In the past, reservoir engineers used numerical simulation or reservoir engineering methods to predict oil production, and the accuracy of prediction depended more on the engineers’ own experience. With the development of data science, a new trend has arisen to use deep learning to predict oil production from the perspective of data. In this study, a hybrid forecasting model (CNN-LSTM) based on a convolutional neural network (CNN) and a Long Short-Term Memory (LSTM) neural network is proposed and used to predict the production of fractured horizontal wells in volcanic reservoirs. The model solves the limitation of traditional methods that rely on personal experience. First, the production constraints and production data are used to form a feature space, and the abstract semantics of the feature time series are extracted through convolutional neural network, then the LSTM neural network is used to predict the time series. The certain hyperparameters of the whole model are optimized by Particle Swarm Optimization algorithm (PSO). In order to estimate the model, some production dynamics from the Xinjiang oilfield of China are used for comparative analysis. The experimental results show that the CNN-LSTM model is superior to traditional neural networks and conventional decline curves.

In this paper, we used a self-developed anisotropic cubic core holder to test anisotropic relative permeability by the unsteady-states method, and introduced the anisotropic relative permeability to the traditional numerical simulator. The oil–water two-phase governing equation considering the anisotropic relative permeability is established, and the difference discretization is carried out. We formed a new oil–water two-phase numerical simulation method. It is clear that in a heterogeneous rock with millimeter to centimeter scale laminae, relative permeability is an anisotropic tensor. When the displacement direction is parallel to the bedding, the residual oil saturation is high and the displacement efficiency is low. The greater the angle between the displacement direction and the bedding strike, the lower the residual oil saturation is, the higher the displacement efficiency is, and the relative permeability curve tends towards a rightward shift. The new simulator showed that the anisotropic relative permeability not only affects the breakthrough time and sweep range of water flooding, but also has a significant influence on the overall water cut. The new simulator is validated with the actual oilfield model. It could describe the law of oil–water seepage in an anisotropic reservoir, depict the law of remaining oil distribution of a typical fluvial reservoir, and provide technical support for reasonable injection-production directions.

Karst cave collapse usually occurs during the process of drilling and reservoir development. Karst cave collapse increases drilling risk and seriously affects oil production. In order to reveal the collapse mechanism of karst caves in fractured-vuggy reservoirs in Tahe oil field, rock mechanic experiments and numerical simulation studies were conducted. In rock mechanic experiments, three types of rock mechanics tests (uniaxial compression, triaxial compression, and Brazilian splitting) were carried out to obtain the basic rock mechanics parameters. In numerical simulation studies, the collapse conditions of a single karst cave and a cave system were studied. Numerical simulation results indicate that for a single karst cave, the size and geometrical features of karst caves are the main factors influencing collapse condition. For a cave system, a combination pattern and distance between caves are two main factors affecting collapse condition. In order to facilitate the application of these research results, the authors present formulas for calculating the critical conditions of karst cave collapse by means of multivariate linear regression method. The calculation formulas and prediction chart were verified as consistent with actual field results. The research results in this paper have great theoretical and practical value for revealing the mechanism of karst cave collapse in fractured-vuggy carbonate reservoirs.