Fracture diagnostic on a cluster scale of multi-stage hydraulic fracturing wells remains challenging but essential to determine the quality of the stimulation operation and the completion strategies for future wells. Since the stimulation fluid is injected at a different temperature compared to the original geothermal, the considerably modified and highly heterogeneous thermal profile after stimulation presents significant potential to serve for fracture diagnostic purposes. In this work, a model to analyze the temperature signal associated with the shut-in period after hydraulic fracturing is presented, along with the pilot testing of two datasets.
The model extends the scope of traditional thermal injection profiling algorithm with fracture diagnostic functions. During the development process, we incorporate the existing warmback model of conventional wells in analyzing shut-in temperature data with a newly developed stimulated region thermal model. Two main outputs of the model, the injection fluid intake and the fracture propagation extent, are estimated and tested. The model is then automated and thoroughly implemented in the software package.
The primary applications of this work are injection fluid intake and fracture propagation extent of each perforation cluster in fractured wells. The spatial resolution of the injection profiling and fracture growth can reach the sub-meter scale (same as the distributed temperature sensing spatial resolution). Compared to the conventional radial warmback model, the temperature signals from the fractured well show a much faster warming trend while taking relatively larger amounts of injection fluid. This behavior can be attributed to the additional heat loss to the unstimulated region and larger contact area between clusters. On the other hand, leak-off fluids create a cooler stimulated region around the fracture plane, which makes the warmback trend slower compared to the linear flow regime model. The model developed in this study considers both behaviors to simulate the actual datasets.
The inverse model estimates the fracture propagation extent in both the stimulated region as well as the fracture plane. Both estimations can jointly infer the leak-off extent of an individual cluster. As a pilot project, this model is tested on warmback temperature data from two datasets. The injection profiling results using the model are consistent with profiles obtained from other data sources, while the estimated fracture propagation extents of individual clusters present different types of fracture geometry (symmetrical, asymmetrical, double peaks, etc.).
Quantitative injection profiling and fracture propagation extent estimations of an individual cluster using warmback analysis have been proven viable and reliable in this field study. It could be the first quantitative warmback analysis applied to fracture wells in the industry.