DCDong Chen

Dong Chen

*This is an algorithmically generated author record based on publicly available information.

Dong Chen has a total of 9 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Generating complete synthetic datasets for high‐resolution amplitude‐versus‐offset attributes deep learning inversion

    by Shuai Sun, Junguang Nie, Benfeng Wang, Luanxiao Zhao, Zhiliang He, Hong Zhang, Dong Chen, Jianhua Geng

    Deep learning has been used in seismic exploration to solve seismic inversion problems, however it requires sufficient and diverse training samples and labels to obtain satisfactory results. Insufficient training labels are a common problem since labels usually come from well‐logging data, which are limited and sparsely distributed. This can result in a trained network with poor generalizability. A novel complete synthetic dataset‐driven method utilizing convolutional neural network is presented for seismic amplitude‐versus‐offset attributes P and G inversion. Gaussian simulation sampling with physical constraints is used to generate a complete elastic parameter dataset by traversing the entire elastic parameter model space. By randomly combining elastic parameters from the full elastic parameters model space, sufficient synthetic pre‐stack amplitude‐versus‐offset gathers and attribute datasets are generated for training the convolutional neural network. Compared with the limited real data‐driven convolutional neural network, the complete synthetic dataset‐driven convolutional neural network has better generalizability. Broadband training labels improve the accuracy and resolution of the complete synthetic dataset‐driven convolutional neural network's inversion results beyond that of the conventional least‐square data‐fitting method. The complete synthetic dataset‐driven convolutional neural network is robust for processing noise‐contaminated seismic data, but if the frequency band of the labels for training the network is too wide and the signal‐to‐noise ratio of pre‐stack amplitude‐versus‐offset gathers is too low, the quality of the inversion results will reduce. The Marmousi II model and field data examples show that the novel complete synthetic dataset‐driven convolutional neural network can extract higher‐resolution amplitude‐versus‐offset attributes P and G.


  • Gaussian mixture model deep neural network and its application in porosity prediction of deep carbonate reservoir

    by Yingying Wang, Liping Niu, Luanxiao Zhao, Benfeng Wang, Zhiliang He, Hong Zhang, Dong Chen, Jianhua Geng

    To estimate the spatial distribution of porosity, model-driven or data-driven methods are usually used to establish the relationship between porosity and seismic elastic parameters. However, due to the strong heterogeneity and complex pore structures of carbonate reservoirs, porosity estimation of carbonates still represents a great challenge. The existing conventional model-driven and data-driven-based porosity estimation methods have high uncertainty. To characterize the complex statistical distribution of porosity, the nonlinear relationship between porosity and seismic elastic parameters, and the uncertainty of porosity estimation, we have used a Gaussian mixture model deep neural network (GMM-DNN) to invert porosity from seismic elastic parameters. We use a Gaussian mixture model to describe the complex distribution of porosity, and we apply a deep neural network to establish the nonlinear relationship among seismic compressional-wave (P-wave) velocity, density, and porosity. The outputs of the GMM-DNN provide an estimated probability distribution of porosity conditioned on the input seismic elastic parameters. The synthetic data example verifies the feasibility of this method. We further apply the GMM-DNN-based porosity inversion method to a deep complex carbonate reservoir in the Tarim Basin, Northwest China. The well-logging data are used to train the GMM-DNN; then, the P-wave velocity and density obtained by prestack amplitude-variation-with-offset inversion are fed into the trained network to reasonably estimate the porosity distribution of the whole target reservoir and evaluate its uncertainties.