Inter-well communication in unconventional reservoirs has received huge attention due to its significant effects on well production. Though it has long been a known side effect of hydraulic fracturing, well interference has become more prominent and frequent as the industry moves to larger completion designs with closer well spacing and infill drilling. Fracturing of infill wells ("child" wells) directly places the older adjacent producing wells ("parent" wells) at risk of suffering premature change in production behavior. Some wells may never fully recover and, in worst cases, permanently stop producing after taking severe frac hits.
This paper presents an automatic data-driven workflow developed to identify inter-well interference events and their impact on EUR (estimated ultimate recovery) based on changes in the well productivity trend. The innovative approach of the workflow is the ability to automatically analyze interference using the complete production history for all wells in a field, using routinely collected data and without introducing human bias in the derivation of the results, instead applying a consistent criteria. The final result is a comprehensive collection of all well interference events occurred in a field, which may be used as a training set for statistical and machine learning based models aiming at predicting such events.
First, the automatic identification of anomalies in the well behavior was developed and criteria set to label the interference events. Next, probabilistic simulations are run to forecast multiple scenarios to quantify the impact of a well interference event reported in terms of change in cumulative oil production. Finally, every event is analyzed in the overall context of field operations, in an attempt to present possible causes which may explain the change of production behavior.