Leveraging Data Physics for Salt Water Disposal Pressure Prediction with ConocoPhillips

In partnership with ConocoPhillips, this study presents a novel approach to modeling saltwater disposal (SWD) pressure in shallow formations to mitigate drilling and completion risks for unconventional wells located below disposal intervals. Proper surveillance of SWD operations is crucial, as uncontrolled pressure buildup can compromise well integrity. Reliable predictive models are essential for optimizing injection rates, selecting ideal injector locations, and minimizing drilling hazards by avoiding high-pressure zones. However, traditional reservoir simulation methods for SWD pressure estimation face challenges due to complex geology and sparse, unreliable pressure data, leading to low-confidence predictions.

To address these limitations, this study introduces a Data Physics-based modeling technique that combines reservoir physics with machine learning, providing a more efficient and reliable alternative to conventional reservoir simulation. This approach simplifies model representations while maintaining physical rigor, allowing for rapid updates and accurate pressure predictions. Additionally, an ensemble-based data assimilation method accounts for uncertainties, ensuring robust results. By leveraging both physics-based constraints and data-driven adaptability, the model enhances pressure forecasting capabilities for SWD formations.

The effectiveness of this technique was validated using a synthetic simulation model, demonstrating its ability to generate accurate pressure predictions with minimal input data. The model was further applied in the Bakken and Permian basins, where it successfully forecasted pressure responses to SWD activities and informed new injection strategies aimed at reducing drilling and completion risks. This innovative approach offers a practical and scalable solution for managing SWD operations, enhancing both safety and operational efficiency in unconventional oil and gas developments.