Summary In an ideal case, the time-lapse differences in 4D seismic data should only reflect the changes of the subsurface geology. Practically, however, undesirable discrepancies are generated because of various reasons. Therefore, proper time-lapse processing techniques are required to improve the repeatability of time-lapse seismic data and to capture accurate seismic information to analyze target changes. In this study, we propose a machine learning-based time-lapse seismic data processing method improving repeatability. A training data construction method, training strategy, and machine learning network architecture based on a convolutional autoencoder are proposed. Uniform manifold approximation and projection are applied to the training and target data to analyze the features corresponding to each data point. When the feature distribution of the training data is different from the target data, we implement data augmentation to enhance the diversity of the training data. The method is verified through numerical experiments using both synthetic and field time-lapse seismic data, and the results are analyzed with several methods, including a comparison of repeatability metrics. From the results of the numerical experiments, we can conclude that the proposed convolutional autoencoder can enhance the repeatability of the time-lapse seismic data and increase the accuracy of observed variations in seismic signals generated from target changes.

No abstract

Accurate estimation of subsurface properties plays an important role in successful hydrocarbon exploration, and a variety of different types of inversion schemes are used to infer earth properties such as velocity or density by analyzing the surface seismic. The Markov-chain Monte Carlo (MCMC) stochastic approach is widely used to estimate subsurface properties. We have used a transdimensional form of MCMC, reversible jump MCMC (RJMCMC), to estimate seismic impedance, which allows the inference of the number of interfaces as well as the interface location and layer impedances. Estimating the uncertainty quantitatively is also very important when performing stochastic inversion. Therefore, the goal of this paper is to apply the transdimensional method to obtain a 3D seismic impedance model and to quantify uncertainty in impedance and interface locations. We also measured the speedup of the proposed algorithm by applying data and task parallelism. To demonstrate the performance and reliability of the proposed RJMCMC impedance inversion, we used seismic data from the E-segment of the Norne field in Norwegian Sea. The results of the quasi 3D transdimensional MCMC approach, which independently inverts data from each common-depth-point location, indicate high velocity contrasts near gas-oil contacts and high uncertainty in impedance near discontinuities. Also, the cross section of the impedance uncertainty volume helps to identify the location of a high-contrast boundary corresponding to the location of the possible gas reservoir. The proposed uncertainty measure can serve as an attribute to identify important reservoir features.