Investigation of imbibition areas during well shut-in based on mercury injection experiment and BP neural network

摘要

Hydraulic fracturing appears to be an important and efficient method in the development of tight oil reservoirs. Well shut-in process after hydraulic fracturing enables the capillary imbibition process in leak-off area thus improving the production in tight oil reservoirs. Therefore, leak-off area is also the imbibition area. However, the range of leak-off area has not been fully understood for production engineers due to the complexity and dynamics of fracturing. In this work, we applied fractal theory to derive the matrix permeability in tight oil reservoirs, which can better describe the microscopic characteristics of reservoirs and manage the reservoir performance. Subsequently, we developed an analytical model to calculate pressure distribution at leak-off areas during the fracturing and well shut-in processes. To avoid solving nonlinear equations, we used Newton-Raphson iterative method to calculate the correlation of time and pressure. The effects of six parameters (e.g., permeability, shut-in time, viscosity, half-length of fracture and height of fracture) on the leak-off area were examined using BP neural network, and a two-dimensional distribution grid of fracturing leak-off was developed. Moreover, we computed pressure distribution of fracturing fluid leak-off area numerically to verify the reliability of the new model. Our results show that the leak-off area of fracturing fluid increases with shut-in time. Based on the BP neural network, permeability and shut-in time play significant influences in the leak-off area of fracturing fluid. This work provides insights to examine the process of hydraulic fracturing and shut-in in tight oil reservoirs, and the proposed model is a useful tool to quantify the leak-off area of fracturing fluid after shut-in stage.