Data Analytics and Machine Learning Predictive Modeling for Unconventional Reservoir Performance Utilizing Geoengineering and Completion Data: Sweet Spot Identification and Completion Design Optimization

Abstract

Abstract Performance of wells in an unconventional reservoir are largely diverse due to different geology, reservoir characteristics, and completion design. A comprehensive method of data analytics and predictive Machine Learning (ML) modeling was proposed to identify production zone "sweet spots", and optimize completion designs from reservoir quality (e.g., geological, geophysical, and geomechanical) data and completion quality data (e.g., frac stage spacing, fluid volume, and proppant intensity, in order to enhance performance of production wells in unconventional reservoirs. Typical data analytics and predictive ML modeling approach utilizes all the reservoir quality data and completion quality data together, which mostly leads to domination of the completion quality data over the reservoir quality data because of higher statistical correlation (i.e., weight) of the completion data to observed production. Hence, resulting predictive ML models commonly underestimate the effects of the reservoir quality on production, and exaggerate the influence of the completion quality data. To overcome the shortcomings, the reservoir quality data and the completion quality data are separated and normalized independently. The normalized reservoir and completion quality data are utilized to identify sweet spots and optimize completion design respectively, through predictive ML modellings. The patent-pending methodology of predictive ML modeling has been exercised in recently developed wells of the Montney unconventional shale gas formation, British Columbia Canada, and identified sweet spots from key controlling reservoir quality data and as well as prescribed optimal completion designs from key controlling completion quality data.