Page 233 - Intelligent Digital Oil And Gas Fields

Workflow Automation and Intelligent Control 183

• Use ANNs to predict short-term (+1, +3, +5, +7, +15, +30days) pro-
duction forecasts and couple with probabilistic analysis.
• Use fuzzy logic and pattern recognition to predict equipment
malfunctions and generate actions to prevent and avoid production
downtime for ESP, GL, PCP, RPs, etc.
• Use pattern recognition with expert rules to build a knowledge-capture
system and generate solutions for troubleshooting and well remediation.
• Track actions and automatically identify well events, that is, automati-
cally recognize that a well test has started by understanding the pressure
buildup without human intervention, or identifying a rig or workover
intervention.

5.5 WELL TEST VALIDATION AND PRODUCTION
PERFORMANCE IN RIGHT TIME

Well test validation is a crucial component of production perfor-
mance, particularly for wells without a multiphase flow meter (MPFM)
and that depend only on physical models, which must be tuned frequently.
If operations and facilities allow it, well tests should be conducted at least
once per month. For best results, the test must be performed with three
or more choke valve or frequency changes for a period of 24h or until sta-
bility of surface pressure is reached.
The test can be conducted using a portable MPFM (integrated with
GPS) or a dedicated flow line test system. The system measures in real-time
wellhead pressure, wellhead temperature, fluid rates, choke size, amperage,
voltage, frequency, and GL injected volume. With the common RTU sys-
tem, the signals are sent to a gathering center.
Workflow automation is recommended to collect data, estimate an aver-
age value flow rate, update the nodal analysis model, and generate recom-
mendations for the virtual meter. Basically, the workflow should run every
day when a flow test is performed. At a minimum, the workflow should per-
form the following tasks and may also perform others:
• Collect, filter, and clean up the data during the flow periods.
• Automatically identify the flow period by changes in choke settings and
measure the response on pressure and rates.
• Automatically identify if the test is running at stable condition (δp/δt ¼ 0).
• Estimate average properties of gas, oil, GOR, water rates, and pressure
for each individual flow period with the corresponding 2σ and 3σ stan-
dard deviation.