Statistical Analysis of Effect of Sensor Degradation and Heat Transfer Modeling on Multiphase Flowrate Estimates from a Virtual Flow Meter

Abstract

Accurate flowrate measurements in petroleum production systems are important for optimization, fiscal metering, and production allocation. Sometimes, Virtual Flow Meters (VFMs) are used for this purpose instead of physical meters to reduce cost. These systems estimate the flowrates using a computational model that represents accurately the production system of interest. Since VFM systems mostly rely on pressure and temperature measurements, it is important to understand how accuracy and degradation of sensors influence the VFM flowrate estimates. In this work, a VFM system for a subsea oil well was created using a transient multiphase model built in a commercial software and controlled from an external computational routine. A statistical analysis of VFM simulation results was performed to quantify the effect of pressure sensors degradation on the VFM flowrate estimates. In addition, the effect of temperature matching and a segmented approach to represent the well heat transfer were evaluated. The analysis showed that the sensor degradation effect should be considered in VFM systems carefully, especially if a high estimation accuracy is required. Measurement drift was found to be the most critical factor of the sensor degradation but high measurement noise can also cause considerable errors of the flowrate estimates. In addition, it was found that a complex representation of the wellbore heat transfer is not required to obtain accurate flowrate predictions and simplified models can be used instead.