This study assesses empirical correlations and time series models for forecasting production in unconventional oil wells, focusing on their effectiveness in complex reservoir dynamics. Utilizing data from 11 unconventional wells in the Wolfcamp Formation, Midland Basin. The research compares 11 empirical decline curve models with modern time series models like ARIMA, LSTM, and GRUs. The results indicate no single model consistently excels, necessitating tailored approaches for each well. Notably, the ARIMA model outperformed the Logistic Growth Model with a significant R2-score of 93.50%, demonstrating its higher accuracy and reliability in capturing data variability for dependable forecasts. This comparison highlights the potential of advanced machine learning models in managing the noisy data typical of these reservoirs, advocating for a shift towards these more sophisticated techniques to enhance accuracy and efficiency in forecasting. This work provides crucial insights into the comparative performance of forecasting models, offering guidance for reservoir engineers to optimize well performance in challenging unconventional environments.
The Wolfcamp formation, located in the Midland Basin of the Permian Basin, has gained significant attention due to its important oil production (EIA, 2022). Unconventional reservoirs, including formations like Wolfcamp, have been crucial in enhancing oil and gas production in the United States. However, primary recovery from these reservoirs typically remains low, often not exceeding 10% (Gupta et al., 2020). Proper modeling and forecasting of production from such reservoirs must consider critical factors like reservoir heterogeneity, such as the presence and impact of natural fractures (Nandlal & Weijermars, 2019b).
While conventional reservoirs have well-established techniques for production forecasting, unconventional reservoirs present unique challenges due to their complex geological characteristics. Accurate prediction and forecasting of production from this unconventional reservoir are crucial for optimizing field development strategies and making informed investment decisions (Cao et al., 2016). The Wolfcamp A formation presents unique challenges due to its geological complexity. It consists of lithologically diverse layers, including silty and organic-rich zones, which can significantly affect fluid flow and, consequently, production behavior (Waite, 2021).
