A systematic investigation into the flowback cleanup of hydraulic-fractured wells in unconventional gas plays

This paper conducts an extensive investigation into fracture cleanup efficiency by considering several pertinent parameters instantaneously over a wide practical range. Injection, shut-in and production stages of the fracturing operation were simulated for 32 sets consisting of 113,072 runs. To perform such a large number of simulation runs, a computer code was utilised to routinely read input data, implement the simulation runs and produce output data. In each set (which consists of 4096 runs), instantaneous impacts of twelve different parameters (i.e., fracture and matrix permeability, Brooks matrix capillary pressure (Pc) parameters, and Brooks-Corey relative permeability parameters) were investigated. To sample the domain of variables, full factorial experimental design (two-level FFS) was employed. The linear surface methodology was used to map the simulation output, which is the loss in gas production (GPL), compared to the clean case (i.e., 100% clean-up) after three production periods of 10, 30 and 365 days. The impact of various combinations of fracture fluid injection volume, fracture length, shut-in soaking time, matrix permeability variation range and drawdown on GPL were studied in different sets. Additionally, more simulation sets were performed to capture the impact of hysteresis, layering and mobile formation water on the clean-up efficiency. Results indicated that in line with some literature data, factors that controlled the mobility of FF inside the fracture had the most significant impact on cleanup efficiency. It was also noted that injecting high volumes of FF, into very tight formations significantly delayed clean-up and impaired gas production. The effect of varying other parameters such as extending soaking time or increasing pressure down in such a case delivered negligible GPL improvement. Introducing hysteresis made clean-up slightly faster in all production periods. The impact of the gravity segregation was discussed in this study. Considering the layered systems, it was indicated that in the top layer, the fracture mobility coefficients were more important than the ones in the bottom layer whist capillary pressure seems to become more important in deeper layers compared to the top layers. Additionally, a slower clean-up was observed for sets with larger initial water saturation compared to those cases with immobile water saturation due to the detrimental effect of mobile water on gas production. In some cases, with significantly high values of water saturation, using chemicals (which IFT reducing agents) to reduce Pc could reduce GPL and improve cleanup efficiency. These findings contribute to the further understanding of the fracture fluid cleanup process and provide practical guidelines to achieve economically successful hydraulic fracturing operations, which are popular but expensive for tight and ultra-tight reservoirs.