Page 238 - Intelligent Digital Oil And Gas Fields

188 Intelligent Digital Oil and Gas Fields

in the control panel. In the absence of an MPFM, VFM validated peri-
odically with a well test are required to evaluate the diagnostic in wells
with ESP or PCP systems.
Constraints: Minimum allowable f BHP , pressure drop, maximum velocity
to avoid early water breakthrough or sand screen out, maximum rate to
avoid coning or cresting, maximum f BHP to avoid fracture pressure, and
maximum operating frequency.
KPI: Typical KPIs used for diagnostic are: pump wear factor, gas inter-
ference, chance of tubing leak, viscous effect in pump, and solid plugged
intake.
Output: Pump head, liquid rate, pressure drop in pump, and pressure
drop in reservoir.
Control: Pump frequency and choke size.

5.6.3 Diagnostic Procedure

The value of diagnostics in real time is to prevent additional workovers or
well interventions. Real-time diagnostics could be the most effective process
to identify pump degradation and impairment in the life of a well. The full
diagnostic process should include the following processes:
Alarm system: Generate a primary data set with minimum, maximum, and
average values and monitor in real time values that exceed the threshold
ranges. Validate if monitoring values are persistent in the next 24h and apply
filtering algorithms to clean false data.
Diagnostics based on a nodal analysis model: The average production and
pump data are sent online to a preexisting well model, which is calibrated
with the latest well test information, pump design, well trajectory, PVT data,
and completion schematic.
Automated model analysis: The model is updated with the latest 24-h
(week, month) average of real-time data, such as pressure and temperature
data, THP, THT, PIP, PDP, PIT MT, flow data, WC, gas rate, and oil rate.
ESP model tuning is performed online daily by comparing the model and
sensor data.
Model match: Fig. 5.21 shows a typical ESP pump gradient showing
f BHP , PIP, PDP, and THP for a well model and sensors. The model com-
putes different values between the calculated and the modelled data. The
tuning process is matched by changing pump wear factor, multiphase
correlation friction factors, multiphase correlation gravitational factor,
and productivity index. In cases where the system cannot find a solution