Abstract

Frac-hit is defined as the communication between an existing horizontal parent well and hydraulic fracturing treatment of the new well called child well. When a parent well is “hit”, it can be very problematic both operationally and economically depending on the severity of the hit. In industry, frac-hit is considered dominantly a function of well spacing and subsequently as the number of wells in a given shale asset increases, probability of interference between parent and child wells increases significantly. However, by increasing the distance between the wells, the recovery of hydrocarbon from the shale asset reduces. Common techniques like Rate Transient Analysis (RTA) and Numerical Reservoir Simulation, inherited techniques from conventional reservoirs, have proven to be unrealistic due to their degree of assumptions and simplifications during modeling and evaluation of unconventional resources [1-5]. In this case-study, AI/ML techniques, which is a pure data-driven, fact-based method without any assumptions, simplifications, and interpretations, is used to predict and mitigate the frac-hit occurrence more accurate than common practices in industry.