Evaluation of the fracturing fluid flowback based on perforation clusters for horizontal shale gas wells using data-mining methods

Abstract

Multi-stage hydraulic fracturing is a crucial technique in the development of horizontal shale gas wells. After fracturing and shut-in for several days, the well would be opened and the fracturing fluid injected into the gas formation would flow back to the ground along with extracted gas. Field experts suggest fracturing fluid flowback plays a key role in the production of shale gas and hope to find the relationship between flowback and production. However, the flowback and production data we collected from 161 wells in one shale gas field in China illustrate that there is no general relationship between flowback and production. The main reason we deduce is that the commonly-used cumulative flowback amount is not a good indicator to represent shale gas wells’ flowback situation. Therefore, we propose a new flowback indicator based on filed data and a shale gas flow model. The proposed indicator is the cumulative flowback ratio per fracturing cluster calculated on production days, where production day is a newly proposed time measurement substituting the calendar day. Further study illustrates that the proposed indicator can be easily predicted using a random forest model with engineering parameters, geological parameters, and well operation parameters. Besides, fracturing scale parameters and brittle mineral proportion are identified as the most important controlling factors of the proposed indicator. With the help of the newly proposed flowback indicator, field experts can have a new way to comprehend each well’s flowback characteristics and make better development plans.

Keywords

• Shale gas
• Flowback ratio
• Gas production
• Machine learning
• Statistical analysis