The nv.analysis team has developed an AI-driven approach for selecting wells that may be potential candidates for perforation workovers . The study aimed to assess the extent of correlation between production, well testing and logging data to select well intervention candidates.

Key tasks of the study included data collection, pre-processing, annotation, as well as AI training and fine-tuning. This led to the development of our own AI algorithm that can facilitate the identification of probable perforation intervals. Real data prove the efficiency of the algorithm, which ensures 88% consistency with opinion of experts studying test samples. In simple terms, 9 out of 10 AI-selected intervals coincide with those selected by expert teams manually.

Such high accuracy shows the potential for large-scale deployment of our solution for the identification of well candidates for perforation, boosting efficiency of scheduling well intervention activities.