised learning framework that integrates data- and model-driven approaches. Under the constraint of the amplitude variation with angle and azimuth reflection coefficient equation, prestack azimuthal seismic data, well-log data, seismic horizon interpretations, and prior models are jointly used to invert elastic and anisotropic parameters in HTI media. Furthermore, the uncertainty of the inversion results is quantified using dropout-based techniques in deep neural networks. Experiments on synthetic and field data show that our joint inversion approach not only improves the accuracy and stability of parameter estimation but also enables effective assessment of inversion uncertainty.
1 January 2026 in IEEE Transactions on Geoscience and Remote Sensing