Data for: Modelling and field testing of back-flow fracturing fluid after acid fracturing in Oil Shale reservoirs

Published: 31 January 2019| Version 1 | DOI: 10.17632/tgy3v7syct.1
Contributors:
LI Qiang,
ZHAO Shuai

Description

In the field experiment site of the in-situ pyrolysis of oil shale, there are three wells. One well is used to inject high temperature and high-pressure nitrogen gas. The other one, named FK-2, is a well that is used to produce oil and gas products. The last well is a monitoring well named M1 for the real-time monitoring of the underground temperature and water inflow. According to the microseismic monitoring at the fracturing site, we have determined that the fracturing may have formed two fractures in the oil shale formation. Because the in situ pyrolysis process of oil shale requires nitrogen injection at a high temperature and pressure, we must monitor the connectivity between the two fractures. There is no need to use high temperature gas to determine the formation connectivity, so we only use high-pressure nitrogen gas. First, the backflow of the fracturing fluid is conducted. After the fracturing fluid backflow, we can determine the connectivity and water inflow of the formation. Based on the microseismic monitoring data, we built a three-dimensional model. The flow field of the high-pressure nitrogen injection into well FK-1 and the gas production from well FK-2 were simulated using the finite element software COMSOL Multiphysics 5.3. The results show that the outlet pressure at the FK-2 well can reach 2.8 MPa when well FK-1 is used as a high-pressure nitrogen injection well and gas is continuously injected at 8.5 MPa. Part of the fracturing fluid flows out of well FK-2, but some of fluid can flow into the formation. When high-pressure nitrogen is injected into the FK-2 well, the result is also the same. To verify the accuracy of the simulation results, experiments were carried out involving high-pressure nitrogen injection into well FK-1. The fracturing fluid volume was monitored while the fracturing fluid flowed back, and the pressure change in well FK-2 was monitored in real time. Finally, the pressure in well FK-2 could reach 2.8 MPa, which proved the connectivity between the two wells. It also confirmed the accuracy of the numerical simulation. In addition, we monitored the pressure changes in the FK-2 well by means of active pressure relief in FK-1 well. The results show that the pressure in the FK-2 well would also decrease, which once again confirms the connectivity status of the two wells.

Files