A Digital Twin Approach to ESP Gas Separation Efficiency Using Pump Discharge Pressure and Physics-Based Modeling

Abstract Gas interference remains a leading cause of ESP underperformance and early failure, especially in unconventional wells where low bottomhole pressure and high gas-liquid ratios are common. This paper presents a method for estimating gas separation efficiency using a physics-based digital twin calibrated with pump discharge pressure (PDP) sensor data. By integrating detailed ESP configuration, well-specific fluid properties, and segmented pressure modeling, the method calculates real-world separation efficiency and links it directly to ESP performance and risk of gas lock. The model was applied to over 200 ESP-equipped wells in the Bakken and Permian basins. In all cases, calculated discharge pressures matched measured PDP within ±5 percent, validating the approach. Results revealed wide variability in actual separator efficiency across well conditions and vendors, with many cases falling below catalog values. This highlights the limitations of fixed assumptions and underscores the value of sensor-calibrated diagnostics. The method offers a practical path for improving ESP design, separator selection, and gas handling strategy across diverse operating environments.