Deep-learning assisted regularized elastic full waveform inversion using the velocity distribution information from wells

Abstract

Summary Elastic full waveform inversion (EFWI) can, theoretically, give high-resolution estimates of the subsurface. However, in practice, the resolution and illumination of EFWI are limited by the bandwidth and aperture of seismic data. The often-present wells in developed fields as well as some exploratory regions can provide complementary information of the target area. We, thus, introduce a regularization technique, which combines the surface seismic and well log data, to help improve the resolution of EFWI. We use deep neural networks to learn the statistical relations between some selected features of the inverted model and the facies interpreted from well logs. The selected features are the means and variances of the inverted velocities defined within Gaussian windows. Using multiple fully-connected layers, we train our neural networks to identify the relation between the means and variances at the well location and those from the inverted model. The network is used to map the means and variances extracted from the well to the whole model domain. We then perform another EFWI in which we fit the predicted data to the observed ones as well as fit the model over a Gaussian window to the predicted means the variances. The tests on synthetic and real seismic data demonstrate that the proposed method can effectively improve the resolution and illumination of deep-buried reservoirs, which often encounter poor illumination from seismic data.

Target-Oriented Time-Lapse Elastic Full-Waveform Inversion Constrained by Deep Learning-Based Prior Model

“…The prior information lead the inversion to a reasonable subsurface model, but the resolution is still limited. To incorporate the high-resolution model information collected in wells, [23] use the velocity distribution information from wells to build a high-resolution prior model to constrain the inversion. In our proposed method, we build a DL algorithm to learn the prior model with high resolution to regularize time-lapse elastic FWI.…”

“…Knowledge of a monitor region can serve as a potential prior information to regularize the time-lapse EFWI scheme [8]. Considering that the injection and production wells are often present in the monitoring zone, well data, containing fine-scale velocity information, can be incorporated into the inversion to enhance the resolution and illumination of inversion [20]- [23].…”

“…A facies is characterized by a group of particular medium properties. Li et al [23] utilized the statistical distribution information in the wells to introduce high-resolution components into the EFWI result assisted by deep learning. To enhance the performance of time-lapse FWI, a prior model for the target monitor zone is constructed based on a DNN that relates an initial seismic estimation and well information [7].…”

Regularized seismic amplitude inversion via variational inference

“…Moreover, the efficiency of the particle-based algorithms depends on the size of the particles. Zhang et al (2021) presented different applications of the variational inference in geophysical inverse problems, such as petrophysical inversion, travel-time tomography and full waveform inversion.…”

“…Li et al. (2021) proposed a neural network to learn the statistical properties, that is, mean and variance, between the elastic velocity and facies, which are then used to regularize the elastic full waveform inversion. Nevertheless, for distributions with multimodal behaviour, a single‐component prior model may infer incorrect statistical information on the model parameters of interest.…”