A Surrogate Model Approach for Fast Estimation of Production Flow Rates in Hydraulically Fractured Wells

Abstract The development of both unconventional and tight oil and gas fields increasingly relies on multistage hydraulic fracturing to be economically viable. While detailed modeling and production simulations are straightforward to develop, rapidly and accurately estimating an asset's production potential remains a challenge. Given the wide variety of geological conditions and operational constraints, there is a pressing need for a universal solution that can optimize the design and execution of fracturing operations. In this paper we propose a machine-learning- (ML-) based technique for the rapid estimation of initial production rates in horizontal wells with multiple fractures. The surrogate model employs comparative predictive modeling using an extensive dataset of numerically resolved simulations. We establish a universal equation that delivers reliable results, providing an efficient tool for asset evaluation and well completion optimization.