After a brief shut-in period following pumping, the DFIT-flowback test (diagnostic fracture injection test with flowback) allows choked production to accelerate fracture closure. DFIT-flowback test analysis is currently limited to empirical approaches. This paper presents a comprehensive model for matching the pressure response for the entire DFIT-flowback test including four behaviors: pumping, shut-in, choked flowback, and rebound when the well is again shut in.
The model couples the material balance function and geomechanics. The material balance function considers injected volume, flowback volume, spurt loss, and matrix leakoff fluid loss and the effect of wellbore fluid expansion on the pressure behavior. The geomechanics model includes wellbore, perforation, and fracture friction losses, fracture tip extension, and elastic fracture closure.
We analyzed each section of a DFIT-flowback field case including injection, brief shut-in, flowback, and rebound. The analysis reveals that injected fluid load recovery is very low, suggesting that fluid remains in primary and secondary fractures opened during fracture injection. The analysis is consistent with recent findings showing fracture swarms in wells cored through Permian Basin and Eagle Ford treatment hydraulic fractures and also considers the implication of joint sets at an angle to the minimum stress direction that have been observed in Marcellus microseismic data and in the Permian Basin cored length.
Besides the Cartesian pressure curve, we tried 3 logarithmic derivative diagnostic plots for flowback and rebound analyses.
The DFIT-flowback test is similar to the pump-in/flow (PI/FB) test developed by Nolte (1979) to accelerate the fracture closing process. The technique has been widely used because of its time efficiency in formation stress estimation, especially in tight formations. Nolte (1979) suggested a flowback rate between 1/6 to 1/4 of the previous injection rate and that it should be at a meaningful fraction of the in-situ leakoff rate. Nolte (1979) initially suggested picking fracture closure pressure at the end of first straight trend in the Cartesian pressure plot (shown in Figure 1). Later, Shlyapobersky et al. (1988) suggested the fracture closure pressure should be picked at the start of the second straight trend in the Cartesian pressure plot because they observed that the pressure response thereafter is dominated by wellbore unloading. Plahn et al. (1997) found that neither method was reliable for determining fracture closure pressure accurately from the and suggested using the intersection of two tangent lines drawn from the first and the second straight trends in the Cartesian pressure plot. Besides, Plahn et al. (1997) observed that when the flowback load recovery is much less than the injected volume, the fracture closure during flowback may represent the closing or pinching of a small fraction of the created fracture near the wellbore, not from global closure of the entire fracture. Instead, most of the fracture remained open during flowback and the initial rebound period. By analogy, this accounts for a fundamental difference between traditional DFIT and the DFIT-flowback tests.