DOI: https://doi.org/10.36719/2663-4619/118/185-188
Agil Ismayilzadeh
Khazar University
Master student
https://orcid.org/0009-0005-9012-5214
aqilismayilzad021@gmail.com
Machine Learning and Deep Learning - Enhanced Production Decline
Curve Analysis for Improved Oil Recovery Forecasting
Abstract
This study outlines a comprehensive methodological framework aimed at improving the accuracy and predictive power of decline curve analysis (DCA) in hydrocarbon reservoir forecasting. Traditional DCA methods, while widely used, often lack the flexibility to adapt to complex reservoir behaviors, especially in unconventional formations. The enhanced framework integrates statistical, empirical and machine learning techniques to overcome these limitations. Key components include data preprocessing, model selection (e.g., Arps, modified hyperbolic, Duong), hybrid model integration and validation through real field data. The approach also emphasizes uncertainty quantification and sensitivity analysis to ensure robust forecasting. By refining the decline curve modeling process, this framework supports more reliable reserve estimation and production planning, ultimately aiding decision-making in reservoir engineering and asset management.
Keywords: machine, deep, model, integration, validation, engineering