A theory and data-driven method for rapid bottom hole pressure calculation in UGS.

Area of Science:

• Petroleum Engineering
• Artificial Intelligence in Energy
• Reservoir Engineering

Background:

• Accurate bottom hole pressure calculation is vital for Underground Gas Storage (UGS) operations, impacting dynamic analysis and production optimization.
• Traditional methods face challenges in efficiency and precision, particularly in dynamic UGS environments involving injection, withdrawal, and shut-in stages.

Purpose of the Study:

• To develop an innovative and efficient method for calculating bottom hole pressure in UGS operations.
• To enhance the operational and maintenance efficiency of UGS through improved pressure calculation.

Main Methods:

• Comprehensive analysis of factors influencing bottom hole pressure based on wellbore flow theory.
• Construction of a neural network model utilizing characteristic variables related to bottom hole pressure.
• Integration of wellbore flow equations with theoretical samples and real-world data for a theory and data-driven neural network model (TDDNN).

Main Results:

• The developed TDDNN model achieves rapid and accurate bottom hole pressure calculations.
• The novel method demonstrates superior performance over traditional techniques across key precision metrics (MAE, MSE, RMSE, MAPE, R²).
• Significant reduction in computational time from seconds to milliseconds compared to traditional theoretical approaches.

Conclusions:

• The TDDNN method offers a significant advancement in bottom hole pressure calculation for UGS.
• This approach provides high prediction accuracy and enhanced computational efficiency, crucial for UGS management.
• The study presents a valuable reference for applying deep learning in sample-limited environments within the energy sector.