Virtual Flow Metering
A virtual flow meter (VFM) is a software-based system that estimates individual well flow rates (oil, gas, water) using existing sensor data and physics-based or data-driven models - without a physical multiphase flow meter or test separator at each well.
Why Virtual Metering?
Cost Savings
Physical multiphase flow meters cost $200K-$500K per well installed. A VFM covering 100 wells costs a fraction of that - typically $50K-$200K total for the software.
Continuous Monitoring
Traditional well testing gives a snapshot every 30-90 days. VFMs provide rate estimates every 5-15 minutes, enabling near-real-time production surveillance.
Better Production Allocation
Accurate individual well rates improve production allocation - critical for fields with multiple equity partners or fiscal metering requirements.
VFM Approaches
Physics-Based (First Principles)
Uses well and fluid models (IPR, VLP, PVT correlations) to calculate rates from pressures and temperatures. Requires good knowledge of well and fluid properties.
Strengths: Interpretable, works from day one, extrapolates well. Weaknesses: Requires tuning, sensitive to input assumptions.
Data-Driven (ML-Based)
Trains ML models on historical well test data paired with concurrent sensor readings. The model learns the relationship between sensor inputs and measured rates.
Strengths: Adapts to complex conditions, no detailed well model needed. Weaknesses: Requires sufficient test data, may not extrapolate beyond training range.
Typical VFM Input Data
THP
Tubing head pressure
THT
Tubing head temperature
CHP
Casing head pressure
Choke %
Choke opening position
GL Rate
Gas lift injection rate
ESP Data
Frequency, current, intake P
Use case: An offshore platform with 40 wells deploys a hybrid VFM that combines physics-based models with ML correction factors. The VFM is calibrated against monthly well tests and provides 15-minute rate estimates with <5% error on oil rate. This eliminates the need for a dedicated test separator ($3M+ capital cost) and reduces well testing frequency from monthly to quarterly - saving 200 hours/year of operator time.