Nonintrusive Rapid Evaluation of Fracture Isolation in Plug-And-Perf Completions Based on Data Science Methodology

Plug-and-perf is a commonly used technique to complete wells in tight reservoirs. Plugs are used to ensure zonal isolation along the lateral for proper stimulation. Plug failure can lead to understimulation of a fracture stage and overstimulation of a previously stimulated zone. Identifying plug failures using only pressure behavior can be time consuming and, in many instances, misleading as pressure drop can be attributed to other aspects of the fracturing process. The objective of this work is to implement a nonintrusive technique to effectively identify plug failure in near real time to support the decision making during fracturing operations. An innovative data science–based methodology is used for high-frequency surface pressure oscillation decomposition and interpretation during the hammer immediately after shut-in. An advanced algorithm identifies the fluid entry depth and estimates the probability of the treatment going to the intended stage or a deeper zone. Before the main fracturing treatment, the hammer of a prior short injection test is analyzed, and fluid entry depth is estimated. Depending on the probabilities of the fluid being injected into the target zone, the decision is made whether to run another plug or continue with the main fracturing. Several hydraulic fracture stages were monitored following this methodology and results were compared to multiple conventional criteria such as ball landing differential pressure, pressure drop magnitude during injection, pressure drop rate, and treating pressure level at the target rate. The ability to deploy a plug or not below the deepest cluster was also used as an indication whether there was a fluid flow down to the previous zone. Twenty-four stages showed proper isolation, with 83% successfully identified using the data science approach. Similarly, six stages showed indication of communication to a previous zone with 67% of these events detected. The analysis proved effective, and interpretations were completed within 15 to 30 minutes, preventing operations delays. The quality of the hammer effect could be detrimental to the interpretation; therefore, a second short injection test may be necessary in some cases. This innovative methodology can benefit completion engineers in evaluating the integrity of the frac plug isolation in a timely and effective manner enabling operation optimization. The paper utilized two separate wells to evaluate the efficacy and usefulness of this methodology to see field operational capability and real world application.