Cost-Effective Production Metering and Allocation in a Mature Offshore Oil Field

BHP Billiton

You have access to this full article to experience the outstanding content available to SPE members and JPT subscribers.

To ensure continued access to JPT's content, please Sign In, JOIN SPE, or Subscribe to JPT

This paper describes how the use of production-modeling tools, coupled with field pressure/temperature-data-acquisition systems and programming software, served as a means to improve production allocation and surveillance on a real-time basis in the Greater Angostura Field. The complete paper discusses a practical approach to obtain accurate real-time well-production rates without the need for purchasing costly flowmeters for each well. It also demonstrates how virtual metering can be used to identify production-optimization opportunities in the production system more readily.


The Greater Angostura Field is located 25 miles offshore Trinidad and Tobago. The field was developed in three phases: Phase 1, which includes the Kairi, Canteen, and Horsts fields; Phase 2, which includes the Aripo gas development; and Phase 3, which was scheduled to have been completed by the second half of 2016.

At the time this paper was written, the Angostura development consisted of 21 producers and seven gas injectors. The facilities and gathering system included one manned central-processing platform (CPP), four unmanned well-­protector platforms (WPPs), and one gas-export platform (GEP). All wells have dry production trees and are distributed as follows:

  • Canteen 1 (C1) WPP: five producers and two injectors
  • Kairi 1 (K1) WPP: three producers and two injectors
  • Kairi 2 (K2) WPP: nine producers and three injectors
  • Aripo (A) WPP: three producers

The production from the wells in each WPP is commingled in each platform’s header and routed as follows:

  • Canteen 1 (C1) WPP: one subsea production flowline to the low-pressure gathering system
  • Kairi (K1) WPP: two subsea production flowlines to either the low-pressure or the high-pressure gathering system
  • Kairi (K2) WPP: two bridge-connected flowlines to either low- or high-pressure gathering system
  • Aripo (A) WPP: one subsea production flowline to GEP

The oil produced from C1, K1, and K2 is processed on the CPP and pumped to shore. The gas produced from these three WPPs is distributed to export gas. The production of the Aripo WPP is sent directly to the GEP for sales.

Historical Well Testing and Production Allocation

The historical allocation methodology was designed to allocate oil, gas, and water production and gas injection to wells and reservoirs and to account for well downtime. Wells are tested periodically in a test separator or a multiphase flowmeter (MPFM), depending upon the well. The C1 wells can be tested only through an MPFM, K1 wells through either an MPFM or a test separator, and K2 wells only through a test separator.

The accuracy of this method of allocation is dependent upon frequent, representative well tests on all of the wells. However, as the field entered its mature phase, liquid rates of many wells fell outside the design range of the test-separator meters. Also, there were periods when issues developed with the MPFM, affecting the accuracy of some well tests. Operational constraints often affected the frequency of well tests available for use, and operating conditions during well tests may be different compared with normal operations for some wells. 

The combination of these issues began to affect the allocation factor (AF) across the fields and reservoirs. AFs are determined over time on the basis of a comparison of the sums of test data of oil, gas, and water production vs. the actual total production measured for fiscal and regulatory purposes. These comparisons were normally conducted on a monthly basis.

Improved Methodology

Accurate production allocation underpins effective production and reservoir surveillance, evaluation of infill and intervention opportunities, and ­production-optimization efforts. Therefore, in order to mitigate some of the uncertainties associated with the allocation approach, a model-based methodology was developed to integrate periodic well tests with ongoing gathering-system operating conditions.

Production Model. The first step was to build a robust production model capable of simulating the pressure/temperature behavior of the wells and gathering system, as well as honoring the material balance for all three streams of production. The model was built with a bottom-up approach from the sandface to the separator.

Individual-well models were built to reflect deliverability on the basis of inflow and outflow relationships. To capture the inflow performance, pressure/volume/temperature data and reservoir parameters were used. Such parameters include pressure, temperature, completion interval, skin, and drainage area. A tank model for the reservoir was sufficient for this purpose.

The well-outflow modeling was based on the well trajectory and completion-equipment dimensions. Each well model was calibrated with known fluid-­property data and pressure/temperature responses obtained from the downhole and wellhead gauges and representative well tests. The model can adjust well performance as system pressure changes and can be updated as additional well tests are obtained over time.

After creating models for each of the wells, a model of the entire production-gathering system was built. The network model was constructed with the physical dimensions of the flowlines using “as-built” schematics. The model was calibrated to simulate flow, pressure, temperature, and choke responses effectively across the system. 

When the well models are combined with the network model, the result is a robust integrated model capable of capturing the physical response of multiphase-flow behavior in the entire production system. The predicted system results can be compared with actual real-time measurements at specific locations. The model is controlled by use of a specific boundary condition for the operating pressure of each gathering system.

Redesigned Well-Testing and Allocation Methodology. After the integrated production model was calibrated, a new well-test methodology was implemented as follows:

  1. Well is routed to its well-test stream.
  2. The gas-rate measurement is recorded. Note that the gas-production rates and pressures are within the accuracy range of the well-test equipment.
  3. Water cut is estimated with the latest fluid sample collected at the well’s production tree. Fluid samples are taken regularly for each well.
  4. Pressure and temperature data at the downhole and wellhead gauges of the well are matched by use of the input data from Steps 2 and 3.
  5. The network model is run on the basis of the boundary conditions of the low-pressure-gathering-system/high-pressure-gathering-system and test-separator pressures at the time of the test.
  6. Simulated results of aggregate production, pressures, and temperatures are compared with the actual production rates and pressure/temperature readings at different points throughout the entire production network where gauges are located. If the simulated results do not match the system’s measured data, the water cut and gas/oil and gas/liquid ratios for the tested well are re-evaluated.
  7. Simulations can be run whenever new well-test data are available. AFs for the system are calculated with each simulation.
  8. A benefit derived from modeling all wells and gathering systems as a single unit is increased confidence in the updated individual-well predictions.  

Virtual-Metering Application. C1, K1, and K2 do not have dedicated flow­meters for each well. Thus, the production rate for a well at any given time is dependent on well uptime and operating conditions downstream of the well. Improved allocation can be achieved if well performance can be modeled or matched on an ongoing basis. 

To accomplish this objective, an event-driven application was coded by use of programming software. The application couples the production-network model with the real-time data-acquisition system; it continuously enters the boundary conditions that control the model (and affect well and system performance). The model then exports user-specified parameters, including production rates, pressures, and temperatures, at specific points in the system.

The application improves production-surveillance capabilities because potential issues can be identified more readily. A graphical user interface was developed to display the parameters calculated by the production model as well as the values retrieved from the real-time data-­acquisition system. Discrepancies between model-calculated and actual parameters alert the engineer to focus on a specific area when performing diagnostics.

Additionally, the virtual metering facilitates the identification of production-optimization opportunities. The performance of any well or the entire system can be evaluated against potential changes in well routing or system-­configuration changes. If an actual change is made, the model prediction can be verified with real-time system data.

Event Handling and Other Modeling Considerations.

  • Shut-ins: The virtual-metering application recognizes when a well is shut in by examining the pressure readings extracted from the manifold and wellhead gauges. When the pressure differential between these two gauges exceeds a user-specified threshold, the tool recognizes that the well is shut in and stops performing the calculations for that well.
  • Wellhead choke position: The application controls the choke size in the model by taking the pressure differential between the pressure gauges upstream and downstream of the wellhead choke and inserting this pressure value in the model.
  • Wells with no downhole gauges: For wells that do not have downhole gauges, pressures and rates are estimated by performing gradient calculations. The calculated rates and pressures/temperatures must tie in with the overall material and pressure/temperature balance of the entire production network.
  • Reservoir pressure: The virtual-metering application is not currently used for production forecasting. It is sufficient to manually adjust the reservoir pressure periodically on the basis of data collected during shut-ins.


The new well-testing methodology resulted in significant improvements in the allocated rates. The AFs for all three streams of production are within ±10%. Deviations are caused by nonsteady-state conditions created by fluctuations in the process equipment on the CPP and by shutdown/startup events.

This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 180887, “A Cost-Effective Methodology for Production Metering and Allocation Using Real-Time Virtual Metering in a Mature Offshore Oil Field: A Case Study of the Greater Angostura Field,” by Italo Raffaele Acuna, BHP Billiton, prepared for the 2016 SPE Trinidad and Tobago Section Energy Resources Conference, Port of Spain, Trinidad and Tobago, 13–15 June. The paper has not been peer reviewed.

Cost-Effective Production Metering and Allocation in a Mature Offshore Oil Field

01 March 2017

Volume: 69 | Issue: 3


Don't miss the latest content delivered to your email box weekly. Sign up for the JPT newsletter.