Page 216 - Intelligent Digital Oil And Gas Fields

170 Intelligent Digital Oil and Gas Fields

pressure depletes over time and reaches the saturation (fluid bubble point or
dew point) pressure, then the situation is more complex. The workflow
should be intelligent enough to predict the dew or bubble point pressure
and detect significant changes in GOR and WCs that could affect f BHP .
Industry’s best practice is to perform periodic well testing to calibrate the
values of VFM against the measured collection system. The well tests are
usually conducted from skid-mounted separator units, commonly scheduled
once every 1–3months per well. However, in DOF systems, well testing
should be scheduled based on an automated well test priority ranking using
real-time data and well events to determine if a well should be tested by
exception.
An automated workflow is needed to auto-calibrate the VFM, so that the
VFM can output production data in real time even if the well is not tested.
This workflow is designed to
Detect automatically the well testing event (using GPS in the mobile tes-
ter truck or pressure detection in separator test).
Capture real-time (24h) gas, oil, and water volumes, and temperature
and pressure data.
Clean and filter the data from frozen, out-of-range, flagged, and other
signal abnormalities.
Estimate average values during 24h.
Feed data into the well performance software.
Perform model calculations and estimate rates and calculate errors
between measured and calculated rate and f BHP .
Provide historical comparisons with previous well tests and VFM
models.
Provide guidance on whether test data is valid based on model adjust-
ment and errors.
Adjust models to reduce error, if any, by changing coefficient and time-
dependent factors such as PI, skin, and other coefficient correlations, for
example, a1, a2, a3, P1, P2, etc., and change stationary data.
Estimate rate and f BHP using calibrated data up to the next well test.
For example, in an unconventional gas well, the gas flow meter data and gas
flow from the VFM were compared to measure the relative error with
changes in choke size. Fig. 5.10 shows the daily gas rate reading for a
multiphase flow meter, nonstationary VFM, stationary VFM, and well test.
The gas flow meter was also compared with well test data, with three
changes in choke size, from 24/64in. to 32/64in. to 48/64in. The non-
stationary VFM responds quickly to the pressure changes; the regular